Modeling Prehistori--(ｾ Ｍ Agricultu ｾ Productivity in Southwes ern Colorado: AGIS Approac Washington State University Department of Anthropology REPORTS OF INVESTIGATlONS 67 published jointly by WASHINGTON STATE UNIVERSITY DEPARTMENT OF ANTHROPOLOGY, Pullman, Washington and CROW CANYON ARCHAEOLOGICAL CENTER, Cortez, Colorado Complete bibliographic format for Repons of Investigations 67: Van West. CarlaR. 1994 Modeling Prehistoric Agricultural Productivity in Southwestern Colorado: A GIS Approach. Repons of Investigations 67. Department of Anthropology, Washington State University, Pullman, and Crow Canyon Archaeological Center, Cortez, Colorado. © 1994 Carla R. Van West ISSN 1059­4000 This repon is printed on acid­free paper. Reports of Investigations is an occasional series published by the Department of Anthropology, Washington State University, to disseminate information resulting from research undertaken by students, staff, and faculty associated with the department Timothy A. Kohler, General Editor. This volume is published jointly with Crow Canyon Archaeological center, an independent. not­for­profit organization devoted to archaeological research and education. ISBN 0­9624640­6­6 -ii- Page Blank in Original To my Father Joseph B. Van West 1922­1987 Page Blank in Original T ABLE OF CONTENTS DEDICATION iii v TABLE OF CON1ENTS vii LIST OF TABLES LIST OF ll.LUSTR.ATIONS . xi PREFACE . xiii xv ACKNOWLEDGMENTS 1. IN1RODUCTION 1 Preview of the Model Organization of the Study Spatial Context Temporal Context Research Context 2 1 5 5 7 13 BUll.DING THE MODEL: METHODS AND DATA ACQUISmON High Resolution Soils Data Modem Crop Yield Data Digital Elevation Data Modeling Soil Moisture Tree­Ring Data GIS Technology Archaeological Data 3. BUll.DING THE MODEL: PRELIMINARY DATA ANALYSES Palmer Drought Severity Index Reconstruction Geographic Information System Processing Linking Soil Moisture Conditions and Levels of Agricultural Productivity 4. BUll.DING THE MODEL: FINAL DATA ANALYSES Final Geographic Information System Analyses Supplemental Analyses with SAS The Complete Model 5. RESULTS 6. SUMMARY AND EVALUATION The Study Area Localities within the Study Area: The Block Survey Areas Site Catchments within the Study Area: Tree­Ring Dated Sites 17 17 40 41 41 43 49 52 55 55 94 97 . 113 113 120 129 131 132 136 147 , 187 Sununary Strengths of the Model Limitations of the Model Reconunendations for Future Research -v- 187 189 190 191 BIBUOGRAPHY 195 APPENDICES A. Eigenvector Amplitudes Generated by a Principal Components Analysis on Seven Expanded Tree­Ring Chronologies (SWOLD7) 215 Historic Commentary on Climate and Agriculture in Montezuma County, A.D. 1894­1970 237 C. Calculating Arumal Population Size: An Example for A.D. 902 241 D. SAS Programs for Calculating Total Annual Maize Productivity (TOTPROD) and Population Density (pOPKM) 243 Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Study Area, A.D. 901­1300 246 Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Sand Canyon Survey Locality, A.D. 901­1300 247 Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Mockingbird Mesa Survey Locality, A.D. 901­1300 248 Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Catchment of 5MT8371, DCA Site. A.D. 901­1300 249 Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the catchment of 5MT8839. Norton House,A.D. 901­1300 250 Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the catchment of 5MT2433, Aulston Pueblo. A.D. 901­1300 251 Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the catchment of 5MT3834. Mustoe Ruin, A.D. 901­1300 252 Total Arumal Maize Productivity and Maximum Potential Population for Three Levels of Storage in the catchment of 5MT6970, Wallace Ruin. A.D. 901­1300 253 Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Catchment of 5MT1566. I..oWty Ruin, A.D. 901­1300 254 Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Catchment of 5MT2149. Escalante Ruin, A.D. 901­1300 255 Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Catchment of 5MT765. Sand Canyon Pueblo, A.D. 901­1300 256 B. E. F. G. R L J. K. L. M. N. O. ­vi- LIST OF TABLES 2.1 Soil Conservation Service Soil Capability Class Definitions 2.2 Primary Soil Data for 113 Soils in Montezuma and Dolores Counties, Colorado, Part 1 . 21 Primary Soil Data for 113 Soils in Montezuma and Dolores Counties, Colorado, Part 2 . 25 . 40 . 44 2.3 _ . 19 2.4 Inter­ and Intra­Coder Variability Rates in Soil Recording 2.5 Weather Stations Used to Reconstruct PDSI 2.6 2.7 Water­Holding Characteristics of Soils _ . 45 Tree­Ring Chronologies Used in the Creation of SWOLD7 _ . 48 Tree­Ring Dated Sites . 53 Soils Selected to' Represent the 11 Classes of Soil Moisture Based on Water Held in the Upper and Lower Soil Layers . 56 Results and Statistical Summary of the Initial Calibration, Verification, and Final, Full­Period Calibration Analyses for 55 Soil Moisture Groups .. 59 . 73 Descriptive Statistics Associated with 55 Fmal, Full Period Calibrations and Long­Telm Reconstructions . 75 Occurrence of Extreme Conditions ($.4.00 or ｾＫＴＮＰ PDSI) in the Years Used to Calculate PDSI for Combinations of Weather Station and Moisture Class _ . 87 Comparative Data from Long­Telm Reconstructions in the Zuni Area and Southwestern Colorado . 91 . 92 2.8 3.1 3.2 3.3 "2.................•................•.....•....................................................... Adjusted r 3.4 3.5 3.6 3.7 _ Values for 55 Full­Period Calibrations Comparison of Historic Comments with Annual PDSI Values for 34 Long­Telm Reconstructions, A.D. 1890­1970 _ _ 3.8 Historic Crop Yields and PDSI Values, A.D. 1931­1960 . 98 3.9 Distribution of Soils for Which Nonirrigated Bean Yield Data Are Available by PDSI Reconstruction Class . 99 . 101 3.10 Results of Linear, Polynomial, and Partial Regression Studies of the Relationship Between PDSI and Crop Yield Values -vii - _ 3.11 3.12 3.13 Table of Productivity Values for All Soil Types within the Study Area Under Five Different Climatic Conditions . 104 Nonirrigated Bean Data from Dolores and Montezuma Counties Used to Generate a Mean Bean Yield Value for Years A.D. 1931­1960 . 107 Nonirrigated Maize Data from Dolores and Montezuma Counties Used to Generate a Mean Maize Yield Value for Years A.D. 1931­1960 . 108 . 109 3.14 Eleven Nominal PDSI Classes 4.1 PDSI Class Correspondence Table . 113 4.2 Productivity Table (pROD.TBL) Used in GIS Analysis . 116 4.3 An Example of the Annual EPPL7 (GIS) Tabular Output: A COUNT File of Agricultural Productivity in A.D. 902 ; . 119 4.4 _ Modem Nonirrigated Bean Yield for 46 Agricultural Soils in Montezuma and Dolores Counties, Colorado . 121 4.5 Summary of Model­Building Steps: Operationalizing the Model _ . 127 5.1 Carrying Capacity Estimates at Three Levels of Storage for A.D. 9011300 for the 1470.36 km 2 Study Area _ . 135 5.2 Periods of Greatest Occupational Attractiveness for the Study Area 5.3 Comparison of Population Values for the Study Area, Sand Canyon Survey Locality, and Mockingbird Mesa Survey Locality, A.D. 901­1300 . 143 Carrying Capacity Values (pOPKM) for the Study Area, Sand Canyon Survey Locality, and Mockingbird Mesa Survey Locality, Based on POnYR Estimates for A.D. 901­1300 . 143 5.5 Tree­Ring Dates from Site 5MI'8371, DCA Site . 148 5.6 Comparison of Population Values (pOPKM) for 5MT8371, DCA Site, for A.D. 901­1300 and A.D. 935­950 _ . 149 Periods of Equal or Greater OCCUpational Attractiveness in the Catchment of 5MT8371 (A.D. 935­950) . 150 5.4 5.7 . 136 5.8 Tree­Ring Dates from Site 5MT8839, Norton House _ . 150 5.9 Comparison of Population Values (pOPKM) for 5MT8839, Norton House, for A.D. 901­1300 and A.D. 1029­1048 _ . 152 5.10 5.11 Periods of Equal or Greater Occupational Atuaetiveness in the Catchment of 5MT8839 (A.D. 1029­1048) Tree­Ring Dates from Site 5MT2433, Aulston Pueblo - viii- 153 _.......... 155 5.12 Comparison of Population Values (POPKM) for 5MT2433, Aulston Pueblo, for A.D. 901­1300 and A.D. 1030­1050 156 Periods of Equal or Greater Occupational Attractiveness in the Catchment of 5MT2433 (A.D. 1030­1050) 157 5.14 Tree­Ring Dates from Site 5MT3834, Mustoe Ruin 159 5.15 Comparison of Population Values (POPKM) for 5MT3834, Mustoe Ruin. for A.D. 901­1300 and A.D. 1173­1231 160 Periods of Greatest Occupational Attractiveness in the Catchment of 5MT3834 (A.D. 1173­1231) 160 Tree­Ring Dates from Site 5MT6970, Wallace Ruin (5.l7a) and from Site 5MT4126, Ida Jean Ruin (5.17b) 163 Comparison of Population Values (POPKM) for 5MT6970, Wallace Ruin, for A.D. 901­1300 and A. D. 1045­1125 164 Periods of Greatest Occupational Attractiveness in the Catclurient of 5MT6970(A.D.1045­1125) 165 5.20 Tree­Ring Dates from Site 5MT1566, Lowry Ruin 168 5.21 Comparison of Populations Values (pOPKM) for 5MT1566, Lowry Ruin, for A.D. 901­1300 and A.D. 1086­1120 171 Periods of Greatest Occupation Attractiveness in the Catchment of 5MT1566 (A.D. 1086­1120) 172 5.13 5.16 5.17 5.18 5.19 5.22 5.23 Tree­Ring Dates from Site 5MT2149, Escalante Ruin 5.24 Comparison of Population Values (pOPKM) for 5MT2149, Escalante Ruin,for A.D. 901­1300 and A.D. 1124­1138 5.25 Periods of Greatest Occupational Attractiveness in the Catchment of 5MT2149 (A.D. 1124­1138) _.......... 172 173 _.......... 175 5.26 Tree­Ring Dates from Site 5MT765, Sand Canyon Pueblo 175 5.27 Suggested Construction Dates (A.D.) for Tree­Ring Dated Structures at 5M1'765 182 Comparison of Population Values (pOPKM) for 5MT765, Sand Canyon Pueblo, for A.D. 901­1300 and A.D. 1252­1274 _.......... 183 Periods of Greatest Occupational Attractiveness in the Catchment of 5MT765 (A.D. 1252­1274) 185 Summary of Population Values (pOPKM) Using POP2YR Estimates for Eight Tree­Ring Dated Sites, A.D. 901­1300 186 5.28 5.29 5.30 - ix- Page Blank in Original LIST OF ILLUSTRATIONS 1.1 Location of the study area _.......... 2 1.2 The conceptual model for reconstructing prehistoric agricultural productivity """"'" 4 2.1 Schematic diagram illustrating the arrangement of DEMs used in the study 38 2.2 Location of the five weather stations _.......... 44 2.3 Location of the seven tree­ring chronology stations _.......... 48 3.1 Distribution of 113 soil map units in the study area by amount of water held in the upper and lower soil layers _.......... 57 Plot of the arumal mean PDSI values by PDSI class for a single group of 34 long­term reconstructions (REC34), A.D. 1890­1970 _.......... 91 Plot of the annual standard deviations associated with the annual mean PDSI value for a single group of 34 long­term reconstructions (REC34), A.D. 1890­1970 _.......... 94 3.2 3.3 3.4 Schematic diagram illustrating the relationship among reconstructed PDSI values, their interpretive categories, and crop yield production categories 110 4.1 4.2 4.3 4.4 Diagram illustrating procedural steps used to process the final GIS­ integrated analysis .. 114 Normal distribution of modern bean yield (lbs/ac) from 46 agricultural soils in Montezuma and Dolores Counties, Colorado .. 120 EPPL7 dotplot image of prehistoric agricultural productivity in the study area as reconstructed for July I, A.D. 902 . 122 Schematic diagram illustrating the relationships among three concepts of carrying capacity .. 127 _ 5.1 Maize production in the study area, A.D. 901­1300 _ . 133 5.2 Supportable population density in the study area, A.D. 901­1300 _ .. 134 5.3 Maize production in the Sand Canyon Survey Locality, A.D. 901­1300 5.4 Supportable population density in the Sand Canyon Survey Locality, A.D. 901­1300 - xi- 138 _ . 139 5.5 5.6 5.7 5.8 5.9 5.10 5.11 5.12 5.13 5.14 5.15 5.16 Maize production in the Mockingbird Mesa Survey Locality, A.D. 9011300 141 Supportable population density in the Mockingbird Mesa Survey Locality, A.D. 90]-1300 ] 42 Comparison of estimates of carrying capacity for the Sand Canyon Survey Locality generated by the model with estimates derived from archaeological survey in the locality, A.D. 901-1300 145 Comparison of estimates of carrying capacity for the Mockingbird Mesa Survey Locality generated by the model with estimates derived from archaeological survey in the locality, A.D. 901-1300 146 Population density supportable within the 1.6-km-radius catchment of 5MT8371, DCA Site, A.D. 901-1300 151 Population density supportable within the 1.6-km-radius catchment of 5MT8839, Norton House, A.D. 901-1300 154 Population density supportable within the 1.6-km-radius catchment of 5Mr2433, Aulston Pueblo, A.D. 901-1300 158 Population density supportable within the 1.6-km-radius catchment of 5MT3834, Mustoe Ruin, A.D. 901-1300 161 Population density supportable within the 1.6-km-radius catchment of 5MT6970, Wallace Ruin, A.D. 901-1300 _.......... 166 Population density supportable within the l.6-km-radius catchment of 5MT1566, Lowry Ruin, A.D. 901-1300 170 Population density supportable within the 1.6-km-radius catchment of 5MT2149, Escalante Ruin, A.D. 901-1300 174 Population density supportable within the 1.6-krn-radius catchment of 5MTI65, Sand Canyon Pueblo, A.D. 901-1300 184 - xii- PREFACE At least since the times of Duff (1904), archaeologists have been making inferences about how culture change in the prehistoric Southwest has been affected by climate change. Over the years the paleoenvironmental data on which these inferences are based have become increasingly refined, with the analysis of pollen and macrobotanical materials from sediment traps, archaeological sites, and packrat middens; the development of alluvial chronologies; and with increasingly sophisticated analyses of tree rings (for a general review see Gumerman, ed., 1988). Because of the high temporal precision they afford, and because of the relative ease with which they can be calibrated to various modem climatic parameters, tree rings have played an increasingly important role Southwestern paleoclimatic studies. The Past Climate of Arroyo Hondo New Mexico Reconstructed from Tree Rings (Rose et al. 1981) demonstrated that both spring precipitation and annual precipitation could be estimated for the prehistoric northern Rio Grande region. Rose et al. (1982) took this analysis one step further by using the reconstructed precipitation values to estimate prehistoric values for soil moisture based on the Palmer Drought Severity Index (see also Sebastian 1992). Van West's work brings these approaches two steps forward by reconstructing soil moisture values (and, through them, estimates for maize production) for specific soil types at specific elevations. She therefore succeeds, for the first time, in coupling the high temporal resolution afforded by dendroclimatologically-based crop retrodictions with a high spatial resolution. Such data in tum make possible detailed studies of the relationship between prehistoric settlement systems-and culture change in general-with annual estimates of potential prehistoric production. model enVan West's ｰ｡ｬ･ｯｾｲ ､ｵｃｴｩｯｮ compasses a 1816 kID area of southwestern Colorado to examine the potential effects of past climatic variation on dryland maize agriculture and sustainable population during the late Mesa Verde Anasazi occupation of the area (A.D. 901-13(0). The data she generates are used to evaluate a question of long standing in the northern Southwest: whether climatic variability was severe enough to disrupt agriculture and promote the abandonment of the northern San Juan region toward the end of the thirteenth century. The model incorporates techniques and data sets that have not been used together before. Long regional dendroclimatic records are used to retrodiet 1070 years (A.D. 9011970) of June Palmer Drought Severity Indices (pDSI}-measures of stored soil moisture-for area-specific soils. Reconstructed PDSls are reexpressed in terms of their local equivalent in potential maize yield. The integration, quantification, and visual display of these productivity values are coordinated through geographic information systems technology. The method results in the production of annual maps depicting the variable character of the potential agricultural environment, and annual values for total maize productivity, which can be translated into the population size and density that could be supported by such yields. The results indicate that there was always enough productive land somewhere in the study area to support thousands of persons (e.g., 31,363 persons or 21 persons/km2 minimum in the 1470.36 km 2 study area) over the 400-year period, even in the dry times of the middle twelfth and late thirteenth centuries. It would seem, therefore, that high-frequency climatic variability was never so extreme in this area that decreased agricultural production can be cited as the sole or even the primary cause of its thirteenth century depopulation. Other implications of the data generated in this study are still being worked through in a series of papers and publications by Van West and coworkers (Kohler and Van West 1994; Van West and Kohler 1994; Van West and Lipe 1992; Van West 1994b, 1995). These papers have generated a great deal of interest in the original dissertation, which itself was awarded Honorable Mention by the 1992 Dissertation Prize Committee of the Society for American Archaeology, who deemed it "well written, well organized, convincingly argued, and a methodological tour­de­force ... an excellent dissenation." This excellent dissenation has been thoroughly revised for this presentation, and we are very pleased make it more readily available to allow funher development of the approaches so ably pioneered here. -xiv - ­ Timothy A. Kohler General Editor, Repons of Investigations Pullman, WA ACKNOWLEDGMENTS Many individuals and institutions have assisted me in the completion of this study. My dissertation committee, consisting of Timothy Kohler, chair, and William Lipe, Jeffrey Dean, and John Bodley, was always patient, supportive, and willing to provide advice. They encouraged my explorations, and frequently suggested a method or strategy that helped me cut a path through the often dark. forest of confusion. lowe a special thanks to Jeff Dean of the University of Arizona, who spent many hours with me discussing data, methods, and interpretations, and who, by means of his calm and thoughtful demeanor, gave me the confidence to keep going when I often felt at a loss. Funding for this project came from several sources. Sigma Xi, The Scientific Research Society, awarded me a grant-in-aid of research for archaeological field work. in southwestern Colorado in 1986. Amaterra, a non-profit educational and environmental service organization, defrayed the costs of keeping me and my survey crew the field in 1986 by feeding and sheltering us in their field camp at Sand Canyon Pueblo. The WelUler-Gren Foundation for Anthropological Research awarded me a predoctoral grant (#4799) in 1987 to partially cover the costs of data analysis. The Laboratory of Tree-Ring Research, University of Arizona, paid the computing costs associated with the dendroclimatic-based reconstructions of Palmer Drought Severity Indices and generously provided support persolUlel to assist me with processing of the data, 1987-1989. The Department of Anthropology, Washington State University, awarded me money to cover computing costs while in residence in Pullman, 1989-1990. Very special support has been provided by several other institutions and individuals. Crow Canyon Archaeological Center, Cortez, Colorado has continually provided logistical support for the project since its inception in 1985. David Wherry, Director of the Digital Image Analysis Laboratory at Washington State University, was a constant source of information, logistic support, and friendship throughout this project. Without his dedication, talent, and interest, this project would not have been as successful. Alan Price, currently the Assistant State Soil Scientist, Soil Conservation Service, Denver, Colorado, was my principal liaison in the Cortez field of the SCS, and durinE his tenure there, kindly provided me with much of the soils data and agricultural data used in this study. Martin Rose, dendroecologist, formerly with the Laboratory of TreeRing Research, University of Arizona, Tucson, and currently with the Desert Research Institute, University of Nevada, Reno, was my first guide into the world of dendroclimatological reconstruction. He suggested the strategy for retrodicting highly localized Palmer Drought Severity Indices from mapped soils data and multiple sets of precipitation and temperature records. He also generously supplied the treering data used in this study from his own as yet unpublished dissertation research. Shela McFarlin, Victoria Atkins, and Susan Baugh, archaeologists with the Bureau of Land Management's Anasazi Heritage Center, Dolores, Colorado granted me access to their archaeological site records and maps and provided me a place to work. when I was gathering information for this project Assistance with computing challenges was provided by Steve Samuels, archaeologist and consultant, Center for North American Archaeology, Washington State University; Dan Vakoch, consultant, Computing Service Center, Washington State University; G. Robert Lofgren, Laboratory of Tree-Ring Research, University of Arizona; and Mark Westergaard, Marvin Landis, and Karen Warren, Center for Computing and information Technology, University of Arizona. Karla Harrington, Pima Community College, Tucson, was an able research assistant and helped me greatly in the mapping of soil data from aerial photographs. Ron Beckwith, Arizona State Museum, Tucson, skillfully drafted the figures in this study. Susan Martin, Statistical Research, Inc., created the cover and the chapter­start artwork. Her interest and good cheer made the design process enjoyable. Over the course of the five years it took to gather data, conduct field work, develop methods, process data, and write up this research, many individuals have been there for me. Without their support and often special gifts, this study would have been more difficult and less rewarding. While it is not possible to list every person who helped in this way, I do want to thank. the following friends and colleagues for their special contributions: E. Charles Adams, Michael Adler, Rick Ahlstrom, Christine Amiss, Edward Berger, Jo Berger, Dale Black, W. Harold Bowling, Robin Bowyer, Bruce Bradley, Sharon Bronson, Geneva Burkhart, Barney Burns, Chiu-an Chang, Ruth Chappell, Jim Colleran, Margie Connolly, Elysa Crum, Norma Dever, Bella Eibensteiner, Carlisle Ellis, Mary Etzkorn, Jim Finley, Steve Gill, Raquel Goldsmith, Judy Han, Brian Harvey, Megg Heath, Ann Hitchcock, Edgar Huber, Wei-Chi Jao, Julie Jefferson, Marge Kennedy, Roberta Leicester, June Lipe, Bill Lipe, Ricky Lightfoot, Nancy Menthe, Shorlene Oliver, Troy Oliver, Roy Paul, Bill Robinson, Angela Schwab, Phil Scuderi, Karen Seger, Ruth Slickman, Allan Smith, Catherine Stanley, Ian Thompson, Micky Thurston, Bill Thurston, Phillip Van West, Mark Varien, Nancy Wall, Dave Whelchel, Jim Whitfield, Ric Windmiller, and Angela Zerdavis. I would also like to acknowledge those individuals, other than my comminee members, who have been instrumental in my academic and professional career, and who have contributed directly and indirectly to my appreciation for the art and science of doing archaeology: F. "Cal" Calabrese, Jonathan Gell, Fekri Hassan, Emil Haury, Julian Hayden, Alden Hayes, Alexander Lindsay, Peter Mehringer, and R. Gwinn Vivian. R. G. Matson provided perceptive comments on the dissertation that I have attempted to address here. I would like to express my gratitude to my parents, Selma Van West and the late Joseph Van West. Although my father died before this study was completed, he saw me begin it and grow through the experience. I have always cherished their unconditional love, support, and confidence in my ability to successfully complete this work, and know they are largely responsible for its outcome. My final thanks must go to my husband, Roger Itwin, who has seen me through an entire decade devoted to the pursuit of this degree. His tolerance, understanding, financial and emotional support were unqualified and abundant. Without him and his gifts born of love, this work would not have been completed. -xvi - - Carla Van West Statistical Research, Inc. Tucson, AZ 1 INTRODUCTION The Anasazi were a prehistoric farming people who inhabited the Northern San Juan Region of the American Southwest more or less continualIy from the beginning or middle of the first millenium B.C. through the thirteenth century A.D. In the late 1200s, however, they permanently abandoned not only the Northern San Juan but also much of their former homeland in the Four Comers area. From there they migrated south and east to fonn the groups ancestral to the Hopi-, Zuni-, Keres-, and Tanoan-speakers colIectively known as the modem Pueblo Indians of Arizona and New Mexico. That they left in the late thirteenth century is well known; what is not known with certainty, however, is why they left, although much archaeological research has been conducted to account for the density and distribution of populations and their settlements, local and regional population movements, and local and regional abandonments. It has long been suspected that such a widespread and seemingly sudden depopulation of this large area must have had its roots in an environmental crisis. Whether precipitated by natural climatic fluctuations, such as the Great Drought of A.D. 1276-1299 (Douglass 1929), or by humanly-induced conditions, such as poor land management practices (e.g. Wycoff 1977; Stiger 1979), the immediate cause has assumed to be connected to agricultural production and a failure to grow adequate supplies of food. The problem is an important one because in its solution and understanding comes an appreciation for our own relationship and responsibility to the environment; the tolerances of our economic adaptations to the nonnal low-frequency and high-frequency fluctuations in climate; and the resilience of our socioeconomic systems to accommodate changing environmental conditions, particularly in semiarid areas. The research described in this monograph helps illuminate this important problem in Anasazi prehistory and cultural process. In attempting to reconstruct the agricultural environment available to Anasazi farmers in the heartland of the abandoned area, J have created a quantitative, high-resolution model of prehistoric agricultural productivity and sustainable population for a 1816 krn 2 (701 mi 2) portion of the dry-farming region of southwestern Colorado (Figure 1.1). My immediate goal was to develop methods, techniques, and the database to construct the model. That portion of this study is a methodological contribution to Anasazi archaeology. However, the tree-ring-based reconstructions resulting from the model also represent a database that can be used to address a large number of substantive questions of current importance to Southwestern archaeologists, including those pertinent to the causes of the final thirteenthcentury abandonment. Insofar as I have begun to explore a few of these questions of settlement attractiveness, agricultural productivity and predictability, and potential human carrying capacity, this research also makes a contribution to theory building and testing with wider implications for American archaeology. PREVIEW OF THE MODEL These chapters summarize my attempt to model past climatic variation and its effect on agricultural productivity in a portion of the Colorado Plateau that was home to the Mesa Verde Anasazi during the last 400 years that they inhabited that region, from A.D. 9001300. The model incorporates the highest resolution environmental data available to archaeologists working on the Plateau. It utilizes several techniques and datasets not used . -' ...ｾＲＭＮ .v / o ..,-'" o I / .. ; . . ］ ］ｾＬ ｾＭＢ 1=1 5 Miles 5 Kilometers If'. / N .. ........... ­ . ­ ' . \, I e7 / .....ｾ ＮＺ ｟ .. \ ... Ｂ ｾ UT CO NM Figure 1.1. Location of the study area. 1. DCA Site; 2. Lowry Ruin; 3. Norton House; 4. Aulstol"! Pueblo; 5. Sand Canyon Pueblo; 6. Mustoe Ruin; 7. Wallace Ruin; 8. Escalante Ruin. -2- ... -, ... .. .... together before, and uses state­of­the­art spatial data management and processing technology to encode, store, retrieve, manipulate, and display data. Spatially, model resolution is 4 ha (200 x 200 m); temporally, its resolution is one year. One important product is a set of 400 arumal reconstructions of the potential agricultural niche available to prehistoric dryland maize farmers. These reconstructions depict the condition of the agricultural habitat as it existed each year on July I, just before the arrival of the summer rains and after the dry spring and early summer typical of the Colorado Plateau. For any single year or series of years, the size and productivity of the niche can be described and quantified and the spatial distribution of the farmlands can be displayed cartographically. In addition, the size of the human population that can be supported from that yield can be estimated, given certain assumptions about dry-fanning locations and consumption rates. Changes through time in these annual estimates can be examined on regional, subregional, and local levels. The model incorporates the following logical elements (Figure 1.2). First, it reconstructs soil moisture conditions for each soil type in the region as they existed on July 1 of each year in the reconstructed sequence. This is accomplished by calculating Palmer Drought Severity Indices (PDSI) for the soils of the study area within the period of instrumented weather records (defined by the availability of appropriate monthly temperature and precipitation records). quantifying the relationship between PDSI values and historic tree-ring width values. and retrodicting prehistoric PDSI values with the preinstrumented and prehistoric portion of the tree-ring record. Second. it translates the soil moisture conditions as expressed by PDSI values into predicted crop yields. first as production in beans and finally as production in maize. Third. it calculates total productivity of the agricultural habitat on a year-by-year basis and provides an estimate of maximum annual food supply potentially available from maize agriculture. Fourth, it estimates the maximum arumal population size and density that could have been supported by the annual maize supply. given certain assumptions about agricultural demand (i.e.• annual consumption rates and potential levels of storage). Finally. it examines the entire 400- year sequence of yield and population values and identifies sustainable population levels over time. expressed as maximum, critical. and optimal carrying capacity values. While efforts to reconstruct climatic variation and its influence on agricultural productivity. population, and settlement in the Mesa Verde area have been attempted before (e.g.. Bums 1983; Cordell 1975; Schlanger 1986b). none has had the opportunity to use both the high-quality environmental data and the spatial data management systems that are now available. Without these newly derived environmental data and this recently developed computing technology, this research could not have been done-the data are too many. the calculations too complex. and the accurate evaluations of options too numerous. Thus. computer technology, especially geographic information systems (GIS). plays a crucial role in this research. Without GIS it would have been impossible to integrate the reconstructed drought indices and the calibrated series of agricultural yields from the actual soils with the spatial locations of these soils. By capturing, co-registering. and evaluating all data layers. as weU as by creating new data layers from reclassifications and transfonnations of the original data, GIS technology made the fast. accurate. and consistent (and therefore replicable) assignments necessary to create the model. display the results. and assess patterning across space and through time. Two raster (grid cell-based) programs were used to capture. store. manipulate. and display the spatial data of the model: mainframe software called VICAR/IBIS and microcomputer software called EPPL7. They made possible the investigation of the study area at a relatively high level of spatial resolution (200 x 200 m or 4 ha cellular units) in an area 40 x 45.4 kIn (or 1816 k.m 2). Input to the GIS was in the form of previously generated computer values entered through floppy disks, digital elevation data purchased from the U.S. Geological Survey on magnetic tape. and newly digitized spatial data and keyboard-entered tabular data. Output from the GIS consisted of color graphic displays on video monitors, black-and-white image output on dot matrix printers. and tabular data that were transferred to mainframe programs for further analysis. -3- Data Collection Preliminary Analyses and Paleoenvironmental ｄ｡ｴ ｾｐｄｓｉ (Tree­Ring Chronologies) Elevational Data (OEMs) ｾ Soils Data (Types, Locations, Depth and ｗ｡ｴ･ｲＭｨｯｬ､ｩｮｧｾ Capacity) I ｾ I Soil Productivity Data ｾ (Agricultural and Natural Plant Productivity) Analysis ｾｌｯｮｧＭｔ･ｮ Preliminary Products PDSI ｾ Reconstructions Elevational Data .... PDSI Plane (s) Reconstruction Data Plane Final Analyses and Final Modeling Products t .. Annual POST Reconstructions .. Annual Productivity Reconstructions .. Annual ｾ Population Estimates ｾ Soils Data Plane(s) Crop Yield and Natural Plant Productivity Studies t Agricultural Productivity and PDSI Calibration Studies Ethnol:raphic Data (Agricultural Practices, Consumption and Per Capita Demand for Yield) Figure 1.2. The conceptual model for reconstructing prehistoric agricultural productivity. I Long­Tenn Population Estimates Equally important to the precision and success of the model is the availability of highquality environmental data for the study area. Most of these have only been available since 1982 and some were created as recently as 1988. These include a contiguous set of 12 7.5-minute Digital Elevation Models (DEMs) available from the National Cartographic Information Center, U.S.G.S., and completed between 1985 and 1987; new maps and classifications of soils in the study area developed by the Cortez Soil Conservation Service and usable since 1988; estimates of potential natural plant productivity under varying climatic conditions for each described soil, as well as modern crop-yield estimates for selected soil types (prepared by the Cortez Soil Conservation Service and at hand in 1988); recently published agricultural yield data that summarize formerly dispersed and unstandardized data (Burns 1983); and unpublished tree-ring data from Martin Rose, related to work published in Rose et al. (1982). A final source of data for this study is the archaeological record. While "real-world" observations were not essential to model implementation, they are extremely useful in preliminary tests of the model's utility. Consequently, two fonns of archaeological data were needed to investigate the usefulness of the model for addressing questions important to settlement analysis: population estimates from high-quality block survey data and site locational information from tree-ring dated sites. High-quality survey data were needed from at least two locales in the region that differed in elevation and exhibited significant settlement during the period of interest. These could be compared against the modeled estimates to explore questions of carrying capacity. Minimum requirements for tree-ring dated sites were at least two cutting dates, at least one cluster of dates, and occupation during a limited period within the 4OO-year range. These data were used to explore questions relating to site catchment characteristics as well as optimal locations and times for site selection. Some of these archaeological data existed prior to the initiation of this research, others did not. Extensive field survey undertaken by the author and others in 1986 and 1987 near Sand Canyon Pueblo and Goodman Point Ruin under the auspices of the Crow Canyon Archaeological Center, Cortez, Colorado, provided the requisite settlement data needed to conduct a preliminary test of the model. Treering dates were obtained from the Laboratory of Tree-Ring Research, University of Arizona, . for sites where complete locational and other information was available from reports, other publications, and site files. ORGANIZATION OF THE STUDY The remainder of this chapter is devoted to placing the study area and research problem in a spatial, temporal, and research context. This is accomplished by a brief review of Puebloan culture history in the Mesa Verde Region; a brief history of archaeological research in the study area; and a brief review of methods used to reconstruct aspects of paleoenvironment on the Colorado Plateau. Chapters 2, 3, and 4 describe the methods used to construct the model and focus on data acquisition, preliminary data manipulation, and final analyses. The methods and data used or created to build the model are described in chapter 2. Preliminary analyses and preliminary products are described in chapter 3. Final analyses are described in chapter 4, and a synopsis of the model-building process is presented. The final products-the long-term reconstructions of annual maize yield and sustainable population at three levels of storage/demand-are provided in the appendices. Chapter 5 describes the results of the model-building effort and provides several test applications of the database created by the model. Chapter 6 summarizes the results of the research and assesses the effectiveness of the model. It includes a discussion of the strengths and weaknesses of the data and the techniques used, and suggests future modifications to test and enhance the reconstructions of agricultural productivity. SPATIAL CONTEXT The study area is located in southwestern Colorado, about 20 miles north of the Colorado-New Mexico border and immediately east of the Utah-Colorado border (Figure 1.1). Notable landmarks within its perimeter are Sleeping Ute Mountain on the south; the -5- northwest escarpment of Mesa Verde on the southeast; the northward bend of the Dolores River on the northeast; the major east­to­west flowing drainage of McElmo Creek in the lower portion; the southwest­trending northern tributaries, canyons, and mesas of McElmo Creek in the lower half; and the gently rolling mesa tops and broad highlands of the Great Sage Plain in the upper half. The horizons beyond but visible from the study area are almost entirely defined by mountains or distinctive volcanic features: the San Juan Range of the Colorado Rocky Mountains (including the La Platas) to the east and northeast, the Mesa Verde highlands to the southeast, the Shiprock in New Mexico to the south-southeast, Ute Mountain immediately to the south, the Carrizos of Arizona to the south-southwest, the Abajos of Utah to the west, and the La Sals of Utah to the north-northwest. The highest elevations in the study area are generally in the northeast. The lowest elevation in the study area is west of Ute Mountain at the level of McElmo Creek at 1,500 m (4,922 ft). The highest elevation is at the peak of Ute Mountain at 3,011 m (9,879 ft). Physiographically, the study area is located in the Navajo section of the Colorado Plateau Province (Hunt 1956). Most of the area is comprised of sedimentary layers of Cretaceous (65-136 million years), Jurassic (136-190 million years), and Jurassicffriassic Period (190-225 million years) sandstones and shales. With the exception of Sleeping Ute Mountain (a volcanic laccolith) and the cuesta-like Mesa Verde, the uppermost formations in the study area are the erosion-resistant Dakota Sandstone and small remnant outcrops of overlying and highly erodable Mancos Shale. Where canyons have eroded the highland landscape, layers of Cretaceous-age Burro Canyon deposits, Jurassic-age Morrison and Summerville Formations, and the Jurassicffriassic-age Navajo Sandstone have been exposed (Ekren and Houser 1965; Haynes et al. 1972). Reddish-hued sand- and silt-sized particles derived from sedimentary rocks and alluvial deposits along the lower San Juan River have been transported by wind from sources in Arizona and Utah to the study area (Price et al. 1988). These eolian sediments were deposited across the entire study area but have been eroded away at canyon rims and slopes (places that are usually characterized by exposed bedrock and colluvial talus). Where present, they have mixed with weathered residual sediments and have formed rich loamy soils (Arrhenius and Bonatti 1965). These reddish soils mantle much of the highland portions of the study area and rest on the Dakota Sandstone substrate. These soils are referred to as eolian, whereas others that were formed in place by the weathering of parent material are referred to as residual soils, and those that were deposited by water are referred to as alluvial soils. Almost 100 soils (or more precisely "soil map units") have been identified to date within the boundaries of the study area. Of these, 44 soil map units, mostly eolian-derived soils on mesas and highland portions of the study area, are currently used for dry-farming of pinto beans, alfalfa hay, and winter wheat. Additional soil map units, principally alluvial in origin, are located in McElmo Creek Canyon, where fruit trees and small grains are grown on floodplain and lower terrace locations. While mineral, gas, and oil exploitation are components of the modem economy in this largely rural area, agriculture is the mainstay. The climate of the study area is typical of that for most places in the northern Southwest west of the Rocky Mountains. It has been described in Koppen's classification as a cold middle latitude, semi-arid climate (Bsk) in which potential atmospheric evaporation exceeds the usual amounts of precipitation available (Trewartha 1954:267). The area has a biseasonal moisture pattern, with the majority falling as snow more evenly over the area during the late winter (January through early March), and a lesser amount falling as rain usually in a patchy pattern as swnmer thunderstorms (July through early September). The Dolores River is the only naturally occurring source of permanent surface water. The McElmo Creek flows more-or-Iess permanently today, but was seasonally ephemeral until the 1890s when it began to collect irrigation water that was artificially routed from the Dolores River (Freeman 1958). Ground water is available, however, in the form of seeps and springs that occur at the contacts between permeable water-bearing sandstones above and impervious shale layers below. These contacts are most apparent in canyonhead settings or within -6- canyon walls, and have a wide distribution across much of the study area. The majority of lands in the study area support an Upper Sonoran Woodland Biotic Pinon Community (Lowe 1964:11, ＳｾＶＩＮ and juniper woodlands occur on the shallow or rockier soils of the mesas and highlands, and sagebrush and other deciduous shrubs occur on the deeper and more level areas at moderate elevations. Woodlands of oaks, Ponderosa pine, and occasionally Douglas­fir occur at higher elevations. Great Basin desert scrub occurs at lower and drier elevations. Riparian vegetation is fOWld in mesic areas, particularly in canyon bottoms near water sources. More detailed descriptions of the flora and faWla in or near the study area may be found in Brandagee (1976), Erdman et al. (1969), Nickens (1977), Petersen et al. (1985), Petersen and Orcutt (1987), and Winter (1976). of southwestern Colorado. The 400-year period is coterminous with the Pueblo II and Pueblo III stages of Anasazi culture (Kidder 1927). and it incorporates notable "events" in Mesa Verde Region culture history that have been attributed to climatic and environmental causes. These include the local abandonment of the Dolores River Valley in the very late 800s or early 900s and the near absence of settlement in this valley in the tenth and early eleventh centuries; the wide geographic dispersal of settlement in the region from the middle eleventh to the early twelfth centuries; the demise of the Chaco presence in the San Juan Basin ca. A.D. 1125-1150; and the final abandonment of Mesa Verde Region at the end of the thirteenth century. These four centuries, then. represent an excellent set of years in which to "observe" the effects of climatic variation on local soils and on their potential productivity. TEMPORAL CONTEXT The Anasazi presence immediately adjacent to the study area can be documented as early as the A.D. 470s (Breternitz 1986), but it is generally accepted that the main arrival and settlement of Formative-level Anasazi Basketmaker III populations (probably from the east) did not occur Wltil the late 500s and early 6005. Until about A.D. 1300, Anasazi populations settled, relocated, and resettled on lands within the study area-their numbers sometimes waxing, sometimes waning, but seemingly always present. Prior to their presence, seasonally-mobile hWlting and gathering Archaic populations (ca. 5500/6000 B.C.- A.D. 1), and before them, migratory big-game hunting and plant gathering PaleoIndian populations (ca. 10,000-5500/6000 B.C.) exploited the resources of the study area. After the agricultural Anasazi left, seasonally mobile foraging populations (first the Numic-speaking Utes and later the Athabascan-speaking Navajos) returned to the study area perhaps as early as the 1500s. The reconstructions of agricultural productivity and sustainable population created by this study are confined to the four centuries beginning in A.D. 901. The lower limit was determined by the availability of tree-ring data necessary to the reconstruction process. The upper limit of A.D. 1300 was selected to correspond with the end of the Anasazi occupation Agricultural populations did not resettle the study area Wltil the late 1800s, with the advent of the Euro-American farmers and ranchers who initially supplied the mines and later established an independent agricultural economy when mining failed to create a dependable demand for crops and meat. Because mobile populations of pre- and post-Anasazi Archaeologically, the study area is located in the Northern San Juan (Le., Mesa Verde) Region of the Anasazi Culture (Bullard 1962). The majority of the study area, with the exception of the northwest portion that incorporates highlands near the Dolores River and the river drainage itself, is contained within a sub-regional setting called the Yellowjacket District by Gillespie (1976), the McElmo Drainage Unit by Eddy et al. (1984), the Core Area of the Northern San Juan by Fuller (1988), or the McElmo-Yellow Jacket District by Varien et al. (1990). In the typological schemes of these authors, the northwest portion would also be assigned to the Yellowjacket District by Gillespie (1976) but would be distinguished as being a portion of the Dolores Drainage Unit by Eddy et al. (1984), the Northern Periphery by Fuller (1988), or the Dolores Valley District by Varien et al. (1990). -7- periods built impennanent shelters and often manufactured and used perishable items, much less of their material culture is preserved and recovered archaeologically. Moreover, until recently, interest in the PaleoIndian. Archaic, Ute, and Navajo culture and land use of the study area has taken a back seat to interest in the better­preserved and often spectacular Anasazi remains. By far, the vast majority of archaeological remains reported for the study area (and for the Mesa Verde Region in general) derives from the sedentary, agricultural Anasazi occupation. Consequently, the brief review of culture history that follows is restricted to the Anasazi Culture of the Northern San Juan (the Mesa Verde) Region, and is provided to place the 400 years of this study in a temporal context. The Anasazi Developmental Sequence in the Northern San Juan Region Basketmaker II This first period of the Pecos Classification system (Kidder 1927) dates from about 500 B.C. to A.D. 450/500 (Matson 1991). Sites attributed to Basketmaker II have only recently been reported for the study area (e.g., Winter 1977). In addition, two sites immediately adjacent to the study area-a Mancos Canyon site and Cougar Springs Cave near the Dolores River-are thought to be affiliated with this period (Hallisy 1974, redocumented by Bretemitz 1986; Gross and Howes 1986). Basketmaker II sites, however, are well documented in portions of the region to the east (e.g., Eddy 1966; Morris and Burgh 1954) and to the west (e.g., Lipe and Matson 1971; Matson and Lipe 1978; Matson et al. 1988). During this time, populations built shallow pit houses that they lived in for at least a portion of the year, built small storage facilities, and cultivated maize and squash. In the early portion of the period, crops may have been grown primarily in floodplain environments, but not by dry-farming teclmiques. In the later portion of the period, around A.D. 100, dry-fanning techniques were also used to grow crops. Habitation sites were located in the open and within rockshelters. Campsites were used as bases for hunting and gathering, which continued to be important sources of food. Populations were small and dispersed, and habitation sites were small, consisting of one or only a few structures. Distinctive material items and traits associated with this period include sophisticated coiled and plaited basketry (but no pottery), atlatls, large corner- or side-notched projectile points, basin metates and one-handed manos, rabbit-fur blankets. and characteristic rock art motifs depicting large anthropomorphic figures associated with shamanistic experiences (Cole 1989; Schaafsma 1980). Dogs were the only domesticated animals. Basketmaker III Just as Basketmaker II culture probably developed out of the Archaic, Basketmaker III cultures continued traditions established in Basketmaker II. This developmental stage dates roughly from A.D. 500 to 750. It is well documented in the study area (e.g., Bretemitz et al. 1986; Fetterman and Honeycutt 1987; Kuckelman and Morris 1988; Lange et al. 1986; Martin 1939; Rolm 1974, 1975; Schlanger 1985; Wheat 1955; Winter 1977) and throughout the Northern San Juan Region. Basketmaker III pit houses were larger and deeper than those in Basketmaker II times and typically included wing-walls, clay-lined hearths, low banquettes, and interior storage cists. Sipapus, interpreted by archaeologists as symbolic places of emergence from the underworld, appeared for the first time. Accompanying pit houses were exterior storage facilities and surface features situated to the north and midden areas located to the south-an internal site configuration and orientation that persisted through the Pueblo III period. Some Basketmaker III settlements were surrounded by wooden stockades, although the function of such features is not clear. Sites occurred in both rockshelters and in the open. Habitation sites were usually small and consisted of one or a few houses, although larger villages in open settings do occur on occasion. The greater investment in architectural facilities, greater attention to storage, more efficient grinding equipment, and addition of beans to the diet in this period, are regarded as signifying a greater dependence on cultivated products and year-round habitation-a trend that continues through Pueblo III times. As with Basketmaker II, campsites and rock art sites are -8- known to be associated with Baskennaker III culture. Although sites were smal1 and moderately dispersed, there is some evidence for the aggregation of populations in particular locations. Great kivas have been identified as being associated with Baskennaker III sites and represent the earliest evidence for the use of public architecture to integrate dispersed populations (unless Baskennaker II rock art panels served a similar purpose). Distinctive material items include the first evidence of fired ceramic vessels in the form of smoothed, coilmade gray jars and bowls; the addition of the bow, arrow, and small comer-notched projectile point (and eventual replacement of the atlatl, dart/spear, and large comer-notched projectile point); and the introduction of trough metates and two-handed manos. Turkeys are believed to have been domesticated by Baskennaker III times. Pueblo I The Pueblo I period dates from about A.D. 750-900. It is wel1 represented in the study area (e.g., Breternitz et al. 1986; Fetterman and Honeycutt 1987; Martin 1939; Varien and Lightfoot 1989; Winter 1977) and in the Northern San Juan (e.g., Brew 1946; Debloois and Green 1978; Haase 1983; Hayes 1964; Hayes and Lancaster 1975; Morris 1939; Roberts 1930; Rohn 1977). Pueblo I populations tended to be clustered in distinct locations and exhibited definite tendencies toward living in larger, aggregated settlements than did their ancestors. A typical Pueblo I habitation site consisted of one or more deep, squarish pit houses located south of an arc or row of contiguous surface rooms and north of a trash midden. Subterranean pit houses continued to have a variety of floor features as part-time domestic structures but had roof entries and ventilator systems rather than antechambers; some ("protokivas") exhibited greater evidence for ceremonial use in the form of complex sipapus and lateral floor vaults (Wilshusen 1989). The surface rooms were usually arranged in two rows and built as "modular units" with two smaller, generally better-made, storage rooms to the north and a large living room to the south with a south-facing door. Both jacal and simple masonry were used to construct walls. These modular units are considered to be the domain of a single household unit, and when joined, are considered to represent the co-residence of several household groups. Pit structures are believed to have been shared by more than one household group. Larger-than-average pit houses and great kivas are documented for some Pueblo I settlements (e.g., Lightfoot 1988) and are thought to symbolize attempts to unify and provide cohesion for local populations, exceeding Baskennaker III attempts. Other than habitations, site types dated to this period include a variety of limited activity locations and field houses. Distinctive material items and traits include the deliberate deformation of the skul1 by means of a hard "pil1ow" positioned at the back of the head of an infant on a cradleboard; neckbanded ceramic jars, widespread trade of redware jars and vessels, small stemmed and tanged projectile points, and the addition of cotton to the list of cultivars in the warmer, better-watered part of the Anasazi world (probably not in the study area). Pueblo II Until recently, the Pueblo II period has been dated to A.D. 900-1100, and subdivided into an early and later period of about 100 years each. At the present time, it is not uncommon to consider that Pueblo n lasts until about A.D. 1150 (Lipe 1989). Populations were larger and sites of this time period were widely distributed across the study area (e.g., Adler 1988; Bradley 1988a; Chandler et al. 1980; Fetterman and Honeycutt 1987; Kuckelman and Morris 1988; Martin 1938; Winter 1976) as well as in the Northern San Juan Region (e.g., Eddy 1966; Hayes 1964; Hayes and Lancaster 1975; Matson et al. 1988; Swannack 1969). In fact, within this period the maximum geographic range of the Anasazi culture was reached. This period is associated with the florescence and the demise of the Chaco sociopolitico-religious expression in the American Southwest. Several sites in and adjacent to the study area (including Wallace, Ida Jean, Lowry, Escalante, Yucca House, and Ansel Hall Ruins) have architectural attributes that are believed to be expressions of Chacoan presence or influence. These Chacoan architectural traits (i.e., Chaco great houses, roads, tower kivas, great kivas, formalized trash middens, or Chaco-style masonry, architectural features, and building proportions) generally occur in association -9- with clusters of somewhat dispersed but apparently contemporary Pueblo II sites believed to be part of a single interacting community. The Chacoan features and associated ideology may have acted as forces of integration for dispersed Pueblo II settlements and communities. In contrast to late Pueblo I settlements, individual Pueblo II settlements were home to fewer households and were dispersed more broadly across the landscape in a wider variety of settings. While the basic architectural arrangement and orientation (the classic "Prudden unit" fonn [Prudden 1903]) persistedroomblock to the north, pit structure in the center, and trash midden to the south­surface structures became more substantial and more commonly built with spalled stone masonry. The pit structures are recognized by fonnal characteristics and are commonly cal1ed kivas rather than pit houses. These fonnal characteristics include a round instead of square plan; a cribbed roof supported on pilasters built above a bench instead of a flat roof supported by four central post supports; inclusion of a banquette or bench; the development of a southern recess in the bench, and a greater use of stone to line the lower chamber. In addition to field houses and limited activity sites of various kinds, water control features (Haase 1985; Rohn 1963) including across­slope terraces, across­drainage check­dams, artificial reservoirs, and shrines were associated with Pueblo II culture. Distinctive items associated with this period include the manufacture of allover corrugated jars, loom­woven textiles, turkey­feather blankets, and small cornernotched points with expanding stems and convex bases. Pueblo III Although the Anasazi sequence continues elsewhere in North American Southwest, this period is the last in the Northern San Juan Region. A dramatic region­wide depopulation at the end of the thirteenth century has led researchers to search for a single powerful cause that could account for this abandonment. Pueblo III is dated from either A.D. 1100 or 1150 to A.D. 1300. Initially it was a time of continuity of tradition in architecture and settlement patterns, with most of the population living in small villages. However, by A.D. 1200, a dramatic coalescence of population into fewer but larger settlements had begun. These sites are often in close proximity to strong springs near mesa rims and canyon heads. In addition, the deliberate management . of water­soil resources continues and apparently increases in this period (Haase 1985; ROM 1963, 1972). Sites in the study area that date at least in part to this period include Yel10w Jacket (Lange et al. 1986), Goodman Point (Fewkes 1919), Sand Canyon Pueblo (Bradley 1986, 1987, 1988a; Kleidon and Bradley 1989), Mud Springs (Fewkes 1919), and the Lancaster Ruin (Martin 1929). In other places in the Northern San Juan, the period is represented by the famous cliff dwellings of Mesa Verde National Park, such as Cliff Palace, Balcony House, Spruce Tree, Long House, and Mug House (Cattanach 1980; Rohn 1971, 1977); Lion and Hoy House in Mancos Canyon (Nickens 1981); and sites like the Nancy Patterson Village in southeastern Utah (Thompson et al. 1986). Pueblo III settlements generally retained their modular configuration of roomblock, . kiva, and trash midden with modules often linked into agglomerated structures with structurally incorporated kivas. By the beginning of the thirteenth century, informal plazas, enclosing walls, and towers were common features of large habitation sites. Kivas were regularly lined with carefully shaped and pecked masonry. New architectural features particular to this time period include masonry cliff dwellings; rare D.shaped, bi­walled, or triwalled structures; and relatively common freestanding towers at canyonheads, such as those of Hovenweep National Monwnent. The function of the D­shaped and bi­ or tri­walled structures is unknown (but see Bradley 1992). Multiple functions have been postulated for free­standing towers and include defense, intersite communication, community storage, religious or astronomical significance (Riley 1950; Schulman 1950; Williamson 1984; Winter 1984). Distinctive items associated with this period include slab metates, flat twohanded manos; tchamahias; new ceramic vessel forms (e.g., canteens, mugs, fluted­handled dippers, kiva jars); elaborate ceramic designs; the predominant use of carbon paints rather 10- than mineral paints to decorate vessels; and small, side­notched, triangular projectile points. Previous Archaeological Research in the Study Alli1 The Northern San Juan Region is one of the most intensely studied archaeological areas in the North American Southwest. Comprehensive reviews of the history of investigations in this region have been published by Brew (1946), Herold (1961), Nickens and Hull (1982), and Eddy et aI. (1984). These works contain valuable chronological summaries as well as commentary on the significance of selected research efforts for advances in archaeological method, theory, and resource management. The brief review that follows identifies the major investigations within the study area. Its intent is to emphasize that the archaeological remains of the study area are abundant, historically well-known, and continue to be investigated as a source of information on prehistoric Pueblo populations and their adaptations. Whether classified as a cultural core area, drainage unit, or archaeological district, the study area has been the location of a remarkable amount of archaeological work, and a large number of well-preserved architectural ruins are contained within it Densities of over 40 sites/km2 are common for those places in the study area that have been systematically surveyed in recent years--an impressive fact considering the intensive land modification that has occurred in the last 100 or so years result of agriculture. energy exploration. and other forms of land development. Early Observations and Investigations The earliest observations recorded on the archaeological remains of the area are those of the Franciscan fathers, Escalante and Dominguez. who passed through the study area in 1776 on their WlSuccessful search for a route to Monterey. California from Santa Fe, New Mexico (Bolton 1972). However, almost 100 years passed before goverrunent officials and the public became aware of the richness of this archaeological area through the explorations and publications of the U.S. Corps of Topo- -11 graphic Engineers (Macomb 1876; Newberry 1876) and the Hayden Survey (Holmes 1878; Jackson 1878). The earliest surveys and excavations in the study area are known through the publications of Prudden (1903, 1914, 1918), Morley (1908), Morley and Kidder (1917), Fewkes (1919), and Roberts (1925). This is also the time when the Wetherills were actively involved in site exploration. Although not systematic or intensive by the standards of today, these early investigations helped establish the basic chronology for the Mesa Verde area (e.g., that the Baskennakers were the cultural predecessors and ancestors to the Pueblo), identified basic architectural configurations (e.g., the modular "Prudden unit" of Pueblo I-III Anasazi habitations). and perceived the enduring settlement and adaptive patterns associated with prehistoric Pueblo culture. Privately Funded and Locality-Based Archaeological Research The first period of rigorous investigations dedicated to exploring a single locality in the study area was initiated by Paul Martin in the Ackmen-Lowry area in 1928. Between 1928 and 1938, Martin and his associates investigated sites representative of seventh century Basketmaker m through thirteenth century Pueblo III occupation in the Mesa Verde Region in the north-central portion of the study area. These included excavations at the well-known ruins of Herren Farm and Charnel House (Martin 1929). Beartooth and Little Dog (Martin 1930). Lowry (Martin 1936) and a number of Basketmaker III through Pueblo n sites in northern portion of the study area (Martin 1938. 1939). Major studies of architecture (Roys 1936) and ceramics (Rinaldo 1950) resulting from the Ackmen-Lowry work. have become part of the classic literature of southwestern archaeology. South and east of Lowry. in the central portion of the study area, a long-term research project including both survey and excavation in the YeUowjacket area was initiated in 1954 by Joe Ben Wheat as annual field schools of the University of Colorado. Boulder (CU). These annual expeditions continue today and have focused on excavation of sites in the "suburbs" of the YelIowjacket Ruin community, the largest Anasazi ruin in the study area (Brown 1975; Lange et al. 1986; Swedlund 1969; Wheat 1955, 1980). Prior to these CU field schools, Hurst and Lotrich from Western State College of Colorado did some limited work in 1931 on the Yellowjacket Ruin itself (Hurst and Lotrich 1932, 1933). Elsewhere in the locality, ROM and students from Wichita State University conducted several seasons of work in the late 1960s and early 1970s. Many sites were recorded on survey and a number of sites were tested, including the stockaded Basketmaker III sites, Payne and Gilliland (ROM 1974, 1975). Locality-based research on Mockingbird Mesa, in the west-central portion of the study area, was undertaken by Ives and students from Fort Lewis College, Durango, from 1970 to 1973. Seventeen sites ranging in age from Basketmaker III to Pueblo III were excavated (Ives 1971a, 1971b, 1972, 1973). The Bureau of Land Management undertook an intensive survey of Mockingbird Mesa between 1981 and 1984 and located 684 sites on 3,976 acres of land (Fettennan and Honeycutt 1987). This survey provided significant data on settlement patterns for regional analyses. In the southwestern portion of the study area, Riley surveyed five units of Hovenweep National Monument (Riley 1948) and published a study of the famous towers from that area (Riley 1950), which was complemented by a spatially extensive study of towers by Schulman (1950) who examined these architectural features across the northern Southwest. Later, problem-oriented research was undertaken in the Hovenweep-Cajon Mesa area by San Jose State University to gain a better understanding of Anasazi farming practices (Hammett and Olsen 1984; Winter 1975, 1976, 1977; Woosley 1977). An offshoot of this project was the systematic and complete survey of Cow Mesa north of the Hovenweep area by Neily (1983). A recent program of research was begun in 1983 in the Sand Canyon locality in the southcentral portion of the study area. In conjunction with major excavations at Sand Canyon Pueblo, the Crow Canyon Archaeological Center, Cortez, is conducting survey and testexcavations in late Pueblo III sites within an approximate 200-km2 locality defined around the Sand Canyon Pueblo and Goodman Point Ruins (Adams 1985; Adler 1988; Bradley 1986, 1987. 1988a; Huber 1989; Huber and Bloomer 1988; Kleidon and Bradley 1989; Lipe 1992; Van West 1986; Van West et al. 1987; Varien 1990). Prior to Crow Canyon's research. only two sites in the Sand Canyon locality had been excavated, one by Colorado Mountain College. Leadville (Bagwell, personal communication, December 1986), and one by Gould (1982). A little to the east of the Sand Canyon Locality. the Crow Canyon Archaeological Center also conducted a major excavation and small locality study around the late Pueblo I Duckfoot Site in the vicinity of Crow and Alkali Canyons (Adams 1984; Lightfoot 1985, 1987, 1992; Lightfoot and Van West 1986; Lightfoot and Varien 1988; Varien and Lightfoot 1989). A final locality-based study located in the far east-central portion of the study area was begun in 1969 by Bradley with the partial excavation of the Wallace Ruin. one of three major sites in the Lakeview Group (the other two being the Ida Jean Ruin and Haney Ruin). This privately-funded research has continued intermittently since 1969 and has produced the strongest evidence to date for the presence or influence of Chaco culture in the heartland of the Mesa Verde Region (Bradley 1974, 1984, 1988b). Federally-Sponsored Archaeological Research and Resource Management A new genre of archaeological research was introduced to the study area in the 1960s with the extensive reconnaissance-level surveys of grazing units managed by the Bureau of Land Management. Hundreds of archaeological sites and areas of high site density were identified during the 1965-1969 surveys (D. Martin 1971) for the purpose of resource inventory and resource management Most archaeological research in the study area since 1960 has been in response to federal and state laws and resource management policies concerned with the identification, protection, and/or interpretation of cultural resources. The Bureau of Land Management (BLM) is the federal land-managing agency with the most extensive holdings in the study area. Subsequent to the reconnaissance surveys of 12 - the 1960s, the BLM was responsible fOr the previously mentioned intensive systematic survey conducted on Mockingbird Mesa; an 8,OOQ-acre quadrat-based survey of lands in the Sacred Mountain Planning Unit of the Montrose district, which produced a statistical sample useful for infening characteristics of the total Plamting Unit (Chandler et al. 1980); and two systematic surveys within lower Sand Canyon (Adler and Metcalf 1991; Gleichman and Gleichman 1992). It has sponsored excavations in the study area at Escalante Ruin (Hallasi 1979), Dominguez Ruin (Reed 1979), Casa de Suenos and another small site near the Anasazi Heritage Center (Douthit 1984), and it has initiated a number of ruins stabilization programs at sites under its care. Stabilization has been undertaken at Lowry Ruin (White and Bretemitz 1976), Escalante and Dominguez Ruins (White and Bretemitz 1979), East Rock Canyon (White 1976), Sand Canyon Pueblo Cliff Dwelling (Metzger and Bretemitz-Goulding 1981), and towers in the McLean Basin and lower Sand Canyon (Tipps 1978). In addition, a stabilization evaluation program was conducted for 49 cliff dwellings in Sand Canyon that provided descriptive data (C. Martin 1976). Finally, the BLM has become the curator of sizable archaeological collections resulting from the Dolores Archaeological Project (described below), the locally renowned Chappell Collection, and other sites on BLM land in the study area. Their collections are housed in the new Anasazi Heritage Center adjacent to the Escalante and Dominguez sites near Dolores. The Center serves as a collections repository and an interpretive museum. In 1972 a major program of data recovery, which included survey, site testing, large-scale excavation, analysis, and model building, was initiated by the Bureau of Reclamation in order to build the McPhee DamlReservoir and its associated sets of canals and laterals on the Dolores River. Early survey in and adjacent to the study area resulted in rePorts that identified the range of cultural resources to be affected by the project (Breternitz and Martin 1973; Kane 1975a, 1975b; Nickens 1977; Toll 1977). In 1978, a major contract was awarded to the University of Colorado to recover a significant sample of data from the project area prior to the filling of the reservoir and construction of related facilities. The project, known as the Dolores Archaeological Project (DAP), became one of the largest archaeological studies ever to have taken place in the United States. The project ran from 1978 through 1985 and generated scores of published reports, major analytic studies, theses, and dissenations. The primary results of the eight-year project are published in a series of Bureau of Reclamation reports too numerous to cite here, with each volume devoted to a particular set of topics or activities. The project is summarized in a final synthetic report (Bretemitz et al. 1986). In addition to the Bureau of Reclamation reports, substantive publications of contracted research undertaken on non-reservoir portiOns of the Dolores Project have been produced. Auxiliary programs of project-related survey, testing, and excavation within the study area were undertaken for the Dove Creek canal (Fuller 1985a, 1985b, 1987; Honeycutt and Fetterman 1985; Ives and Enickson 1983; Ives and Orcutt 1981; Monis 1986b), South canal (Kuckelman and Monis 1988), Towaoc canal (Kuckelman 1986), Fairview laterals (Walkenhorst et al. 1983; Monis 1986a), Cahone laterals (Walkenhorst 1983), and the Ute Mountain Ute inigated lands (Fuller 1988). A great deal of contracted archaeological research in the study area has been conducted in conjunction with energy exploration and resource development. The mining and processing of carbon dioxide gas in southwestern Colorado for use as a tertiary recovery agent in the oil fields of West Texas has resulted in surveys and excavations within proposed facilities and along rights-of-ways for pipelines and powerlines (e.g., CASA 1981; Dykeman 1986; Fetterman and Honeycutt 1982a; Hammack 1983, 1984; Honeycutt and Fetterman 1982; Whitten et al. 1986; Woods Canyon Archaeological Consultants 1985). Similarly, construction of a liquid hydrocarbon pipeline for El Paso-Northwest Gas companies initiated many miles of intensive survey, test excavations, and full-scale excavations along proposed rightsof-way (Fetterman and Honeycutt 1980, 1982b). RESEARCH CONTEXT Paleoenvironmental research is well established in the American Southwest and may be 13- classified as to the sources of its data: geomorphological, palynological, dendroclimatological, macrobotanical, and faunal studies (Cordell 1984:35). Of these, the first three are the major sources of information for reconstruction of Anasazi environments. Each of the five specialties has its strengths and limitations as well as appropriate levels of resolution and confidence (Cordell 1984:35-45; Dean 1988b), but underlying all attempts at paleoenvironmental reconstruction is the principle of uniformitarianism, which assumes that processes responsible for the variation in the natural world that can be observed in the present also operated in the past. Geomorphological studies generally have focused on identifying and dating episodes of deposition and erosion along drainages, identifying the fluctuations in ground-water tables, and finally suggesting the climatic and/or cultural correlates for these cutting and filling events (e.g., Bryan 1925, 1941, 1954; Hack 1942; Hall 1977; Karlstrom 1988; Schoenwetter and Eddy 1964). Palynological studies generally have focused on the recovery and identification of fossil pollen and spores from cultural and noncultural contexts to reconstruct the composition and distribution of former vegetation communities. From these data are inferred the prevailing temperature and precipitation patterns (e.g., Dean et al. 1985; Euler et al. 1979; Hevly 1988; Martin and Byers 1965; Petersen 1981, 1988; Petersen and Mehringer 1976; Schoenwetter 1970; Schoenwetter and Eddy 1964) or human activities (e.g., Betancourt and Van Devender 1981; Wycoff 1977) that likely produced these plant assemblages. Pollen recovered from archaeological sites is less often used for climatic reconstruction than it once was because the biases introduced by human agency are now recognized, but it is frequently used to identify the presence of plants of economic importance and cultural significance (Bohrer 1981). Dendroclimatological studies in the Southwest use knowledge of the particular annual growth characteristics of conifers such as pifton, juniper, Douglas-fir, and ponderosa to infer annual and seasonal patterns in precipitation and temperature that account for that growth. Sophisticated biological and mathematical models have been developed to discriminate the macroclimatic components of tree growth from the microclimatic and genetic components. Tree-ring chronologies of 1,500 to 2,000 years or more have been constructed by the merging of archaeological tree-ring chronologies with living-tree chronologies to create an absolutely-dated record of local and regional climatic conditions with a temporal resolution of one year (Dean and Robinson 1978). The modem end of these chronologies can be calibrated with various climatic variables to estimate (or "retrodict") values for these same variables for the full length of the chronologies. In this way, long records of annual or seasonal values of precipitation and temperature (Hogan 1987; Rose et al. 1981, 1982; Sebastian 1988), stream flow (Graybill 1989; Larson and Michaelsen 1990; Stockton 1975; Stockton and Boggess 1980), annospheric pattern anomalies (Blasing 1975; Fritts et al. 1971), crop yields (Bums 1983), and droughts or meteorological conditions as measured by Palmer Drought Severity Indices (Meko et al. 1980; Mitchell et al. 1979; Rose et al. 1982) have been created. Macrobotanical studies generally focus on the identification of seeds and plant parts recovered from archaeological contexts to reconstruct economic practices of former populations rather than paleoclimates or paleoenvironments (e.g., Adams 1980; Bohrer 1981; Bohrer and Adams 1977). However, packrat midden analysis-a type of macrobotanical study-is increasingly being used to reconstruct the distribution of vegetation communities and their associated climatic regimes (Betancourt and Davis 1984; Van Devender and Spaulding 1979). Like macrobotanical studies, faunal studies generally focus on the identification of animal remains recovered from archaeological contexts to infer past subsistence behavior. However, knowledge of the specific ecological adaptations of certain species and observation of changes in the frequency of one indicator species to another have been used to infer local vegetation types and changing environmental conditions that are likely to have been present in the past (e.g., Flint and Neusius 1987; Harris 1970; Mackey and Holbrook 1978). 14 - While it is known that major patterns of climate and vegetation patterns have been stable in the Southwest for some 8,000 years (Van Devender and Spaulding 1979), fluctuations in temperature and precipitation that result in "minor" shifts of water supply and the distribution of plants and animals have been identified. Some of the changes in environmental conditions are caused by natural processes that occur slowly, some even regularly, over many centuries. Generally they occur in periodicities of greater than 25 years. Changes in alluvial water tables and episodes of arroyo cutting and filling are often-cited examples. Other changes in environmental condition are caused by natural processes that occur quickly, on the scale of years, seasons, and even shorter periods of time. Changes in the seasonal patterns of precipitation and temperature, and the rapid changes in the size and distribution of certain plant and animal populations are examples. Dean (1988a) has termed the former Low Frequency Processes (LFP) and the latter High Frequency Processes (HFP) Paleoenvironmental reconstructions drawn from geomorphology and some palynological studies generally reconstruct LFP changes, whereas paleoenvironmental reconstructions drawn from dendroclimatology, some palynological studies, macrobotanical studies, and faunal studies generally reconstruct changes conditioned by HFP. Both are important to document because they have different behavioral and adaptational consequences for human populations. Dean contends that humans generally do not perceive LFP because they often occur so slowly (although there are predictable exceptions to this; for a complete discussion see Dean 1988a), and that populations consider these conditions to be the prevailing and stable condition of their environment. When LFP changes do occur, they are severe and require major behavioral adaptations on the part of human populations to survive. Alternatively, he proposes that the environmental changes caused by HFP are quite perceptible and are considered to be within the range of normal variation experienced by the population. Consequently, it is to these variations in climate and resource availability that human populations regularly respond with - behavioral mechanisms such as water-control technology, food-storage systems, and exchange networks. For additional discussion of this topic, the reader is referred to Dean's full treatment of tttis model of "Anasazi behavioral adaptation" (Dean 1988a). The significance of ttis distinction between LFP and HFP here, however, is that both types of processes and their effects on the environment and the subsistence and settlement patterns of prettistoric populations must be reconstructed in order to fully comprehend cultural stability and change. Thus, the goal of the archaeological efforts to reconstruct past climate and environments in the Southwest should be to document as fully as possible these fluctuations and their effects on human populations. The research described in this study is based on dendroclimatological data. Its goal is to reconstruct local high-frequency variation in regional climate, as it has been modeled by tree-ring data, and to observe its effect on the distribution of arable land and potential human carrying capacity. The work has been inspired by four studies in particular: the conceptual model of Anasazi adaptive behavior proposed by Dean et at. (1985; see also Euler et al. 1979; Gumerman 1988); the methods proposed in Rose et al. (1981) for high-resolution reconstructions of past climate; Bums' (1983) long-term reconstruction of crop yields for Southwestern Colorado; and Petersen's reconstruction of the dryland farm belt in the Dolores Archaeological Project Area (Petersen 1988). One of the conclusions drawn by Dean and his colleagues in the 1985 paper was that no matter how compelling their regional reconstruction of paleoenvironment, demography, and human behavior, it is still necessary to create local reconstructions of environment, population, and behavioral variability to explain sociocultural stability and change on the Colorado Plateaus (Dean et al. 1985:550). The method used here for creating highresolution reconstructions of annual and seasonal precipitation and temperature and Palmer Drought Severity Indices (PDSI) was described by Rose et al. (1981) in a major study for the southeastern portion of the Colorado Plateaus. Of these reconstructions, those employing the PDSI produced the best results. This was attributed to the fact that-like tree-ring 15- width­PDSI (an index of stored soil moisture) is a measure that integrates the cumulative effects of both precipitation and temperature (Rose et al. 1981 :244). The attempts by Bums (1983) to reconstruct annual yields of bean and maize for southwestern Colorado at varying levels of storage and to make inferences about specific intervals of food shortfalls and food surplus and their correspondence to periods of major building episodes, was made possible by the existence of appropriate long-term tree-ring chronologies. Bums' study led me to believe that high-frequency reconstructions of climate could have direct applications to the interpretation of the archaeological record. Finally, Petersen's reconstruction of changes in the location and width of the potential dryingfanning belt in southwestern Colorado from A.D. ＵＲＳＱｾ (Petersen 1988; Petersen and Clay 1987) led him and some other DAP researchers to conclude that fluctuations in temperature and precipitation were of sufficient magnitude and duration that populations were forced to migrate from the area because of their inability to grow adequate supplies of food. This was cited as the probable cause of the depopulation of the Dolores area in the - early 900s and of the entire Mesa Verde Region in the late 1200s. Yet, the Dolores area was relatively attractive from an agricultural perspective after the early tenth century local abandonment and the late thirteenth century regional abandonment. Therefore, it appeared to me that the role of climatic fluctuations on potential agricultural productivity and on sustainable population size and spatial distribution was unresolved, and that further investigation of this relationship was warranted. In retrospect, Dean et al. (1985) validated the need for locally relevant, high-frequency reconstructions of climatic variability by demonstrating climatic influence on imponant aspects of Anasazi adaptations. Rose et al. (1982) described how high-frequency climatic reconstruction could be done and suggested use of the Palmer Drought Severity Index as a particularly effective measure of climate. Bums (1983) demonstrated the archaeological value of high-frequency, long-term reconstructions of crop production in particular. Finally, Petersen's (1988) findings challenged me to examine this question in a new way with recently available data and methods. 16- 2 A model is a simplified representation of some multivariate, dynamic, real­world phenomenon. The ability of a model to faithfully portray and ultimately predict behavior is largely a function of two conditions. First, the variables that account for the greatest amount of variation in the targeted phenomenon must be correctly identified and included. Second, the observations or measurements taken on those key variables must be of sufficient quality and quantity that the dynamic character of the phenomenon can be perceived. The phenomenon under investigation here is agricultural productivity of lands used for dryland fanning of maize in southwestern Colorado. For the purposes of this research, the critical environmental variables accounting for the most variation in pre-industrial agricultural production are arable soil quality, soil moisture, and adequate growing season length (see Hogan [1987 :29--60] for a detailed discussion of growth requirements of maize). Arable soil quality is measured on a number of dimensions, but texture, depth, waterholding capacity, and natural productivity are among the most telling attributes. The average crop yield of a specific soil type is, of course, an excellent measure of productivity, but these data are not available for all soils. Natural productivity measures the ability of a soil to produce new native plant growth on an annual basis. It is most influenced' by plant nutrients and moisture supply and to a lesser degree by soil acidity, salt content. and a seasonal high water table (Soil Conservation Service 1951). The Soil Conservation Service measures this property on rangeland as "natural potential productivity" under favorable, normal, or unfavorable climatic conditions. It is expressed as the weight, in pounds, of armually produced air-dried understory vegetation per acre of land. In this study. this natural potential productivity is used as an integrative index of a particular soil's ability to produce plant growth. whether native or domestic. Soil moisture is largely a function of soil texture and water-holding potential, coupled with precipitation, evaporation, and runoff. It can be calculated by the Palmer Drought Severity Index, which requires data on waterholding capacities of each subsurface layer, total depth of the soil profile, and a source of information on precipitation and temperature input for that specific soil and location. Historic records of at least 30 years in length on monthly precipitation and temperature from various weather stations in the Four Comers area provide the climatic data used to calculate soil moisture values in this study. Adequate growing season length, in this area, is predominantly a function of elevation, mediated through the adaptive capability of the plant to withstand temperature extremes. Elevation is used as a proxy for the general zone in which the number of growing days (or alternatively expressed, the number of killing frost-free days) is sufficient to grow and mature a crop of maize. The location of each soil type can be recorded from maps and photographs and georeferenced ｾｩｴｨ digital elevation maps. mGH RESOLUTION SOILS DATA Descriptive and Locational Infounation Currently Available Information on soil was obtained from the Cortez field office of the U.S. Department of Agriculture, Soil Conservation Service (SCS). This office is responsible for the classification and inventory of land in Montezuma and Dolores Counties, southwestern Colorado. and the provision of land use information to the public. Field survey by SCS staff produces soil maps and soil descriptions published in Soil Survey manuals for specific areas. The reports for Montezuma and Dolores Counties are not yet available, but by 1988, when data were being gathered for this study, most of the land in the study area had been mapped as to soil type, and preliminary descriptions of those soils had been written. Certain data on soil productivity and water-holding characteristics also had been amassed. Therefore, it was possible to record soil location from aerial photographs and gather data on soil attributes from materials on file in that office. The SCS recognizes a total of 113 different soil map units in the two-county area. These units are either consociations of a single soil taxon representing a soil series (e.g., Witt Loam, 1-3% slope) or complexes of heterogeneous soil taxa (e.g., Romberg-Cragola-Rock Outcrop Complex, 25-80% slope) that are given distinct names and unique field map symbols (e.g., ROB and M2CE, respectively). Descriptions of these soil map units include information on topography, vegetation, range or woodland association, elevational range, average annual precipitation and temperature, and length of the killing frost-free period. Where relevant, all identifiable components of that map unit are described and their relative contribution to the soil complex estimated. Information on soil depth, permeability, shrink -swell potential, and rooting depth are given, and all subsurface units of the individual soils are described. The potential natural plant community on each named soil map unit is estimated in pounds per acre of arumal production of air-dried understory vegetation for a normal growing year. For each soil, the range of current land uses and an ev'aluation of agricultural capability are reported. Comments on agricultural land management practices are also included. Each soil map unit is classified on the basis of a number of inherent and situational characteristics. Inherent characteristics include: composition, arrangement, and thickness of subsur- face horizons; soil structure, color, texture, reaction, and consistence; content of humus, rock fragments, carbonates, and other salts; and mineralogical composition. These inherent properties govern such agriculturally important qualities as productivity, tillage, rooting depth, and available water capacity. Situational characteristics are also used to classify soil types and include slope, elevation, and topographic position. Archaeologists working in southwestern Colorado have often used the SCS capability classification value of a soil map unit as a proxy for its appropriateness for prehistoric agriculture and to a lesser extent, its agricultural productivity. While this is a convenient measure, it may not be sufficient for identifying a soil's potential for prehistoric agricultural success. The system of capability classification was designed to indicate the suitability of a soil for commercially growing most but not all modem field crops using modem technology. In this typological system, soils are evaluated on 11 dimensions (including both physical properties and climatic characteristics) but assigned a capability class based on their limitations for yielding crops and resisting deterioration. In fact, a soil is assigned the value equivalent to the lowest value on all dimensions for which it is evaluated. Physical property limitations include: depth, surface texture, percentage of coarse fragments, maximum slope, minimum permeability, minimum available water capacity, minimum drainage class and . growing season water table depth, maximum flood hazard, maximum salinity, and maximum erosion hazard by water and wind. Climatic limitations include the combination of average annual precipitation and the number of killing frost-free growing season days (Hamon 1983). Capability classes.(Table 2.1) are minimally designated by roman numerals I through VIII where Class I has the fewest limitations and is the best for successful modem agriCUlture, whereas Class VIII has the most severe limitations and is regarded as totally unsuitable for agriculture. In the Montezuma and Dolores County area, the best nonirrigated soils used for commercial agriculture are Class III and IV. No Class I, II, or V nonirrigated soils exist. Only two soil map units of the many that 18 - Table 2.1. Soil Conservation Service Soil Capability Class Definitions. Class Class I Class II Class III Class IV Class V Class VI Class VII Class VIII subclass "e" subclass "w" subclass "s" subclass "c" Definition Soils with few limitations that restrict their use. Soils with moderate limitations that reduce the choice of plants or that require special conservation practices. Soils with severe limitations that reduce the choice of plants or that require special conservation practices or both. Soils with very severe limitations that reduce the choice of plants or that require very careful management or both. Soils not likely to erode but that have other limitations, impractical to remove, that limit their use. Soils with severe limitations that make them generally unsuitable for cultivation. Soils with very severe limitations that make them unsuitable for cultivation. Soils with limitations that nearly preclude their use for commercial crop production. The main limitation is risk of erosion unless close-growing plant cover is maintained. The main limitation is water in or on the soil that interferes with plant growth. The main limitation is shallowness or stoniness. The main Iimhation is climate, either too cold or too dry. are classed as VI are ever used for commercial agriculture. No Class VII or Class VIII soil is reported as having been used for commercial agriculture. Often the main limitation or risk to successful plant growth is indicated in the capability classification system by adding a small letter e, w, s, or c to the roman numeral (see definitions in Table 2.1). By this method, a soil receives class designations on each of the 11 dimensions above. By invoking "Liebig's Law of the Minimum," the final land capability class assigned to a soil type is based on the presence of even a single higher (Le., more limiting) value. . record them for other purposes. Instead, I modeled prehistoric agricultural potential from known characteristics of and yield measurements associated with each named soil map unit in the study area. PDSI modeling was used to calculate moisture inputs and outputs to a given soil and the soil moisture retained in the profile. For each soil, it required historic values on monthly precipitation and temperature and unit specific values on the potential waterholding capacity on the uppermost six inches of soil and the potential water-holding capacity of the remaining lower soil section. Whereas the capability classification system provides broad guidelines for successful modem crop production, which is dependent on mechanization and often on chemical applications and minimum levels of production, it is less satisfactory for approximating prehistoric agricultural potential. The adaptation of prehistoric crops to climatic or soil conditions, differences in the maximum allowable slope (where mechanized equipment is not an issue), differences in risk associated with erosion due to (presumably) less extensive land clearance and lack of deep plowing, are not taken into account. Therefore, I decided not to use the capability classes as major indices of prehistoric agricultural productivity although I did The resulting PDSI value, calculated on an annual basis, approximates stored soil moisture available to a plant and incorporates the effects of moisture surplus or deficits to a soil over many months and even years. Thus, it can be used as an integrative index of potential dryland productivity because it takes into consideration soil depth and texture, available waterholding capacity, moisture input (precipitation and runoft), and moisture output (temperaturecontrolled evapotranspiration). After calculating PDSI values, potential natural plant productivity values-estimated under a variety of climatically controlled growing conditionswere linked to actual measurements of nonirrigated agricultural yields. This resulted in a set of estimated yield values that were quantitative 19 - and realistic under varying cHmatic conditions. These were used as an alternative method for estimating prehistoric potential agricultural capability. Descriptive and Locational Infonnation Recorded Soils data acquired from the SCS included descriptions of each soil type in the two-county region and information on the distribution of each of the mapped soil types in the study area. Most relevant for building the model were data on soil depth, available water-holding capacity, potential natural plant productivity estimates under varying climatic conditions, associated crop yields (when available), and whether or not the soil unit was artificially wet or saline due to irrigation. Other information included data on nonirrigated and irrigated land capability classification, elevational range, and range/woodland affiliation as they were reported in the individual soil descriptions on file in the Cortez SCS office (Tables 2.2 and 2.3). Soil scientists in the Cortez office have located these soil map units on aerial photographs enlarged to plainly identify local natural and cultural landmarks. The central portion of each photograph depicts a horizontal ground distance of one mile on four inches of photograph (approximately 1: 15,625). Section comer markers and section numbers have been identified on each photo. Final soil maps printed as soil distributions overlaid on 7.5-minute orthophoto quads are not yet available from the SCS. Had they been, the mapped soil units could have been digitized from these as vector data sets and encoded directly into a computer file to be converted to raster data. Since they were not, I was forced to take a different approach to recording soil distributions. This consisted of manually creating a raster data set, cell by cell, from the aerial photographs using each photograph in the study area. In order to do this, several interrelated decisions had to be made. The first decision was how to record the soil locational data, either in vector format (Le., as a multitude of irregular polygons created by linking lines and nodes and points) or in raster format (Le., as a uniform set of cells created by the regular gridding of the study area into rows and columns of a predetermined size). Given that the GIS systems available to me use the cell as the basic unit of analysis, and given that the DEMs were already in the form of a raster-like array, it seemed logical to record the locations of the soils in a raster format directly. without having to convert them for analysis. At this point I also firmly established the bound-aries of the study area; decided on a coordinate system to provide the geographic reference points; determined the cell size that would be the basic unit of analysis and gridded the entire study area into these cellular units; located each cell on the orthophoto quad and its equivalent position on the aerial photograph; ascertained the dominant soil type within each cell; and recorded the map unit symbol so that each map unit code is linked to its Cartesian coordinates. Boundary Delimitation Several factors were taken into account in determining the size of the study area. The region had to be large enough to allow significant spatial variability in soils and human settlement patterns. However, the effective study area had to be comprised of areas for which DEMs and soil mapping and descriptions were available. A rectangular study area incorporating highlands and lowlands and archaeological areas of potential interest (Mockingbird Mesa, Sand Canyon Survey LoCality, and the Dolores Archaeological Project, among others) satisfied these requirements (Figure 1.1). Its maximum extent is a rectangle that surrounds the perimeter of a mosaic of 12 7.5-minute DEMs arranged in a pattern of four maps (west-to-east) by three maps (north-to-south). It incorporates some 1,943.12 km 2 (750.24 mi 2) representing 30 minutes of long-itude (the four 7.5-minute maps joined west-to-east) and 22.5 minutes of latitude (the three 7.5minute maps joined north-to-south) de-icted on the 12 DEMs (Figure 2.1). Coordinate System DEMs are georeferenced in both latitude/longitude and UTM systems. Soils are referenced to township, range, and section -20- Table 2.2. Primary Soli Data for 113 Solis In Montezuma and Dolores Counties, Colorado. Part 1. ­ !'.) Soil Code 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Soli Name and Percent Slopes Torrifluvents Fluvaquents complex. 0-3% Sill silty clay loam, 0-3% Recapture variant sandy loam, 0-6% Panitchen loam. 0--3% Panitchen clay loam, 0-3% Ustic Torrlfluvents, gravelly substratum, 0--3% Youngston clay loam, 0-6% Sill silty clay loam, 3-6% Sill silty clay loam, 6-12% sm variant silty clay loam, 3-9% Umbarg-Winn-Tesajo complex, 0-2% Lillings silt loam, sodic, 1--3% Lillings silty clay loam, 1--3% L1l1ings silty clay loam, 3-8010 Poganeab loam, 0-2% Torrlorthents-Badland complex, 25-100% Zyme variant channery clay loam, 3-25% Zyme gravelly clay loam, dry, 3-12% Zyme-Sm complex. 25-65% Zyme very channery clay loam dry, 12-65% lies loam, moist, 3-12% lies loam, moist, 12-25% Uzona-Zwlcker-Claysprings complex, 3-12% SIII·Zyme complex. 3-25% NI Classb 6w 3s 6s 3c 3c 6s 6c 3e 4e 4e 4w 6s 6s 6s 6w 8e 6e 6s 7e 7e 4e 6e 6s 6e I Soil Map Unita A1B A2B (A2A) A3B A4 A41 A4A A4B A5C A5D (C9D) A5Wd A6A (J1B) AOAB AOB AOC AOd BD C2CE C2D C2F (C7E) C2V C3CD (S3C) C3E C5C C7D Class c 6w 3s 6s 3c 3c 4s 3e 3e 4e 4e 4w 6s 4s 4s 6w 8e 6e 6s 7e 7e 4e 6e 6s 6e Low Elev (tt) 5,000 5,400 5,000 5,400 5,400 5,400 5,000 5,400 5,400 5,400 5,400 5,400 5,400 5,400 5,400 5,000 5,400 5,400 5,400 5,400 7,100 7,100 5,000 5,400 High Elev (tt) 7,400 7,400 5,700 7,400 7,400 7,400 5,700 7,400 7,400 7,400 7,400 7,400 7,400 7,400 7,400 7,400 7,400 6,200 7,400 6,200 8,500 8,500 5,700 7,400 25 26 27 28 29 30 31 32 CP DOCE GP H1CD H1DC H6D (H6CD) M1CE M2C Shaly Pits Circleville variant cobbly loam, 3-25% Gravelly Pits Sill clay loam, shale substratum, 3-12% Belmear silty clay, 3-12% Romberg very stony loam. 6-25% Falcon gravelly fine sandy loam, stony, 3-25% Romberg-Cragola. comj)lex,very stony. 6-25% 8e 7s 8e 4e 4e 7s 6e 7s 8e 7s 8e 4e 4e 7s 6e 7s 5,000 7,100 5,000 5,400 7,100 5,400 7,100 5,400 8,000 8,500 8,500 7,400 8,500 7,400 8,500 7,400 Range/Woodland Classification semi-wet salt streambank alkali bottom alkali bottom loamy bottom loamy bottom loamy bottom alkali bottom clayey foothills clayey foothills salt meadow wet meadow alkali bottom alkali bottom alkali bottom wet meadow NA piFlon-juniper shaly knobs pinon-juniper shaly knobs mountain loam mountain loam clayey salt desert pinon-juniper and clayey foothills pit ponderosa pine pit clayey foothills ponderosa pine pinon-juniper pine grassland pinon-juniper Table 2.2 (Continued). Nt Low I Soli Name and Percent Slopes Classb Class c Elev (tt) Romberg-Cragola-Rock Outcrop complex, 25-80% 7e 7e 5,400 High Elev (ft) 7,400 Gladel-Pulpit complex, 3-9% 6s 6s 6,200 7,400 Gladel-Pulpit complex, dry, 3-9% 6s 6s 5,400 6,200 36 M2001 (ROHC1) M5E (S52) lIes-Granath Ioams, 2-6% 4c 4c 7,100 8,500 37 M6C Ormiston very stony-Nortez complex, 3-12% 7s 7s 7,100 8,500 38 M7E Ormiston variant-Granath complex, 3-12% 7s 7s 7,100 8,500 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 M80 (5S5, 54C) OAB PiC RiC RiO R1HB R1HC R1HO R3B R3B1 R3C (R9B, R5C) R3C1 R30 R301 (ROO1) R4C (R4B) R4C1 (R4B1, R7C1, ROLC1) R40 R7C (RC) R70 R701 (ROH01) R8B IIes-Pramlss-Faicon complex, 3-20% Binton clay loam, 0-3% Sheppard fine sand, 1-6% Granath loam, 3-6% Granath loam, 6-12% Rlcot loam, 1-3% Ricot loam, 3-6% Ricot loam, 6-12% Cahona very fine sandy loam, 1-3% Cahona very fine sandy loam, dry, 1-3% Cahona very fine sandy loam, 3-6% Cahona very fine sandy loam, dry, 3-6% Cahona very fine sandy loam, 6-12% Cahona very fine sandy loam, dry, 6-12% Cahona-Sharps-Witt complex, 2-6% Cahona-Sharps-Witt complex, dry, 2-6% 6e 7s 7s 4c 4e 4c 4c 4e 3c 4c 3e 4c 4e 4e 3e 4c 6e 4s 4s 4c 4e 4c 4c 4e 3c 2e 3e 3e 4e 4e 3e 3e 7,100 5,000 5,000 7,100 7,100 7,100 7,100 7,100 6,200 5,400 6,200 5,400 6,200 5,400 6,200 5,400 8,500 5,700 5,700 8,500 8,500 8,500 8,500 8,500 7,400 6,200 7,400 6,200 7,400 6,200 7,400 6,200 Range/Woodland Classification pinon-juniper and rock outcrop pinon-juniper and loamy foothills pinon-juniper and semidesert loam mountain clay and mountain loam ponderosa pine and pine grassland pine grassland and mountain loam ponderosa pine alkali bottom desert sand mountain loam mountain loam mountain loam mountain loam mountain loam loamy foothills semi-desert loam loamy foothills semi-desert loam loamy foothills semi-desert loam loamy foothills semi-desert loam Sharps-Cahona complex, 6-12% Sharps-Pulpit loams, 2-6% Sharps-Pulpit loams, 6-12% Sharps-Pulpit loams, dry, 6-12% Recapture fine sandy loam, 0-6% 4e 3e 4e 4e 6s 4e 3e 4e 4e 6s 6,200 6,200 6,200 5,400 5,000 7,400 7,400 7,400 6,200 5,700 loamy foothills loamy foothills loamy foothills semi-desert loam alkali flat Soil Code 33 34 35 N N 55 56 57 58 59 Soil Map Unita M2CE (H5E, H50E) M200 Table 2.2 (Continued). Soli Code 60 61 62 63 64 65 66 67 68 69 70 71 72 N w Soil Map Un/ta Nt Classb 4e 8e 3c 4c 3e 4c 4e 3e 4e 3e 4e 4e 6c I Soli Name and Percent Slopes Cahona-Pulpit complex, 3-9% Rock Outcrop Witt loam, 1-3% Witt loam, dry, 1-3% Witt loam, 3-6% Witt loam, dry, 3-6% Witt loam, 6-12% Pulpit loam, 3-6% Pulpit loam, 6-12% Sharps loam, 3-6% Sharps loam, 6-12% Sharps loam, dry, 6-12% Mack-Redlands fine sandy loams, 0-6% Class c 4e 8e 3c 2e 3e 3e 4e 3e 4e 3e 4e 4e 3e Low Elev (ft) 6,200 NA 6,200 5,400 6,200 5,400 6,200 6,200 6,200 6,200 6,200 5,400 5,000 High Elev (ft) 7,400 NA 7,400 6,200 7,400 6,200 7,400 7,400 7,400 7,400 7,400 6,200 5,700 Range/Woodland Classification loamy foothills rock outcrop loamy foothills semi-desert loam loamy foothills semi-desert loam loamy foothills loamy foothills loamy foothills loamy foothills loamy foothills semi-desert loam alkali flat Acree loam, sandstone substratum, 0-6% Acree loam, 2-6% lIes-Granath Ioams, 6-12% 4c 4c 4e 4c 4c 4e 7,100 7,100 7,100 8,500 8,500 8,500 4e 7s 4e 7s 7,100 7,100 8,500 8,500 4c 4c 7,100 8,500 mountain loam mountain loam mountain clay and mountain loam mountain loam mountain clay and mountain loam mountain clay and mountain loam mountain clay pine grassland and mountain loam mountain clay and mountain loam loamy foothills pine grassland NA NA loamy foothills loamy foothills loamy foothills 73 74 75 RHC RL ROB ROB1 ROC ROC1 ROD ROHC ROHD ROLC (C4C) ROLD (C4D) ROLD1 (R4D1) S1C (S2B, S2C, MD,DD,MOS) S5(S5C) S51 S53 76 77 S5D S6(S60,S6C) 78 S6C. Acree loam, 6-12% Ormiston-very stony Acree, sandstone substratum complex, 1-12% Pramlss-Acree Ioams, 3-9% 79 80 S6E S7 Pramlss very cobbly loam, very stony, 6-25% Nortez-Acree sandstone substratum loams, 0-6% 7s 4c 7s 4c 7,100 7,100 8,500 8,500 81 S8C lIes-Pramiss Granath Ioams, 2-6% 4c 4c 7,100 8,500 82 83 84 85 86 87 88 T1B (T1C) T2C T3E T4B T6 T6B T6C Weber loam, 1-3% Ormiston variant very cobbly loam, 0-9% Torrlorthents,very stony, 12-65% Purcella loam, 0-3% Collbran clay loam, 3-6% Collbran-Collbran variant complex, 0-2% Collbran-Collbran variant ｣ｯｭｰｬ･ｸＬｾﾰ ____ 3c 7s 7e 6s 3e 3c 3e 3c 7s 7e 6s 3e 3c 3e 5,400 7,100 5,000 5,400 5,400 5,400 5,400 7,400 8,500 7,400 7,400 7,400 7,400 7,400 Table 2.2 (Concluded). Soil Code 89 90 91 92 93 94 95 96 97 IV ｾ 98 99 100 101 102 103 104 105 106 107 Soil Map Unita T6D V1B V1C V1D V2B (W1B) V2C (V3C) V2S (VOAB) V2Vd (W1BB) v2wd (W1CB, AOBB, VOBB) V4C VOB VOC VOCB VOCC (VOBC) WOB WOBBd WOC XC6D (C5E) XC6E (XM3F) 108 109 110 XH2D XH2E XH3E (M4E) 111 XM1F Fughes loam, 2-9% Yarts fine sandy loam, 1-3% Yarts fine sandy loam, 3-6% Yarts clay loam, 1-3% Yarts clay loam, 3-6% Ackmen loam, 1-3% IraJ< loam, G-3% Ackmen loam, 3-6% Claysprings clay loam,very stony, 12-65% Typic Torriorthents-bouldery Rock Outcrop complex, 12-80% Sheek-Archuleta complex,very stony, 6-25% Sheek-Archuleta Rock Outcrop complex, 25-80% Sheek-Archuleta-Rock Outcrop complex, 25-80%, north slopes Falcon-Rock Outcrop complex, 25-65% 112 XM3D Farb-Rock Outcrop complex, 3-12% Soli Name and Percent Slopes Collbran clay loam, 6-12% Hesperus loam, 1-3% Hesperus loam, 3-6% Hesperus loam, 6-12% Mldm clay loam, 1-3% Miklm loam, 3-6% Miklm clay loam, sodle, G-3% Miklm clay loam, slightly wet, G-3% Midm variant clay loam, 0-3% NI Class b 4e 4c 4c 4e 4s 4s 6s 4s 7s I Class c 4e 4c 4c 4e 3s 3e 6s 3s 7s Low Elev (ft) 5,400 7,100 7,100 7,100 5,400 5,400 5,400 5,400 5,400 High Elev (ft) 7,400 8,500 8,500 8,500 7,400 7,400 7,400 7,400 7,400 4e 3c 3e 3c 3e 3c 3w 3e 7e 7e 4e 3c 3e 3c 3e 3c 3w 3c 7e 7e 7,100 5,400 5,400 5,400 5,400 5,400 5,400 5,400 5,000 5,000 8,500 7,400 7,400 7,400 7,400 7,400 7,400 7,400 5,700 5,700 7s 7e 7e 7s 7e 7e 7,100 7,100 7,100 7e 7e 7,100 7s 7s 5,000 7e 5,400 Ustic Torriorthents-Gullled Land complex, 25-60% 7e Z 113 aEarlier, alternative soil map unit symbols, now officially subsumed under a final soil map unit designation, are shown bNonirrigated land SCS Capability Class; Roman numerals converted to Arabic numbers. Clrrigated land SCS Capability Class; Roman numerals converted to Arabic numbers. dArtificial1y wet and possibly saline from irrigation water. Range/Woodland Classification loamy foothills loamy pari< loamy pari< brushy loam alkali flat alkali flat alkali flat NA salt meadow loamy pari< semi-desert loam semi-desert loam semi-desert loam semi-desert loam loamy bottom wet meadow loamy bottom salt desert breaks salt desert breaks and rock outcrop 8,500 ponderosa and Douglas-fir 8,500 ponderosa and Douglas-fir 8,500 ponderosa pine, Douglas-fir, and rock outcrop 8,500 ponderosa pine and rock outcrop 5,700 shallow desert and rock outcrop 7,400 gullied land in parentheses. Table 2.3. Primary Soil Data for 113 Soils In Montezuma and Dolores Counties, Colorado. Part 2. Soli Code 1 Soil Map Unit A1B 2 A2B Soil Unit and Percent a Torrifluvents (62.5) Fluvaquents (37.5) Sill (100) 3 A3B Recapture (100) 4 A4 Panitchen (100) 5 A41 Panitchen (100) 6 A4A Ustic Torrifluvents (100) 7 A4B Youngston (100) 8 A5C Sill (100) 9 A5D· Sill (100) 10 A5We Sill Variant (100) N UI SoJI Unit Depth (in) NA NA 0-6 6-60 0-6 6-13 13-17 17-38 38-60 0-3 3-60 0-3 3-60 0-3 3-11 11-27 27-60 0-10 10-43 43-60 0-6 6-60 0-6 6-60 0-4 4-32 32-60 AWC (inlln)b NA NA .18-.20 .16-.18 .08-.10 .15-.17 .08-.10 .12-.14 .07-.09 .15-.18 .14-.17 .14-.17 .13-.16 .05-.06 .13-.15 .04-.08 .08-.10 .15-.18 .15-.18 .15-.18 .18-.20 .16-.18 .18-.20 .16-.18 .15-.17 .13-.16 .13-.15 NI Beand (Ib/ac) Favorable NPpc (Ib/ac) 2,000 2,500 875 Normal NPpc (Ib/ac) 1,200 2,000 625 Unfavorable NPpc (Ib/ac) 500 1,500 450 900 650 450 1,200 900 700 1,200 900 700 950 800 600 900 700 500 1,200 900 600 275 1,200 900 600 250 2,500 (1,200) 2,000 (900) 1,500 (600) Table 2.3 (Continued). Soil Code 11 Soil Map Unn A6A Soil Unn and Percent a Umbarg (40) Winn (35) Tesajo (25) to­) 12 AOAB L1l1ings (100) 13 AOB L1l1ings (100) 14 AOC L1l1ings (100) 15 AOe Poganeab (100) 16 BO 17 C2CE Torriorthents (62.5) Badlands (37.5) Zyme (100) 18 C20 Zyme (100) 19 C2F Zyme (67.5) 0\ Sill (32.5) 20 C2V Zyme (100) 21 C3CO lies (100) Soil Unn .Depth (in) 0-2 2-5 5-12 12-42 42-60 0-4 4-31 31-60 0-3 3-36 36-60 0-2 2-60 0-2 2-60 0-2 2-60 0-2 2-60 0-60 0-60 0-1 1-19 0-1 1-19 0-1 1-19 0-6 6-60 0-1 1-19 0-2 2-22 22-60 AWC (inlin)b .15-.18 .17-.20 .17-.20 .15-.18 .08-.11 .18-.20 .18-.20 .07-.10 .11-.13 .04-.06 .04-.06 .18-.20 .18-.20 .18-.20 .18-.20 .18-.20 .18-.20 .15-.18 .17-.20 .00-.00 .00-.00 .08-.10 .15-.17 .12-.14 .15-.17 .08-.10 .15-.17 .18-.20 .16-.18 .09-.10 .15-.17 .08-.18 .19-.21 .14-.16 Favorable NPpc (Ib/ac) 3,000 Normal NPpc (Ib/ac) 2,600 Unfavorable NPpc (Ib/ac) 2,300 3,000 2,600 2,300 2,500 2,000 1,500 775 600 450 1,000 700 500 1,000 700 500 Nt Beand (Ib/ac) 2,000 (667) 200 4,000 (1,333) 500 0 600 3,000 (1,000) 350 0 400 500 300 200 600 400 300 700 500 300 500 300 200 1,800 1,500 1,200 a 300 400 Table 2.3 (Continued). Soil Code 22 Soil Map Unit C3E Soil Unit and Percent a lies (100) 23 C5C Uzona (40) Zwicker (35) Claysprings (25) 24 C7D Sill (52.5) Zyme (47.5) N -...I 25 26 CP DOCE Pits (100) Circleville Variant (100) 27 28 GP H1CD sm (100) 29 H1DC Belmear (100) 30 H6D Romberg (100) 31 M1CE Falcon (100) 32 M2C Romberg (52.5) Pits (100) Cragola (47.5) 33 M2CE Romberg (40) Cragola (35) Rock Outcrop (25) Soil Unit Depth (in) 0-2 2-22 22-60 0-1 1-45 45-60 0-1 1-4 4-32 0-3 3-18 0-6 6-60 0-1 1-19 0-60 0-9 9-24 0-6 0-6 6-42 0-10 10-30 0-2 2-20 0-4 4-13 0-2 2-60 0-2 2-18 0-2 2-60 0-2 2-18 0-60 AWC (inlin)b .16-.18 .19-.21 .14-.16 .16-.18 .15-.17 .17-.19 .17-.21 .17-.21 .14-.21 .10-.12 .16-.18 .18-.20 .16-.18 .08-.10 .15-.17 .00-.00 .10-.12 .10-.12 .00-.00 .17-.19 .14-.16 .14-.16 .12-.15 .12-.14 .07-.08 .07-.10 .06-.08 .12-.14 .07-.08 .06-.09 .07-.10 .12-.14 .07-.08 .06-.09 .07-.10 .00-.00 Favorable NPpc (Ib/ac) 1,800 Normal NPpc (Iblac) 1,500 Unfavorable NPpc (Ib/ac) 1,200 625 450 250 500 350 200 375 250 150 1,200 900 600 600 400 300 0 700 0 500 0 300 0 1,100 0 800 0 500 900 500 300 500 350 200 900 650 550 500 350 200 400 200 100 500 350 200 400 200 100 0 0 0 NI Beand (Ib/ac) 250 Table 2.3 (Continued). Soil Code 34 Soil Map Unit M2DD Soil Unit and PercentS Gladel (55) Pulpit (45) 35 M2DD1 Gladel (55) Pulpit (45) 36 M5E lies (67.5) Granath (32.5) N 00 37 M6C Ormiston (52.5) Nortez (47.5) 38 M7E Ormiston Variant (57.5) Granath (42.5) 39 M8D lies (40) Pramlss (35) Falcon (25) Soil Unit Depth (In) 0-5 5-15 0-10 10-20 20-36 0-5 5-15 0-10 10-20 20-36 0-2 2-22 22-60 0-13 13-60 60-80 0-3 3-7 7-32 32-44 0-2 2-6 6-31 0-9 9-13 13-60 0-13 13-60 60-80 0-2 2-22 22-60 0-3 3-31 0-4 4-13 AWC (inJin)b .11-.13 .09-.12 .16-.18 .19-.21 .16-.18 .11-.13 .09-.12 .16-.18 .19-.21 .16-.18 .16-.18 .19-.21 .14-.16 .18-.20 .18-.20 .15-.19 .06-.11 .07-.13 .05-.11 .11-.17 .15-.18 .17-.20 .15-.18 .06-.09 .09-.11 .07-.10 .18-.20 .18-.20 .15-.19 .16-.18 .19-.21 .14-.16 .07-.12 .16-.18 .07-.10 .06-.08 Favorable NPPC (Ib/ac) 500 Normal NPpc (Ib/ac) 350 Unfavorable NPpc (Ib/ac) 200 NIBeand (lb/ac) 300 1,125 800 600 500 350 200 750 600 450 250 1,300 1,100 800 450 1,700 1,400 1,100 450 1,300 1,000 800 1,400 1,200 900 1,200 900 750 1,700 1,400 1,100 1,500 1,000 800 1,100 800 550 900 650 550 Table 2.3 (Continued). Soil Code 40 Soil Map Unit OAB Soil Unit and PercentS Blnton (100) 41 P1C Sheppard (100) 42 R1C Granath (100) 43 R1D Granath (100) 44 R1HB Rlcot (100) 45 R1HC Rlcot (100) 46 R1HD Rlcot (100) 47 R3B· Cahona (100) 48 R3B1 Cahona (100) 49 R3C Cahona (100) 50 R3C1 Cahona (100) IV \0 Soli Unit Depth (in) 0-9 9-60 0-7 7-60 0-13 13-60 60-80 0-13 13-60 60-80 0-8 8-16 16-34 34-60 0-8 8-16 16-34 34-60 0-8 8-16 16-34 34-60 0-11 11-24 24-60 0-11 11-24 24-60 0-11 11-24 24-60 0-11 11-24 24-60 AWC (Inlin)b .14-.17 .12-.15 .05-.07 .06-.08 .18-.20 .18-.20 .15-.19 .18-.20 .18-.20 .15-.19 .16-.18 .16-.18 .12-.15 .10-.14 .16-.18 .16-.18 .12-.15 .10-.14 .16-.18 .16-.18 .12-.15 .10-.14 .14-.16 .15-.17 .13-.16 .14-.16 .15-.17 .13-.16 .14-.16 .15-.17 .13-.16 .14-.16 .15-.17 .13-.16 NI Beand (Ib/ac) Favorable NPpc (Ib/ac) 800 Normal NPpc (Ib/ac) 600 Unfavorable NPpc (Ib/ac) 400 900 550 350 1,700 1,400 1,100 450 1,700 1,400 1,100 400 1,800 1,500 1,200 1,800 1,500 1,200 1,800 1,500 1,200 1,500 1,000 800 350 800 700 400 275 1,500 1,000 800 350 800 700 400 275 Table 2.3 (Continued). Soil Code 51 Soil Map Unit R3D Soil Unit and Cahona (100) 52 R3D1 Cahona (100) 53 R4C Cahona (40) Percent 8 Sharps (35) Witt (25) w 54 R4C1 Cahona (40) 0 Sharps (35) Witt (25) 55 R4D Sharps (52.5) Cahona (47.5) Soli Unit Depth (in) 0-11 11-24 24-60 0-11 11-24 24-60 0-11 11-24 24-60 0-9 9-19 19-30 0-7 7-48 48-60 0-11 11-24 24-60 0-9 9-19 19-30 0-7 7-48 48-60 0-9 9-19 19-30 0-11 11-24 24-60 AWC (inJin)b .14-.16 .15-.17 .13-.16 .14-.16 .15-.17 .13-.16 .14-.16 .15-.17 .13-.16 .16-.18 .15-.17 .13-.15 .15-.18 .18-.21 .16-.19 .14-.16 .15-.17 .13-.16 .16-.18 .15-.17 .13-.15 .15-.18 .18-.21 .16-.19 .16-.18 .15-.17 .13-.15 .14-.16 .15-.17 .13-.16 NI Beand (Ib/ac) 300 Favorable NPpc (Ib/ac) 1,500 Normal NPpc (Ib/ac) 1,000 Unfavorable NPpe (Ib/ac) 800 800 700 400 250 1,500 1,000 800 350 1,200 900 700 350 1,500 1,200 800 450 800 700 400 275 750 600 450 250 1,200 750 500 300 1,200 900 700 300 1,500 1,000 800 300 Table 2.3 (Continued). Soil Code 56 Soil Map Unit R7C Soli Unit and Percent a Sharps (52.5) Pulpit (47.5) 57 R7D Sharps (52.5) Pulpit (47.5) 58 R7D1 Sharps (52.5) Pulpit (47.5) .... VJ 59 R8B Recapture (100) 60 RHC Cahona (57.5) Pulpit (42.5) 61 62 RL ROB Rock Outcrop (100) Witt (100) 63 ROB1 Witt (100) Soli Unit Depth (in) 0-9 9-19 19-30 0-10 10-20 20-36 0-9 9-19 19-30 0-10 10-20 20-36 0-9 9-19 19-30 0-10 10-20 20-36 0-7 7-26 26-60 0-11 11-24 24-60 0-7 7-25 25-36 0-60 0-7 7-48 48-60 0-7 7-48 48-60 AWC (inlin)b .16-.18 .15-.17 .13-.15 .16-.18 .19-.21 .16-.18 .16-.18 .15-.17 .13-.15 .16-.18 .19-.21 .16-.18 .16-.18 .15-.17 .13-.15 .16-.18 .19-.21 .16-.18 .13-.15 .15-.17 .14-.16 .14-.16 .15-.17 .13-.16 .16-.18 .19-.21 .16-.18 .00-.00 .15-.18 .18-.21 .16-.19 .15-.18 .18-.21 .16-.19 NI Beand (Ib/ac) 350 Normal NPpc (Iblac) 900 Unfavorable NPpc (Ib/ac) 700 1,125 800 600 350 1,200 900 700 300 1,125 800 600 300 750 600 450 250 750 600 450 250 900 650 450 1,500 1,000 800 350 1,125 800 600 350 0 1,500 0 1,200 0 800 450 1,200 750 500 300 Favorable NPpc (Ib/ac) 1,200 ｾ Table 2.3 (Continued). Soli Code 64 Soil Map Unit ROC Soil Unit and PercentS Witt (100) 65 ROC1 Witt (100) 66 ROD Witt (100) 67 ROHC Pulpit (100) 68 ROHD Pulpit (100) 69 ROLC Sharps (100) 70 ROLD Sharps (100) 71 ROLD1 Sharps (100) 72 S1C Mack (52.5) N Redlands (47.4) 73 S5 Acree (100) 74 S51 Acree (100) Soil Unit Depth (In) 0-7 7-48 48-60 0-7 7-48 48-60 0-7 7-48 48-60 0-10 10-20 20-36 0-10 10-20 20-36 0-9 9-19 19-30 0-9 9-19 19-30 0-9 9-19 19-30 0-13 13-60 0-7 7-46 46-60 0-10 10-42 42-50 0-10 10-42 42-60 AWC (inJln)b .15-.18 .18-.21 .16-.19 .15-.18 .18-.21 .16-.19 .15-.18 .18-.21 .16-.19 .16-.18 .19-.21 .16-.18 .16-.18 .19-.21 .16-.18 .16-.18 .15-.17 .13-.15 .16-.18 .15-.17 .13-.15 .16-.18 .15-.17 .13-.15 .13-.18 .13-.16 .13-.16 .16-.18 .15-.17 .16-.18 .15-.21 .16-.18 .17-.21 .13-.16 .13-.16 Favorable NPpc (Ib/ac) 1,500 NI Beand (Ib/ac) 450 Normal NPpc (Iblac) 1,200 Unfavorable NPpc (Ib/ac) 800 1,200 750 500 300 1,500 1,200 800 400 1,125 800 600 350 1,125 800 600 300 1,200 900 700 350 1,200 900 700 300 750 600 450 250 800 700 500 800 700 500 1,800 1,500 1,200 450 2,000 1,600 1,400 450 Table 2.3 (Continued). Soli Code 75 Soil Map Unit S53 Soli Unit and Percent a lies (67.5) Granath (32.5) 76 S5D Acree (100) 77 S6 Ormiston (57.5) ｾ Acree (42.5) 78 S6C Pramlss (55) Acree (45) 79 S6E Pramiss(100) 80 S7 Nortez(52.5) Acree (47.5) Soil Unit Depth (in) 0-2 2-22 22-60 0-13 13-60 60-80 0-10 10-42 42-60 0-3 3-7 7-32 32-44 0-10 10-42 42-50 0-3 3-31 0-10 10-42 42-60 0-3 3-31 0-1 1-9 9-25 0-10 10-42 42-50 AWC (inlin)b .16-.18 .19-.21 .14-.16 .18-.20 .18-.20 .15-.19 .17-.21 .13-.16 .13-.16 .06-.11 .07-.13 .05-.11 .11-.17 .16-.18 .15-.21 .16-.18 .15-.18 .15-.18 .17-.21 .13-.16 .13-.16 .07-.12 .16-.18 .15-.18 .17-.20 .15-.18 .16-.18 .15-.21 .16-.18 Favorable NPPC (Ib/ac) 1,300 Nonnal NPpc (Iblac) 1,100 Unfavorable-------NIB-eand NPpc (Ib/ac) (Ib/ac) 800 400 1,700 1,400 1,100 400 2,000 1,600 1,400 400 1,200 900 600 1,800 1,500 1,200 1,100 900 700 400 2,000 1,600 1,400 450 1,100 800 550 1,400 1,200 900 400 1,800 1,500 1,200 450 Table 2.3 (Continued). Soil Code 81 Soil Map Unit S8C Soil Unit and Percent a lies (40) Pramiss (35) Granath (25) 82 11B Weber (100) 83 T2C Ormiston Variant (100) 84 T3E Torrlorthents (100) 85 T4B Purcella (100) 86 T6 Collebran (100) 87 T6B Collebran (52.5) Yo) .f>. Collebran Variant (47.5) Soil Unit Depth (in) 0-2 2-22 22-60 0-3 3-31 0-13 13-60 60-80 0-8 8-32 32-60 0-9 9-13 13-60 0-3 3-7 7-16 0-4 4-11 11-41 41-60 0-10 10-40 40-60 0-10 10-40 40-60 0-2 2-8 8-45 45-60 AWC (inlin)b .16-.18 .19-.21 .14-.16 .15-.18 .15-.18 .18-.20 .18-.20 .15-.19 .16-.19 .16-.20 .03-.06 .06-.09 .09-.11 .07-.10 .04-.06 .07-.08 .07-.08 .15-.18 .12-.15 .05-.08 .03-.05 .18-.20 .16-.19 .13-.15 .18-.20 .16-.19 .13-.15 .15-.18 .18-.21 .15-.18 .07-.10 NI Beand (Ib/ac) 450 Normal NPpc (Ib/ac) 1,100 Unfavorable NPpc (Ib/ac) 800 1,100 900 700 400 1,700 1,400 1,100 450 1,500 1,200 800 1,200 900 750 500 350 200 NA NA NA NA NA NA NA NA NA NA NA NA Favorable NPpc (Ib/ac) 1,300 Table 2.3 (Continued). Soil Code 88 SoitMap Unit T6C Soil Unit and Pereent a Collebran (52.5) Collebran Variant (47.5) w VI 89 T6D Collebran (100) 90 V1B Hesperus (100) 91 V1C Hesperus (100) 92 V1D Hesperus (100) 93 V2B Miklm (100) 94 V2C Mikim (100) 95 V2S Mikim (100) 96 V2Ve Mikim (100) Soil Unit Depth (in) 0-10 10-40 40-60 0-2 2-8 8-45 45-60 0-10 10-40 40-60 0-11 11-44 44-60 0-11 11-44 44-60 0-11 11-44 44-60 0-3 3-15 15-32 32-60 0-3 3-15 15-32 32-60 0-3 3-15 15-32 32-60 0-3 3-15 15-32 32-60 AWC (inlin)b .18-.20 .16-.19 .13-.15 .15-.18 .18-.21 .15-.18 .07-.10 .18-.20 .16-.19 .13-.15 .16-.19 .16-.19 .16-.19 .16-.19 .16-.19 .16-.19 .16-.19 .16-.19 .16-.19 .14-.17 .14-.17 .14-.17 .14-.17 .12-.15 .14-.17 .14-.17 .14-.17 .12-.15 .12-.15 .12-.15 .12-.15 .15-.17 .15-.17 .15-.17 .15-.17 NI Beand (Ib/ae) Favorable NPpc (Iblae) NA Normal NPpc (Ib/ae) NA Unfavorable NPpc (Ib/ae) NA NA NA NA NA NA NA 2,500 1,800 1,000 500 2,500 1,800 1,000 500 3,000 2,000 1,500 450 875 625 450 875 625 450 675 475 350 2,000 (875) 1,600 (625) 1,200 (450) Table 2.3 (Continued). Vo) Soil Code 97 Soil Map Unit V2We Soil Unit and PercentS Miklm Variant (100) 98 V4C Fughes (100) 99 VOB Yarts (100) 100 vac Yarts (100) 101 VOCB Yarts (100) 102 VOCC Yarts (100) 103 WOB Ackmen (100) 104 WOBBe IraJ< (100) 105 woe Ackmen (100) 106 XC6D Claysprlngs (100) 107 XC6E Typic Torrlorthents (67.5) 0\ Rock Outcrop (32.5) 108 XH2D Sheek (57.5) Archuleta (42.5) Soil Unit Depth (In) 0-8 8-35 35-60 0-4 4-50 50-60 0-9 9-13 13-60 0-9 9-13 13-60 0-9 9-13 13-60 0-9 9-13 13-60 0-20 20-60 0-8 8-60 0-20 20-60 0-3 3-18 0-3 3-7 7-16 0-60 0-4 4-18 18-60 0-8 8-17 AWC (inlin)b .15-.17 .09-.15 .09-.15 .16-.18 .15-.17 .14-.16 .13-.15 .11-.12 .11-.12 .13-.15 .11-.12 .11-.12 .18-.20 .11-.12 .11-.12 .18-.20 .11-.12 .11-.12 .17-.19 .17-.19 .14-.17 .15-.18 .17-.19 .17-.19 .10-.12 .16-.18 .04-.06 .07-.09 .07-.09 .00-.00 .06-.08 .11-.14 .11-.14 .08-.10 .16-.18 Favorable NPpc (Ib/ac) 1,800 (675) Normal NPpc (Ib/ac) 1,600 (475) Unfavorable NPpc (Ib/ac) 1,200 (350) NI Beand (Ib/ac) 2,300 1,800 1,100 900 700 500 900 700 500 900 700 500 900 700 500 1,000 800 600 500 3,000 (1,000) 1,000 2,500 (800) 800 2,000 (600) 600 250 500 350 200 500 350 200 0 500 0 400 0 200 500 350 200 500 500 Table 2.3 (Concluded). Soil Code 109 Soli Map Unit XH2E Soil Unit and Percent a Sheek (40) Archuleta (30) 110 XH3E Rock Outcrop (20) Sheek (45) Archuleta (30) w 111 XM1F Rock Outcrop (25) Falcon (62.5) 112 XM3D Rock Outcrop (37.5) Farb (62.5) 113 z ....,J Rock Outcrop (37.5) Ustlc Toniorthents (52.5) Gullied Land (47.5) Soil Unit Depth (in) 0­4 4­18 18­60 0­8 8­17 0­60 0­4 4­18 18­60 0­8 8­17 0­60 0­4 4­13 0­60 0­3 3­16 0­60 0­5 5­60 0­60 ­ AWC (in/in)b .06­.08 .11­.14 .11­.14 .08­.10 .16­.18 .00­.00 .06­.08 .11­.14 .11­.14 .08­.10 .16­.18 .00­.00 .07­.10 .06­.08 .00­.00 .08­.13 .06­.13 .00­.00 .10­.13 .06­.18 .00­.00 Favorable NPpc (Ib/ac) 500 ­ ­ ­ ­ ­ Normal NPpc (Ib/ac) 400 Unfavorable NPpc (Ib/ac) 200 500 350 200 0 700 0 550 0 300 700 500 300 0 900 0 650 0 550 0 550 0 400 0 275 0 500 0 300 0 200 0 0 0 NI Beand (Ib/ac) aThis value is used as a weighted multiplier in computing total available water capacity per soil profile. It is derived from estimates of soil type composition provided by the SCS. Each soil type Is actually composed of the major components listed plus some small fraction, usually 20 percent or less, of other named and unnamed soil types. The multiplier provided above is the sum of the actual percentage of that component, as estimated by the SCS, plus a portion of the remaining percent representing the unspecified soils. For example, Soil Type A6A Is comprised of three major components. SCS soil descriptions indicate that approximately 35% of this "soil type" is Umbarg, 30% is Winn, 20% is Tesajo, and 15% are other less prevalent soil types. This remaining 15% is divided equally among the dominant soil types to croduce the weighting of 40:35:25 used to calculate soil moisture for this study. Available Water Capacity per subsoil unit given in inches of water per inch of soil (SCS 6/21/88). CPotential Natural Plant Productivity estimate provided for a soil under favorable, normal, and unfavorable growing season conditions (SCS 3/3/89). dAverage yield per given soil type for pinto beans grown on nonirrigated soils in Montezuma and Dolores Counties (SCS 3/3/89). eArtificially wet and possibly saline from irrigation water. Values reduced to reflect this; guidelines suggested by SCS Soil Scientist, Alan Price (personal communication, March 1989). lI NW Corner e ll ":' ( 101. 37°37'0" long. 109°0'0" 4165886 m N 676499 m E 11 RUIN CANYON (CEDAR 3 SW) All NEGRO CANYON (MOQUI NW) I YELLOWJACKET TRIMBLE POINT WOODS CANYON ARRIOLA DOLORES WEST 37° 15'0" long. 109°0'0" 4166944 m N 720630 m E II 11 0 11 S I1 CORTEZ "\;) SW Corner V 101 37°37'30" long 108°30'0" 101. PLEASANT VIEW BOWDISH BATT LEROCK MUD CREEK CANYON (MOQUI SE) (MOQUI SW) NE Corner SE Corner 101 37° 15' 0" long. 108°30'0" 4125332 m N 721734 m E 4124277 m N 677382 m E Figure 2.1. Schematic diagram illustrating the arrangement of OEMs used in this study. boundary lines depicted on the available aerial photographs. Whereas the Public Land Survey System of township, range, section, quarter section and 40­acre parcels proves to be useful in legal matters of ownership and property decription, it is not useful for consistent georeferencing because of its low resolution, often deformed grid, and inconsistent depiction on standard topographic or orthophoto quads. Given that both latirudenongitude and UTM systems work well to accurately locate grid cells, it was decided that UTM would be preferable because it is a plane grid rather than a spherical grid, and the other forms of descriptive measures--distances and elevations-were in meters rather than in degree, minutes, and seconds. Thus, the UTM coordinate system was selected for georeferencing. Cell Size Raster-based spatial analysis systems demand that cells (or "pixels", short for picture elements) be rectangular units described by their location within columns and rows in a numeric array. Although they need not be, these cells are generally square. The entire area contained within the boundary of the cell is assigned a single value, regardless of the size of the area being represented. Thus, the issue of cell size is important (Wehde 1982). If the cell is too large, it underrepresents the true variability of the phenomenon being considered and reduces the accuracy of the model. If the cell is too small, however, it may impede recording or storing information on an area large enough for meaningful analysis. Given these competing considerations, the precision and accuracy of the soil mapping, and the general goals of the research, a cell of 200 x 200 m (4 har-believed to be the smallest cell possible given the large site of study area-was selected as the standard unit of observation. This cell size would require 48,578 observations (214 rows by 227 columns of cells) to cover the study area. 38 - Gridding the Study Area and Locating Cells Accurately locating the individual 200 x 200 m cells on photographs using the UTM system was the next step. This was facilitated by hypothetically gridding the entire study area into 2 x 2 kID blocks and identifying each by the UTM coordinate of its northwest comer (e.g., Block I = UTM Zone 12,4,166,000 m N, 676,000 m E; Block 2 = 4,166,000 m N, 678,000 m E; etc.). Each block was subdivided into 10 rows and 10 columns depicting 100 200 x 200 m cells. I therefore created 2 x 2 kID grids on large, clear mylar sheets that completely overlaid each 7.5­minute orthophoto quad. The first 2 x 2 kIn block that encompassed the northwestenunost areas on the northwestenunost orthophoto quad was placed so that its northwestern comer fell on U.S.G.S-ploned UTM tick marks, which were easy to track across the entire study area. A second, smaller mylar grid scaled to fit the 2 x 2 kID blocks on the orthophoto quad was subdivided into 10 rows and 10 columns representing the 100 200 x 200 m cells. The exact location of each cell was then easily determined by reference to the coordinates of its northwest comer in relation to that of the larger 2 x 2 kIn block. A third, small mylar sheet was gridded into the 100 cell format but drawn at the scale of the aerial photograph so that the area bounded by one 200 x 200 m cell on the orthophoto quad was equivalent to the area bounded by a cell on the photograph. In this way, 502 2 x 2 kIn blocks were examined across the full extent of the study area, representing a maximum of 50,200 200 x 200 m georeferenced cell locations. Observations in excess of 48,578 (214 rows by 227 columns) were made to create a soils data plane for the study. Ninety-eight different soil map units were recorded within the study area, representing 87% of the 113 soil map units recorded for the combined Montezuma and Dolores County area. a computer data entry sheet. Township, range, and section values associated with that 2 x 2 kID block were used as general guides to selecting the proper aerial photograph from the coded set of photos on file with the SCS. Once found, distinct cultural and natural features visible on the orthophoto quad were used to position the lOO-cell mylar grid on the aerial photo and match it with its 100-celI mylar grid on the orthophoto. The evaluation process began with the cell in the first row and first column, and proceeded eastward then southward until all cells within the 2 x 2 kID block were examined. The dominant soil type within the cell was determined and its soil map unit code (Table 2.2) entered in its proper geographic position within the rows and columns on the data entry sheet. Many times, one soil type completely filled a cell, but frequently, two, three, four, and even five different soils fell within a single cell. In these situations subjective judgment was exercised to select one soil type as representing the greatest area in the cell. Of the 502 2 x 2 kIn blocks covering the entire study area, 76 were nearly or completely empty (Le., no soil data were available), and 32 had fewer than 50 cells with soil information. This left a total of 394 2 x 2 kIn blocks that were completely recorded. The replicability of the subjective method was tested informally by comparing my "call" with that of my recording assistant after we had spent approximately seven 4O-hour weeks performing these evaluations. As a test, 36 out of the 394 (9.4% of the total number of blocks) were reexamined. The person who had made the original assigrunent checked the record sheets and the alternate evaluator reestablished the placement of the 100-cell grid on the aerial photograph and determined the dominant soil type in each cell. The precision rate (Le., the number of identical calls) for anyone block ranged from 77-99% with rates per 7.5-minute orthophoto quad ranging from 85.3-98.5%, for an overall mean recorder precision for the 36-block sample of 92.9 ± 5.0% (Table 2.4). Determining and Recording Soil Types Once this system was established, the actual recording of the soils could begin. First, a 2 x 2 kIn block was identified on one of the 12 7.5-minute orthophoto quads. Its number and northwest comer coordinates were recorded on - In this way, 50,200 cells were located and evaluated as to predominant soil type and their associated values entered into a computer file. Eventually the original image, which was comprised of the set of 12 mosaiced OEMs (214 rows x 227 columns representing 1943 kIn 2 or 39- Table 2.4. Inter- and Intra-Coder Variability Rates in Soil Recording. Orthophoto Quad Map Name Ruin Canyon Negro Canyon Bowdish Canyon Pleasant View Woods Canyon Battle Rock Yellowjacket Arriola Mud Creek Trimble Point Dolores West Cortez Resampled Block Numbers and Percent of Cells in Agreement 10=95%, 11 =96%, 20=87%, 33=93%, 39=85% 57=95%, 61=94% 74=95% 92=96%, 105=96%, 122=88% 144=93%, 146=91 %, 162=92% 178=96%,199=91%,211=87% 225=96%, 227=91 %, 229=93% 275=96%, 286=92%, 291=95% 300=77%, 301=89%, 317=80%, 325=95% 341=99%, 369=98% 389=95%, 402=99%, 414=94% 421 =95%, 438=97% 468=96%, 487=99% 750 mi 2) with 48,578 cells, was trimmed along its southern margin to reflect the absence of soil data in that portion of the study area. The final image contained 45,400 cells (200 rows by 227 columns representing 1,816 km 2 or 701 mi 2). However, 7,175 cells were "empty" (Le., no soils were mapped in those locations). Some cells lacked information on natural plant productivity or water­holding capacity. Others were located on the margins of the soil plane where information on the elevation plane was lacking. Ultimately, some 8,641 cells could not be analyzed, leaving a final total of 36,759 cells within the 200 x 227 image that could be analyzed, representing 1,470 km 2 or 567 mi 2 of data MODERN CROP YIELD DATA To determine what effect soil moisture would have on agricultural productivity, data were sought on historic yields for dryland crops grown in the study area. These data were available from two sources: records of historic bean and com yield in Montezuma and Dolores counties compiled by Burns (1983:313, 315, 388), and dry bean crop yield values for 46 of the 113 soil types in Dolores and Montezuma Counties available from the Cortez area Soil Conservation Service (see Table 2.3, column 9). Corroborative data were found in environmental studies undertaken by the Dolores Archaeological Project (petersen 1987:220, 221), and an early study undertaken by the Mean Percentage in Agreement Per Quad 91.2% 94.7% 93.3% 92.0% 91.3% 93.3% 94.3% 85.3% 98.5% 96.0% 96.0% 97.5% Colorado State College (University) Agricultural Experimental Station (Brown 1938). These data helped to solve two methodological problems. The first was to establish a relation between PDSI values and agricultural yield. What is the shape, direction, and strength of any relationships between bean yield and PDSI, and between maize yield and PDSI? The second problem was to assign crop yield values to soil type and PDSI values and thereby create a quantitative "calibration" between soil type, soil moisture conditions, and agricultural productivity. Solving the second problem required solving the first. In order to address the first problem, relationships between historic bean yield, actual PDSI values, and reconstructed PDSI values; the relationships between com yield, actual PDSI values, and reconstructed PDSI values; and the relationship between bean yield and com yield needed to be explored. In order to address the second, a correlational study needed to be conducted on historic crop yield data available for selected soils with natural plant growth yield data obtained under a variety of climatic conditions for those same soils within the study area. This would establish a transfer function that could be used to assign potential crop production values to all soils, regardless of the availability of modem yield data. In sum, the crop yield and PDSI data and regression studies established the direction and strength of the relationship, and the crop -40- yon 1965 [1971 PI] (formerly Moqui NE 1955), Arriola 1965 [1973 PI], Dolores West 1965, Bowdish Canyon 1979 (formerly Moqui SW 1955), Battle Rock 1979 (formerly Moqui SE 1955), Mud Creek 1979 (formerly Cortez SW 1955), and Cortez 1965. yield/soil type/potential natural productivity data and correlation studies provided distributional information and the values used to calibrate absolute crop yield values with PDSI moisture classes. DIOITAL ELEVATION DATA MODELINO SOIL MOISTURE Digital Elevation Models Palmer Drought Severity Index Digital maps are numeric representations of cartographic data used in the analysis of spatially distributed phenomena Particularly useful are digital maps encoding elevation within the area of a 7.5­minute (1 :24,000 scale) U.S.O.S. topographic map. Such digital data sets, called Digital Elevation Models (DEMs), have become standard cartographic products of the U.S.O.S. National Mapping Program. They consist of an array of numeric values spaced at 30­m intervals georeferenced to the UTM system. This results in each numeric value having three locational attributes, an x, y, and z coordinate, representing easting and northing (horizontal) values and an elevation above sea level (vertical) value. They are often created as by­products of orthophoto quads from aerial photographs taken at an altitude of 12,192 m or 40,000 ft (l:80,OOO scale). Each digital map is tested for critical accuracy before it is included in the National Digital Cartographic Data Base and made available to the public. Vertical accuracy must be within ± 7 m of the true elevation for that point when created by orthophotographic methods (McEwen et al. 1983; Elassal and Caruso 1983) or within one­half of a contour interval when created from topographic maps (U.S. Bureau of the Budget 1979). The Palmer Drought Severity Index (PDSI) is a temporally sensitive measure of soil moisture. It was designed as an index of meteorological drought, defined as "a period of prolonged and abnormal moisture deficiency" (palmer 1965:1). Its creator, Wayne Palmer, a meteorologist working for the U.S. Weather Bureau, developed a quantitative method for calculating moisture­departure values that can be compared through time and across space. (It is important to note that PDSI values represent departures from the long­term mean condition of a place as recorded by a single weather station. As such, the absolute amount of water in a soil at one place is not being compared directly to the absolute amount of water in another soil in a different place; rather it is the magnitude of the departure from normal local conditions that is being compared. Thus, a PDSI indicating moderate drought for a soil in Iowa may be based on more soil moisture than a PDSI indicating a wet period for a soil in southwestern Colorado). Palmer assumed that local economies were adapted to the long­term mean climatic conditions of a place and that it was only Significant departures from this mean that presented adaptational problems. His immediate goal was to evaluate the frequency, duration, and magnitude of prolonged periods of abnormally dry or wet weather in the past and ultimately to forecast the likelihood of either condition in the immediate future. Today the calculation, reporting, and graphic depiction of the PDSI has become standard for the U.S. Departments of Commerce and Agriculture (e.g., NOAA's Weekly Weather and Crop Bulletin) because it is a reasonably accurate and integrative measure of the effects of past temperature and precipitation on current soil moisture conditions. Whereas it is primarily a meteorological For many portions of the U.S., DEMs may be purchased from the U.S.O.S. National Cartographic Information Center. They are stored on magnetic tape and various formatting options may be specified by the user; I used an unlabeled 9­track tape, at a density of 6250 BPI, with a fixed blocksize of 7200, in ASCII character mode. This study required purchase of 12 adjacent DEMs: Ruin Canyon 1979 (formerly Cedar 3 SW 1955), Pleasant View 1965 [1971 PI], Yellowjacket 1965 [1973 PI], Trimble Point 1965 [1973 PI], Negro Canyon 1979 (formerly Moqui NW 1955), Woods Can­ 41 index, it has significant implications for plant growth. hydrological supplies, and human economic systems. including agriculture. Palmer utilized a water-balance or hydrological accounting approach to modeling soil moisture conditions (palmer 1965:6). Both sup-ply and demand of water are considered. Water supply to a given soil is modeled by precipitation (from snow. rain. and recharge from other water sources) and water already stored in a soil profile. Water demand for a given soil is modeled by potential evapotranspiration (the process of transferring soil moisture to the atmosphere by evaporation of water and transpiration from plants). by potential moisture needed to recharge soils to their normal waterholding levels. and by runoff required to maintain water in rivers. lakes. and reservoirs. Calculation of the PDSI requires specific climatic. locational. and soil moisture data These include monthly total precipitation and monthly mean temperature values consistently recorded at known stations near the area of interest. These climatic records must also incorporate enough time to embrace characteristic trends as we)) as significant variation occurring in the area so that reasonably accurate mean climatic values can be computed. The PDSI calculations also require information on the global location of the study area. in the form of monthly daylight adjustment values, so that potential evapotranspiration rates can be computed. Fmally, data on the available waterholding capacity of soils used in the study area must be estimated or provided. Palmer's model divides a soil profile into an upper soil layer equivalent to the plow zone, and a lower soil layer that includes the effective root zone of economic plants. In general, researchers who derive PDSI consider this upper zone to be six inches thick and assume it contains one inch of water at field capacity. The depth of the remaining lower soil section normally is not specified but it is assumed to contain five inches of water. A schematic diagram of the Palmer waterbalance model would portray incoming moisture filling the upper soil layer until it reaches its field capacity, then filling or recharging the lower layer until it reaches its field capacity. Only then can excess water run off the ground surface. Similarly. evapotranspiration takes place first at the surface layer and removes all of its stored moisture before any moisture is withdrawn from the lower layer. Details of the method used to derive PDSI are contained in Palmer (1965), but in broad outline the procedure is as follows. PDSls are expressed as negative (dry) and positive (wet) real numbers (e.g.• -3.23, +1.75) and result from a number of steps. First, all relevant values in the hydrological accounting are calculated on a monthly basis for an adequately long and representative period. The minimum number of years is not specified by Palmer. but 30 years is considered minimum by researchers who regularly use PDSI in paleoenvironmental reconstructions (Martin Rose, personal com-munication, April 1989). Second, certain der-ived mean values are compared to create ratios or coefficients that weight the key hydrologic variables in such a way that subsequent values that are "Climatically Appropriate For Existing Conditions" (CAFEC) are produced. These CAFEC values reflect the long-term mean conditions for evapotranspiration. recharge, runoff, and so on, and therefore reflect the average climate of the area. Third, the entire time series is reanalyzed using these coefficients and CAFEC values to derive a CAFEC value for overall moisture required to match "normal" climatic conditions for each month in the time series. Fourth, departures (d) from the CAFEC value are calculated by subtracting the CAFEC value for precipitation for a given month (now used as the theoretical or expected value) from the actual value of precipitation recorded for that month. Fifth, these departures are converted to indices of moisture anomaly (z) by multiplying this quantity by a climatic constant (k) that represents the ratio of local moisture demand to local moisture supply for a given time period. Finally, these moisture anomaly indices are converted into Palmer Drought Severity Indices (x) by incorporating the effects of the previous month's soil moisture index on the current month's index through an equation that simulates lag time. In this way, the effects of antecedent meteorological and soil moisture conditions are reflected in subsequent conditions. -42- While PDSI is used frequently by dendro- climatologists (Cook 1982; Cook andJacoby 1977; Meko et al. 1980; Mitchell et al. 1979) and dendroecologists (Graybill 1989; Zahner et al. 1989), it has only recently been used in archaeological applications (Rose et al. 1982). The research described in this dissertation is a new use of the PDSI to link long­tenn tree­ring chronologies and estimates of agricultural productivity in the prehistoric Southwest. Data Sources for Calculating the PDSI Palmer's PDSI procedure has been implemented as a 800+ line FORTRAN computer program calIed PDSI or DROUGHT by the Laboratory of Tree-Ring Research, University of Arizona This program was used to generate the PDSI values used in this study. The requisite climatic data were obtained from Climatological Data Annual Summaries for Colorado and Utah, published by the National Oceanic and Atmospheric Administration (l987a, 1987b). Instead of using a single weather station to model the precipitation and temperature conditions in the study area, a region exhibiting significant relief and ranging in elevation from 1,500-3,012 m (4,921-9,882 ft), I decided to use multiple stations with each representing a different range of elevations in the study area. Each selected station had to be near the study area, have precipitation and temperature records of adequate length, and be representative of the regional climate of the southeastern portion of the Colorado Plateau. Data from all of the stations initially considered were statistically examined by the Laboratory of Tree-Ring Research prior to the commencement of this research for the purpose of detecting "inhomogeneities in the records and unusual trends" (Graybill 1989:9). Some (e.g., YelIowjacket) were found to be too short; some (e.g., Dolores and Mancos) possessed either precipitation or temperature records but not the other; and others (e.g., Durango) had significantly deviant trends. For the usable data sets any missing values were estimated with linear regression (Graybill 1989:9). Eventually, five stations were selected on the basis of contrasting elevation, proximity to the center of the study area, and the proportion of the study area that could be modeled by that weather station. Dis- play capability and tabular data output functions of the GIS made this an easy task. While several combinations ranging from three to five stations were acceptable, the five stations fmally selected were Ft Lewis, Mesa Verde, Ignacio, and Cortez, alI in Colorado, and Bluff, Utah (Table 2.5, Figure 2.2). Daylight adjustment values appropriate to these five stations were calculated by dividing the "mean possible monthly duration of sunlight in the northern hemisphere expressed in units of 12 hours" (Mather 1977:364, Table 19) by the number of days in a month. Available water capacity values (AWC) for the upper six inches and the lower soil section for each soil were calculated for all 113 soils within SCS jurisdiction (Table 2.6). SCS data on average soil depth and water-holding capacity per subsoil unit (Table 2.3) were used to derive the values required by the PDSI model. This derivation of actual AWC values is generally not done by the meteorologists, paleoclimatologists, and dendrochronologists who use the PDSI to model soil moisture. Instead they usually use the one inch and five inch approximations (e.g., Meko et al. 1980) suggested by Palmer. I hoped that more accurate estimates of AWC would lead to more realistic modeling of moisture stored and withdrawn from individual soils in the study area and that these could be portrayed easily with GIS technology. This seemed to be especially desirable since there are marked differences in soil quality, texture, depth and distribution across the area Further, the purpose of reconstructing PDSI values for the study area was not only to identify any episodes of abnormally dry or wet weather but also to identify the places where these occurred to form the basis for future models of settlement This ｰｲｯ｣･､ｵｾｷｩｴｨ actual waterholding values computed for soils in the study area--produced markedly different PDSls than would have been obtained if only the standard one inch and five inches of water were assumed for the upper and lower soil layers. -43- TREE-RING DATA The tree-ring data were eigenvector Table 2.5. Weather Stations Used to Reconstruct PDSI. NOAAa Index No. 3016 Distance to Precip. Temp. Weather Study Area Data Data Elevation Center Longitude Begins Latitude Station Begins 03' 2,316 m 95km 1912 108 Ft. Lewis, La 37" 14' 1921 (7,600 ft) (GO mi) Plata Co., CO 2,169 m 40km 1923 108 29' 37 12' 1923 5531 Mesa (7,115 ft) (25 mi) Verde, Montezuma Co., CO 168 km 1,969 m 1914 107 38' 1914 37 08' 4250 Ignacio, La (105 mi) (6,460 ft) Plata Co., CO 18 km 1,893 m 1931 108 33' 1930 37" 22' Cortez, 1886 (11 mi) (6,212 ft) Montezuma Co., CO 1,315 m 105 km 1927 1927 109 33' 37 17' Bluff, San 0788 (4,315 ft) (66 mi) Juan Co., UT aNational Ocenanic and Atmospheric Administration (U.S. Environmental Data and Information Service, National Climatic Data Center). 0 0 0 0 0 0 0 0 UT co + + c ++ NM Figure 2.2. Location of the five weather stations. Lewis. a, Bluff; C, Cortez; I, Ignacio; M, Mesa Verde; F, Ft. -44- Table 2.6. Water-Holding Characteristics of Soils. Soil Code 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 Soil Map Unit A1B A2B A3B A4 A41 A4A A4B A5C A5D A5Wa A6A AOAB AOB AOC AQa BD C2CE C2D C2F C2V C3CD C3E C5C C7D CP DOCE GP H1CD H1DC H6D M1CE M2C M2CE M2DD M2DD1 M5E M6C M7E M8D OAB P1C R1C R1D R1HD R1HC R1HD R3B R3B1 Est. Soil Depth (in.) .2:,60 .2:,60 ｾＶＰ ＰＶｾ .2:,60 ｾＶＰ .2:,60 .2:,60 ＰＶｾ ＰＶｾ .2:,60 ｾＶＰ .2:,60 .2:,60 ＰＶｾ 0-60 6-20 6-20 6-60 6-20 ＰＶｾ ｾＴＰ 18-60 10-60 NA 20-40 NA 40-60 20-40 ＰＶｾ 20-40 10-60 10-60 10-40 10-40 ｾＸＰ 40-60 40-80 10-60 ｾＶＰ ｾＸＰ ＰＸｾ ＰＸｾ ＰＶｾ ＰＶｾ ＰＶｾ .2:,60 .&:.60 Upper Soil Layer (6 in.) NA 1.14 0.54 0.96 0.90 0.57 0.99 1.14 1.14 0.93 0.96 1.14 1.14 1.14 1.07 0.00 0.89 0.93 0.97 0.89 1.14 1.14 0.99 1.02 0.00 0.66 0.00 1.08 0.90 0.56 0.48 0.53 0.40 0.85 0.85 1.14 0.80 0.80 0.85 0.93 0.36 1.14 1.14 1.02 1.02 1.02 0.90 0.90 Lower Soil Layer Total Field (6-80 in.) Capacity (in.) NA NA 9.18 10.32 5.97 6.51 8.37 9.33 7.83 8.73 4.63 5.20 8.91 9.90 9.18 10.32 9.18 10.32 7.69 8.62 6.31 7.27 10.26 11.40 10.26 11.40 10.26 11.40 9.99 11.06 0.00 NA 2.08 2.97 2.08 3.01 4.39 5.36 2.08 2.97 8.90 10.04 8.90 10.04 5.68 6.67 5.81 6.83 0.00 NA 1.98 2.64 0.00 NA 5.40 6.48 3.24 4.14 4.05 4.61 0.53 1.01 2.61 3.14 1.98 2.38 2.95 3.80 2.95 3.80 10.45 11.59 3.94 4.74 7.98 8.78 5.17 6.02 7.35 8.28 3.77 4.13 13.66 14.80 13.66 14.80 7.25 8.27 7.25 8.27 7.25 8.27 8.05 8.95 8.05 8.95 ­45 - AWCClass (1-11 ) NA 9 1 8 7 1 8 9 9 7 6 10 10 10 9 NA 2 2 4 2 8 8 5 5 NA 1 NA 5 3 1 1 1 1 2 2 10 3 7 5 7 1 11 11 7 7 7 8 S Table 2.6. (Continued). Soil Code 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 Soil Map Unit R3C R3C1 R3D R3D1 R4C R4C1 R4D R7C R7D R7D1 R8B RHC RL ROB ROB1 ROC ROC1 ROD· ROHC ROHD ROLC ROLD ROLD1 S1C S5 S51 S53 S5D S6 S6C S6E S7 S8C T1B T2C T3E T4B T6 T6B T6C T6D V1B V1C V1D V2B V2C V2S V2Va V2Wa Est. Soil Depth (in.) 2,60 2,60 2,60 2,60 20-60 20-60 20-60 20-40 20-40 20-40 2,60 20-60 NA 2,60 2,60 2,60 2,60 2,60 20-40 20-40 20-40 20-40 20-40 2,60 40-60 2,60 2,80 2,60 40-60 20-60 20-40 20-60 20-80 2,60 2,60 10-60 2,60 2,60 2,60 2,60 2,60 2,60 2,60 2,60 2,60 2,60 2,60 2,60 2,60 Upper Soil Layer (6 in.) 0.90 0.90 0.90 0.90 0.97 0.97 0.96 1.02 1.02 1.02 0.84 0.95 0.00 0.99 0.99 0.99 0.99 0.99 1.02 1.02 1.02 1.02 1.02 0.90 1.02 1.14 1.14 1.14 0.75 1.06 0.80 1.06 1.09 1.05 0.45 0.38 0.93 1.14 1.13 1.13 1.14 1.05 1.05 1.05 0.93 0.87 0.81 0.96 0.96 Lower Soil Layer Total Field (6-80 in.) Capacity (in.) 8.05 8.95 8.05 8.95 8.05 8.95 8.05 8.95 7.06 8.03 7.06 8.03 5.74 6.70 4.48 5.50 4.48 5.50 4.48 5.50 8.28 9.12 6.92 7.87 0.00 0.00 10.26 11.25 10.26 11.25 10.26 11.25 11.25 10.26 10.26 11.25 5.40 6.42 5.40 6.42 4.67 3.65 4.67 3.65 4.67 3.65 9.33 8.43 7.80 8.82 9.15 8.01 10.45 11.59 8.01 9.15 5.51 6.93 5.87 6.93 4.25 5.05 5.40 6.46 9.51 8.42 6.98 5.93 5.07 4.62 0.75 1.13 3.39 4.32 8.81 9.95 9.45 8.32 9.45 8.32 8.81 9.95 9.45 10.50 10.50 9.45 9.45 10.50 9.30 8.37 8.37 9.24 8.10 7.29 8.47 9.43 7.52 6.56 -46- AWCClass (1-11) 8 8 8 8 7 7 5 4 4 4 8 6 NA 10 10 10 10 10 5 5 3 3 3 8 7 8 10 8 5 5 4 5 8 5 1 1 3 8 8 8 8 9 9 9 8 8 7 8 6 Table 2.6. (Concluded). Soil Code 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 Soil Map Unit V4C vas VOC vocs VOCC was wassa wac XC6D XC6E XH2D XH2E XH3E XM1F XM3D Z Est. Soil Depth (in.) ｾＶＰ ｾＶＰ ｾＶＰ ｾＶＰ ＰＶｾ ｾＶＰ ｾＶＰ ｾＶＰ 10-20 4-60 ＰＶｾ 10-60 10-60 10-20 10-20 ｾＶＰ Upper Soil Layer (6 in.) 1.00 0.84 0.84 1.14 1.14 1.08 0.93 1.08 0.84 0.39 0.53 0.37 0.40 0.30 0.38 0.70 Lower Soil Layer Total Field (6-80 in.) Capacity (in.) 8.54 9.54 6.29 7.13 6.29 7.13 6.44 7.58 6.44 7.58 9.72 10.80 8.89 9.82 9.72 10.80 2.04 2.88 0.80 1.19 4.61 5.14 3.58 3.21 3.55 3.95 0.31 0.61 0.97 0.59 7.18 6.48 AWCClass (1-11) 8 6 6 6 6 9 8 9 2 1 1 1 1 1 1 1 aArtificially wet and possibly saline from irrigation water. amplitudes (factor scores) that resulted from a principal components analysis of seven expanded tree-ring chronologies from the Colorado Plateau. The data set, "SWOLD7" (Southwest Old Seven), was created by Martin Rose in conjWlction with research Wldertaken by the Laboratory of Tree-Ring Research. The original data, as well as the data manipulation and reduction techniques used to create the 107D-year (row) by seven variable (column) amplitude matrix, are described by Rose et al. (1982). The tree-ring chronologies used by Rose et al. (1982) are called "expanded chronologies" because they result from merging chronologies created with archaeological specimens with chronologies created from living trees (Dean and Robinson 1977). The merging was possible because of a temporal overlap between the recent end of the archaeological series and the older end of the living tree series. Rigorous statistical tests were performed on the period of overlap to ascertain whether the archaeological .and living-tree sequences represented a single statistical population, and the tests confirmed the acceptability of merging these chronologies (Rose et al. 1982). Methods for constructing tree-ring chronologies are discussed in detail in Stokes and Smiley (1968); Fritts (1976); Hughes et al. (1982). The tree-ring chronologies themselves are already standardized to facilitate comparison and to emphasize the macroclimatic factors that affect tree growth. Standardization is accomplished in several ways. Dated, annual ring-width measurements are fitted to growth curves, removing the biological effects of aging on growth rings, and are then converted to departures from the growth curve for that sample. Thereafter, ring-width indices for all samples for a particular site or locale are averaged to create a mean index value per year for that location. Thus, a tree-ring chronology, from living trees or archaeological samples, is comprised of the standardized, averaged, ringwidth indices for a location or region; each value represents the mean index for all specimens included in that chronology for a given year. Rose examined many chronologies before settling on seven (Table 2.7, Figure 2.3). A battery of statistical techniques and tests was used to identify chronologies that would emphasize tree growth controlled by the macroclimatic variables of precipitation and temperature over growth affected by nonclimatic sources of variation. Rose then subjected the tree-ring indices of the seven selected chronologies to a principal components analysis -47- Table 2.7. Tree-Ring Chronologies Used in the Creation of SWOLD7 (Rose et al. 1982; years AD). Archaeoloaical Chronologies End Number Date Name 799900 1930 Santa Fe 890009 1864 Jemez Mountains 1751 803999 Gobernador Cibola 400999 1900 Chama Valley 900009 1835 Mesa Verde 799199 1276 Cebolleta Mesa 666003 1885 Living-Tree Chronologies Beginning & Number End Dates Name 1556-1972 Glorieta Mesa 343000 Paliza Canyon 383000 1658-1972 Pueblito Canyon Turkey Springs Echo Amphitheatre Bobcat Canyon and Schulman Old Tree Cebolleta Mesa Expanded Tree-Ring Chronoloaies 073000 1594-1971 Name Santa Fe Jemez Mountains Gobemador 272000 1595-1972 Cibola 400009 171000 1362- 1972 900005 061099 and 641000 1390-1971 and 1200-1963 Chama Valley Mesa Verde 303000 1662-1972 UT Cebolleta Mesa Number 700000 890000 803000 772100 666000 co + + 2 + + 3 +5 6 +i +7 NM Figure 2.3. Location of the seven tree-ring chronology stations. 1, Mesa Verde; 2, Gobernador; 3, Cibola; 4, Cebolleta Mesa; 5, Chama; 6, Jemez; 7, Santa Fe. -48- (PCA), following conventions established by the Laboratory of Tree­Ring Research (Stockton 1975; Fritts 1976; Fritts et al. 1981). This was done to further reduce the data and create a set of statistically independent (orthogonal) predictor variables that could be used for subsequent climatic reconstructions. PCA is a type of factor analysis that reexpresses original multivariate data as axes ("principal components" or "eigenvectors") of variability in the same number as the original variables. It is done in such as way that each axis is orthogonal to the others and the first axis accounts for as much of the total variance as possible, with subsequent axes accounting for decreasing amounts of the total variability. Given that the seven chronologies used were all created with an emphasis on macroclimate, and that they are all from a fairly restricted geographic region, Rose anticipated that the original variables (the seven chronologies) would be correlated to some degree. large data sets and for displaying results were essential to this research. The term "geographic information system" refers to a particular type of computer system incorporating software programs and hardware to capture, store, manipulate, analyze, retrieve, and display data referenced to geographic locations (Avery and Berlin 1985; Berry 1989a, 1989b; Burrough 1986; Clarke 1986; Kvamme 1988; McDonald and Crain 1985; Marble 1984; Parker 1988; Tomlinson et al. 1976). In contrast to computer-automated mapping and design programs (Cowen 1988; Dueker 1987), which generally focus on the manipulation and display of a single image, GIS have the ability to interrelate multiple spatial and nonspatial data sets concurrently; to create new information through combinations or transformations of original data (Collins and Moon 1981; Monmonier 1982:76-79); and to produce analytic products such as tables, charts, and graphs in addition to map-like images. Therefore, performing a PCA would produce a set of uncorrelated axes that would not only meet the statistical requirements of later manipulations but also would likely produce a few axes accounting for most of the variance and a larger number of axes that would account for small amounts of variance. Later, the axes depicting the majority of the variability could be used in stepwise, multiple regression-based climatic reconstructions and the lesser contributors ignored. In this way, a reduced and more powerful data base is produced, particularly since it has been found that the largest axes are generally those that can be directly attributed to climate (Stockton 1975; Fritts et al. 1981). There are two basic types of GIS, which are classified on the basis of how the geographic information is spatially referenced and organized. Vector-based systems encode spatially referenced data as points (e.g.• a spring), lines (e.g., a stream), or polygons (e.g., the perimeter of a lake) to defme the specific location and boundaries of spatial phenomenon. The boundaries of the polygons imply uniformity of their interiors (Berry 1989b:34). Recording data in vector format facilitates production of images that resemble ordinary maps; they are ideal for bounding discrete spatial data; they use less computer storage space than do rasterbased systems; and they are easy to "rasterize" at a variety of cell sizes. However, they require more complex algorithms to encode and manipulate data and hence operate more slowly, are less appropriate for representing continuous spatial data (e.g., elevations) and often must be converted to raster format to perform even simple analyses, such as determining area (Maffini 1987). Appendix A contains the final output generated for the PCA on SWOLD7. The values listed are eigenvector amplitudes, "the products of eigenvectors and the data from which the eigenvectors were derived expressing the importance of each eigenvector in each observation set" (Fritts 1976:531). This table was the source of tree-ring data for this study. GIS TECHNOLOGY Geographic Information System hardware and software for managing and analyzing these Raster systems, on the other hand, encode spatially referenced data to specific cell locations on a regular grid of columns and rows arbitrarily placed over a study region. Raster systems approximate the location of spatially distributed information in a probabilistic fashion with a dominant or mean value assigned to -49- the entire cell. In contrast to vector­based systems, raster systems depict interiors and only imply boundaries (Berry 1989b:34). Consequently, they are better at representing continuous spatial data that in the real world have no visible edge (e.g., slope, climatic zones). Because raster systems are cellular, the map-like images they produce are less "realistic" representations than vector display images, they do not convert accurately to vector form, and they require more computer storage space. However, they require much simpler mathematics to store, overlay, and analyze data, and they more readily produce statistical and numeric output than do vector systems (Maffini 1987). GIS technology has existed since the 1960s but it has only become a mainstream data management tool since the mid-1980s for land-, facilities-, resource-, and records-management agencies and as a scientific tool for the professional community. At least 10 recent fulllength texts and reference books have been written about the history and use of geographic information systems (Aronoff 1989; Bracken et al. 1989; Burrough 1986; Forrest et al. 1990; GIS World 1990; Huxold 1990; Ripple 1987, 1989; Star and Estes 1989; Tomlin 1990) and many articles are published arumally in a variety of professional journals, conference proceedings, and trade publications that report on the analytic capabilities and limitations of the technology, new methods of modeling, and the substantive results of specific studies. GIS technology is increasingly being used by archaeologists in the management and analysis of archaeological and environmental data. The first book-length treatment of GIS has been published (Allen et al. 1990). Recent review articles by Kvamme (1986a, 1986b, 1989) and Kvamme and Kohler (1988) detail the history of use and the potential future uses of this technology in archaeological research. Additional examples of GIS use in cultural resource inventories and database management can be found in, for example, Briuer (1988), Brown and Rubin (1982), Calamia (1986), Harris (1986, 1988); Hasenstab (1983a), and Parker (1986). For predictive models of site location and land use, see Bailey et al. (1985), Brown and Rubin (1982), Carmichael and Christensen (1988), Creamer (1985), Farley et al. (1990), Hasenstab (1983b), Kvamme (1983a, 1983b, 1984, 1985, 1986b, 1988), Kvamme and Jochim (1989), Marozas and Zack (1990), Martin and Scurry (1988), Parker (1985), Parker et al. (1986), Scholtz (1981), Wansleeben (1988), Warren (1988), and Warren et al. (1987). GIS approaches are also well-suited to simulation models focused on archaeological methods, as can be seen for the study of sampling strategies and statistical models in Kvamme (1986b); studies of surface artifact distributions (Gill and Howes 1985); and for simulation models focused on behavioral processes, such as movement and size of populations (Zubrow 1990) or agricultural productivity (Darsie 1982). This study might be described as a simulation of prehistoric agricultural potential. A dynamic model of climatic variation and soil moisture conditions has been applied to a set of spatially distributed environmental variables in order to simulate the variation in agricultural production over time and across space in order to "observe" its effect on human populations. The model affords "dynamism" by virtue of the time-depth of the annual tree-ring data from the study region. The model suggests methods and generates data that allow the size and shape of the annual agricultural niche to be delimited on a number of criteria, and it permits sustainable population size to be estimated in a number of ways. Consequently, a series of scenarios or "submodels" could be generated and explored for their congruence with known archaeological data, although this is not done here. What the simulation model has been able to do is provide insight into the role that climatic variation plays in determining soil moisture and plant growth conditions, which in tum affect the spatial and temporal distribution of farmable lands and set limits on the size of human populations over given periods of time. Without GIS it would have been impossible to model these phenomena, depict them spatially at anyone point in time, or simulate them dynamically from the perspective of a specific geographic location over time. VICAR/IBIS and EPPL7 Washington State University (WSU) is among a number of institutions that has devel- -50- mental) status for the upgraded Release 2.0. It runs under MS­DOS and was developed by the Minnesota State Planning Agency of S1. Paul. Minnesota as a commercial product for IBMcompatible machines with either EGA or VGA display capabilities. EPPL7 has four major program components that permit encoding and capture of digital data (DIGITIZE), analysis and manipulation of spatial data (EPPL). display of spatial data (DISPLAY) and output of spatial data (DOTPLOT). It can also import files from and export files to other computer systems. It is easier to use than VICAR and has more GIS capability in most functional areas. Although normally an interactive system, EPPL7 can also perform batch­type operations in which lengthy instruction sets are read from disk. In EPPL7, data can be entered into the system as stored files on diskettes, entered directly from the keyboard, or entered via digitizing programs linked to external digitizing equipment on a mouse. Data are stored as files and may be manipulated and transformed in a variety of ways by evoking various executable EPPL7 commands tailored to user requirements through specification of parameters and options. Spatial images can be immediately displayed on color monitors and modified interactively. Tabular data output and digital images may be plotted on standard personal computing printing devices. Although no internal photographic equipment is installed to record spatial images in color on either EPPL7 or VICAR, it is quite easy to photograph images on a color monitor screen using standard camera and tripod setups. oped a considerable facility devoted to image processing and geographic analysis. WSU's Digital Image Analysis Laboratory (DIAL) has obtained a number of image processing and spatial analysis systems for which instruction and research support are available. DIAL was established in the late 1970s with the installation of a single, powerful mainframe batchprocessing system developed at NASA's Jet Propulsion Laboratory in Pasadena California called VICAR/IBIS (Hart and Wherry 1984). These are acronyms for "Video Image Communication and Retrieval" and a subset appli.cation called "Image Based Information Systern." Since that time. DIAL has upgraded its capabilities for interactive image analysis and GIS software and hardware, but at the time of this study, only VICAR/IBIS and EPPL7 were available. From DIAL's system offerings, an IBM DOS­based GIS software package was selected to be the primary system for use in this research. EPPL7 was selected because of its greater accessibility (it is installed on multiple IBM­compatible machines), its simplicity, and its appropriateness to the tasks required. VICAR/IBIS was used only to process and mosaic the OEMs. VICAR/IBIS is a set of computer programs designed specifically to process digital image data It is raster­based. A special set of programs called ffiIS was added later to analyze previously processed earth imagery (Hart and Wherry 1984:1). It is considered a user­oriented system in which FORTRAN and Assembler language programs are invoked by specifying VICAR control language statements and program options. VICAR runs on IBM­environment mainframe computers and at the present time resides within WSU's IBM 3090­200 system. Input to the system may take the form of digital data stored on tape or entered by keyboard. Manipulation of data takes place through batch programming. Image display is accomplished by transmitting the image data to image monitors such as the International Imaging Systems (I2S) color monitor. Permanent output is routed to the mainframe printers and related plotting devices. Data storage is accomplished through disk and tape on the IBM mainframe. VICAR and EPPL7 not only made this research possible by their ability to capture, store, and manipulate vast quantities of spatial and nonspatial data, but they also established the conditions under which the data had to be structured. Three examples illustrate this second point First, being raster or cell­based systems, both VICAR and EPPL7 ultimately must use data in raster format. TIlis was an important consideration in the way soils data were collected. Second, OEMs when purchased from the USGS are stored digitally on magnetic tape. This necessitates access to a tape drive that can load and transmit data to the appropriate program. VICAR is a mainframe program with pre­established connections to mainframe data transfer programs and tape drives; EPPL7 is a raster­based GIS that, at the time of use, was in beta release (i.e., develop- ­ 51 this was clearly the system of choice with regard to capturing digital elevation data and the converting them to fonnats readable by EPPL7. Third, EPPL7 runs on widely available personal computers and is much easier to use than VICAR, but it has limitations on the number of classes of data in anyone thematic layer that can be manipulated at one time. The maximum number of classes is 255. Therefore, some data needed to be reclassified in order to be analyzed and displayed; for example, elevations depicted on the mosaiced set of 12 OEMS ranged from a low of 1,500 m to a high of 3,012 m, representing 1,512 different elevation values. In order to use the full range and display these data, these 1,512 different values had to be reduced to fewer than 255. In the end, 189 elevation classes were established with each class representing eight meters. In sum, geographic infonnation systems were an essential aspect of this research. The capabilities and limitations of the available systems were also important. EPPL7 was selected as the major system of spatial analysis with VICAR providing auxiliary assistance in the initial capture and organization of the elevation and soil data. EPPL7 was relatively easy to learn and was more accessible given its residence on multiple personal computer systems. Sufficient as a spatial analysis system, it has limitations as a database management and statistical analysis system. Consequently, large data sets containing nonspatial tabular information created by EPPL7 were transferred to an ffiM mainframe for editing and statistical analysis. The use of GIS technology depended on the existence of other computer systems for the creation of input data sets and for subsequent analysis of data using statistical software. ARCHAEOLOGICAL DATA Block Survey Data Archaeological information from extensive block surveys in the study area was sought to generate popUlation estimates. These archaeological estimates of actual population would be compared with the environmentally-conditioned estimates of potential population derived from the model. The comparison would permit an assessment of how close the prehistoric population estimated from the archaeological data was to the periodic and long-term values for optimum and maximum carrying capacity estimated from data generated by the environmental model. This would provide a preliminary test of the model's utility for providing insight into issues relating to population/resource relationships. Within the study area, three large-scale, contiguous block surveys had taken place: on Mockingbird Mesa (survey conducted 19811984, reported by Fetterman and Honeycutt [1987]); in the Dolores Archaeological Project area (conducted 1972-1973 and 1978-1981, reported by Breternitz and Martin [1973], Orcutt and Goulding [1986], Schlanger [1986a], and Schlanger and Harden [1986]); and in the Sand Canyon Locality (1986-1987, reported by Van West [1986], Van West et al. [1987], and Adler [1988]). Other programs of intensive block surveys having portions of their survey areas within the study area were considered but not used. Sur- \ vey has been conducted in the Hovenweep National Monument area on NPS and BLM monument parcels and within samples of 40acre quadrats between Montezuma Creek/Cross Canyon on the west and McElIno Canyon/Yellow Jacket Canyon on the east (Winter 1975, 1976). Contiguous block survey has occurred on Squaw Point and Cow Mesa between Papoose Canyon on the west and Ruin Canyon on the east (Neily 1983). Survey also has taken place in 100 8o-acre quadrats selected by a stratified random sample of locations within the BLM's Sacred Mountain Planning Unit in Dolores and Montezuma Counties (Chandler et al. 1980). These last three datasets were rejected as sources of data for this research because of lack of complete records on sites or survey coverage, lack of data on room counts or rubble area, lack of precise dating of archaeological components, noncontiguous survey coverage accompanying quadrat-based sample surveys, and redundant elevational range coverage. The Dolores Archaeological Project data set was also excluded because so few of the more than 1,600 sites recorded were assigned -52- to the A.D. 900­1300 period of interest, and so few of these provided data useful for estimating population (Schlanger 1985: 142, 152154, 158-164, 168, 174, 181, 193-195). Given the large number of sites on both Mockingbird Mesa and in the Sand Canyon Locality dating to the A.D. 900-1300 period for which population estimates have already been derived, and the contrasting elevational settings of these two areas-Mockingbird Mesa at 1,859-1,951 m (6,lO()",{),400 ft) and Sand Canyon area at 2,012-2,064 m (6,600-7,100 ft)-it was decided that these two surveys would provide sufficient data for the preliminary test (Figure 1.1). Schlanger's estimates of population in the Mockingbird Mesa Survey Locality (Schlanger 1985:194, 199) and Adler's estimates of population for the Sand Canyon Survey Locality (Adler, personal communication 1989; Adler 1990) were used as given. ｾｊｄｩｒＭ･ ｲｔ Dated Sites The files of the Laboratory of Tree-Ring Research were searched for evidence of well- documented and excavated habitation sites dating to the A.D. 900-1300 period in the study area that had produced sets of tree-ring dates that included at least two cutting dates and at least one cluster of dates. In addition, supporting evidence that the site was likely to have been occupied during the tree-ring dated interval was sought from discussions of chronology in published reports. These sites would provide locational and temporal information that might be useful in exploring the role that agricultural potential in the site's immediate "catchmnent area" might have played in the decision to establish or expand a habitation in that location at a given time. The study area is entirely encompassed within the Colorado "V" quadrangle designated by the Laboratory of Tree-Ring Research. Within the 1816 km 2 study area, 46 archaeological sites associated with datable tree-ring samples were identified. Twelve of these were reported in Robinson and Harrill (1974) (TRL# 1, 2,3,4,5,7,15,21,22,24, 125, 126); the remaining 34 (TRL# 169, 170, 176, 177, 178, 179, 181, 182, 183, 184, 187, Table 2.8. Tree-Ring Dated Sites. TRL No.a 231 237 Site No. 5MT8371 5MT8839 235 5MT2433 182 5MT4126 b 183 5MT6970 b 2 5MT1566 170 5MT2149 187 233 5MT3834 5MT765 c Site Name None Norton House Aulston Pueblo Ida Jean Ruin Wallace Ruin Lowry Ruin Escalante Ruin Mustoe Site Sand Canyon Pueblo Occupation Period (Years A.D.) 935-950 1029-1048 No. of Tree-Ring Samples 5 19 No. of Cutting Dates 2 8 25 5 32 24 1045-1071 44 15 1086-1120 53 33 1124-1138 25 9 1173-1231 1252-1274 24 329 9 124 References Dykeman 1986 Fuller 1987; Kuckelman 1988 Kane 1975a; Monis 1030-1050 1986a Brisbin and Brisbin 1973 1124 Bradley 1974, 1984, 1988b Martin 1936; White and Breternitz 1976 Hallasi 1979; White and Bretemitz 1979 Gould 1982 Adams 1985; Bradley 1986, 1987, 1988a, 1992; Kleidon and Bradley 1989 aSite number of the Laboratory of Tree-Ring Research, University of Arizona, Tucson. bSMT4126 and SMT6970 are combined as a single site in a later step to represent an A.D. 1045-1124 occupation. CData as of April 1990; SMT76S currently under investigation by Crow Canyon Archaeological Center, Cortez, CO. Note: Tree-ring dates are listed by specimen for each site in Tables 5.5, 5.8, 5.11, 5.14, 5.17, 5.20, 5.23, and S.26. -53 - 188, 194, 195,211, 216, 217, 228, 231, 232, 233,234,235,236,237,238,239,240,241, 242, 243, 247, 253, 254) are documented in unpublished files at the lab. For the most part, the earliest cutting date from a site was used as the beginning occupation date. The last date, regardless of its status as a cutting date or noncutting date, was used as the end date of the occupation/use period. Within this bracketed period, a cluster of dates was sought. Nine of the 46 sites met all these criteria. Four of these may be characterized as small villages, hamlets, or homesteads; (5MT8371, 5MT8839, 5MT2433, and 5MT3834), and five as larger villages or "central places" (5MT6970, 5MT4l26, 5MT1566, 5MT2149, and 5M1765) (Table 2.8, Figure 1.1). It was hoped that these sites would possess the following characteristics. They would be associated with occupational dates that are reasonably representative of the occupational history of the region. Also, they would be dispersed geographically and located in various elevational settings. For the most part, these expectations were met. The early 900s were represented by a single site, the middle 1000s by three, the late 10005 by two, the early 1100s by three, the late 1100s by one, the early 1200s by one, and the middle 1200s by one site. Lacking representation are the middle and late 900s, the early WOOs, middle 1100s, and the very late 1200s. The sites were somewhat dispersed across the study area, although all were essentially mesa top sites, and none was located in deeply entrenched canyons. None of the nine sites was located in the lowest elevational range modeled by Bluff climatic reconstructions «1,605 m/5,266 ft, representing 2.7% of the study area), or in the highest elevational range modeled by Ft. Lewis climatic reconstructions (>2,244 ,362 ft, representing 2.4% of the study area). Three of the nine were located within the elevational range modeled by Cortez climatic reconstructions (1,6051,932 m/4,26(r-6,339 ft, representing 41.0% of the study area), two were within the elevations modeled by Ignacio climatic reconstructions (1,933-2;068 m or 6,442-6,785 ft, representing 33.4% of the study area), and four were within the elevational range modeled by Mesa Verde climatic reconstructions (2,069-2,244 mor 6,788-7,363 ft, representing 20.5% of the study area). While all sites contained tree-ring samples with cutting dates, the clusters of dates were not always as clear as would have been liked; however, architectural and artifactual data also support the dates assigned in Table 2.8. mn In sum, the nine sites that met the minimum criteria include four small habitations and five larger habitations. Collectively they are reasonably representative of the occupational history of the area. Whereas the highest and lowest elevations taking in only 5% of the study area are not represented in the sample of sites, the middle ranges that take in some 95% of the elevations in the study area are represented. Finally, the sites are somewhat dispersed geographically. -54- 3 BUILDING THE MODEL: PRELIMINARY DATA ANALYSES This chapter explains the process of reconstructing Palmer Drought Severity Indices for representative soils in the study area using modem precipitation and temperature data and expanded tree­ring chronologies. It also describes the processing of the elevation and soils data for use in the GIS­integrated analysis. Finally, it details the assumptions and methods used to link reconstructed annual soil moisture levels, as modeled by the POSI, with levels of agricultural production. which the historic record is divided into two blocks of time. One block is used for building the regression model and the other is used for testing the initial regression equation. This second step represents the "initial calibration" of the POSI to the tree­ring data. PALMER DROUGHT SEVERITY INDEX RECONSTRUcnON Third, the "verification" of the initial calibration equation takes place. Correlation coefficients and probability tests are used to assess the strength of the initial equations to faithfully predict the actual values generated by the instrumented POSI data for years not used in the creation of the original calibration equations. The reconstruction of soil moisture by the Palmer Drought Severity Index (POSI) is a multi­step process (Fritts 1976; Graybill 1989; Hughes et al. 1982; Meko et al. 1980; Rose et al. 1982). Subsequent to data acquisition, the process begins with calculation of POSI values for specific soils using historic climatic data from a specific weather station. In this step, actual POSI values are determined for every month and every year in the instrumented series. Fourth, a "full calibration" period regression equation is created for the entire period of instrumented record that overlaps with the modem end of the expanded tree­ring data. The product of this step is the final transfer function to be used to retrodict the pre­instrumented POSI values. Because this regression uses more years of data and often incoIporates more climatic variation, it always produces higher correlation coefficients, than does the initial calibration. Second, POSI values from a selected month are correlated with tree­ring values for a common period of time in order to generate an initial multiple regression equation that can be used as a transfer function to predict (or in this case, "retrodict") POSI values in the prehistoric and pre­instrumented time period. Here the tree­ring data are treated as independent variables and the POSI value as the dependent variable. In this second step, a portion of the historic record is used to create the initial regression equation. Both split­record and random year methods have been used, but the more common is the split­record approach in Last, the retrodiction of the entire POSI series is accomplished by applying the transfer function to the full set of tree­ring values. Here the tree­ring data are used as the predictor or independent variables and the POSI is the predicted or dependent variable. PDSI Calculation Input parameters for program "POSI" include a) the total precipitation and mean temperature values expressed as inches and average degrees Fahrenheit per month for each Table 3.1. Soils Selected to Represent the 11 Classes of Soil Moisture Based on Water Held in the Upper and Lower Soil Layers. Representative Soil Type XH2D M2DD ROLC R7C ROHC VOCC S5 V4C V1B ROB R1C Nonirrigated Capability Upper Soil Class Layer (in.) 7s .53 6s .85 3e 1.02 3e 1.02 1.02 3e 1.14 3c 1.02 4c 1.00 4e 1.05 4c 0.99 3c 1.14 4c Lower Soil Layer (in.) 4.61 2.95 3.65 4.48 5.40 6.44 7.80 8.54 9.45 10.26 13.66 weather station used, b) the total number of years to be used in calculating PDSI values, c) the monthly daylight adjustment values for the latitude of the weather station, and d) the amounts of water held in the upper soil layer and lower soil layer expressed as inches. Experimentation with varying the amount of water potentially held in the upper and lower soil layers, using actual values for available water capacity for each of the 113 soil map units, revealed that quite different results would obtain by using estimates other than the one inch and five inch approximations suggested by Palmer (1965). Therefore, it was decided to use actual values for a reduced number of soils rather than the standard model. This was done by assigning each soil to a class based on water held in the upper and lower layers, plotting these values graphically, and selecting the single soil that fell closest to the center of its class as being the representative of that group of soils. In this manner, a total of 11 different soils was selected to represent the 113 original soils recorded in the Montezuma-Dolores County area (Table 3.1, Figure 3.1). The PDSI program was run 55 times, calculating PDSI values for each of the 11 AWC soil classes with each of the five sets of climatic data (Bluff, Ignacio, Cortez, Mesa Verde, and Ft. Lewis). An example output from one of these 55 PDSI calculations is presented in the original study as appendix B (Van West 1990:384- Upper Soil Class (in.) 0.50 ±.25 1.00 ±.25 1.00 ±.25 1.00 ±.25 1.00 ±.25 1.00 ±.25 1.00 ±.25 1.00 ±.25 1.00 ±.25 1.00 ±.25 1.00 ±.25 Lower Soil Class (in.) 0.00-6.50 2.00-2.99 3.00-3.99 4.00-4.99 5.00-5.99 6.00-6.99 7.00-7.99 8.00-8.99 9.00-9.99 10.00-10.99 13.00-13.99 AWC Class 1 2 3 4 5 6 7 8 9 10 11 408). A single monthly PDSI table from each of these programs is used as input to the next program described below. Calibration. verification. and Creation of the Transfer Function The program that performs the initial splitperiod calibration and verification as well as the final full-period calibration, called "TERIFY," was developed by Donald Graybill and Michael McCarthy of the Laboratory of TreeRing Research. Input parameters include: a) identification of the named source of tree-ring values used as the independent data set, b) a beginning and ending date (year) for the independent data, which will also be the range of the retrodicted data set, c) identification of which variables in the tree-ring data set will be used, d) specification of the maximum number of independent variables from the tree-ring data set that could be incorporated into multiple regression equations, e) specification of the percentage of explained variance that must be present in a variable before it is allowed to enter a multiple regression equation, f) identification of the named instrumented data source of PDSI values used as the dependent data set and which month is to be reconstructed, g) identification of the beginning and ending dates for the calibration process, h) selection of the verification method to be used-split-record or random, and i) selection of the number of years to be used in the initial calibration. -56- Available Water Capacity Of The Lower Soil Layer (6-80 inches) 0.0" 1.0" 1.25" 2.0" 3.0" 4.0" /2 /3 14 I I I I I I I I I I , I I I I 5.0" 6.0" 5 I6 I I ROLe I R7C I ROLO I RRTTOOI I I I ­ ; 100" Q) ­ ..c o g 0 0. ­ o .... U I-C20 I I C2eE ,C600 I I I 0 I ,0 C2v ｾｷ ·u I ｒｏｌｾＱ ｒ＿ｈｧｾ｣ｯ e2F I 1 ｾｧ I ｈｉｃｾ 5T- ｔｉｂｾＱ AOHe e 56C I, .1 I I I -HIDe I \ 1 I I I I I oM80 I MGC.' ｾ .S6E I ｾ __ｯｳｾ ｾ ﾰＱｾＰ I I ­ ­ / Ｍｾ A28 ·S8C I V4C I I .. Ｑﾷ ｓｾＳＱ I ＺｧｾＮ A41· 1• I ｾｯｶＲｃ ｂＲｾ 12.0" 13.0" AD8 I RIC RIO ｍｾ｛ｉ 0 I .. I I "'48 VIO ｂ ｾＧ［ I" I I I I ｾｧｬ ］ｧｾＬ AOO ·SIC , I I , I I I I oR88 I I I I I I I I VlS I -M7E- R)B I A3Bl I ­ ­ _!­ ｾＬｭ oZ 14.0" "V18 jAO VIC I I A4: V2V I OAS ＧｷｾＺ ＺｧｾＸＧ I I ａｾｏ ｓｾ･ I ｯｾｧ 1' ｾＮａ ｃ I .1 -R4CI RHCj 160 r 11.0" /0 I I I I ｾ R4C 10.0" 9 ｃＳ｛ｾＧ I ＢＸＰ｣ｈｾＧＺｒ • I CT,O V2W I M2r% M2001' I 8 I 1 ｳｾﾷ (t 9.0" I I R40· A6A-· _148 8.0" I I _____ l __ J. __ .75" e ｾ 70" -- -- ---- _! Ｍｾ I ｾ J _ R30, Q) >. .COCE .... 0 ｾＮｊ 0_ ｾ ·0 Q) Cf) ..0 .... ­ o ｾｘｈＲＰ .50" eAIB .MICE Q) 0. 0. ·0 .M2CE >:::> <1 • PIC • XMIF Q) ｾ .XH)[ XH2E- .25" 80 CP GP 00" RL Soil AlB not plotted .. /-/1 no dalo available Soi I used 10 represent Qroup Class desiQnatlon Figure 3.1. Distribution of 113 soil map units in the study area by amount of water held in the upper and lower soil layers. This program was run 55 times, once for each of data sets generated by the PDSI program. Except for specification of the name of the weather station (Bluff, Cortez, Ignacio, Mesa Verde, FtLewis) and the first year of dependent data .available from the station, (1928, 1931, 1915, 1924, 1922, respectively) all other parameters were the same. All used SWOLD7, the matrix of eigenvector amplitude values, for the tree­ring­based independent data. The full range of values available for SWOLD7 was applied to the reconstruction process, A.D. 901­1970. Only axes 1­5 (Appendix A, columns 4­8) were considered since they collectively represented the vast majority of the variation in the PDSI data, where axis 1 (variable 5) is contained in the last column on the right, and axis 5 (variable 1) is contained by the fifth colwnn from the right (IMSL program output of principal component analysis). In all cases, a variable was allowed to enter the stepwise multiple regression equation only if it explained a minimum of five percent of variance in the dependent variable, otherwise, it ­57- was dropped. In all cases, a split­record calibration was selected where the most recent 30 years were used to derive an initial calibration and the remaining portion was used to verify the results. The full period calibration had a beginning date equivalent to the earliest date used for the PDSI calculations and an end date always at 1970, equivalent to the end date for the tree-ring data. Example output from TERIFY for one of the 55 programs is presented in the original study as appendix C (Van West 1990:409447). The regression coefficients and Y intercept for the final full-period calibration output by TERIFY are used as input for the final program that predicts PDSI values. Evaluating the Calibrations Table 3.2 summarizes the results of the calibration process following conventional procedures (Graybill 1989). The topmost section of the three parallel sections summarizes the results of the initial calibration period. It provides the parameters generated by the multiple regression (the regression coefficients for the variables used in the equation and the Y intercept), the correlation coefficients (the Pearson's product-moment correlation coefficient. r; the coefficient of detennination, r2; and a more conservative coefficient of determination, the adjusted ,.1., "adjusted" for degrees of freedom), and the F value with its associated probability that the correlation between the two sets of data is zero (Le., there is no correlation). The r value can range from -1.00 to 1.00. High absolute values indicate a strong relationship between the predictor treering series and the predicted PDSI series. A positive r indicates that they covary positively, a negative r indicates that they are inversely related. Both ,.1. and adjusted ,.2 are measures of how much variance in the PDSI values is explained by the tree-ring values, and thus their power to predict the PDSI values. Adjusted r 2 values are always lower than unadjusted ,.1. values and provide a more conservative estimate of the explained variance. The higher these values, the better the fit between the estimated modern PDSI values for particular soil types and weather stations and the modern climate data, as filtered through the tree-ring data and reduced through principal compon- ents analysis. The F value is a ratio of the explained variance to the unexplained variance, adjusted for degrees of freedom. This F value is interpreted using the probability of F(P > F) which provides the probability that the true correlation is zero. The smaller this value, the better the chance that the locally estimated PDSI values and the filtered. reduced regional tree-ring data are in fact related in the population from which these samples were drawn. The utility of the equation is questioned if the Probability of F (P > F) value is greater than .01. The middle section of Table 3.2 summarizes the verification period tests of the initial calibration. It assesses the ability of the initial calibration model to estimate the actual PDSI values for a portion of the historic record not used in the creation of the initial calibration. Thus, it is an independent test of the reliability of the initial regression equation. The mean of the actual PDSI values during that period (ACT mean) and the mean reconstructed by the initial regression equation (REC mean) are provided, as are the correlation coefficients (r, r 2, and the Probability of r, the P > r) associated with the reconstruction during the verification period. High r and ,.2 values are associated with low probability (p > r)of r values, meaning they have a low probability of not being correlated. Conversely, low rand r 2 values are associated with probability of r (P > r) values that are higher, indicating a greater chance that the reconstructed PDSI values and the actual PDSI values are uncorrelated. Reconstructions with a (P > r) value greater than .05 (greater than one chance in 20 that the data sets are uncorrelated) indicate a poor correlation, probably caused by the small number of years used in the verification period relative to the calibration period as well as by markedly different climatic conditions characteristic of these two periods. Another test is the T-test for differences in means. It compares the mean of the actual PDSI values (ACT mean) and the mean of the reconstructed PDSI values (REC mean) using a T statistic to assess the probability (P > that both series derive from the same statistical population. However, it may be questioned if the sample size is small « 30). The smaller the T value and the larger the P > T value, the -58- n Table 3.2. Results and Statistical Summary of the Initial Calibration, Verification, and Final, FullPeriod Calibration Analyses for 55 Soil Moisture Groups. RECONSTRUCTION 1 B 0.53 4.61 * 1941-1970 Calibration Period 30 (70%) No. of Years I Percent Regression -.974 Coefficients : V5 -1.036 V4 V3 V2 V1 -1.060 .479 Y Intercept .778 f ,2.606 .591 Adjusted ,213.325 F P>F <.000 RECONSTRUCTION 1 B. 0.53 4.61 Verification Period 1928-1940 No of Years I Percent 13 (30%) .588 f .346 ,2P>f .033 -.57 ACT mean REC mean -.22 T .536 P>T .607 Ws -.594 P>Z .276 RECONSTRUCTION 1 B. 0.53 4.61 1928-1970 Calibration Period No. of Years I Percent 43 (100%) Regression -.954 Coefficients : V5 V4 -.900 V3 -1.560 V2 V1 Y Intercept -.556 .781 .610 ,2.600 Adjusted ,2F 20.352 P>F <.000 , Represe ntation in Study Area 1.52% INITIAL CALIBRATION 2 B 0.85 2.95 3 B 1.02 3.65 1941-1970 1941-1970 30 (70%) 30 (70%) 4 B 1.02 4.48 1941-1970 30 (70%) -.932 -1.290 -.898 -.930 -.921 -.849 -1.116 -1.232 -.422 .793 .628 .614 14.656 <.000 -.524 .743 .552 .535 16.626 <.000 -.607 .769 .591 .575 12.525 <.000 VERIFICATION 2 B 0.85 2.95 3 B 1.02 3.65 1928-1940 1928-1940 13 (30%) 13 (30%) .485 .702 .493 .235 .090 .007 -1.30 -.62 -.18 -.46 1.932 .194 .075 .842 -1.852 -.314 .377 .032 4 B 1.02 4.48 1928-1940 13 (30%) .812 .659 .001 -.48 -.58 .115 .905 -.384 .350 FINAL CALIBRATION 2 B 0.85 2.95 3 B 1.02 3.65 1928-1970 1928-1970 43 (100%) 43 (100%) 4 B 1.02 4.48 1928-1970 43 (100%) -.739 -1.419 -.937 -.912 -1.504 -.948 -.865 -1.647 -.787 .719 .517 .505 21.387 <.000 -.559 .770 .606 .596 20.011 <.000 -.571 .781 .610 .600 20.359 <.000 0.42 -59- NA NA Table 3.2 (Continued). RECONSTRUCTION 5 B 1.02 5.40· Calibration Period 1941­1970 No. of Years / Percent 30 (70%) Regression ­.883 Coefficients : V5 ­.888 V4 V3 V2 V1 ­.631 Y Intercept .742 f .551 ,2 .535 Adjusted ,2 16.578 F <.000 P>F RECONSTRUCTION Verification Period No of Years / Percent f ,2 P>f ACT mean REC mean T P>T Ws P>Z 5 B 1.02 5.40 1928­1940 13 (30%) .672 .451 .012 ­.41 ­.56 .170 .861 ­.035 .486 RECONSTRUCTION 5 B 1.02 5.40 1928­1970 Calibration Period 43 (100%) No. of Years / Percent Regression ­.952 Coefficients : V5 V4 ­1.916 V3 V2 V1 ­.481 Y Intercept .746 f .557 ,2.546 Adjusted ,2 F 22.125 P>F <.000 Representation in Study Area .27% INITIAL CALIBRATION 6 B 1.14 6.44 7 B 1.027.80 1941­1970 1941­1970 30 (70%) 30 (70%) 8 B1.008.54 1941­1970 30 (70%) ­.872 ­.849 ­.883 ­.766 ­1.051 ­.874 ­.753 ­1.046 ­.636 .742 .551 .534 16.555 <.000 ­.652 .767 .588 .572 12.382 <.000 ­.664 .767 .589 .573 12.401 <.000 VERIFICATION 7 B 1.027.80 6 B 1.14 6.44 1928­1940 1928­1940 13 (30%) 13 (30%) .738 .640 .410 .545 .004 .018 ­.19 ­.39 ­.56 ­.62 .463 .196 .653 .841 ­1.013 ­.035 .486 .155 8 B1.008.54 1928­1940 13 (30%) .722 .521 .005 ­.16 ­.63 .502 .626 ­1.013 .155 FINAL CALIBRATION 6 B1.14 6.44 7 B 1.027.80 1928­1970 1928­1970 43 (100%) 43 (100%) 8 B 1.008.54 1928­1970 43 (100%) ­.935 ­.910 ­.898 ­1.929 ­1.916 ­1.924 ­.484 .742 .551 .540 24.546 <.000 ­.440 .733 .538 .526 23.254 <.000 ­.442 .729 .532 .520 22.700 <.000 NA -60- .04% .47% Table 3.2 (Continued). RECONSTRUCTION 9 B 1.05 9.45* Calibration Period No. of Years I Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept 1941-1970 30 (70%) f ,2Adjusted ,2F P>F -.864 -.742 -1.036 -.856 -.733 -1.027 -.904 -.831 -1.184 -.677 .767 .588 .572 12.373 <.000 -.683 .767 .588 .572 12.368 <.000 -1.072 -.685 .814 .663 .650 12.306 <.000 RECONSTRUCTION 9 B 1.05 9.45 Verification Period No of Years I Percent 1928-1940 13 (30%) .699 .489 .008 -.12 f ,2- P>f ACT mean REC mean T P>T Ws P>Z -.64 .555 .592 -1.013 .155 RECONSTRUCTION 9B1.059.45 Calibration Period No. of Years I Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept 1928·1970 . 43 (100%) f ,2Adjusted ,2- F P>F Representation in Study Area INITIAL CALIBRATION 11 B 1.14 10 B 0.99 10.26 13.66 1941-1970 1941-1970 30 (70%) 30 (70%) VERIFICATION 11B1.14 10 B 0.99 10.26 13.66 1928-1940 1928-1940 13 (30%) 13 (30%) .411 .682 .169 .465 .010 .160 -.09 .05 -.65 -.47 .596 .496 .565 .630 -1.013 -.314 .155 .377 FINAL CALIBRATION 10 B 0.99 11 B 1.14 10.26 13.66 1928-1970 1928-1970 43 (100%) 43 (100%) 12 C 0.53 4.61 1941-1970 30 (75%) -.858 -1.254 -.073 .766 .587 .572 19.188 <.000 12 C 0.53 4.61 1931-1940 10 (25%) .600 .361 .065 -.47 -.05 .602 .568 -.866 .193 12 C 0.53 4.61 1931·1970 40 (100%) -.884 -.875 -.835 -1.929 -1.932 -1.924 -.441 .722 .522 .510 21.821 <.000 -.436 .717 .514 .502 21.169 <.000 -.443 .692 .479 .466 18.403 <.000 -1.167 .736 .541 .529 21.848 <.000 NA NA NA 18.38% - 61- -.852 -1.130 Table 3.2 (Continued). INITIAL CALIBRATION RECONSTRUCTION 13 C .85 2.95* 14 C 1.02 3.65 15 C 1.02 4.48 16 C 1.02 5.40 1941-1970 Calibration Period 1941-1970 1941-1970 1941-1970 30 (75%) No. of Years I Percent 30 (75%) 30 (75%) 30 (75%) Regression -.850 -.874 -.732 Coefficients : V5 -.871 -1.266 -1.090 V4 -1.286 -1.065 V3 V2 V1 -.118 -.074 -.306 -.128 Y Intercept .751 .770 .751 .743 f .593 .563 .564 .553 ,2.578 .547 .548 .536 Adjusted ,219.637 17.419 17.491 16.685 F <.000 <.000 <.000 <.000 P>F VERIFICATION RECONSTRUCTION 13 C 0.85 2.95 14 C 1.02 3.65 15 C 1.02 4.48 16 C 1.02 5.40 1931-1940 1931-1940 1931-1940 1931-1940 Verification Period 10 (25%) 10 (25%) 10 (25%) 10 (25%) No of Years I Percent .598 .609 f .473 .590 .370 .348 .224 .358 t2 .060 .066 .070 .165 P>f -.47 -.44 -.38 ACT mean -.70 -.06 -.05 -.06 -.33 REC mean .592 .541 .498 .426 T .607 .634 .574 .681 P>T -.561 -.459 -.866 -.459 Ws .193 .323 .288 .323 P>Z FINAL CALIBRATION RECONSTRUCTION 13 C 0.85 2.95 14 C 1.02 3.65 15 C 1.02 4.48 16 C 1.02 5.40 1931-1970 1931-1970 1931-1970 1931-1970 Calibration Period 40 (100%) 40 (100%) 40 (100%) 40 (100%) No. of Years I Percent Regression -.718 -.844 -.867 -.868 Coefficients : V5 V4 -1.128 -1.138 -.981 -.948 V3 V2 V1 -.165 -.203 -.195 -.383 Y Intercept .739 .724 f .703 .714 .494 .546 .524 ,2.510 .534 .511 .480 .497 Adjusted ,222.238 20.358 F 18.030 19.269 <.000 <.000 P>F <.000 <.000 Representation in Study Area 7.41% .67% -62- .50% 1.23% Table 3.2 (Continued). RECONSTRUCTION Calibration Period No. of Years / Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept f ,2Adjusted ,2F P>F 17 C 1.14 6.44* 1941-1970 30 (75%) -.928 -.915 -.907 -.899 .225 .677 .459 .439 23.730 <.000 .191 .663 .440 .420 21.967 <.000 .180 .660 .435 .415 21.577 <.000 .118 .657 .432 .412 21.322 <.000 RECONSTRUCTION 17 C 1.14 6.44 Verification Period No of Years / Percent f ,2P>f ACT mean REC mean T P>T Ws P>Z 1931-1940 10 (25%) .619 .384 .055 -.29 .55 .829 .575 -1.172 .121 RECONSTRUCTION 17 C 1.14 6.44 Calibration Period No. of Years / Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept f ,2Adjusted ,2F P>F Representation in Study Area INITIAL CALIBRATION 18 C 1.02 19 C 1.00 8.54 20 C 1.05 9.45 7.80 1941-1970 1941-1970 1941-1970 30 (75%) 30 (75%) 30 (75%) 1931·1970 40 (100%) VERIFICATION 18 C 1.02 19 C 1.00 8.54 20 C 1.05 9.45 7.80 1931·1940 1931-1940 1931-1940 10 (25%) 10 (25%) 10 (25%) .594 .584 .573 .353 .341 .329 .068 .074 .081 -.25 -.22 -.18 .51 .43 .50 .744 .698 .592 .526 .502 .569 -.968 -.866 -.663 .166 .193 .254 FINAL CALIBRATION 18 C 1.02 19 C 1.00 8.54 20 C 1.05 9.45 7.80 1931-1970 1931-1970 1931-1970 40 (100%) 40 (100%) 40 (100%) -.921 -.915 -.908 -.902 -1.563 -1.410 -1.383 -1.350 .025 .716 .512 .499 12.607 <.000 .017 .682 .466 .451 16.113 <.000 .017 -.018 .674 .454 .439 15.395 <.000 -.n3 .57% 2.40% -63 - .6n .459 .444 15.678 <.000 5.11% .67% Table 3.2 (Continued). RECONSTRUCTION Calibration Period No. of Years I Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept f ,2 Adjusted ,2 F P>F RECONSTRUCTION Verification Period No of Years I Percent f ,2 P>f ACT mean REC mean T P>T Ws P>Z RECONSTRUCTION Calibration Period No. of Years I Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept , ,2Adjusted ,2F P>F Representation in Study Area 21 C .99 10.26* 1941-1970 30 (75%) INITIAL CALIBRATION 22 C 1.14 23 I 0.53 4.61 13.66 1941-1970 1941-1970 30 (75%) 30 (54%) 24 I 0.85 2.95 1941-1970 30 (54%) -.889 -.855 -.724 -1.226 -1.219 -.692 .091 .654 .428 .407 20.911 <.000 .033 .646 .417 .397 20.056 <.000 -.453 .770 .593 .578 19.686 <.000 -.446 .772 .596 .581 19.892 <.000 21 C 0.99 10.26 1931-1940 10 (25%) .565 .319 .087 -.15 .40 .531 .609 .663 .254 VERIFICATION 22 C 1.14 23 I 0.53 4.61 13.66 1931-1940 1915-1940 10 (25%) 26 (46%) .541 .677 .292 .459 .104 .000 .06 .21 .33 .36 .260 .229 .794 .815 -.357 -.343 .361 .366 21 C 0.99 10.26 1931-1970 40 (100%) FINAL CALIBRATION 22 C 1.14 23 I 0.53 4.61 13.66 1931-1970 1915-1970 40 (100%) 56 (100%) -.864 -.051 .632 .399 .383 25.212 <.000 4.03% -.842 ­.645 24 I 0.85 2.95 1915-1940 26 (46%) .663 .440 .000 .07 .34 .421 .679 -.622 .267 24 I 0.85 2.95 1915-1970 56 (100%) -1.729 -.612 -1.633 -1.217 -1.233 -.037 .624 .389 .373 24.228 <.000 -.548 .782 .611 .604 27.263 <.000 -.574 .778 .606 .598 26.606 <.000 NA 10.69% -64- 5.97% Table 3.2 (Continued). RECONSTRUCTION Calibration Period No. of Years I Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept f ,2Adjusted ,2F P>F 25 I 1.02 3.65* 1941-1970 30 (54%) INITIAL CALIBRATION 26 I 1.02 4.48 27 I 1.02 5.40 1941-1970 30 (54%) 1941-1970 30 (54%) 1941-1970 30 (54%) -.698 -1.186 -.710 -1.132 -.715 -1.062 -.714 -1.048 -.384 .754 .568 .552 17.763 <.000 -.392 .743 .553 .536 16.685 <.000 -.496 .723 .523 .505 14.790 <.000 -.624 .718 .516 .498 14.381 <.000 VERIFICATION RECONSTRUCTION 25 I 1.02 3.65 26 I 1.02 4.48 27 I 1.02 5.40 1915-1940 1915-1940 1915-1940 Verification Period 26 (46%) 26 (46%) 26 (46%) No of Years / Percent .674 .676 .678 f .454 .457 .459 ,2.000 .000 .000 P>f ACT mean .16 .20 .28 .40 .27 REC mean .38 .274 .017 T .365 .717 .781 .984 P>T -.597 Ws -.394 -.063 .275 .347 .475 P>Z FINAL CALIBRATION RECONSTRUCTION 25 I 1.02 3.65 26 I 1.02 4.48 27 I 1.02 5.40 Calibration Period 1915-1970 1915-1970 1915-1970 No. of Years / Percent 56 (100%) 56 (100%) 56 (100%) Regression -.620 -.636 Coefficients : V5 -.656 V4 -1.703 -1.693 -1.676 V3 V2 -1.236 -1.204 -1.172 V1 -.522 Ylntercept -.516 -.546 r .773 .767 .759 .598 ,2 .589 .576 .590 .581 .567 Adjusted ,2 F 25.753 24.790 23.503 P>F <.000 <.000 <.000 Representation in Study Area 28 I 1.14 6.44 .35% .33% -65 - 1.22% 28 I 1.14 6.44 1915-1940 26 (46%) .671 .450 .000 .27 .13 .196 .840 -.140 .444 28 I 1.14 6.44 1915-1970 56 (100%) -.691 -1.448 -.653 .711 .505 .496 27.056 <.000 .90% Table 3.2 (Continued). RECONSTRUCTION Calibration Period No. of Years / Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept f ,2 Adjusted ,2 F P>F 29 I 1.02 7.80· 1941-1970 30 (54%) ,2 P>f ACT mean REC mean T P>T Ws P>Z -.697 .719 .517 .499 14.447 <.000 1915-1940 26 (46%) .653 .427 .000 .42 .06 .523 .610 ­.648 .259 RECONSTRUCTION 29 I 1.02 7.80 Calibration Period No. of Years / Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept f ,2Adjusted ,2F P>F Representation in Study Area 1941-1970 30 (54%) 1941-1970 30 (54%) 32 10.99 10.26 1941-1970 30 (54%) -.704 -.980 -.834 -.888 -1.562 -1.859 -1.250 -.349 .722 .521 .504 14.712 <.000 -.195 .746 .557 .540 10.907 <.000 -.713 -1.032 RECONSTRUCTION 29 , 1.02 7.80 Verification Period No of Years / Percent f INITIAL CALIBRATION 30 I 1.008.54 31 I 1.05 9.45 1915-1970 56 (100%) -.701 .707 .500 .481 13.483 <.000 VERIFICATION 30 I 1.00 8.54 31 I 1.05 9.45 1915-1940 26 (46%) .641 .411 .000 .45 .03 .610 .553 -.698 .242 1915-1940 26 (46%) .572 .327 .002 .47 .14 .483 .637 -.292 .385 FINAL CALIBRATION 30 I 1.008.54 31 I 1.05 9.45 1915-1970 56 (100%) 1915-1970 56 (100%) 32 10.99 10.26 1915-1940 26 (46%) .544 .296 .004 .44 .18 .348 .729 -.394 .347 32 10.99 10.26 1915-1970 56 (100%) -.805 -1.202 -1.576 -1.016 -.728 -1.597 -1.284 -1.057 -.771 -1.362 -1.712 -1.203 -.627 .742 .550 .541 21.178 <.000 -.565 .768 .590 .582 18.361 <.000 -.473 .738 .545 .537 20.784 <.000 -.382 .749 .560 .552 16.254 <.000 -.656 -1.768 1.32% 1.18% ­ 66- .56% 10.93% Table 3.2 (Continued). RECONSTRUCTION Calibration Period No. of Years I Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept f ,2Adjusted ,2F P>F RECONSTRUCTION Verification Period No of Years I Percent f ,2P>f ACT mean REC mean T P>T Ws P>Z RECONSTRUCTION Calibration Period No. of Years I Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept f ,2Adjusted ,2- F P>F Representation in Study Area 33 I 1.14 13.66· 1941-1970 30 (54%) -.866 INITIAL CALIBRATION 34 M 0.53 4.61 35 M 0.85 2.95 36 M 1.02 3.65 1941-1970 30 (64%) 1941-1970 30 (64%) 1941-1970 30 (64%) -.660 -1.194 -.598 -1.097 -.651 -1.182 -.317 .824 .680 .668 28.654 <.000 -.239 .766 .587 .572 19.178 <.000 -.292 .828 .686 .674 29.482 <.000 -1.822 -1.298 .289 .753 .568 .551 11.375 <.000 33 I 1.14 13.66 1915-1940 26 (46%) .509 .259 .008 .45 .07 .526 .607 -.521 .301 33 I 1.14 13.66 1915-1970 56 (100%) VERIFICATION 34 M 0.53 4.61 35 M 0.85 2.95 36 M 1.02 3.65 1924-1940 17 (36%) .619 .383 .008 -.27 .00 .374 .712 -.497 .310 1924-1940 17 (36%) .604 .365 .010 -.32 .04 .538 .601 -.686 .246 1924-1940 17 (36%) .606 .367 .010 -.29 .02 .432 .673 -.450 .326 FINAL CALIBRATION 34 M 0.53 4.61 35 M 0.85 2.95 36 M 1.02 3.65 1924-1970 47 (100%) 1924-1970 47 (100%) 1924-1970 47 (100%) -.749 -1.260 -1.627 -1.136 -.622 -1.406 -.566 -1.286 -.609 -1.393 -.403 .728 .530 .521 14.364 <.000 -.454 .747 .558 .548 27.787 <.000 -.410 .703 .494 .482 21 ＷＸ ｾ <.000 -.443 .742 .551 .541 27.007 <.000 NA 2.94% -67- 1.63% .11% Table 3.2 (Continued). RECONSTRUCTION Calibration Period No. of Years / Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept f ,2 Adjusted ,2 F P>F 37 M 1.02 4.48· 1941-1970 30 (64%) INITIAL CALIBRATION 38 M 1.02 5.40 39 M 1.14 6.44 1941-1970 1941-1970 30 (64%) 30 (64%) -.674 -1.183 -.692 -1.086 -.751 -1.152 -.810 -1.042 -.264 .832 .691 .680 30.251 <.000 -.218 .819 .671 .659 27.569 <.000 -.220 .851 .724 .714 35.382 <.000 -.201 .865 .748 .739 40.050 <.000 VERIFICATION 39 M 1.14 RECONSTRUCTION 37 M 1.02 4.48 38 M 1.02 5.40 6.44 1924-1940 1924-1940 1924-1940 Verification Period 17 (36%) 17 (36%) 17 (36%) No of Years / Percent .628 .628 f .622 .394 .395 ,2 .387 .007 .007 P>f .007 -.36 -.34 -.28 ACT mean .12 .09 .05 REC mean .614 .676 .461 T .512 .551 P>T .653 -1.065 -1.018 .497 Ws .143 .154 P>Z .310 FINAL CALIBRATION 39 M 1.14 RECONSTRUCTION 37 M 1.02 4.48 38 M 1.02 5.40 6.44 1924-1970 1924-1970 1924-1970 Calibration Period 47 (100%) 47 (100%) 47 (100%) No. of Years / Percent Regression -.625 -.661 -.628 Coefficients : V5 -1.442 -1.510 V4 -1.428 V3 V2 V1 ­.444 -.431 -.467 Y Intercept .754 .781 .753 f .567 .569 .610 ,2 .601 .557 .559 Adjusted ,2 34.340 28.989 28.785 F <.000 <.000 P>F <.000 Representation in Study Area 40 M 1.02 7.80 1941-1970 30 (64%) .70% 2.14% ­68 - .57% 40 M 1.02 7.80 1924-1940 17 (36%) .612 .375 .009 -.38 .14 .727 .520 -1.207 .114 40 M 1.02 7.80 1924-1970 47 (100%) -.704 -1.454 -.474 .790 .624 .616 36.535 <.000 .95% Table 3.2 (Continued). RECONSTRUCTION Calibration Period No. of Years I Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept r ,2 Adjusted ,2 F P>F RECONSTRUCTION Verification Period No of Years I Percent r ,2- P>r ACT mean REC mean T P>T Ws P>Z 41 M 1.00 8.54* 1941-1970 30 (64%) r ,2Adjusted ,2- F P>F Representation in Study Area 44 M 1.14 13.66 1941-1970 30 (64%) -.798 -1.033 -.766 -1.057 -.765 -1.045 -.733 -.961 -.321 .844 .712 .702 33.414 <:.000 -.393 .823 .677 .665 28.292 <:.000 -.419 .820 .672 .659 27.616 <:.000 -.410 .758 .575 .559 18.233 <:.000 41 M 1.00 8.54 1924-1940 17 (36%) .615 .378 .008 -.37 .02 .539 .600 -1.018 .154 RECONSTRUCTION 41 M 1.00 8.54 Calibration Period No. of Years I Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept INITIAL CALIBRATION 42 M 1.05 43 M. 0.99 9.45 10.26 1941-1970 1941-1970 30 (64%) 30 (64%) 1924-1970 47 (100%) VERIFICATION 42 M 1.05 43 M. 0.99 10.26 44 M 1.14 13.66 9.45 1924-1940 1924-1940 1924-1940 17 (36%) 17 (36%) 17 (36%) .624 .620 .651 .389 .384 .424 .007 .008 .005 -.37 -.37 -.17 -.06 -.09 -.10 .433 .397 .096 .672 .696 .921 -.876 -.876 -.308 .191 .191 .379 FINAL CALIBRATION 42 M 1.05 43 M. 0.99 9.45 10.26 1924-1970 1924-1970 47 (100%) 47 (100%) -.6n -.700 -1.449 -1.480 -.545 .780 .609 .600 34.250 <:.000 -.589 .769 .591 .582 31.838 <:.000 1.28% .73% -69- -.675 -.1482 -.609 .767 .588 .578 31.352 <:.000 9.13% 44 M 1.14 13.66 19221924-1970 47 (100%) -.686 -1.471 -.530 .734 .539 .528 25.696 <:.000 .28% Table 3.2 (Continued). RECONSTRUCTION Calibration Period No. of Years I Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept f ,2Adjusted ,2F P>F 45 F 0.53 4.61* 1941-1970 30 (61%) INITIAL CALIBRATION 46 F 0.85 2.95 47 F 1.02 3.65 48 F 1.02 4.48 1941-1970 30 (61%) 1941-1970 30 (61%) 1941-1970 30 (61%) -.613 -.811 -.585 -.751 -.604 -.785 -.614 -.806 .100 .672 .451 .431 11.088 <.000 .075 .671 .450 .430 11.059 <.000 .100 .675 .455 .435 11.271 <.000 .118 .674 .454 .433 11.211 <.000 VERIFICATION RECONSTRUCTION 45 F 0.53 4.61 46 F 0.85 2.95 47 F 1.02 3.65 48 F 1.02 4.48 1922-1940 1922-1940 1922-1940 1922-1940 Verification Period 19 (39%) 19 (39%) 19 (39%) 19 (39%) No of Years I Percent .478 .483 .483 f .486 .233 .228 .237 .233 ,2.035 .037 .035 P>f .033 -.31 -.38 -.37 -.38 ACT mean .28 .31 REC mean .32 .33 1.273 1.940 T 1.963 1.983 .210 .065 P>T .063 .060 -1.730 -2.012 Ws -1.972 -2.012 .042 .022 .024 .022 P>Z FINAL CALIBRATION RECONSTRUCTION 45 F 0.53 4.61 46 F 0.85 2.95 47 F 1.02 3.65 48 F 1.02 4.48 1922-1970 1922-1970 1922-1970 1922-1970 Calibration Period 49 (100%) 49 (100%) 49 (100%) No. of Years I Percent 49 (100%) Regression -.548 -.521 -.538 Coefficients : V5 -.547 -.644 -.665 V4 -.630 -.638 V3 V2 V1 -.148 -.149 -.150 -.133 Ylntercept .607 .609 .608 .609 f .370 .369 .371 .370 ,2.355 .357 .356 .357 Adjusted ,213.524 13.454 F 13.486 13.574 <.000 <.000 P>F <.000 <.000 Representation in Study Area .29% NA -70- .03% .37% Table 3.2 (Continued). RECONSTRUCTION Calibration Period No. of Years / Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept ( ,2Adjusted ,2F P>F 49 F 1.02 5.40· 1941-1970 30 (61%) -.636 -.832 -.680 -1.010 -.693 -1.026 -.703 -1.014 .083 .674 .454 .434 11.239 <.000 -.002 .703 .495 .476 13.227 <.000 -.022 .697 .486 .467 12.751 <.000 0.036 .696 .484 .465 12.661 <.000 RECONSTRUCTION 49 F 1.02 5.40 Verification Period No of Years / Percent ( ,2- P>( ACT mean REC mean T P>T Ws P>Z 1922-1940 19 (39%) .471 .222 .040 -.36 .31 1.902 .070 -1.972 .024 RECONSTRUCTION 49 F 1.02 5.40 Calibration Period No. of Years / Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept ( ,2 Adjusted ,2 F P>F Representation in Study Area INITIAL CALIBRATION 50 F 1.14 51 F 1.027.80 52 F 1.008.54 6.44 1941-1970 1941-1970 1941-1970 30 (61%) 30 (61%) 30 (61%) 1922-1970 49 (100%) VERIFICATION 50 F 1.14 51 F 1.027.80 52 F 1.008.54 6.44 1922-1940 1922-1940 1922-1940 19 (39%) 19 (39%) 19 (39%) .469 .477 .482 .220 .228 .233 .041 .037 .035 -.32 -.34 -.32 .24 .26 .24 1.529 1.629 1.499 .117 .140 .148 -1.690 -1.650 -1.610 .045 .049 .054 FINAL CALIBRATION 50 F 1.14 51 F 1.027.80 52 F 1.008.54 6.44 1922-1970 1922-1970 1922-1970 49 (100%) 49 (100%) 49 (100%) -.557 -.646 -.593 -.769 -.604 -.783 -.608 -.785 -.155 .604 .371 .358 13.585 <.000 -.200 .636 .405 .392 15.657 <.000 -.206 .635 .403 .390 15.533 <.000 -.217 .635 .404 .391 15.566 <.000 .11% NA -71- .04% .57% Table 3.2 (Concluded). INITIAL CALIBRATION RECONSTRUCTION 53 F 1.04 9.45* 54 F 0.99 10.26 55 F 1.14 13.66 Calibration Period 1941-1970 1941-1970 1941-1970 No. of Years / Percent 30 (61%) 30 (61%) 30 (61%) Regression -.707 Coefficients : V5 -.706 -.697 V4 -1.000 -.986 -.930 V3 V2 V1 -.055 -.045 -.077 Y Intercept .686 .690 .665 f .477 .470 ,2.442 .451 .457 .422 Adjusted ,211.988 10.701 F 12.296 <.000 P>F <.000 <.000 53 F 1.04 9.45 1922-1940 19 (39%) .456 f ,2.208 P>f .048 ACT mean -.32 REC mean .22 T 1.459 P>T .159 Ws -1.569 P>Z .058 RECONSTRUCTION Verification Period No of Years / Percent RECONSTRUCTION Calibration Period No. of Years / Percent Regression Coefficients : V5 V4 V3 V2 V1 Y Intercept r ,2 Adjusted ,2 F P>F Representation in Study Area 53 F 1.04 9.45 1922-1970 49 (100%) VERIFICATION 54 F 0.99 10.26 55 F 1.14 13.66 1922-1940 1922-1940 19 (39%) 19 (39%) .435 .300 .189 .090 .060 .209 -.26 -.34 .20 .17 1.217 1.276 .238 .216 -1.248 -1.167 .106 .122 FINAL CALIBRATION 54 F 0.99 10.26 55 F 1.14 13.66 1922-1970 1922-1970 49 (100%) 49 (100%) -.606 -.760 -.604 -.748 -.559 -.687 -.224 .626 .392 .378 14.798 <.000 -.205 .620 .384 .371 14.334 <.000 -.250 .575 .331 .316 11.361 <.000 .05% .86% . .11 % *indicates Awe values used in reconstruction; the letter represents the station used and the numbers represent the potential inches of water that could be stored in upper and lower soil layers. NA. not appropriate, reconstruction does not actually occur in final study area. -72- better (Le.• the closer the means are to each other. the higher the probability that they both derive from the same statistical population of values and the better the equation that generated the reconstructed mean). If the P > T value is less the .05 then it is considered to be a poor reconstruction. An alternative test that is less sensitive to sample size is the Wilcoxon match-pairs signed-rank test (Ws). The magnitude of the Ws statistic is uninterpretable without its probability. the Probability of Z (P > Z.) The larger the P > Z value. the better the calibration equation is at replicating the actual mean of verification period. With this test. an initial calibration equation can be rejected as being inadequate to predict PDSI values if the P > Z value is less than .05 (Le.• the REC mean is statistically different than the ACT mean). In sum. it is the results of the three tests. the probabilities of r, T, and Z, along with a consideration of the number of years included in the verification period, that are used to assess the reliability of the initial model. The lowermost section of Table 3.2 summarizes the results of the final, full-period calibration. The same values are presented here as for the initial calibration period and are interpreted by the same criteria. Because more years are incorporated in the sample and they embody a more representative set of climatic Table 3.3. Adjusted ｾ conditions, the correlation coefficients are consistently higher. their probabilities of not being uncorrelated are greater. and their ability to predict actual PDSI values are more substantial. Likewise, their F values are consistently higher than their initial calibration counterparts. Results All 55 equations produced adequate initial calibrations (P > F of less than .000 were obtained without exceptions). When tested with independent data during the verification period, the vast majority (45/55 or 62%) passed all three tests for replicative adequacy. None failed all three tests and only one failed two tests (Reconstruction 2. Bluff .534.61). This is the single reconstruction with the lowest total available water capacity (AWe) of any of the 55. Nine failed to meet one of the three probability levels suggested as acceptable margins of error (Reconstructions 11, 12, ＬＹｾＵＴ 54, 55), although none seriously. In each of these cases, the reconstruction represents extremes in water-holding capacity for its elevation. either as very low amounts of water or very great amounts of water in the combined upper and lower soil layers. The reconstructions representing the middle range of AWC values consistently passed every verification period in all five elevational ranges. Values for 55 Full-Period Calibrations. Bluff AWC AWC Values Cortez Ignacio Mesa Verde Ft. Lewis 1315 m 1893 m 1969 m 2169 m Class (Su +SI)a 2317 m 60 53 55 0.53 4.61 60 36 1 51 48 0.85 2.95 48 36 2 60 1.02 3.65 60 53 59 54 3 36 4 4.48 60 51 58 56 1.02 36 55 50 57 5 1.02 5.40 56 36 54 50 50 1.14 6.44 60 6 39 53 45 54 1.02 7.80 62 7 39 44 1.00 8.54 52 8 58 60 39 44 51 54 1.05 9.45 58 9 38 10.26 50 55 10 0.99 38 58 37 47 1.14 13.66 52 11 37 53 32 54.82 46.63 Mean 56.09 56.36 36.73 S.D. 5.19 5.55 3.33 3.91 2.05 aAvaiiable water-.holding capacity of upper six inches of soil and lower soil profile. -73 - Mean±S.D. 52.80 ±9.88 48.60 ± 8.59 52.40 ±9.66 52.20 ±9.65 50.80 ±8.70 50.60 ± 7.67 50.60 ±8.85 50.60 ±8.99 49.00 ±8.00 47.60 ±9.66 44.20 ±9.31 49.95 8.49 Finally, all 55 full­period calibrations, without exception, produced significant correlations, again with the probability being less than .000 that the correlation was zero. Table 3.3 lists the conservative estimate for the explained variance (the adjusted ,2) for each of these 55 regression equations. In this data set, the adjusted ,2 consistently is about 1% less than the unadjusted ,2. As a group, they range from 32-62% with an overall mean of 50%. This indicates that about 50% of the variation in a PDSI value can be explained by the treering data, which is here considered to be a proxy for stochastic variation in climate. Summarized by elevation group, the five sets had adjusted ,2 values that ranged from 37-56%. Ignacio and Mesa Verde reconstructions proved to be the best predictors of PDSI as these were the stations that had the longest sets of climatic records (56 and 47 years, respectively). Bluff is next, not only by virtue of the number of years in the sample (43 years), but also because it is the lowest elevation and tree-ring based reconstructions in the American Southwest are most sensitive in arid settings. Cortez placed fourth primarily because of the limited time depth of its climatic records (40 years) and Ft Lewis placed last, not unexpectedly, because as the highest elevational zone it is less sensitive. Summarized by the 11 classes of soil moisture, the adjusted Tl ranged from 44-53%. The middle and lower groufs produced equations with higher adjusted, than the highest groups, not surprisingly. a mean of 58% (Rose et all 1982:252). Retrodiction of PDSI Values A.D. 901-1970 The program that reconstructs the entire time series of PDSI values from A.D. 901 to 970 is called "RECON" and was written by G. Robert Lofgren of the Laboratory of TreeRing Research at the University of Arizona. In the case of the present study, it uses the 1070year series of tree-ring values contained in SWOLD7 as the annually varying independent variables and the regression coefficients and the Y intercept from the full-period calibration derived by TERIFY as parameters. It also generates some descriptive statistics as a check on both input and output data and creates a file usable for plotting reconstructed values. The output from this program was formatted in such a way that it would be readily acceptable to the EPPL7 GIS program. Consequently, the output files were created with 55 rows, each representing one of the long-term PDSI reconstructions, and 20 columns, each column representing a single year (e.g., A.D. 901-920, 921-940, 941-960, etc.). A total of 54 files were created, all with the 55 row by 20 column format with the exception of the last file, which contains the final 10 years of data (A.D. 1961-1970). These data are presented in the original study (Van West 1990:448-555) as appendix D. Evaluating the Reconstructions This range compares favorably with adjusted ,2 values obtained by Rose in a reconstruction of July PDSI values for the Zuni area (using Black Rock and the combined Southwest Mountains weather stations: Ft Wingate, EI Morro, Bluewater, Grants, San Fidel, and Marquez) and Southwestern Colorado (using the combined Southwest Colorado weather stations: Cortez, Mancos, Mesa Verde, Ft Lewis, Ignacio, and Pagosa Springs) using similar tree ring data for essentially the same l,07Q-year period. For the Zuni area Rose obtained adjusted ,2 values ranging from 4554% with a mean of 50% (Martin Rose personal communication, April 1989) and for the combined reconstructions of Southwestern Colorado he obtained a range of 48-71 % with Following conventions established at the Laboratory of Tree-Ring Research (Graybill 1989), Table 3.4 summarizes for each of the 55 reconstructions the descriptive statistics associated with the final, full-period calibrations and the long-term reconstructions. The final, full-period calibrations include statistics for the actual June PDSI values calculated by the original PDSI programs as well as statistics for the reconstructed series generated by the final mUltiple regression equation that was used as the transfer .function. It summarizes descriptive statistics associated with the 1070-year reconstruction (A.D. 901-1970) and the first 400year subset of this long-term reconstruction (A.D. 901-13(0), the focus of this study. -74- Table 3.4. Descriptive Statistics Associated with 55 Final, Full-Period Calibrations and Long-Term Reconstructions. Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.63 2.84 ­4.52 8.66 .95 3.91 Full-Period Calibration, Actual PDSI Mean Standard Deviatio n Minimum Value Maximum Value Skewness Kurtosis ­.77 2.61 ­4.49 8.00 1.11 4.60 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.63 2.80 ­4.53 8.38 .91 3.74 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.65 2.83 ­4.52 8.65 .99 3.96 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis -.65 2.85 -4.48 8.88 1.06 4.18 Reconstruction 1: B 0.53 4.61 Full-Period Calibration, Reconstruction A.D. 901­1970 Recon PDSI ­.63 2.22 ­5.29 4.17 .02 2.44 ­.55 2.45 ­7.87 6.65 ­.11 ­.13 Reconstruction 2: B 0.85 2.95 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.77 1.88 ­4.87 3.63 .15 3.05 ­.78 1.98 ­6.10 5.05 ­.07 ­.28 Reconstruction 3: B 1.02 3.65 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.63 2.18 ­5.24 4.11 .02 2.45 ­.55 2.40 ­7.74 6.46 ­.11 ­.13 Reconstruction 4: B 1.02 4.48 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.65 2.21 ­5.29 4.09 .02 2.41 ­.56 2.45 ­7.89 6.81 ­.11 ­.12 Reconstruction 5: B 1.02 5.40 FUll-Period Calibration, Reconstruction Recon PDSI A.D. 901-1970 -.65 2.13 -4.29 3.66 .07 2.45 -75 - -.48 2.45 -7.55 7.15 -.09 -.13 Reconstruction A.D. 901­1300 ­.57 2.42 ­7.20 6.65 ­.14 .01 Reconstruction A.D. 901­1300 ­.82 1.96 ­6.10 4.94 ­.16 ­.30 Reconstruction A.D. 901­1300 ­.57 2.38 ­7.06 6.46 ­.15 .00 Reconstruction A.D. 901­1300 ­.58 2.43 ­7.24 6.81 ­.13 .03 - Reconstruction A.D. 901-1300 -.48 2.49 -7.31 7.15 -.12 -.03 Table 3.4 (Continued). Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.65 2.82 ­4.43 8.77 1.06 4.16 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.60 2.79 ­4.38 8.42 .97 3.89 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.60 2.78 ­4.35 8.25 .95 3.80 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.60 2.77 ­4.32 8.08 .94 3.73 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.59 2.77 ­4.28 7.96 .94 3.68 Reconstruction 6: B 1.14 6.44 Full-Period Calibration, Reconstruction Recon POSt A.D. 901­1970 ­.65 2.10 ­4.21 3.60 .07 2.45 ­.48 2.42 ­7.46 7.13 ­.09 ­.12 Reconstruction 7: B 1.02 7.80 Full-Period Calibration, Reconstruction A.D. 901­1970 Recon PDSI ­.60 2.05 ­4.06 3.54 .08 2.44 ­.43 2.37 ­7.25 7.06 ­.09 ­.11 Reconstruction 8: B 1.00 8.54 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.60 2.03 ­4.00 3.50 .08 2.44 ­.44 2.35 ­7.19 7.04 ­.08 ­.11 Reconstruction 9: B 1.05 9.45 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.60 2.00 ­3.93 3.45 .08 2.44 ­.44 2.33 ­7.11 7.01 ­.08 ­.10 Reconstruction 10: B 0.99 10.26 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.59 1.98 ­3.88 3.43 .08 2.43 -76- ­.43 2.31 ­7.05 7.00 ­.08 ­.10 Reconstruction A.D. 901­1300 ­.48 2.46 ­7.22 7.13 ­.11 ­.02 Reconstruction A.D. 901­1300 ­.44 2.41 ­7.01 7.06 ­.10 ­.01 Reconstruction A.D. 901­1300 ­.44 2.39 ­6.95 7.04 ­.10 ­.00 Reconstruction A.D. 901­1300 ­.44 2.37 ­6.86 7.01 ­.09 .01 Reconstruction A.D. 901­1300 ­.43 2.35 ­6.80 7.00 ­.09 .01 Table 3.4 (Continued). Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.59 2.76 ­4.17 7.58 .98 3.75 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.20 2.78 ­5.44 5.73 .43 2.53 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maxiroom Value Skewness Kurtosis ­.39 2.54 ­5.35 5.84 .63 3.65 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maxiroom Value Skewness Kurtosis ­.19 2.75 ­5.50 5.65 .40 2.56 Full-Period Calibration, Actual PDSI Mean Standard Deviation· Minimum Value Maxiroom Value Skewness Kurtosis ­.25 2.80 ­5.50 5.59 .46 2.47 Reconstruction 11: B 1.14 13.66 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.59 1.91 ­3.75 3.27 .09 2.42 ­.44 2.24 ­6.81 6.85 ­.07 ­.09 Reconstruction 12: C 0.53 4.61 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.20 2.04 ­4.49 4.49 .09 2.69 ­.16 2.07 ­5.90 5.83 ­.11 ­.31 Reconstruction 13: C 0.85 2.95 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.39 1.79 ­4.08 3.69 .11 2.76 ­.38 1.81 ­5.31 4.92 ­.10 ­.30 Reconstruction 14: C 1.02 3.65 .Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.19 2.03 ­4.45 4.46 .09 2.70 ­.16 2.06 ­5.85 5.80 ­.11 ­.31 Reconstruction 15: C 1.02 4.48 Full-Period Calibration. Reconstruction Recon PDSI A.D. 901­1970 ­.25 2.03 ­4.55 4.40 .07 2.64 -77- ­.20 2.05 ­5.94 5.65 ­.13 ­.32 Reconstruction A.D. 901­1300 ­.44 2.28 ­6.56 6.85 ­.08 .03 Reconstruction A.D. 901­1300 ­.20 2.06 ­5.59 5.26 ­.21 ­.35 Reconstruction A.D. 901­1300 ­.41 1.80 ­5.20 4.60 ­.19 ­.33 Reconstruction A.D. 901­1300 ­.20 2.05 ­5.56 5.25 ­.21 ­.35 Reconstruction A.D. 901­1300 ­.24 2.04 ­5.48 4.93 ­.23 ­.36 Table 3.4 (Continued). Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.25 2.82 ­5.46 5.55 .49 2.43 Full-Period Calibration, Actual POSt Mean Standard Deviation. Minimum Value Maximum Value Skewness Kurtosis ­.15 2.97 ­5.40 5.49 .50 2.22 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.16 2.99 ­5.39 5.77 .56 2.28 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value . Skewness Kurtosis ­.16 2.99 ­5.37 5.85 .57 2.29 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.19 2.98 ­5.35 5.91 .60 2.31 Reconstruction 16: C 1.02 5.40 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.25 2.02 ­4.53 4.38 .07 2.63 ­.19 2.04 ­5.92 5.62 ­.13 ­.32 Reconstruction 17: C 1.14 6.44 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.15 2.13 ­4.11 4.52 .05 2.45 .03 2.31 ­6.72 7.28 ­.11 ­.07 Reconstruction 18: C 1.02 7.80 Full-Period Calibration, Reconstruction A.D. 901­1970 Recon PDSI ­.16 2.04 ­3.78 3.84 .08 2.36 .02 2.21 ­6.47 6.27 ­.13 ­.19 Reconstruction 19: C 1.00 8.54 Full-Period Calibration, Reconstruction A.D. 901­1970 Recon PDSI ­.16 2.02 ­3.76 3.81 .07 2.37 .02 2.19 ­6.41 6.19 ­.14 ­.19 Reconstruction 20: C 1.05 9.45 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­970 ­.19 2.01 ­3.78 3.75 .07 2.37 -78- ­.01 2.17 ­6.39 6.05 ­.14 ­.19 Reconstruction A.D. 901­1300 ­.23 2.04 ­5.43 4.86 ­.24 ­.37 Reconstruction A.D. 901­1300 .03 2.37 ­6.57 7.28 ­.08 ­.02 Reconstruction A.D. 901­1300 .01 2.24 ­6.31 6.27 ­.19 ­.14 Reconstruction A.D. 901­1300 .01 2.22 ­6.25 6.19 ­.19 ­.14 Reconstruction A.D. 901­1300 ­.02 2.19 ­6.23 6.05 ­.19 ­.14 Table 3.4 (Continued). Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis Full-Period Calibration, Actual PDSI -.20 2.97 -5.33 5.92 .61 2.31 Reconstruction 21: C 0.99 10.26 Full-Period Calibration, Reconstruction A.D. 901-1970 Recon PDSI -.05 -.20 1.88 1.88 -4.07 -5.29 4.64 3.71 -.18 -.03 -.37 2.3.6 Reconstruction A.D. 901-1300 -.07 1.93 -5.29 4.64 -.25 -.40 Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis Full-Period Calibration, Actual PDSI -.19 2.93 -5.22 5.94 .64 2.32 Reconstruction 22: C1.14 13.66 Full-Period Reconstruction Calibration, A.D. 901-1970 Recon PDSI -.03 -.19 1.83 1.83 -5.14 -3.95 4.54 3.63 -.18 -.03 -.37 2.36 Reconstruction A.D. 901-1300 -.06 1.88 -5.14 4.54 -.25 -.40 Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis Full-Period Calibration, Actual PDSI -.14 2.66 -6.57 6.23 .02 2.75 Reconstruction 23: 10.53 4.61 Full-Period Calibration, Reconstruction A.D. 901-1970 Recon PDSI -.14 -.54 2.11 2.08 -6.30 -4.66 3.75 6.03 .03 -.08 2.72 -.16 Reconstruction A.D. 901-1300 -.58 2.09 -6.30 5.09 .03 -.19 Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis Full-Period Calibration, Actual PDSI -.19 2.55 -6.33 6.14 .13 2.80 Reconstruction 24: 10.852.95 Full-Period Calibration, Reconstruction A.D. 901-1970 Recon PDSI -.19 -.57 1.98 2.01 -4.51 -6.09 3.57 5.69 -.09 .04 -.15 2.73 Reconstruction A.D. 901-1300 -.60 2.00 -6.09 4.73 .04 -.18 Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis Full-Period Calibration, Actual PDSI -.13 2.63 -6.59 6.14 -.02 2.75 Reconstruction 25: 11.02 3.65 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901-1970 -.13 -.52 2.03 2.06 -4.52 -6.16 3.69 5.95 -.08 .04 2.73 -.15 Reconstruction A.D. 901-1300 .55 2.05 -6.16 4.97 .04 -.18 -79- Table 3.4 (Continued). Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.12 2.67 ­6.71 6.12 ­.05 2.72 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.15 2.73 ­6.78 6.14 .00 2.62 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.22 2.77 ­6.80 6.14 .10 2.59 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.20 2.82 ­6.69 6.11 .14 2.48 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.19 2.82 ­6.59 6.07 .15 2.45 Reconstruction 26: I 1.02 4.48 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.12 2.05 ­4.57 3.72 ­.08 2.72 ­.51 2.07 ­6.18 5.94 .03 ­.16 Reconstruction 27: 11.02 5.40 Full-Period Reconstruction Calibration, Recon PDSI A.D. 901­1970 ­.54 2.09 ­6.23 5.90 .03 ­.17 ­.15 2.07 ­4.68 3.72 ­.07 2.71 Reconstruction 28: 11.146.44 Full-Period Reconstruction Calibration, A.D. 901­1970 Recon PDSI ­.65 1.91 ­5.79 4.98 ­.06 ­.27 ­.22 1.97 ­4.64 3.57 .08 2.46 Reconstruction 29: 11.027.80 Full-Period Calibration, Reconstruction A.D. 901­1970 Recon PDSI ­.20 2.09 ­4.81 3.68 ­.03 2.68 ­.62 2.11 ­6.32 5.98 .02 ­.18 Reconstruction 30: 11.00 8.54 Full-Period calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.19 2.17 ­5.38 4.11 ­.15 2.59 ­ 80- ­.56 2.31 ­6.86 6.44 .01 ­.03 Reconstruction A.D. 901­1300 ­.55 2.06 ­6.18 5.02 .03 ­.19 Reconstruction A.D. 901­1300 ­.58 2.07 ­6.23 5.03 .02 ­.21 Reconstruction A.D. 901­1300 ­.68 1.89 ­5.79 4.98 ­.14 ­.27 Reconstruction A.D. 901­1300 ­.66 2.09 ­6.32 5.23 .01 ­.22 Reconstruction A.D. 901­1300 ­.58 2.29 ­6.21 6.44 .14 .14 Table 3.4 (Continued). Full-Period Calibration, Actual PDSJ Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.11 2.68 ­6.48 6.04 .17 2.34 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.07 2.96 ­6.41 6.03 .22 2.25 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.11 2.90 ­6.20 5.78 .28 2.27 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.28 2.29 ­5.11 5.34 .38 2.98 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.25 2.22 ­4.54 5.41 .57 3.03 Reconstruction 31: I 1.05 9.45 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.11 2.13 ­5.26 4.04 ­.07 2.43 ­.47 2.28 ­7.07 6.51 ­.07 ­.10 Reconstruction 32: 10.99 10.26 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.07 2.21 ­5.18 4.26 ­.16 2.52 ­.38 2.42 ­7.53 7.81 .02 .08 Reconstruction 33: 11.14 13.66 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.11 2.11 ­4.95 4.01 ­.15 2.52 ­.40 2.31 ­7.22 7.41 .02 .07 Reconstruction 34: M 0.53 4.61 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.28 1.71 ­4.19 3.44 .11 2.86 ­.45 1.77 ­5.22 4.85 ­.05 ­.26 Reconstruction 35: M 0.85 2.95 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.25 1.56 ­3.82 3.14 .11 2.86 ­ 81- ­.40 1.62 ­4:76 4.43 ­.05 ­.26 Reconstruction A.D. 901­1300 ­.48 2.25 ­6.53 6.51 ­.05 .09 Reconstruction A.D. 901­1300 ­.39 2.43 ­6.35 7.81 .16 .31 Reconstruction A.D. 901­1300 ­.41 2.32 ­6.12 7.41 .15 .30 Reconstruction A.D. 901­1300 ­.48 1.75 ­5.22 4.85 ­.13 ­.25 Reconstruction A.D. 901­1300 ­.44 1.60 ­4.76 4.43 ­.13 ­.24 Table 3.4 (Continued). Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.27 2.27 ­4.87 5.31 .41 2.94 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.26 2.30 ­5.22 5.31 .30 2.95 Full-Period Calibration, Actual PDSI Mean Standard Deviation. Minimum Value Maximum Value Skewness Kurtosis ­.27 2.30 ­5.43 5.34 .26 3.10 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.28 2.34 ­5.52 5.36 .24 2.92 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maxirrum Value Skewness Kurtosis ­.30 2.37 ­5.59 5.39 .26 2.81 Reconstruction 36: M 1.02 3.65 Full-Period Calibration, Reconstruction Recon POSt A.D. 901­1970 ­.27 1.68 ­4.13 3.38 .11 2.86 ­.44 1.75 ­5.14 4.79 ­.05 ­.25 Reconstruction 37: M 1.02 4.48 Full-Period Reconstruction Calibration. A.D. 901­1970 Recon PDSI ­.43 1.80 ­5.26 4.95 ­.05 ­.26 ­.26 1.73 ­4.22 3.51 .11 2.86 Reconstruction 38: M 1.02 5.40 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.27 1.73 ­4.24 3.49 .11 2.86 ­.44 1.80 ­5.28 4.96 ­.05 ­.25 Reconstruction 39: M 1.14 6.44 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.28 1.83 ­4.46 3.68 .11 2.86 ­.46 1.89 ­5.56 5.21 .05 ­.26 Reconstruction 40: M 1.02 7.80 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.30 1.87 ­4.51 3.82 .10 2.82 ­ 82- ­.47 1.04 ­5.68 5.23 ­.06 ­.27 Reconstruction A.D. 901­1300 ­.47 1.73 ­5.14 4.79 ­.12 ­.24 Reconstruction A.D. 901­1300 ­.46 1.78 ­5.26 4.95 ­.13 ­.24 Reconstruction A.D. 901­1300 ­.47 1.78 ­5.28 4.96 ­.12 ­.24 Reconstruction A.D. 901­1300 ­.50 1.87 ­5.56 5.21 ­.12 ­.24 Reconstruction A.D. 901­1300 0.51 1.92 ­5.68 5.21 ­.14 ­.28 Table 3.4 (Continued). Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis Full-Period Calibration, Actual PDSI -.38 2.39 -5.61 5.37 .33 2.77 Reconstruction 41: M 1.00 8.54 Full-Period Calibration, Reconstruction Recon PDSI A.D.901-1970 -.38 -.54 1.86 1.93 -5.73 -4.56 5.13 3.76 -.06 .10 -.27 2.82 Reconstruction A.D.901-1300 -.58 1.91 -5.73 5.12 -.14 -.28 Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis Full-Period Calibration, Actual PDSI -.41 2.39 -5.60 5.36 .38 2.78 Reconstruction 42: M 1.05 9.45 Full-Period Reconstruction Calibration, Recon POSt A.D. 901-1970 -.41 -.58 1.84 1.91 -4.59 -5.71 3.60 5.07 -.05 .10 -.26 2.84 Reconstruction A.D. 901-1300 -.62 1.88 -5.71 5.07 -.13 -.26 Mean Standard Deviation Minimum Vatue Maximum Value Skewness Kurtosis Full-Period Calibration, Actuat POSt -.43 2.40 -5.59 5.37 .41 2.78 Reconstruction 43: M 0.99 10.26 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901-1970 -.43 -.60 1.84 1.90 -4.60 -5.73 3.58 5.05 .10 -.05 2.85 -.26 Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis Full-Period Calibration, Actual POSt -.36 2.52 -5.54 5.37 .37 2.48 Reconstruction 44: M 1.14 13.66 Full-Period Calibration, Reconstruction Recon POSt A.D. 901-1970 -.36 -.52 1.85 1.92 -4.54 -5.68 3.69 5.14 .10 -.06 2.84 -.27 Reconstruction A.D. 901-1300 -.56 1.89 -5.68 5.14 -.14 -.26 Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis Full-Period Calibration, Actual PDSI -.11 2.07 -4.21 6.43 .57 3.82 Reconstruction 45: F 0.53 4.61 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901-1970 -.11 -.14 1.26 1.30 -2.91 -3.79 2.78 3.59 .07 -.13 2.67 -.31 Reconstruction A.D. 901-1300 -.17 1.30 -3.52 3.15 -.23 -.36 ­83 - Reconstruction A.D. 901-1300 ­.64 1.88 -5.73 5.05 -.13 -.26 Table 3.4 (Continued). Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.10 2.00 ­3.57 6.16 .59 3.54 Full-Period Calibration, Actual POSt Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.11 2.03 ­3.95 6.34 .61 3.79 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.10 2.06 ­4.27 6.37 .54 3.82 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.12 2.09 ­4.55 6.41 .48 3.84 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.15 2.17 ­4.86 6.47 .41 3.66 Reconstruction 46: F 0.85 2.95 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.10 1.21 ­2.79 2.68 .08 2.69 ­.14 1.26 ­3.63 3.48 ­.12 ­.31 Reconstruction 47: F 1.02 3.65 Full-Period Reconstruction Calibration, A.D. 901­1970 Recon PDSI ­.11 1.23 ­2.85 2.72 .07 2.67 ­.14 1.28 ­3.72 3.52 ­.13 ­.31 Reconstruction 48: F 1.02 4.48 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.10 1.25 ­2.88 2.78 .07 2.67 ­.13 1.29 ­3.76 3.59 ­.13 ­.31 Reconstruction 49: F 1.02 5.40 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.12 1.27 ­2.95 2.81 .07 2.67 ­.15 1.32 ­3.85 3.63 ­.13 ­.31 Reconstruction 50: F 1.14 6.44 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.15 1.38 ­3.20 3.03 .08 2.69 ­ 84- ­.19 1­43 ­4.18 3.95 ­.12 ­.31 Reconstruction A.D. 901­1300 ­.17 1.25 ­3.42 3.11 ­.22 ­.35 Reconstruction A.D. 901­1300 ­.17 1.27 ­.3.45 3.08 ­.23 ­.36 Reconstruction A.D. 901­1300 ­.15 1.29 ­3.49 3.15 ­.23 ­.36 Reconstruction A.D. 901­1300 ­.17 1.32 ­3.57 3.18 ­.23 ­.36 Reconstruction A.D. 901­1300 ­.22 1.43 ­3.95 3.54 ­.22 ­.35 Table 3.4 (Concluded). Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.15 2.22 ­5.16 6.48 .34 3.63 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.16 2.23 ­5.31 6.44 .30 3.63 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.17 2.24 ­5.47 6.40 .27 3.63 Full-Period Calibration, Actual PDSI Mean Standard Deviation Minimum Value Maximum Value Skewness Kurtosis ­.16 2.25 ­5.59 6.37 .21 3.61 Full-Period Calibration, Actual POSt Mean Standard Deviation Minimum Vatue Maximum Value Skewness Kurtosis ­.21 2.24 ­5.99 6.22 .16 3.86 Reconstruction 51: F 1.02 7.80 Full-Period Reconstruction Calibration, A.D. 901­1970 Recon PDSI ­.20 1.46 ­4.26 4.02 ­.12 ­.31 ­.15 1.41 ­3.26 3.08 .08 2.69 Reconstruction 52: F 1.008.54 Full-Period Reconstruction Calibration, A.D. 901­1970 Recon PDSI ­.21 1.47 ­4.29 4.03 ­.12 ­.31 ­.16 1.42 ­3.30 3.09 .08 2.69 Reconstruction 53: F 1.05 9.45 Full-Period Reconstruction Calibration, A.D. 901­1970 Recon POSt ­.17 1.40 ­3.28 3.05 .08 2.68 ­.22 1.46 ­4.27 3.98 ­.12 ­.31 Reconstruction 54: F 0.99 10.26 Full-Period Calibration, Reconstruction Recon PDSI A.D. 901­1970 ­.16 1.40 ­3.25 3.05 .07 2.68 ­.20 1.45 ­4.24 3.97 ­.12 ­.31 Reconstruction 55: F 1.14 13.66 Full-Period Calibration, Reconstruction Recon POSt A.D. 901­1970 ­.21 1.29 ­3.07 2.76 .07 2.68 ­ 85- ­.25 1.34 ­3.98 3.60 ­.12 ­.31 Reconstruction A.D. 901­1300 ­.23 1.45 ­4.02 3.60 0.22 ­.35 Reconstruction A.D. 901­1300 ­.24 1.46 ­4.05 3.61 ­.22 ­.35 Reconstruction A.D. 901­1300 ­.25 1.45 ­4.02 3.53 ­.22 ­.36 Reconstruction A.D. 901­1300 ­.23 1.44 ­3.97 3.52 ­.22 ­.36 Reconstruction A.D. 901­1300 ­.27 1.33 ­3.73 3.18 ­.22 ­.36 An examination of these data reveal consis- structed calibration data set has both minimum and maximum values that are less extreme than their counterparts in the actual calibration period data set This is also generally expected with regressionbased estimation given its tendency to create relatively normal distributions with predicted values closer to the mean. tent patterning in these statistics from set to set in each of the 55 reconstructions• The means of the reconstructed calibration period data sets are always equivalent to the means of the actual calibration period data sets. This is expected since mean values of the x and y variables will fall directly on the regression line and thus will be predicted by the regression equation, as well. • In all cases, the reconstructed calibration period data set produces distributions that are less skewed (i.e.• closer to 0.00) than those ｾｯ the actual calibration period data Ｌｴｾ whIch tend to be characterized by positIVely skewed distributions. • The calibration period mean is never zero, although it often approaches zero. In part this is due to the limited number of years used to calibrate any reconstruction, which results in a skewed sample that does not fully capture the long-term climatic variation characteristic of a location. However, the greater part of this phenomenon results from using a slightly extended set of years to calculate the original PDSI values (i.e., all instrumented climatic data were taken through 1983), whereas the calibration period must end at 1970 because that is the final year of treering data. Because no reconstruction has joint precipitation and temperature data earlier than 1915, it is important to include as many years as possible in the original PDSI calculations in order to derive more accurate long-term means used in the PDSI equations. If, however, the post-calibration years used to determine the actual PDSI values were characterized by precipitation and temperature conditions somewhat different than that of the calibration period, the distribution of those values and their mean will be different Although it varies for each of the 55 reconstructions for the 1971-1983 period, the post-1970 period was generally wetter with more years characterized by extremes in wetness than extremes in dryness (fable 3.5). Consequently, the calibration period means differ from zero or near-zero and are consistently on the negative side. • In all cases, the reconstructed calibration period data set produces distributions that better approximate normal kurtosis or "peakedness" (normal=3.00) than the actual calibration period data sets. • The means of both the full, long-term reconstruction (A.D. 901-1970) and the first 400-year subset (A.D. 901-1300) are somewhat different from the calibration period means, reflecting the variation associated with a large number of years. • The reconstruction period standard devia- tions are similar to the calibration period standard deviations and fall between those generated for the actual full period calibration and the reconstructed full period calibration. • The minimum and maximum values differ from those of the calibration period, reflecting more extreme conditions over the 107Q-year reconstruction period than in the reconstructed calibration period. • Both reconstructed series tend to be normally distributed as judged by skewness, and exhibit platykurtic distributions. • The distribution of values within the 400 years used in this study exhibits characteristics quite similar to that of the full long-term reconstruction. Small patterned differences OCCW" from elevationaJ. zone to elevational zone. • The standard deviation of the reconstructed calibration period is always less than the actual calibration, and again this is an expected result of using a least-squares ｾ gression line to predict or reconstruct PDSI values. The long-term June PDSI reconstructions presented in this study compare favorable with Rose's Zuni area (Martin Rose, personal communication, April 1989) and Southwestern • In most cases (50/55 or 91 %) the recon- - 86- Table 3.5. Occurrence of Extreme Conditions estimated for the month of June (s -4.00 or 2+4.00 PDSI) in the Years Used to Calculate PDSI for all Combination of Weather Station and Moisture Class. Weather Station Bluff AWca Class 1 Extreme Dry Years (ore-1941 ) 1934 (-4.52) 1934 (-4.49) 1934 (-4.53) Extreme Dry Years (1941-1970) and r1971-19831 b 1951 (-4.40); 1956 (-4.06): 1959 (-4.02) 1951 (-4.28); 1959 (-4.08) 1947 (-4.05); 1951 (-4.32); 1959 (-4.03) 1934 (-4.52) 1934 (-4.48) 1934 (-4.43) 1934 (-4.38) 1947 (-4.07); 1951 (-4.31) 1947 (-4.07); 1951 (-4.30) 1947 (-4.05); 1951 (-4.25); 1956 (-400) 1947 (-4.05); 1951 (-4.24); 1956 (-4.01) 1934 (-4.35) 1934 (-4.32) 1934 (-4.28) 1934 (-4.17) 1947 (-4.05); 1951 (-4.23)' 1956 (-403) 1947 (-4.06); 1951 (-4.21)' 1956 (-4.05) 1947 (-4.04); 1951 (-4.21)' 1956 (-4.07) 1947 (-4.06); 1951 (-4.16)' 1956 (-4.15) 1951 (-4.93); 1959 (-4.13); [1977] Bluff 2 Bluff 3 Bluff 4 Bluff 5 Bluff 6 Bluff 7 Bluff 8 Bluff 9 Bluff 10 Bluff 11 Cortez 1 Cortez 2 Cortez 3 Cortez 4 1951 (-5.50); 1959 (-4.84); [1977] Cortez 5 1951 (-5.46); 1959 (-4.69); [1977] Cortez 6 1951 (-5.40); 1959 (-4.47); [1972,1977] Cortez 7 1951 (-5.39); 1959 (-4.27); [1972,1977] 1951 (-5.35); 1959 (-4.95): [19771 1951 (-5.50); 1959 (-4,93); [1977] - 87- Extreme Extreme Wet Years (1941Wet Years 1970) and [1971-1983Jb (pre-1941) 1941 (8.66); 1957 (4.25); r19741979,1980,1983] 1941 (8.00); 1957 (5.00); £1978 1979 1980 1983] 1941 (8.38); 1957 (4.36); [1973,1978,1979,1980, 19831 1932 1941 (8.65); (4.17) r1973 1979 1980,19831 1932 1941 (8.88); (4.53) f1973 1979 1980 19831 1932 1941 (8.77); (4.28) £1973,1979 1980 19831 1928 1941 (8.42); (4.68); [1973,1979,1980,1983J 1932 (4.08) 1941 (8.25); 1928 (4.99) £1973 1979 1980 19831 1941 (8.08); 1928 (5.36) £197319781980,19831 1928 1941 (7.96); (5.63) f1973 1979 1980 1983]; 1928 1941 (7.58); (6.56) f1973 1979 1980 19831 1932 1941 (5.16); 1942 (4.71); (4.21 ) 1957 (4.80); £19731978 19791 1932 1941 (5.38); 1957 (5.84); (4.01) 1965 (4.93); f1973 19791 1941 (5.19); 1942 (4.54); 1932 (4.15) 1957 (5.65); 1965 (4.73); £1973.19791 1932 1941 (5.15); 1942 (4.87); (4.28) 1957 (5.59); 1965 (4.61); £1973 19791 1932 1941 (5.10); 1942 (5.18); (4.49) 1957 (5.55); 1965 (4.51); (1973 19791 1932 1941 (5.05); 1942 (5.48); (4.75) 1957 (5.49); 1958 (5.04); 1965 (4.42); £1973 1979 19801 1932 1941 (5.12); 1942 (5.n); (4.92) . 1957 (5.48); 1958 (5.38); 1965 (4.18); £1973 1979 19801 Table 3.5 (Continued). Weather Station Cortez AWCa Class 8 Extreme Dry Years Extreme Dry Years (1941-1970) and (pre-1941 ) r"1971-19831 b 1951 (-5.37); 1959 (-4.12); [1972,1977] Extreme Wet Years (pre-1941) 1932 (5.03) Cortez 9 1951 (-5.35); [1972,1977] 1932 (5.19) Cortez 10 1951 (-5.33); [1972,1977] 1932 (5.31); Cortez 11 1951 (-5.22); [1972,1977] 1932 (5.94) Ignacio 1 1934 (-6.57) 1946 (-4.40); 1950 (-4.54); 1959 (-4.53); [1977] Ignacio 2 1934 (-6.33) 1946 (-4.19); 1950 (-4.47); 1959 (-4.28); [1977] Ignacio 3 1934 (-6.59) 1946 (-4.37); 1950 (-4.56); 1959 (-4.41); [1977] Ignacio 4 1934 (-6.71 ) 1946 (-4.44); 1950 (-4.54); 1959 (-4.47); [1977) Ignacio 5 1934 (-6.78) 1946 (-4.44); 1950 (-4.67); 1951 (-4.19); 1959 (-4.44); [1977] Ignacio 6 1934 (-6.80) 1946 (-4.36); 1950 (-4.44); 1951 (-4.36); 1959 (-4.28); [1977] Ignacio 7 1934 (-6.69) 1946 (-4.28); 1950 (-4.40); 1951 (-4.49); 1959 (-4.08); [1977] 1926 (5.04); 1927 (4.49) 1926 (4.85); 1927 (4.26)' 1926 (4.93); 1927 (4.45) 1926 (4.89); 1927 (4.56) 1917 (4.05); 1926 (4.81); 1927 (4.74) 1917 (4.29); 1926 (4.76); 1927 (5.04) 1917 (4.51); 1926 (4.n); 1927 (5.28) ­88 - Extreme Wet Years (19411970) and [1971-1983]b 1941 (5.06); 1942 (5.85); 1957 (5.41); 1958 (5.40); 1965 (4.09); r1973 1979 19801 1941 (4.94); 1942 (5.91); 1957 (5.36); 1958 (5.26); r1973 1989 1980) 1941 (4.87); 1942 (5.92); 1957 (5.30); 1958 (5.13); r1973 1979 19801 1941 (4.69); 1942 (5.57); 1957 (5.27); 1958 (4.62); r1973 1979 1980) 1941 (6.23); [1973,1979] 1941 (6.14); [1973] 1941 (6.14); [1973] 1941 (6.12); [1973,1979] 1941 (6.14); [1973,1979] 1941 (6.14); [1973,1979] 1941 (6.11); [1973,1979] Table 3.5 (Continued). Extreme Dry Years (Dre-1941 ) 1934 (-6.59) Extreme Dry Years (1941-1970) and r1971-19831 b 1946 (-4.25); 1950 (-4.37); 1951(-4.52); [1977J 9 1934 (-6.48) 1946 (-4.22); 1950 (-4.34); 1951 (-4.54); [1977] Ignacio 10 1934 (-6.41 ) 1946 (-4.18); 1950 (-4.31); 1951 (-4.56); [1977] Ignacio 11 1934 (-6.20) 1946 (-4.07); 1959 (-4.29); 1951 (-4.56) Mesa Verde 1 1951 (-5.11); [1977] Mesa Verde Mesa Verde 2 1951 (-4.54); [1977] 3 1951 (-4.87); [1977] Weather Station Ignacio AWCa Class 8 Ignacio Mesa Verde Mesa Verde Mesa Verde Mesa Verde Mesa Verde Mesa Verde Mesa Verde Mesa Verde Ft. Lewis 4 5 6 7 8 9 10 11 1 1934 (-4.04) 1934 (-4.14) 1934 (-4.18) 1934 (-4.15) 1934 (-4.14) 1934 (-4.14) 1934 (-4.13) 1934 (-4.20) 1951 (-5.22); [1977198f1 1951 (-4.85); [197719811 1951 (-5.52); [197719811 1951 (-5.59); (19771981) 1951 (-5.61); f197719811 1951 (-5.60); [1977 1981'1 1951 (-5.59); [1977] 1951 (-5.54); [1977] 1964 (-4.21): r19771 -89- Extreme Wet Years (Dre-1941) 1917 (4.62) ; 1926 (4.77); 1927 (5.41) 1917 (4.71); 1926 (4.78); 1927 (5.54) 1917 (4.78); 1926 (4.81); 1927 (5.64) 1917 (5.10); 1926 (4.61 ); 1927 (5.71) 1926 (4.03); 1938 (4.39) 1938 (4.32) 1926 (4.00); 1938 (4.43) 1938 (4.46) 1938 (4.45) 1938 (4.47) 1938 (4.49) 1938 (4.51 ) 1938 (4.57) 1938 . (4.61 ) 1938 (5.09) Extreme Wet Years (19411970) and [1971-1983Jb 1941 (6.07); [1973,1979J 1941 (6.04); 1942 (4.56); [1973.1979J 1941 (6.03); 1942 (4.65); 1958 (4.84); [1973.1979] 1941 (5.78); 1942 (4.72); 1958 (4.59); [1973,1979,1980J 1941 (5.34) 1941 (5.41); 1957 (4.01); [19791 1941 (5.31) 1941 (5.31); [1979] 1941 (5.34); [1973,1979] 1941 (5.36); [1973.1979 19831 1941 (5.39); [1973 1979 19831 1941 (5.37); [1973 1979 19831 1941 (5.36); f1973 1979 19831 1941 (5.37); r19731979.19831 1941 (5.37); r1973 1979 19801 1941 (6.43) Table 3.5 (Concluded). Weather Station Ft. Lewis Ft. Lewis Ft. Lewis Ft. Lewis AWCa Class 2 3 4 5 Ft. Lewis 6 Ft. Lewis 7 Ft. Lewis 8 Ft. Lewis 9 Ft. Lewis 10 Ft. Lewis 11 Extreme Dry Years (pre-1941) .. Extreme Dry Years (1941-1970) and 1971-19831b 19771 19771 1964 (-4.27)· r19771 1959 (-4.06); 1964 (-4.55); r19n1 1959 (-41.2); 1964 (-4.86)· r19n1 1959 (-4.16); 1963 (-4.06); 1964 (-5.16); r19771 1959 (-4.17); 1963 (-4.17); 1964 (-5.31); (1977) 1959 (-4.13); 1963 (-4.29); 1964 (-5.47); r19771 1959 (-410); 1963 (-438); 1964 (-559); f19771 1963 (-472); 1964 (-599); f19771 Extreme Extreme Wet Years (1941Wet Years 1970) and [1971-1983]b (ore-1941) 1941 (6.16 1941 {6.34 1941 (6.37 1941 (6.41) 1941 (6.47); 1949 (4.07); r19731 1941 (6.48); [1973] 1941 (6.44); 1949 (4.25); [1973] 1941 (6.40); 1949 (4.32); [1973] 1941 (637); 1949 (436); [1973] 1941 (622); 1949 (453); f19731 ­ Inches ­ of SOIl and In the lower aSee Table 3.3 for descnptlOn of the amount of water held In upper SIX sod profile for each AWC class. bYears between 1971 and 1983 that were used in calculating CAFEC values necessary for deriving POSt, but not used in calibration procedures. Years 1941·1970 used in calibration. Pre-1941 years used in verification. Colorado July PDSI reconstructions (Rose et al. 1982:385). Table 3.6 summarizes data from these two studies for comparison with data in Table 3.4. Historic Commentary The agreement between actual and reconstructed PDSI values may be assessed in a qualitative manner, as well. While historic commentary may be regarded as anecdotal, it does provide a source of information that can be used to evaluate the general robustness of reconstructed PDSI values. This should be particularly noticeable in years when more extreme climatic conditions that affected local populations and their perceptions of "abnormality" were recorded. Historical remarks on weather conditions in Montezuma and Dolores counties have been assembled from-three sources: an oral history of the Goodman Point area recently assembled by MaJjorie Connolly of the Crow Canyon Archaeological Center, Cortez (Connolly, 1990); printed commentary in Cortez's newspaper, the Montezuma Valley Journal (Ian Thompson, personal communication, January 1990); and newspaper commentary gleaned from the files of the Mancos Times, Mancos Times-Tribune, and the Montezuma Valley Journal as reported in a published history of Montezuma County (Freeman 1958:207-317). The earliest remarks begin in 1894 and are contained in appendix B. Table 3.7 lists the annual mean reconstructed PDSI values of the 34 reconstructions associated with soils known to be used for agriculture in the Montezuma and Dolores County area for the period of 1890-1970 (see section on Historic Crop Yield and Total Regression, this chapter, for more detail on the identification of these 34 reconstructions). It also notes -90- concurrence with available historic commentary, for those years where comments were found. Figure 3.2 plots these annual mean PDSI values. Figure 3.3 plots the standard deviations for these mean values for the same 1890-1970 time period. Whereas the plot of annual mean values tracks the prevailing condition noted across the region during a given year, the plot of standard deviations provides infonnation on the spatial variability present in any year. The lower the standard deviation. the more consistent and widespread the condition. The higher the standard deviation, the more localized and discontinuous the condition. Table 3.6. Comparative Data from Long-Term Reconstructions in the Zuni Area and in Southwestern Colorado. SW coa Mean Standard deviation Minimum value Maximum value Skewness Kurtosis -1.07 1.85 -6.03 4.54 -0.04 -0.25 SW Mountains b -0.94 1.78 -6.29 4.16 -0.17 -0.30 Black Rockc -0.42 2.08 -6.58 5.31 -0.05 -0.21 aSouthwest Colorado reconstruction combines instrumented climatic data from Cortez, Mancos, Mesa Verde, Ft. Lewis, Ignacio, and Pagosa Springs, Colorado (Rose et al. 1982:237,385). bSouthwest Mountains reconstruction combines instrumented climatic data from Ft. Wingate, EI Morro, Bluewater, Grants, San Fidel, and Marquez, New Mexico (Martin Rose, personal communication, April 1989). cBlack Rock reconstruction uses climatic data from Black Rock, New Mexico (Martin Rose, personal communication, April 1989). Historic remarks (Appendix B) are associated with 41/81 (51%) years between 18901970. Of these, 37/41 (90%) unquestionably support the reconstructed PDSI annual mean values. 3/41 (7%) do not contradict the PDSI values but do not strongly support it, and only 1/41(2%) seem not to support the PDSI value at all. However, it is noted that in that year, 1947, there was much variability about the 11 Wet 10 9 ｾ ­-= ｾ ｣Ｎ＾ｾ ｖｊｾ ｃｬｾ ｾ 8 ­ ｾ 1 J ｾ 7 6 j 5 ｾ 4 1 3 2 Dry 1 1890 1910 1930 1950 1970 Years (A.D.) Figure 3.2. Plot of annual mean PDSI values by class for 34 long­term reconstructions (REC34), A.D. 1890­1970. Class 1 represents conditions of extreme drought, Class 6 represents conditions that are near normal. and Class 11 represents conditions of extreme wetness' re­ maining­classes are intermediate. • -91- Table 3.7. Comparison of Historic Comments with Annual PDSI Values for 34 Long-Tenn Reconstructions, A.D. 1890-1970. Year A.D. 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 Mean REC34 125.32 136.85 111.15 94.15 94.53 125.15 78.91 134.53 119.82 73.24 100.24 124.65 73.41 137.38 72.06 145.21 141.71 146.08 133.41 139.71 126.21 146.88 133.32 98.38 150.24 152.32 149.41 139.65 97.32 138.15 133.06 123.06 129.47 103.15 135.21 105.59 147.21 127.91 119.24 113.71 116.97 102.41 137.44 103.77 93.12 132.24 110.06 134.15 127.15 111.50 114.24 156.09 128.4,1 S.D. REC34 9.93 3.00 8.44 4.61 4.69 5.10 6.52 9.79 1.83 5.58 4.06 4.69 8.74 6.84 6.46 3.58 9.43 3.90 13.99 14.14 6.94 7.97 7.16 4.26 4.16 4.92 4.87 9.69 9.26 7.87 6.32 2.64 5.18 4.21 9.27 5.62 5.49 3.58 6.17 1.43 4.48 8.61 10.22 5.67 7.01 3.64 3.58 5.03 3.97 2.72 4.47 3.97 11.13 PDSI Class 7 8 5 3 3 7 1 8 6 1 3 6 1 8 1 9 9 9 8 8 7 9 8 3 10 10 9 8 3 8 10 6 7 4 8 4 9 7 6 5 6 4 8 4 3 8 4 8 7 5 5 10 7 PDSI Class Condition incipient wet slightly wet incipient drought moderate drought moderate drought incipient wet extreme drought slightly wet near nannal extreme drought moderate drought near nannal extreme drought slightly wet extreme drought moderately wet moderately wet moderately wet slightly wet slightly wet incipient wet moderately wet slightly wet moderate drought unusually wet unusually wet moderately wet slightly wet moderate drought slightly wet unusually wet near nonnal incipient wet slight drought slightly wet slight drought moderately wet incipient wet near nonnal incipient drought nearnonnal slight drought slightly wet slight drought moderate drought slightly wet slight drought slightly wet incipient wet incipient drought incipient drought unusually wet incipient wet -92- Historic Agreement yes yes yes yes yes yes yes yes yes yes yes yes no, not in Cortez ok yes ok yes yes ok yes yes yes Table 3.7 (Concluded). Year A.D. 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 Mean REC34 104.06 118.35 120.71 88.71 130.15 130.41 142.21 97.24 79.94 145.62 102.4'7' 108.68 110.35 98.50 133.03 122.09 80.27 126.27 108.62 120.88 108.56 110.68 145.18 119.77 106.62 120.15 112.53 115.59 S.D. REC34 6.80 3.69 4.23 5.17 6.95 1.78 4 37 6.58 5.00 9.14 7.12 2.14 5.84 5.53 6.83 7.05 8.04 5.33 5.38 5.69 3.68 7.93 4.37 7.06 5.41 7.83 6.49 5.80 PDSI Class 4 6 6 2 8 8 9 3 1 9 4 4 4 3 8 6 1 7 4 6 4 4 9 6 4 6 5 5 PDSI Class Condition slight drought near normal near normal severe drought slightly wet slightly wet moderately wet moderate drought extreme drought moderately wet slight drought slight drought slight drought moderate drought slightly wet near normal extreme drought incipient wet slight drought near normal slight drought slight drought moderately wet near normal slight drought near normal incipient drought incipient drought Historic Agreement ok yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes Note: For the entire 81 -year period, the overall PDSI class mean and standard deviation is 6.03 ± 2.50. Historic remarks are listed in Appendix B by year. "Yes' indicates clear agreement with available historic commentary; "ok" indicates no contradiction but no strong relationship; "no' indicates contradictory conditions indicated from available remarks. The absence of a comment means no historic remarks were found for that year. mean, and meteorological drought varied considerably from place-to-place and soil-to-soil. Figures 3.2 and 3.3 illustrate that the turnof-the-century years 1896-1904 were notably dry, especially 1896, 1898, 1902, and 1904. The period between 1905-1920 was notably wet, especially 1914, 1915, and 1920, where only 2/16 years (1913, 1918) were ever below the long-term average (i.e., PDSI Class 6). The years 1921-1930 were rather average as a whole but exhibited much year-to-year variation. The decade between 1931 and 1940 was a notably dry period with 6/10 years drier than the long-term average, especially 1934. The 1941-1949 period was quite variable with than was usual, especially more extreme ｶｾｵ･ｳ the years 1942, a year of high spatial variability; 1941, a very wet year, and 1946, a very dry year. The years 1950-1959 were notably dry, particularly 1951 and 1959. The final years in the series, 1960-1970, were moderate with little inter-annual variation. The 1905-1920 wet period and the 19311940 dry period were widely experienced in the western United States and show up in numerous climatic reconstructions (Fritts 1965). It is also interesting to note that the modem practice of dry-farming in Montezuma and Dolores Counties began in 1905 and 1906 near Mancos (Freeman 1958:215) and 19091910 near Yellowjacket (Freeman 1958:144) in years with greater than average soil moisture. Experimentation continued through -93- 16 14 12 '0 "'= " '0 ­ e = 10 tI) fIl file U·ｾｯ ＭＮｾｴｉＩｾ ｑｾ ­ =.oQ 8 6 4 ｾ ｾ • n ｾ l1 11 ｾ ｾ :II 3I!Il II 2 X­ ｾ U I!I IrJ • ｲｾ ｉｾ o 1890 1930 1910 1950 1970 Years (A.D.) Figure 3.3. Plot of annual standard deviations for mean PDSI values for 34 long-term reconstructions (REC34), A.D. 1890­1970. The mean standard deviation is 6.03 ± 2.50; ± one standard deviation ranges between 3.53­8.53, ± two standard deviations ranges between 1.03­11.03. Thus, three years, 1908, 1909, and 1942 are beyond two standard deviation units from the normal amount of spatial variation exhibited in the Montezuma County and Dolores County area. the wetter­than­average years prior to and during the First World War. It was found to be very successful and was adopted by many farmers in the area by 1926­1927 (e.g., within Dove Creek, and the communities of ａ｣ｬｾｭ･ｮＬ Cahone [Freeman 1958:148, 152]), again in a period when soil moisture conditions were almost never unfavorable to this type of nonirrigated agriculture. April 1989), using very similar tree­ring data as predictor variables. Further, it appears that the reconstructed series of PDSI values closely match the historic record available for Montezuma and Dolores Counties. This adds qualitative support for the general accuracy of the reconstructed series and their usefulness for paleoenvironmental reconstruction. Thus, all reconstructions were judged to be acceptable and were used in the final GIS analyses. Conclusion The 55 long­term reconstructions based on the 55 full­period calibration transfer functions appear to be quite reasonable and compare well against other long­term reconstructions for locales on the southeastern Colorado Plateau. Reconstructions based on June PDSI do not seem to be markedly different than the July PDSI reconstructions made by Rose et al. (1982) and Rose (personal communication, GEOGRAPillC INFORMATION SYSTEM PROCESSING Elevation Data Digital Elevation Models (OEMs) are not instantly usable when they are obtained from the USGS. There is usually some preprocessing to read in, transform, mosaic, correct, and rescale images before they can be used in ­94- analysis. In this case, the ASCII tape containing the 12 DEMs was mounted on a drive connected to the IBM 3090-200 mainframe at WSU and processed by the VICAR command VTOPA. VTOPA logs in the digitized data and records selected header information about each DEM file, including map identification number, labeling information, and geographic coordinates of the four comers of the image. These were used thereafter as descriptive labels for succeeding VICAR programs. VTOPA also examines the data records of each DEM and checks for possible errors in the elevation files. Finally, it creates a binary 16-bit pixel rectilinear image with one DEM elevation sample per pixel location in VICAR format. Next, each image was rotated 90 degrees counter-clockwise by VICAR command R..OT so that the top of each image is north. Following rotation, the individual rectangular images were pieced together to create a single uninterrupted image. This process is called "mosaicing" and was accomplished by VICAR commands FASTMOS, which matches the edges and joins the pieces, and by SAR, which interpolates missing values. Subsequent VICAR commands reexpressed the data so that they could be processed by EPPL7 programs. Command C converted the 16-bit pixels to 8-bit significance thereby compressing the original number of elevational classes, representing altitudes from I,SOO3,012 m, to some number between ｾＲＵｓＬ the maximum class range for EPPL7 (28= 2S6). The 1512 different elevational classes or "DNs" (Density Number, a term borrowed from remote sensing) were reduced to 189 by grouping elevational classes in groups of eight meters. Thus, DN#1 equals elevations 1500IS07 m, DN#2 equals 1508-1515 m, etc. Commands TIECONM, MGEOM, and STRETCH resampled the 3D-m cell resolution of the original DEM to match the 200-m cell resolution of the soils data. Here, the original image with 1,425 lines (rows) and 1,508 samples (columns) and a 30-m cell size was "degraded" by a factor of about 6.6 to produce an output image with only 214 lines (rows) and 227 samples (columns) and a 2oo-m cell size. Other programs created histograms of the DN values and line printer maps, permitting visual inspection of the data and facilitating error detection. The final mosaiced. corrected, and rescaled output image, now formatted as a binary raster file, was downloaded to an mM PC on which the EPPL7 GIS program was loaded. This EPPL7 file (CRVDEM.EPP) was the basic elevational data plane for this study. Later, a "trimmed" or "windowed" version of this image was created with EPPL command WINDOW. Here the 14 bottom-most rows were deleted from the analysis leaving an image with 200 rows and 227 columns (DEMWIND.EPP). This was done largely because so little land in Utat part of the study area had been mapped for soils; much of this land is Ute Mountain Ute Indian Reservation and had not been examined by SCS personnel. Finally, using EPPL's RECLASS command, this second "trimmed" elevational data plane was used to create a third elevational data plane (SELEV.EPP) that grouped the 189 elevation classes into five elevational classes. Here each class was the range of elevations to be modeled by each of the weather stations. Thus, Class 1 included those cells on the image modeled by the set of Bluff reconstructions (DNs# 1-13) representing elevations 1,500-1,604 m, which covered 2.71 % of the study area. Class 2 included those cells modeled by the Cortez reconstructions (DNs# 14-54) representing elevations 1,605-1,932 m, which covered 40.97% of the study area. Class 3 included the Ignacio reconstructions (DNs# 5S-71) representing elevations 1,933-2,068 m, which covered 33.44% of the study area. Class 4 included cells modeled by the Mesa Verde reconstructions (DNS# 72-93) representing elevations 2,069-2,244 m, which covered 20.45% of the study area. Class 5 included cells modeled by the Ft Lewis re-constructions (DNs# 94-189), representing elevations 2,245-3,012, which covered 2.32% of the study area. Soil Data As described in Chapter 2, information on soil type and location was collected and recorded by hand for each 200 x 200 m cell in the study area. Cells were grouped in blocks of 100 cells, 10 rows (west to east) by 10 columns -95 - (north to south). Each block was identified by the UTM coordinate of the northwest comer of the northwestenunost cell. This point corresponds to some readily visible tick mark on a 7.5-minute USGS topographic map or orthophoto quad for ease in location. In this manner, a total of 502 blocks representing 50,200 cells were recorded in the Soil Conservation Service field office. The first task, then, was to input these data into a computer program that could be read by VICAR and would allow each block to be inserted into a pre-specified "picture window" in a fashion similar to placing puzzle pieces into an image space. A mainframe editing program, WYLBUR, accepted the original soil data input as numeric and alphanumeric codes. Later, all alphanumeric codes were reexpressed as 113 numeric codes, each representing a different soil map unit, to facilitate computer processing. Next, an overall image or "picture window" was created with VICAR program GEN to encompass the full set of soil<oded blocks in their proper geographic positions. The image's overall dimension were taken from the maximum dimensions of the three (north to south) by four (west to east) grid of topographic maps equivalent to the 12 mosaiced DEMs. The VICAR program QSAR pieced each l00<ell block into its appropriate location within the larger picture window. After extensive checking for code errors and block mismatches, the finished 214 row x 227 column image was downloaded to the IBM microcomputer and assigned an EPPL7 filename (CRVSOD....EPP) and was ready for EPPL7 processing. The initial EPPL7 processing task was to register the soil plane with the elevational plane making sure they were georeferenced to the same earth coordinates and expressed the same extent of areal coverage. An auxiliary computer system, International Imaging Systems (l2S), was used to check for basic registration. An I2S command, FLICKER, alternately displayed both elevation and soils planes in such quick succession that it was easy to see whether or not map features fell into alignment A need for a small adjustment to the position of the soils plane was "detected by this method and was easily made with EPPL7. EPPL command WINDOW permitted the coregistration and windowing of the soil plane that would match the elevation plane and resulted in the creation of a 200 row x 227 column soil plane (SOILWIND.EPP). A third soils plane (lIAWC.EPP) was then created with command RECLASS that grouped all 113 potential soils into the study area into II groups of soils based on their available water holding capacity. This was the same grouping of soils used in the calculation of PDSI values and the long-term PDSI reconstructions. New Data After the creation of 5ELEV.EPP, which identified the elevational zones to be modeled by the five weather stations, and I IAWC.EPP, which grouped all soils in the study area into one of II soil groups based on soil moisture characteristics, a new data plane was created which combined these two. This data plane, PDSIMAP.EPP, was created with EPPL command EVALUATE. It created 55 classes where each class corresponds to a long-term PDSI reconstruction and any given cell was associated with its assigned available water holding class and its assigned elevational zone. In the case of this study, only 44 of the 55 long-term reconstructions were actually represented in the study area, but their numbers ranged from 155, despite nonrepresented classes. The creation of PDSIMAP.EPP greatly facilitated the processing of the final data planes and provided the link between the elevation and soils data and the long-term PDSI reconstructions. Arcbaeolo&ical Data Block Survey Areas The perimeters of the two survey areas selected to test the model's effectiveness, the Mockingbird Mesa and Sand Canyon localities, were digitized with the EPPL7 DIGITIZE program. These survey areas bad been plotted on USGS 7.5-minute maps and were digitized with standard digitizer table and mouse equipment connected to the microcomputer containing -96- the digitizing program. The digitized files. MOCK..DOT and SAND.DGT were used to create survey area­shaped "windows" (MOCK..EPP and SAND.EPP) within the larger study area. These windowed areas were irregular in shape. MOCK.EPP depicts the Mockingbird Mesa Survey Locality and incorporates 17.96 km 2 of land surface. SAND.EPP portrays the Sand Canyon Survey Locality and encompasses 26.08 km 2 of land. They were used in separate runs of the final model to depict contrasting agricultural productivity in two subregions of the study area and to compare the results with that of the entire study area. Tree­Ring Dated Sites Nine sites were selected to test the model's ability to discern spatial and temporal patterning in agricultural productivity on a local scale. Two of these nine, 5MT6970 (Wallace Ruin) and 5MT4126 (Ida Jean Ruin), are very close to each other in time and in space and may, in fact, be part of the same architectural complex. Consequently, they were combined for the purpose of this analysis, and a point halfway between the two chosen to represent the center of the site. The site number for Wallace Ruin was retained and the site number for Ida Jean dropped. Therefore, the UTM coordinates of the eight sites were entered into an EPPL7 data file via a microcomputer editing program, KEDIT. The EPPL commands GRIDPOINT, BUFFER, and EVALUATE were used to create an eight­cell (1.6 km) radius aroWld the cell containing a given site. This created a roughly circular, 3.4­km­diameter (7.88 km 2), "site catchment" area for each of the eight sites (5MT8371 5MT8839, 5MT2433, 5MT3834, 5MT6970, ＵｾＴＱＲＶＬ 5MT2149, and 5Mf765). These were used later in eight separate iterations of the final model and generated annual data on productivity within the area of the catchment. LINKING SOn.. MOISTURE CONDmONS AND LEVELS OF AGRICULTURAL PRODUCTIVITY As discussed in the section of this chapter on data acquisition, information on historic yields of drylaIid bean and maize was gathered to address two methodological problems. The first problem was to establish the nature and the strength of the relationship between modem crop yield and modem soil moisture conditions as modeled by Palmer Drought Severity techniques. If a relationship could be demonstrated, it would provide the shape of the function that must be fitted to the PDSI values generated by the long­tenn PDSI reconstructions and to the yield estimates of prehistoric agricultural production. The second problem was to develop a method whereby specific yield values­currently available for only about half of the soils in the study area ­could be estimated for all soils. If these two problems could be solved, crop yield estimates could be associated with each soil type and each soil linked to its appropriate 1,070­year PDSI reconstruction via GIS technology. Historic Yield and Soil Mojsture Conditions: Establishing a Relationship Yield values for pinto bean and maize production grown by dry farming techniques have been compiled by Bums (1983) for five COWlties in southwestern Colorado. Since the vast majority of the study area is located in Montezuma County, only those records for Montezuma COWlty (Bums 1983:315) were used for this particular study. These yield figures are presented as poWlds of beans per harvested acre or bushels of maize per harvested acre and are presented as mean values (Table 3.8, columns 2 and 3). Bums' values for 1931­1960 were used in this analysis. The selection of 1931 as a begirming date was conditioned by the availability of joint monthly precipitation and temperature records for Cortez, Colorado, which were required for the calculation of Palmer Drought Severity Indices during the historic period. The ending date of 1960 was determined by the lack of maize data after that time, even though bean data continued to be recorded. As a commercial venture, dryland beans grown in southwestern Colorado were successful and quite competitive in regional, national, and even international markets, but maize grown in southwestern Colorado was not Consequently, maize was decreasingly grown as a commercial product and official records end at 1960. Thus, only 30 years of jointly available crop yield data, which were ­97- Table 3.8. Historic Crop Yields and PDSI Values. A.D. 1931-1960. Bean Yield Maize Yield Year ACT ROC ACT ROC REC ROC A.D. (OR1G) PDSlc (GIS) PDSld (lbs/ac)8 (bu/ac)b PDSt e 1931 145 115 270 2.53 9 5.31 173 151 1932 350 12 -1.18 108 1933 12 113 240 1934 87 -3.34 4 99 170 101 -1.94 138 1935 240 10 88 -3.15 114 1936 280 9 128 143 1937 0.80 12 300 143 117 2.32 12 1938 360 106 -1.41 112 1939 260 9 106 125 -1.39 12 1940 510 169 157 4.87 1941 400 20 179 139 5.92 15 1942 350 144 2.41 120 1943 16 740 150 127 3.03 1944 12 380 118 125 -0.23 1945 16 430 92 -2.14 99 18 1946 420 123 -1.95 100 20 1947 600 133 0.63 126 1948 410 13 134 145 1.35 1949 480 19 97 83 -2.30 16 1950 350 67 -5.33 79 1951 15 170 146 131 2.61 1952 15 520 94 123 14 0.26 1953 480 99 108 1954 14 -2.08 330 103 96 14 -1.71 1955 530 101 -1.87 93 1956 304 18 173 115 1957 5.30 692 19 171 129 1958 16 5.13 350 96 81 -3.87 1959 220 15 100 132 16 -2.00 1960 280 aBean yield from Burns (1983:315). bMaize yield from Bums (1983:310). cACTROC, computed POSI values for ROC soil (reconstruction #21; C 0.9910.26 in). dACTROC. in reexpressed POSI values in GIS units ranging from 1-240. eRECROC, reconstructed POSI values (GIS units) for ROC soil. fREC34, reconstructed mean POSI value for 34 soil group reconstructions (GIS units). coterminous with the historic PDSI data, were used in this particular study. Historic Crop Yield and Total Regression The Soil Conservation Service provides nonirrigated bean yield data in the form of pounds of beans per harvested acre for 46 dif­ ferent soil types in Montezuma and Dolores Counties (Table 2.3, column 9). Forty­four of these 46 occur in the study area. Of these 44 the soil type mapped as ROC, Witt Loam 3­6% slope is by far the most abundant agricultural soil in the study area. A total of 5,868 cells, representing 15·.96% of the 36,759 analyzable REC34 PDSlt 102 137 104 93 132 110 134 127 112 114 156 128 104 118 121 89 130 130 142 97 80 146 102 109 110 99 133 122 80 126 cells, was assigned this soil type. For the purpose of this study, a combined measure of the 44 soils currently evaluated for modern dry bean farming, as well as two mea­ sores of the ROC soil as it is modeled at the Cortez elevation, were used to examine the re­ lation between crop yield and soil moisture conditions. In order to create a combined measure that might better reflect the mean yield values for Montezuma County compiled by Burns (1983) (Table 3.8, columns 2 and 3), nested frequency counts of the 55 PDSI re­ constructions and all of the soil types in the study area were examined. Often a soil occur- -98 - red in more than one elevational range. In addition, different soil map units, because of similarity in their water­holding capacities, fell into the same POSI reconstruction (Table 3.9). The 44 different soil types for which yield data had been compiled were associated with 34 of the 55 reconstructions. The POSls for each of the 34 were summed annually from 19311960 and an annual mean was taken. This mean index value was called "REC34" (Table 3.8, column 7). Similarly, but more simply, the annual POSI value for reconstruction #21 (Le., those soils that fell into the soil class defmed on its available water holding capacity of 0.99 inches of water in the upper six inches of soil and 10.26 inches of water in the lower soil profile using the Cortez climatic reconstruction), which includes ROC (Witt Loam, 3-6% slope), was recorded for the 1931-1960 period (Table 3.8, column 6), as was the original POSI calculation for the historic period (Table 3.8, columns 4 and 5). First it was necessary to establish that a strong relationship existed between the reconstructed POSI estimates and the actual POSI calculations for the 30 years under consideration so that a case could be made for the usefulness of the reconstructed series for modeling soil moisture. A plot of the reconstructed POSI values for the ROC soil and the actual POSI calculations for the ROC soil during the 1931-1960 period produced a positive linear regression line with an r = .65 (significant at the .01 level). Values greater than .80 are considered high, around .50 considered moderate, and below .30 considered low (Downie and Heath 1970:101). While not perfect, regression indicates that these two data sets are moderately to highly correlated and that the reconstructed values are reasonably accurate predictors of the actual calculated values. Second, the actual POSls and the reconstructed POSIs for 1931-1960 were regressed against bean yield, first for the ROC soil and then for the 34 soil groups (Table 3.10, middle column, top half). The highest linear correlation was between actual POSI values for ROC soil and bean yield, with an r = .41. Although this is significant at the .05 level, the correlation is rather low. AU of these plots were better represented by second-degree polynomial equations in which an inverted U or a parabolic, curvilinear shape was fitted through the points. Again, the actual POSI data for ROC soil when plotted against the bean yield data produced the best correlation of the test attempts (r .51). = Third, linear and curvilinear regression lines were fitted to the maize data in the same Table 3.9. Distribution of Soils for Which Nonirrigated Bean Yield Data Are Available by PDSI Reconstruction Class. Recon. Weather Station Bluff Bluff Bluff Bluff Bluff Bluff Bluff Bluff Bluff 10 Bluff Bluff 11 Cortez 12 Cortez 13 Cortez 14 Cortez 15 Cortez 16 17 Cortez # 1 2 3 4 5 6 7 8 9 AWCValues (Su + SI)8 0.53 4.61 0.85 2.95 1.02 3.65 1.02 4.48 1.02 5.40 1.14 6.44 1.02 7.80 1.00 8.54 1.05 9.45 0.99 10.26 1.14 13.66 0.53 4.61 0.85 2.95 1.02 3.65 1.02 4.48 1.02 5.40 1.14 6.44 Soil Map Unit M2DD, M2DD1 ROLC, ROLD, ROLD1 R7C,R7D,R7D1 H1CD, R4D, ROHe, ROHD RHC -99- Table 3.9 (Concluded). Recon. Weather # Station 18 Cortez 19 Cortez 20 Cortez 21 Cortez 22 Cortez 23 Ignacio 24 Ignacio 25 Ignacio 26 Ignacio 27 Ignacio 28 Ignacio 29 Ignacio 30 Ignacio 31 Ignacio 32 Ignacio 33 Ignacio 34 Mesa Verde 35 Mesa Verde 36 Mesa Verde 37 Mesa Verde 38 Mesa Verde 39 MesaVerde 40 Mesa Verde 41 Mesa Verde 42 Mesa Verde 43 Mesa Verde 44 Mesa Verde Ft. Lewis 45 46 Ft. Lewis 4 7 F t . Lewis Ft. Lewis 48 Ft. Lewis 49 50 Ft. Lewis 51 Ft. Lewis 52 Ft. Lewis 53 Ft. Lewis 54 Ft. Lewis 55 Ft.Lewis AWC Values (Su + Sl)a 1.02 7.80 1.00 8.54 1.05 9.45 0.99 10.26 1.14 13.66 0.53 4.61 0.85 2.95 1.02 3.65 1.02 4.48 1.02 5.40 1.14 6.44 1.02 7.80 1.00 8.54 1.05 9.45 0.99 10.26 1.14 13.66 0.53 4.61 0.85 2.95 1.02 3.65 1.02 4.48 1.02 5.40 1.14 6.44 1.02 7.80 1.00 8.54 1.05 9.45 0.99 10.26 1.14 13.66 0.53 4.61 0.85 2.95 1.02 3.65 1.02 4.48 1.02 5.40 1.14 6.44 1.02 7.80 1.00 8.54 1.05 9.45 0.99 10.26 1.1413.66 Soil Map Unit R4C, R4C1 R3B, R3C, R3D, R3B1, R3C1, R3D1, WOBB A5C, WOB ROB, ROC, ROD, ROB1, ROC1 M2DD ROLC, ROLD R7C,R7D H1CD,R4D,ROHC,ROHD RHC R4C,R4C1 R3B, R3C, R3D, R3C1, WOBB ASC, A5D, WOB ROB, ROC, ROD M2DD ROLC, ROLD R7C,R7D H1CD, R4D,ROHC, ROHD,S6C,S7 RHC R4C, S5 R3C,R3D,S51,S5D,S8C,V4C,WOBB ASC, ASD, V1 C, WOB, WOC M5E, ROB, ROC, ROD, SSE R1C,R1D S6C, S7 S5 S5D, V4C, S8C V1B, V1C M5E, SSE R1C,R1D Note: This list associates 44 different soil types with 34 different PDSI reconstructions. Some soil types occur with more than one reconstruction because they occur in different elevational ranges. Two soil types listed in Table 3.3, C3CD and V1 D, are not present in the study area but are located elsewhere within Montezuma and Dolores Counties and are associated with Reconstructions #52 and 53, respectively. aAWC Values (Su +SI): Available water­holding capacity of upper six inches of soil and lower soil profile. way they had been to the bean data (Table .3.10, middle column, bottom half). Again, actual and reconstructed PDSI values for the single reconstruction for ROC soil and the combined reconstruction for 34 long-term reconstructions were used. Unlike the bean data, the best linear equation was between the com- bined 34 reconstructions instead of with the single ROC. reconstruction. Despite this, none of the linear equations are considered significant (at the .05 level), the best having only an r = .25, which is considered to be low. As with the bean data, all the maize plots were better modeled by second-degree polynomial 100 - Table 3.10. Results of Linear, Polynomial, and Partial Regression Studies of the Relationship Between PDSI and Crop Yield Values. For the 30-year period, the mean bean yield = 380.53 ± 141.16 pounds/acre and the mean maize yield = 14.07 ± 3.69 bushels/acre. Beans Actual PDSI for ROC Reconstructed PDSI for ROC Mean for REC34 Simple Linear Regression and Polynomial Regression y = 152.2844 + 1.8683x f = .41 ,2 = .17 (significant at .05) Partial Regression y = ­9782.55180 + 5.09523678x1 + 2.04829764x2 f= .52,2= .27 (significant at .01) y = -566.7831 + 13.8224x-0.0468x2 f= .51 ,2 = .26 (significant at .01) y = ­10.971.18229 + 5.71318142x1 + y = 238.016 + 1.2064x f= .17 ,2 = .03 (not significant at .05) 2.00384746x2 f= .38 ,2 = .14 (Significant at .05) y = -724.6224 + 18.1378x -0.0723x 2 f=.29,2 = .08 (not significant at .05) y = ­9667.04731 + 4.99434456x1 + y = 82.1425 + 2.5672x f= .35,2 = .12 (not significant at .05) 2.84843676x2 f = .47 ,2 = .22 (significant at .01) y = -633.0555 + 15.2679x 0.0549x 2 f= .38 ,2 =.14 (significant at .05) Actual PDSI for ROC Simple Linear Regression and Polynomial Regression y = 10.8752 + 0.0261x f= .22 ,2 = .05 (not significant at .05) Reconstructed PDSlforROC y = 15.8618 ­ 0.0568x + 3.243e­4x2 f= .23 ,2 =.05 (not significant at .05) y 12.3279 + 0.0147x f= .08 ,2 = .01 (not significant at .05) Maize = y Mean for REC34 Partial Regression y = ­525.61342 + 0.27517129x1 + 0.03546922x2 f=.69,2 = .47 (significant at .001) y = ­580.67051 + 0.30224351x1 + 0.05690542x2 f= .69 ,2 = .47 (significant at .001) =44.5n2 ­ 0.5525x + 0.0024x2 f= .30,2 = .09 (not significant at .05) y = 8.5048 + 0.0479x f= .25 ,2 .06 (not significant at .05) = =45.0841 ­ 0.6017x + 0.0028x2 f= .41 ,2 = .17 (significant at .05) y = ­532.58051 + 0.27718880X1 + 0.06346171x2 f =.70 ,2 =.49 (significant at .001) y equations. As with the linear model, the combined 34 reconstruction PDSI model had the highest correlation coefficient (r .41) and was significant at the .05 level, but is still considered weak. However, this curvilinear . equation, as were the other two, was fitted with an upright V­shaped curve. The results did not = ­ fulfill expectations about how the yield data were related to soil moisture conditions. Historic Crop Yield and Partial Regression In general the results of the above analyses were not very convincing. However, another 101- course of action still remained to search for pattern in the data that might provide the relationship needed to relate soil moisture conditions and crop yield. This was to use partial linear regression on the data with, "time" also considered as an independent variable. With partial regression techniques, the contribution to crop production by the first independent variable, time, can be controlled for and the unique contribution made by the second independent variable, soil moisture, assessed. This was thought to be potentially very useful because time might also incorporate the influence of modem technological changes on crop yield. Burns (1983:70-73) refers to this as the "technology trend" and includes the role that technological components such as tractors, fertilizer, pesticide, fungicide, seed selection, and cultivation practices play in effecting crop yield. He used the application of fertilizer to crops as a proxy measure of the general influence of technology on crop production, although he recognized that the use of chemical fertilizers on dry farming lands may be questioned (Bums 1983:75-76). The partial regression study was accomplished by specifying regression options in the Statistical Analysis Systems (SAS) program, REG. It was performed first on the bean data and then on the maize data; summary statistics and residual scatter plots were produced for each. Three models were specified for each crop type. The first specified that the dependent variable (crop yield) would be regressed on independent variables year (time) and actual PDSI values for the single ROC soil group for 1931-1960. The second specified that the regressors were year and reconstructed PDSI values for the ROC soil. The third specified that the regressors were year and the mean value for the 34 soil groups, which represented the 44 yield-documented agricultural soils. The data used in the study were those listed in Table 3.8 and also used in the linear and polynomial regression studies. For the purpose of comparability between the bean and maize data, the 30-year period of 1931-1960 was used for both crops. The results of the analysis are summarized in Table 3.10 (right column). The equations and regression coefficients generated for the same data for 1931-1960 period using simple linear regression and the second-degree polynomial regression are included for comparison. The use of partial regression was successful in increasing the correlation coefficient (r) and the index of determination (r 2 ) in every case. Where only one simple linear equation for bean yield and the actual PDSI calculations was successful in producing a significant correlation, all six of the equations derived under the partial regression method produced statistically significant correlations. Under both simple linear and partial regression, the bean yield was better explained by the actual PDSI calculations for ROC soil than by the reconstructed PDSIs for ROC or the mean value of the combined 34 soil groups. Also, under both simple linear and partial regression, the combined 34 reconstructions model explained more variation in maize yield than either of the ROC data sets. In comparing the partial regressions on the bean data versus those on the maize data, the correlation coefficients of the maize models increased dramatically and moderately strong, positive, linear relationships appeared between soil moisture and maize yield. The "technology trend", as proxied by the variable time, was much more strongly reflected in the maize than in the bean data. Once the contribution of this variable was removed, the relationship between increasing amounts of soil moisture and increasing crop production was visible graphically and numerically in the correlation coefficients. In contrast, the bean data seemed to take on a more curvilinear function with increasing soil moisture correlated with increasing bean yields up to a point and then decreasing yield with the highest levels of soil moisture. In sum, partial regression analysis succeeded in demonstrating a relationship between PDSI values and crop yields. Of most importance was the partial regression for maize yield and the 34 combined reconstructions, which is thought to best reflect the soil types in the study area from which the reported maize yield was derived. This regression indicated that a positive linear relationship exists between reconstructed PDSI values and modem maize production. The partial regression not only demonstrated a significant relationship between soil moisture and crop yield (r = .70, r 2 = .49, significant at .(01), but it also correctly Historic Crop Yield and Natural Plant. Productiyity: Estimating Prehistoric Crop Yields for IndiYidual Soils The second methodological problem was to develop a procedure whereby potential crop yield values could be assigned to each soil type in the study area and could vary under different climatic and soil moisture conditions. Such a method was developed and it involved the use of modem bean yield data for specific soils compiled by the SCS (Table 2.3, colwnn 9) and potential natural plant productivity data for all soils estimated under favorable, unfavorable, and normal growing conditions (Table 2.3, columns 6-8). Ultimately, these estimates of productivity expressed as bean units would have to be reexpressed as maize units, for the reconstruction of maize production is one of the goals of this research. The mean yield estimates for pinto beans grown on nonirrigated lands in Montezuma and Dolores Counties have been compiled by the Soil Conservation Service. These estimates have been made by SCS agents on the basis of reported harvests by individual farmers who commercially grew beans on these soils for a number of years. No comparable maize data exist. These values are associated with specific soil types. Of the 113 different soil map units in their two-county jurisdiction, the SCS has compiled data on mean yield from 46 different soil map units. Forty-four of these 46 are present in the study area and represent 45% of all the 98 soil map units (by number) that actually occur in the study area. The soils for which yield data have been compiled represent the majority of local soil types that modem dry farmers use to grow pinto beans in commercial quantities. These farmers use mechanized equipment and regularly apply pesticides and fungicides but less often use fertilizers to grow beans on nonirrigated lands. Because their capital investment is relatively great and the price of harvested beans relatively low, they must plant and harvest a significant area to make a profit Consequently, their fields are generally on interior mesatop locations where the soil is deeper and can hold greater soil moisture, where slope does not present major problems for equipment and slope erosion, and where large patches of arable land occur. All this is to say that these 46 soils are not the only ones that can successfully grow nonirrigated crops, particularly at the scale and technological level believed to have been employed by Anasazi farmers. Therefore, it is not unreasonable to suppose that additional soil types were also used by prehistoric farmers of the area, some not included on the list of 46. Therefore, a general method for estimating crop yield values on all soils in the study area was devised. First, a linear regression equation derived from the relationship between bean yield and potential natural plant productivity on the 44 known soils was used to estimate crop yield values for the entire set of 98 soils in the study area. The values were plotted, the least-squares regression line calculated for the scatter (y = -116.4617 + 3.1745x), and the correlation coefficients determined to be significant (r .77, r 2 .59, significant at the .01 level) and valid for the purposes of linear estimation. Using this model, the natural productivity values for all soils are used to derive predicted bean yields for all soils. This results in a series of estimated crop yields for all the soils in the study area grown under "normal climatic conditions" (taken to mean average precipitation and temperature conditions). Next, the estimates of natural productivity grown under favorable and unfavorable climatic conditions for all soils were used to estimate crop yields for all soils. Although this could have been accomplished again through linear estimation from the regression equations, an alternative method was used. Here the percentage change from favorable to normal conditions and again from unfavorable to normal conditions was calculated, and this percent used as a multiplier to the normal value. For example, the natural productivity for one soil is 1,200 pounds per acre in favorable years, 900 pounds per acre in normal years, and 700 pounds per acre in unfavorable years. If the crop yield for normal years was eStimated to be 320 pounds of beans per acre, then 320 pounds is multiplied by 1,200/900 or 1.33 to derive 427 pounds of beans per acre for favorable years, and by 700/900 or .78 to derive 249 pounds of bean per acre for unfavorable years. Subsequent to calculating the favorable and unfavorable categories of yield, two intermediate sets of values representing a condition between normal and favorable, and another between normal and unfavorable, were created -103 - = = so that five different "states" or "choices" were possible, and these five could be fit to the PDSI values. This was accomplished simply by halving the difference between normal and favorable yield estimates to provide normal-tofavorable category and again by halving the difference between normal and unfavorable estimates creating a normal-to-unfavorable category. Table 3.11 provides bean yield estimates for the five categories. The Relationship Between Bean Yield and Maize Yield The values created by the above steps create a matrix of values where the crop yield reconstructed is dry bean production. This was necessary because only values for dry bean production per soil type were available from the Soil Conservation Service for lands in Montezuma and Dolores Counties. However, the estimation of bean production for all soil types under varying climatic and soil moisture conditions is a necessary but intermediate step toward the estimation of maize production under varying climatic and soil moisture conditions, for maize is really the crop of interest relative to prehistoric Pueblo subsistence. Two points must be emphasized with regard to these bean yield estimates. First, as individual values these numbers are only best­guess estimates based on the least­squares regression formula; a real world value might be somewhat different. Second, these values are relative estimatesproduction values relative to the same soil under different growing conditions, and relative to all other soil types present in the area. Likewise, when the relationship is established between bean growth and maize growth, and when these relative production units are converted from beans to com, these qualities of approximation and relativity must also be seen to be associated with the resulting maize yield estimates. The relationship between bean and maize yield was established by taking the mean value of production of beans and for maize for the combined Montezuma and Dolores County area for the 1931-1960 period (Tables 3.12 and 3.13). The combined data were thought to provide a larger, acreage-weighted sample that might better reflect trends in the data. The crop yield means were set equal to each other: 368 pounds per acre for nonirrigated beans and 13.9 bushels per acre (equivalent to 778 pounds per acre, where one bushel weighs 56 pounds) for nonirrigated maize. These numbers compare favorably with the long-term Table 3.11. Table of Productivity Values For All Soil Types Within the Study Area Under Five Different Climatic Conditions. Soil Code 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Map Unit A1B A2B A3B A4 A41 A4A A4B A5C A5D A5W* A6A AOAB AOB AOC AO* BD C2CE C2D Favorablea FlNb NormaJC NlUd Unfavorable 9 328 334 427 427 343 330 427 427 281 288 374 374 316 294 374 374 234 241 320 320 289 257 320 320 201 204 285 285 253 221 267 267 168 167 249 249 217 184 213 213 948 292 367 367 469 0 245 218 878 259 312 312 411 0 204 175 808 226 257 257 352 0 163 131 751 198 221 221 293 0 143 109 693 170 184 184 234 0 122 87 - 104- Number of Cellsf 296 108 55 315 134 56 201 293 35 0 36 230 168 103 93 42 377 44 Percent of Area9 .77 .28 .14 .82 .35 .15 .53 .77 .09 .00 .09 .60 .44 .27 .24 .11 .99 .12 Table 3.11 (Continued). Soil Code 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 Map Unit C2F C2V C3CD C3E C5C C7D CP DOCE GP H1CD H1DC H6D M1CE M2C M2CE M2DD M2DD1 M5E M2C M7E M8D OAB P1C R1C R1D R1HB R1HC R1HD R3B R3B1 R3C R3C1 R3D R3D1 R4C R4C1 R4D R7C R7D R7D1 R8B RHC RL ROB ROB1 ROC ROC1 ROD ROHC ROHD ROLC ROLD ROLD1 Favorable a 253 218 F/Nb 213 175 Normalc 173 131 NJUd 147 109 Unfavorabies 120 87 216 351 0 272 0 397 184 299 0 233 0 343 152 246 0 194 0 289 119 211 0 155 0 235 86 176 0 116 0 181 210 334 203 167 292 242 494 470 491 433 301 344 580 580 179 288 164 135 252 213 454 426 439 368 264 277 529 529 147 241 125 103 211 183 414 382 387 302 226 210 478 478 116 223 97 80 176 154 362 339 350 268 189 172 427 427 84 204 69 56 141 124 310 296 313 233 151 134 376 376 528 294 528 294 528 294 489 326 475 416 416 283 334 483 0 519 437 519 437 519 406 406 427 427 283 440 276 440 276 440 276 423 288 405 361 361 255 288 404 0 467 355 467 355 467 348 348 374 374 255 352 257 352 257 352 257 356 250 335 305 305 226 241 325 0 415 273 415 273 415 289 289 320 320 226 317 202 317 202 317 202 312 207 300 269 269 198 204 292 0 346 228 346 228 346 253 253 285 285 198 282 147 282 147 282 147 268 163 264 233 233 170 167 258 0 277 182 277 182 277 217 217 249 249 170 -105 - Number of Cellsf 104 49 0 0 387 43 9 10 36 200 0 67 36 3617 6322 3150 1093 369 25 15 21 59 45 116 29 0 0 0 3 60 137 199 462 75 1075 375 938 153 102 61 75 500 413 982 54 5868 152 1058 130 69 219 130 75 Percent of Area9 .27 .13 .00 .00 1.01 .11 .02 .03 .09 .52 .00 .18 .09 9.46 16.54 8.24 2.86 .97 .07 .04 .05 .15 .12 .30 .08 .00 .00 .00 .01 .16 .36 .52 1.21 .20 2.81 .98 2.45 .40 .27 .16 .20 1.31 1.08 2.57 .14 15.35 .40 2.77 .34 .18 .57 .34 .20 Table 3.11 (Concluded). Map Soil FavorUnfavorNumber Percent NlUd F/Nb Unit able a Nonnalc Code able 9 of Cells f of Areag 72 S1C 294 257 276 221 184 365 .95 73 S5 611 560 509 458 407 8 .02 74 S51 609 541 507 676 473 40 .10 454 414 345 75 494 276 S53 250 .65 541 507 473 76 676 609 S50 64 .17 77 405 351 455 296 S6 504 61 .16 419 385 350 78 S6C 519 469 16 .04 289 244 199 343 79 397 278 S6E .73 409 357 460 80 S7 544 502 14 .04 339 293 385 81 424 S8C 463 204 .53 82 T1B 0 .00 294 267 374 320 83 427 2 .01 T2C 116 147 179 84 14 84 210 T3E .04 0 T4B 85 .00 0 86 T6 .00 87 0 .00 T6B 0 88 T6C .00 0 89 T60 .00 470 604 722 336 4 839 90 V1B .01 470 722 604 336 839 34 91 V1C .09 92 V10 0 .00 281 234 201 328 168 93 V2B 552 1.44 234 201 94 V2C 281 168 123 328 .32 137 225 186 162 95 V2S 264 59 .15 V2V· 328 281 234 201 168 96 248 .65 137 V2W* 225 162 .97 264 186 146 .38 487 772 369 31 98 V4C 688 604 .08 257 221 184 VOC 330 294 5 .01 99 257 184 100 VOC 330 294 221 49 .13 101 VOCB 294 257 221 184 3 330 .01 VOCC 294 257 221 184 14 102 330 .04 103 WOB 361 289 217 253 325 173 .45 WOBB· 104 361 289 217 325 .42 253 159 105 WOC 361 325 289 253 217 7 .02 XC60 210 179 147 116 106 84 1083 2.83 107 XC6E 158 135 111 87 63 1572 4.11 119 XH20 206 181 156 4 108 82 .01 XH2E 170 145 120 94 109 68 275 .72 XH3E 217 186 155 124 110 93 23 .06 111 ><M1F 0 .00 158 137 115 97 ><M30 79 537 112 1.40 143 115 72 86 57 55 .14 113 Z Note: All productivity values are Soil Produdivity Units Based on Beans (SPUBs) and are given in pounds per acre. aFavorable Growing Season (assigned to PDSI Classes 10 and 11, Unusually Wet and Extremely Wet). bFavorable­to­Normai Growing Season (assigned to PDSI Classes 8 and 9, Slightly Wet and Moderately Wet). cNormal Growing Season (assigned to PDSI Classes 5, 6, and 7). dNormal­to­Unfavorable Growing Season (assigned to PDSI Classes 3 and 4, Moderate and Slight Drought). eUnfavorable Growing Season (assigned to PDSI Classes 1 and 2, Extreme Drought and Severe Drought). fEach cell is 200 x 200 m (4 hal in area. gThis is the proportion of a given soil type within the analyzable portion of the study area. The total study area contains 45,400 cells. Of these, 7175 have not yet been mapped for soil associations. Therefore, only 38,225 cells have been assigned a soil value. •Artificially wet and possibly saline from irrigation water; consequently the natural potential productivity values used to generate these yield estimates were reduced to better approximate the original, nonirrigated conditions (SCS Soil Ｌｴｾｩｮ･｣ｓ Alan Price, personal communication, March 1989). ­106 - (A.D. 652­1968) "adjusted" (Le., technology-free) mean estimates made by Bums (1983) who used 374 pounds per harvest acre for dry beans and 12.7 bushels per harvest acre for nonirrigated maize (711 pounds per acre). Additionally, the maize value compares favorably with Petersen's (1987) 1919-1960 estimate for mean maize production as 13.98 bushels (783 pounds per harvested acre). Finally, the bean value compares favorably with the mean bean production value of 362 pounds per acre taken from the 46 different soils as reported by the SCS. Second, the bean yield values, which ranged from 0-948 pounds per acre (Table 3.11), were reclassified as soil productivity values that must not exceed the 0-255 maximum range defined by the GIS. This was accomplished simply by dividing each value by 10 and truncating the number to the right of the decimal (e.g., 948 becomes 94.0) and assigning each to a new set of 94 classes. In this way, a new range of soil productivity class values ranging from 0-94 was created. Each class preserves its relative measure of productivity and is usable by the GIS. For example, for soil #2 (A2B, Sili Silty Clay Loam, 0-3% slope), the unfavorable through favorable growing season yields expressed in bean units are 168, 201,234,281, and 328 pounds per acre. When reexpressed these become 16, 20, 23, 28, and Table 3.12. Nonirrigated Bean Data (in pounds) from Dolores and Montezuma Counties Used to Generate a Mean Bean Yield Value for Years A.D. 1931-1960. Mean bean yield for this 30-year period is 368.10 ± 126.14 pounds per acre. Dolores Year Co. Yield 1931 250 1932 300 1933 320 1934190 1935 230 1936 260 1937 300 1938 350 1939 320 1940 370 1941 400 1942 350 1943 770 1944 380 450 1945 1946 390 1947 590 1948 400 1949 450 1950 330 1951 250 1952 500 1953 530 1954 340 1955 470 1956 160 1957 700 1958 320 1959 180 1960 280 Montezuma Co. Yield 270 350 240 170 240 280 300 360 260 510 400 350 740 380 430 420 600 410 480 350 170 520 480 330 530 304 692 350 220 280 Dolores Co. Acres Harv. 1180 910 3,230 2,484 7,520 5,060 9,770 13,890 17,440 23,950 20,400 20,710 31,000 NA NA NA NA 51,000 44,590 35,650 27,750 29,820 39,640 42,780 40,960 40,020 41,900 42,950 36,130 36,250 Montezuma Co. Dolores Co. Acres Harv. Total Production 8,190 295,000 5,090 273,000 6,300 1,033,600 6,774 471,960 14,010 1,729,600 12,510 1,315,600 20,110 2,931,000 30,250 4,861.500 42,670 5,580,800 35,920 8,861,500 28,150 8,160.000 30,100 7,248,000 39,440 23,870,000 NA NA NA NA NA NA NA NA 49,000 20,400,000 40,520 20,065,000 41,680 11,764,500 25.100 6,937,500 29,200 14,910,000 33,670 21,009,200 51,230 14,545,200 39,660 19,251,200 54,560 6,403,200 58,420 29,330,000 59,620 13,744,000 51,800 6,503,400 52,950 10,150,000 Montezuma Co. Mean Total Production Per Acre 2,211,300 267.5 1,781,500 342.4 1,512,000 267.1 1,151,580 175.4 3,362,400 236.5 3.502,800 274.2 6,033.000 300.0 10,890,000 356.9 11,094,200 277.4 18,319,200 454.0 11,260,000 400.0 10,535,000 350.0 29,185,600 753.2 NA *380 NA *440 NA *405 NA *595 20,090,000 404.9 19,449,600 464.3 14,588,000 340.8 4,267,000 212.0 15,184,000 509.9 16,161,600 507.1 16,905,900 334.6 21,019,800 499.5 16,586,240 243.1 40,426,640 432.2 20,867,000 337.4 11,396,000 203.6 14,826,000 280.0 Note: All of these dry bean values for Dolores and Montezuma Counties are derived from Bums (1983:313, 315). *designates averaged but unweighted yields for years 1944-1947 when complete records were not kept. -107 - Table 3.13. Nonirrigated Maize Data (in bushels) from Dolores and Montezuma Counties Used to Generate a Mean Maize Yield Value for Years A.D. 1931-1960. Mean maize yield for this 30-year period is 13.9 ± 3.2 bushels per acre (or n8 ± 179 pounds per acre where one bushel equals 56 pounds). Year 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1969 Dolores Montezuma Dolores Co. Montezuma Dolores Co. Co. Yield Co. Yield Acres HaN. Co. Acres Harv. Total Prod. 9 9 1,030 4,100 9,270 13 12 1,700 5,540 22,100 11 12 1,700 4,690 18,700 5 4 1,950 3,440 9,750 27 10 2,613 3,073 70,551 10 9 2,380 3,010 23,800 11 12 2,830 3,260 31,130 12 12 2,510 3,230 30,120 12 9 2,070 3,050 24,840 14 12 2,610 2,350 36,540 17 20 2,540 1,970 43,180 10 15 1,230 1,480 12,300 15 16 1,470 1,230 22,050 17 12 NA NA NA 15 16 NA NA NA 17 18 NA NA NA 17 20 NA NA NA 14 13 960 910 13,440 19 19 1,040 810 19,760 15 16 260 380 3,900 14 15 100 260 1,400 12 15 90 320 1,080 14 14 60 350 840 13 14 360 350 4,680 13 14 310 110 4,030 12 18 250 70 3,000 17 19 320 60 5,440 15 16 230 30 3,450 13 15 240 150 3,120 17 16 30 30 510 Montezuma Mean Co.Total Prod. Per Acre 36,900 9.0 66,480 12.2 56,280 11.7 13,760 4.4 30,730 17.8 27,090 9.4 39,120 11.5 38,760 12.0 27,450 10.2 28,200 13.1 39,400 18.3 22,200 12.7 19,680 15.5 NA *14.5 NA *15.5 NA *17.5 NA *18.5 11,830 13.5 15,390 19.0 6,080 15.6 3,900 14.7 4,800 14.3 4,900 14.0 4,900 13.5 1,540 13.3 1,260 13.3 1,140 17.3 480 15.1 2,250 13.8 480 16.5 Note: All of these maize values for Dolores and Montezuma Counties are derived from Bums (1983:308, 310). *designates averaged but unweighted yields for years 1944-1947 when complete records were not kep1. 32 soil productivity units (SPUs), respectively. Next, the mean bean production (368 pounds per acre) is associated with soil productivity unit 36 (which includes any value between 360 and 369 pounds of beans) as is the mean maize production of 778. Thus, the soil productivity unit 36 which is based on bean production (SPUB) is equivalent to the mean value of beans and the mean value of maize. Finally, an equation converting each.SPUB into a SPUM (soil productivity unit based on maize) is developed, so that each and every soil productivity unit is associated with maize production. The SPU is multiplied by 10 and to this the number four is added. This number produces a value that exists at the midpoint of its class and approximates the original bean yield estimate. For example, SPU 36 represents bean yields of 360-369 pounds for which the range midpoint is 364; it is this number that is now used to represent bean yield. 1bis number is then divided by the ratio of 368 to 778 or .473 (Le., the mean bean yield to the mean maize yield) to produce the equivalent value in pounds per acre of maize. As an equation, this may be written as SPUM (pounds/acre) = «SPUB *10)+ 4) / 0.473. Ultimately, this SPUM value must be reexpressed as yield in kilograms per hectare to be consistent with the unit of analysis used in the GIS, which is a 4-ha cell, 200 m on a side. To convert from pounds 108- per acre to kilograms per hectare, the maize yield is multiplied by 1.12. Thus, the fmal equation is SPUM (kg! ha) = «(SPUB * 10) + 4) / O. 473) * 1.12. In order not to loose additional resolution in these maize estimates, it was decided to apply the SPUB to SPUM conversion in the post-GIS processing of data with SAS programs, after the GIS had analyzed and summed all the SPUBs for each year in the 400-year study. Calibratinl: Palmer Drought Severity Index Values with Yield Values To graphically portray contrasting areas of agricultural production within the study cu:ea, and to generate absolute yield values that m tum could be used to estimate populations that could be sustained by that yield, it was necessary to assign crop yield value.s to PDSI ｾｵ･ｳＮ Depicting differential soil mOisture condltlOns is not enough; this must be "translated" in1:O equivalent measures of agricultural production. Further, this had to be accomplished in such a way that the GIS could manage and display the PDSI and crop yield information. The first step toward achieving this calibration was data reduction of the PDSI information into fewer, more interpretable categories. The second was reexpression of original positive and negative PDSI values into GIS-compatible units that would preserve the meaning and rank order of the original. The third was to "fit" or "calibrate" the positive linear function created by the partial regression study to e five categories created by the crop yield and natural plant productivity study. Once this was carried out, it was a relatively simple task for the GIS to integrate these tabular data as "look-up tables" with the spatial data already resident in the system. Data Reduction In Palmer's (1965:28, Table 11) original description of his method, he created a classification scheme to interpret the PDSI values that resulted from his computations. Using Palmer's original classification system as a guide, a number was assigned to each of the 11 interpretative classes that would retain their rank order and meaning (Table 3.14). Here Class 1 Table 3.14. Eleven Nominal PDSI Classes. Original PDSI Values PDSI Class Extremelv Wet 3.00-3.99 Unusuallv Wet 2.00-2.99 Moderatelv Wet 1.00-1.99 Sliahtly Wet 0.50-0.99 Incipient Wet (-0.49)-(0.49) Near Normal (-0.50)-(-0.99) Incipient Drought (-1.00)-(1.99) Slight Drouaht (-2.00)-(2.99) Moderate Drought (-3.00)-(3.99) Severe Drouaht s;(-4.00) Extreme Drouaht Ｐ ＮＴｾ PDSI Nominal Class 11 10 9 8 7 6 5 4 3 2 1 represents the category "Extreme Drought" G;-4.00 PDSI), Class 11 represents the category PDSI), Class 6 rep"Extremely Wet" ｾＫＴＮＰ resents the category "Near Normal" (-.49+.49 PDSI), with all other values falling into intermediate positions (Figure 3.4, rows a and b). Data Reexpression As explained previously, the original PDSI values are expressed as both negative and positive real numbers. These values had to be reexpressed as positive whole numbers with a maximum range from 0-255 so as to be manipulated by EPPL7. At the time this was done, the exact range of reconstructed PDSI values was not known, but it was known that values below -6.00 PDSI and above +10.00 PDSI did occur in the actual calculations for PDSI for the 55 soil groups used in the study. To assure that all PDSI values would fall within the true range, a wide hypothetical range of -12.00 to +12.00 was selected. These values were reexpressed in GIS Units ranging from 0-240. This was accomplished by first rounding the value one decimal place (Le., "rounding up" where the value is equal or greater than five, and "rounding down" where the value is less than five); multiplying the number by 10 (to produce a nondecimal number); and adding 120 to the number to insure that it be positive and within the woIkable range of the GIS. For example, an original PDSI value of 1.64 -109 - PDSI = 0.00 GIS = 120 Ｍ Ｍ Ｍ Ｍ ｾ WET DROUGHT A. 2 3 EJ.1teme Severe Maderale SliQht Incip- N.ar Incipient Narmal ienl SliQht ­3.00 ­2.00 ­1.00 ­.50 PDSI Categories 8. Values in Original PDSI Units C. Reexpressed Values in GIS Units D. I 4 Agricultural Productivity Categories ｾ ­4.00 ｾ 80 to to 5 10 10 ­3.99 ­2.99 ­1.99 ­99 8190 91100 101110 111115 6 .49 to ­.49 7 8 9 10 II Madera'. Severe Ellrem. 3.00 .50 1.00 2.00 10 10 10 10 .99 1.99 2.99 3.99 116- 125124 129 130139 140149 155 159 4.00 ｾ ｾ 160 PRODUCTIVITY Unfovorable Normol/ Unfavorable Normal Fa vorable / Normal Favorable Figure 3.4. Schematic diagram illustrating the relationship among reconstructed POSI values, their interpretive categories, and crop yield production categories. becomes 1.6 when rounded, then 16 when multiplied by 10, and finally is reexpressed in GIS units as 136 when 120 is added to 16. Similarly, an original PDSI value of ­3.28 becomes, ­3.3, then ­33, and finally reexpressed as 87 in GIS units (Figure 3.4, rows b and c). Calibrating Yield Categories to PDSI Categories The goal of this step was to assign the 11 moisture categories (Le., extreme drought to extremely wet) to the five crop yield categories (i.e., unfavorable to favorable) using the knowledge that increasing crop yields are reason ably well correlated with increasing amounts of stored soil moisture. This was done subjectively, but guided by the knowledge gained from the partial regression study. The central ­ PDSI categories 5, 6, and 7, representing the incipient drought, near normal, and incipient wet, were combined and assigned to the "normal growing conditions," the central category of the crop yield estimates. PDSI categories 1 and 2, extreme drought and severe drought, were assigned to the lowest crop yield estimate category, "unfavorable growing conditions." Correspondingly, the highest PDSI categories, 10 and II, representing very wet and extremely wet, were assigned to the highest crop yield category, "favorable growing conditions." PDSI categories 3 and 4, moderate and slight drought, were assigned to low intermediate "normal­to­unfavorable growing conditions" category of yields. Finally, PDSI categories 8 and 9, slightly and moderately wet, were assigned to the high intermediate category of "normal­to­favorable growing conditions" 110- category of estimated yields (Figure 3A, rows a and d). In sum. historic crop yield data. data on natural plant productivity, and reconstructed Palmer Drought Severity Indices. were used in a variety of pattern­recognition studies that allowed quantifiable estimates of prehistoric crop production associated with climatically controlled soil moisture conditions to be made. These studies provided the data to link the treering based PDSI reconstructions with the data manipulation and display capabilities of geographic information systems. Collectively, the data and techniques made possible the visual discrimination of agricultural variability as well as the quantitative estimates of crop production and human carrying capacity. -111 - Page Blank in Original 4 BUILDING THE MODEL: FINAL DATA ANALYSES This chapter completes the description of the data, assumptions, and analytic techniques used to create this model of prehistoric agricultural productivity. It details the use of the EPPL7 geographic information system in integrating the environmental data and the results of the long term reconstructions and agricultural productivity studies. Further, it identifies the numeric output required from the GIS analysis for subsequent analysis. It also explains the methods used to calculate total agricultural productivity and to estimate the number and density of human beings that could be sustained on the available yield. The chapter concludes with a tabular synopsis of the model-building process described in chapters 2,3, and 4. FINAL GEOGRAPlDC INFORMAnON SYSTEM ANALYSES Introduction This final stage of GIS analysis can be divided into several steps (Figure 4.1). First, a PDSI value (PDSI values ranging from Ｑｾ 240, or 0 if no data exists for that cell) is assigned to each cell in the 45,400 cell image for each year of the analysis. Second, this PDSI value is reassigned to a PDSI Nominal Oass (value ranging from 1-11). Third, the PDSI Oass value is linked to one of five newly created data planes called favorability planes. These planes depict the specific soil productivity unit values (SPUs) that have been estimated for each soil type under different growing season conditions as discussed in chapter 3. Once linked, the soil productivity unit value (SPU value ranging from 5-94) is assigned to a final agricultural productivity data plane for a given year. Fourth, and last, a tabular file is produced that lists the range of SPU values assigned to a yearly data plane and provides counts and percentages for those values. AssiWin2 a PDSI Value First, EPPL command INT ABLE creates a temporary data plane that depicts the reconstructed annual PDSI value for each cell in PDSlMAP.EPP (the data plane that assigned each cell to its appropriate long-term PDSI reconstruction). It uses the appropriate long-term PDSI reconstruction table (see Van West 1990:448-555 for an example) and locates the reconstruction number (a row value, ranging from 1-55 for a given year and a column value ranging from 1-20, where the maximum number of years in a single table is 20). The conjunction of row and column is the reconstructed June PDSI value for a given soil class in a given year. A different version of this table is used by EPPL, since it can only process positive intergers between ｾＲＵ (recall Fig. 3.4). Table 4.1. PDSI Class Correspondence Table. PDSI Nominal Class Range of PDSI Values (in GIS Units) 1 ＱｾＰ 2 3 4 5 6 7 8 9 81-90 91-100 101-110 111-115 116-124 . 125-129 130-139 140-149 150-159 160-240 10 11 PDSISoil Moisture Class Extreme Drought Severe Drought Moderate Drought Slight Drought Incipient Drought NearNonnal Incipient Wet Slightly Wet Moderately Wet Unusually Wet Extremely Wet PRELIMINARY ANALYSES EPPL COMMANDS: RECl.ASS FINAL ANALYSES EVAWA"ffi Ir=r EVALUATE JNfARLE PDSlyyyy.TBL r+ 20 tbls, 20 yrs eh DEMWIND.EPP.5ELEV.EPP ­ ­ . (1­189) (1­5) • ｓｏｉｌｗ ｎｄＮｅｐ ｾｉ ａｗｃＮｅｐ Ｍｬ PDSIMAP.EPP --+- (TEMP) (I­55) (0·240) --+- FAVNORM.EPP­, (POSI 8.9) • PDSI.CCT..­TEMP.EPP­+ NORM.EPP .... PRODyyyy.EPP ｾ (0.240= (I­II) (POSt 5.6,7) (5·94) L.. ｴｾｐｅＧｆｎｕｍｒｏ 1·11) (1·113) (1·11) PDSI Rccon.Table PDSt090I.TBL (AD 901 ­ 920) ｾ DEMWIND.EPP.5ELEV.EPP ｾ 168 13 (2044m) (IGNACIO) . POSIMAP.EPP 1#32 SOILWlND.EPP'­IIAWC.EPP 1#62 1# 10 (ROB) (.99 10.26") t -+- ! (TEMP] ｾ 57 (POSt = .6.35) Class CorY. Table PROOyyyy.CNT (5­94) (POSI 3,4) Ｍｾ -- I FAV.EPP (POSI 10.11) UNFAV.EPP (PDSt 1,2) ｆ｡ｶｯｲ ｢ｩｬ ｴｙ Ａ ｕ ･ｾ Productivity Planes Prod. Cm. Files FAV.EPP (PDSt 10.11) FAVNORM.EPP (PDSI 8,9) PDSI.CCT..­TEMP.EPP . NORM.EPP 0·80 = post (PDSt 5.6.7) Class t Class I ｾ LNORMUNF.EPP (POSt 3.4) PRODyyyy.EPP SPU 27 ｾ PROOyyyy.CNT SPU 27,# ,%, HA UNFAV.EPP (post 1.2) Figure 4.1. Diagram illustrating procedural steps used to process the final GIS-integrated analysis. The top half is a generalized representation. Notations in parentheses denote the range of class values associated with each data plane or file. The bottom half illustrates a hypothetical example for one cell of the 45,400 cell image containing Witt Loam Soil (ROB) at 2044 m elevation for the year A.D. 901. ａｳｩｬｭｏｾ a PDSI Class Second, command INTABLE reassigns the PDSI values into 11 PDSI Classes (ranging from Class 1, which is equivalent to extreme drought, to Class 11, which is equivalent to extremely wet, see Table 3.14). Compression into the 11 classes is guided by a "class correspondence table" in which the reclassification rules are stored (pDSI. CCf). A reduced set of PDSI classes is created by grouping the values in Table 4.1. This not only makes the reconstructed values more interpretable, but it also simplifies the data manipulation. Thus, a named but temporary data file, TEMP.yyyy.EPP is created for every year, in which every cell is assigned a soil moisture (PDSI) class. The "yyyy" designates the year A.D. Assilming a Soil Productivity Unit Value Before EPPL command EVALUATE could be used to assign agricultural productivity values to each cell of a final output image for each year, five new data planes had to be created using data in Table 3.11. These five data planes-FAV.EPP, FAVNORM.EPP, NORM.EPP, NORMUNF.EPP, and UNFAV.EPP-are called "favorability" planes. They represent the varying levels of agricultural productivity (expressed as soil productivity units, SPUs) that are associated with individual soil types as they are affected by different growing season conditions. Each favorability plane was created with EPPL command !NTABLE, which linked one of the five levels of productivity from Table 3.11 (and reexpressed in an EPPL7-compatible version of this table called PROD.TBL, Table 4.2) with the soil appropriate soil type in SOn.WIND.EPP. Class values (the SPUs) on the favorability data planes range from 5 to 94. As a set, they represent the range of potential yields from all soils in the study area. Favorability layer FAV.EPP assigns the most favorable productivity values for a given soil (pROD.TBL, Table 4.2, column 2) to the cells in its image. These values range from 14 to 94, and depict the agricultural situation in the study area should every soil be producing at its optimum level. This, of course, never occurs, but it creates a data plane that can be used for evaluation purposes. Likewise, FAVNORM.- EPP depicts a study area where every soil uses values from the third column of data in PROD.TBL, with values ranging from 11 to 87. NORM.EPP uses the fourth column of data in PROD.TBL with values from 8-80. NORMUNF.EPP uses the fifth column of data in PROD.TBL with values from 7-75. Finally, UNFAV.EPP uses the sixth and final column of data in PROD.TBL with values from 5 to 69, and depicts the agricultural situation in the study area should every soil be at its minimum productive level. Once these five data planes were created, the third analytic step could be performed. Via a series of logical "if-then" operations, EVALUATE was used to examine each cell in TEMPyyyy.EPP, which has been assigned a PDSI Class, and match it with the appropriate favorability plane. While it is rare that more than eight or fewer than three different PDSI Class values occur over the surface of the study area during anyone year, it is quite usual for at least four or five different PDSI classes to exist simultaneously for different soil classes at different elevations. If the PDSI Class value assigned to a cell in TEMP0934.EPP for year A.D. 934, for example, is either a Nominal Class 1 or 2 (extreme or severe drought), then command EVALUATE uses the UNFAV.EPP favorahility plane, looks up the SPU value for that cell in that plane, and as-signs that SPU value to the same cell location in the final output data plane, PR0D0934.EPP. Similarly, if the PDSI value associated with a cell in TEMP.0934.EPP for AD. 934 is Class 3 or 4 (moderate drought or slight drought) the program uses the NORMUNF.EPP agricultural productivity plane and assigns those values to the output data plane. PDSI Class Values 5, 6, and 7 (incipient drought, near normal, and incipient wet) are assigned to the NORM.UNF productivity plane; PDSI Class Values 8 and 9 (slightly wet and moderately wet) are assigned to the FAVNORM.EPP productivity plane; and PDSI Class Values 10 and 11 (unusually wet and extremely wet) are assigned to the FAV.EPP productivity plane. Thus, the final data set for year AD. 934 will have integrated the values contained in selected cells of the five different productivity planes. This evaluation and assignment process is carried out for every year from AD. 901-1300 and consequently creates 400 PRODyyyy.EPP files. 115 - Table 4.2. Productivity Table (PROD.TBL) used in GIS Analysis. Soil Code 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 FAV 0 32 33 42 42 34 33 42 42 0 94 29 36 36 46 0 24 21 25 21 0 0 21 35 0 27 0 39 0 21 33 20 16 29 24 49 47 49 43 30 34 58 58 0 0 0 52 29 52 29 52 29 48 32 47 FAVNORM 0 28 28 37 37 31 29 37 37 0 87 25 31 31 41 0 20 17 21 17 0 0 18 29 0 23 0 34 0 17 28 . 16 13 25 21 45 42 43 36 26 27 52 52 0 0 0 44 27 44 27 44 27 42 28 40 NORM 0 23 24 32 32 28 25 32 32 0 80 22 25 25 35 0 16 13 17 13 0 0 15 24 0 19 0 28 0 14 24 12 10 21 18 41 38 38 30 22 21 47 47 0 0 0 35 25 35 25 35 25 35 25 33 ­ 116- NORMUNF 0 20 20 28 28 25 22 26 26 0 75 19 22 22 29 0 14 10 14 10 0 0 11 21 0 15 0 23 0 11 22 9 8 17 15 36 33 35 26 18 17 42 42 0 0 0 31 20 31 20 31 20 31 20 30 UNFAV 0 16 16 24 24 21 18 21 21 0 69 17 18 18 23 0 12 8 12 8 0 0 8 17 0 11 0 18 0 8 20 6 5 14 12 31 29 31 23 15 13 37 37 0 0 0 28 14 28 14 28 14 26 16 26 Table 4.2 (Continued). Soil Code 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 FAV 41 41 28 33 48 0 51 43 51 43 51 40 40 42 42 28 29 61 67 49 67 50 51 39 54 46 0 42 21 0 0 0 0 0 83 83 0 32 32 26 32 26 77 33 33 33 33 36 36 36 21 15 20 17 FAVNORM NORM 30 30 22 24 32 0 41 27 41 27 41 28 28 32 32 22 25 50 54 41 54 40 41 28 46 38 0 32 14 0 0 0 0 0 60 60 0 23 23 18 23 18 60 25 25 25 25 28 28 28 14 11 15 12 36 36 25 28 40 0 46 35 46 35 46 34 34 37 37 25 27 56 60 45 60 45 46 34 50 42 0 37 17 0 0 0 0 0 72 72 0 28 28 22 28 22 68 29 29 29 29 32 32 32 17 13 18 14 ­ 117- NORMUNF 26 26 19 20 29 0 34 22 34 22 34 25 25 28 28 19 22 45 50 34 50 35 38 24 40 33 0 29 11 0 0 0 0 0 47 47 0 20 20 16 20 16 48 22 22 22 22 25 25 25 11 8 11 9 UNFAV 23 23 17 16 25 0 27 18 27 18 27 21 21 24 24 17 18 40 47 27 47 29 35 19 35 29 0 26 8 0 0 0 0 0 33 33 0 16 16 13 16 13 36 18 18 18 18 21 21 21 8 6 8 6 Table 4.2 (Concluded). Soil Code 110 111 112 113 FAV FAVNORM 21 0 15 14 NORM 18 0 13 11 15 0 11 8 Creating Count files The final step is the creation of a tabular output file that contains the values used later to calculate total annual maize productivity and sustainable human population size. EPPL command COUNT produces a file that lists classes, cell counts, individual percentages, cumulative percentages. areas, and a legend that briefly describes the classes. Here the classes listed are the soil productivity classes (SPUBs) present in year yyyy with class values ranging from 5 to 94. The cell counts are the number of 200 x 200 m cells included in each class and the individual percentage is the proportion of those cells to the total number of cells counted. The cumulative percentage is the running total of percentages calculated class by class. The area is the total amount of land measured in hectares that is included in that particular class. The legend identifies the productivity class. Table 4.3 is an example of a GIS tabular output "count file" for year A.D. 902. Production of the count file, particularly the values associated with class and area, is essential to the model-building process for it embodies the numeric values needed to estimate annual maize production and ultimately the number of people that could be supported on that yield. With creation of the count files for each of the 400 years in the analysis, the GIS portion of the analysis is largely concluded, although GIS technology continues to be used for displaying final products (the annual "maps"). Hereafter, these data are transferred to a mainframe statistical analysis program for processing, since these particular capabilities do not reside within EPPL7. Mapping the Agricultural Niche Once the integrated file for a given year has been created by the above analysis, it can - NORMUNF 12 0 9 7 UNFAV 9 0 7 5 be viewed with EPPL7 DISPLAY programs. In DISPLAY, the locations of cells representing the various levels of agricultural productivity are visible. The many soil productivity classes (5-94) may be regrouped into fewer classes (1-5) that are easier to interpret. To retain the original detail of differential productivity reconstructed for each cell in the 45,400 cell image, simplification of productivity classes (SPUBs) was performed via selective colorcoding during the display of the data only. In the color-classification scheme that was devised, five colors are arranged in a rainbow-like palette representing five general categories of agricultural productivity: SPUB classes 5-18 represents "low productivity" locations (purple), SPUB classes 19-27 represents "lowto-moderate productivity" locations (blue), SPUB classes 28-44 represents "moderate productivity" locations (green), SPUB classes 45-52 represents "moderate-to-high productivity" locations (yellow), and SPUB classes 53-94 represents "high productivity" locations (red). These divisions are based on the distribution of bean yield values for the 46 documented soils in Montezuma and Dolores Counties. Using this distribution of values, any value within one standard deviation of the mean is considered normal or "moderate" yield. A value falling between one and two standard deviations above the mean is considered "moderate-to-high" yield. A value greater than two standard deviations above the mean is "high" yield. Conversely, a value falling between one and two standard deviations below the mean is considered "moderate-to-Iow" yield, and a value lower than two standard deviations below the mean is considered "low" yield (Table 4.4, Figure 4.2). Using this classification and color scheme it is easy to visually identify which locations represent the most productive lands in any given year. Both large and small patches of productive land can be discerned and precisely 118- Table 4.3. Example EPPL7 Tabular Output: COUNT File of Agricultural Productivity in A.D. 902. Classa 8 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 27 28 29 30 31 32 33 34 35 36 37 38 40 41 42 43 44 45 46 47 50 52 54 56 60 68 80 87 Count b 17 2,713 521 2,273 5,047 205 122 1,863 1,045 315 10 70 3,677 276 1,020 152 2,252 225 651 32 123 54 1,193 219 313 1,118 150 185 201 925 8,494 452 2 284 41 18 40 8 103 153 7 144 10 20 16 Percent C 0.05 7.38 1.42 6.18 13.73 0.56 0.33 5.07 2.84 0.86 0.03 0.19 10.00 0.75 2.77 0.41 6.13 0.61 1.77 0.09 0.33 0.15 3.25 0.60 0.85 3.04 0.41 0.50 0.55 2.52 23.11 1.23 0.01 0.77 0.11 0.05 0.11 0.02 0.28 0.42 0.02 0.39 0.03 0.05 0.04 Cumulative d 0.05 7.43 8.84 15.03 28.76 29.32 29.65 34.72 37.56 38.42 38.44 38.63 48.64 49.39 52.16 52.57 58.70 59.31 61.08 61.17 61.51 61.65 64.90 65.49 66.35 69.39 69.80 70.30 70.85 73.36 96.47 97.70 97.70 98.48 98.59 98.64 98.75 98.77 99.05 99.46 99.48 99.87 99.90 99.96 100.00 Area (ha) 68 10,852 20,840 9,092 20,188 820 488 7,452 4,180 1,260 40 280 14,708 1,104 4,080 608 9,008 900 2,604 128 492 216 4,772 876 1,252 4,472 600 740 804 3,700 33,976 1.808 8 1,136 164 72 160 32 412 612 28 567 40 80 64 Legend (Prod. Level)e 8D-89 SPUs; Low 100-109 SPUs; Low 110-119 SPUs; Low 120-129 SPUs; Low 130-139 SPUs; Low 140-149 SPUs; Low 150-159 SPUs; Low 160-169 SPUs; Low 170-179SPUs;Low 180-189 SPUs: Low 190-199 SPUs; Low/Moderate 200-209 SPUs; Low/Moderate 210-219 SPUs; Low/ Moderate 220-229SPUs;Low/Moderate 230-239 SPUs; Low/Moderate 240-249 SPUs; Low/Moderate 250-259 SPUs; Low/Moderate 270-279 SPUs; Low/Moderate 280-289 SPUs, Moderate 290-299 SPUs; Moderate 300-309 SPUs; Moderate 310-319 SPUs; Moderate 320-329 SPUs; Moderate 330-339 SPUs; Moderate 340-349SPUs; Moderate 350-359SPUs;Moderate 360-369 SPUs; Moderate 370-379 SPUs; Moderate 380-389 SPUs; Moderate 400-409 SPUs; Moderate 410-419 SPUs; Moderate 420-429 SPUs; Moderate 430-439 SPUs; Moderate 440-449SPUs;Moderate 450-459 SPUs; Moderate/High 460-469 SPUs; Moderate/High 470-479 SPUs; Moderate/High 500-509 SPUs; Moderate/High 520-529 SPUs; Moderate/High 540-549 SPUs; High 560-569 SPUs; High 600-609 SPUs; High 680-689 SPUs; High 800-809 SPUs; High 870-879 SPUs; High Notes: The variables used for subsequent analysis with SAS are class (SPUB) and area (HA). The others are not used. 36,759 cells are onsite; 8,641 cells are offsite. ｾｨ･ SPUB value (soil productivity unit, initially expressed in bean units). hthe number of cells exhibiting this class value. Each cell is 4 ha in area. Cpercent of onsite cells or analyzable cells. Offsite refers to cells that cannot be analyzed. A cell may be offsite because soils have not yet been mapped for that cell or because available water capacity or natural productivity information were not available at the time of analysis. Alternatively, a cell may be near the margin of the image in a p'lace where incomplete overlap between data planes exists. Cicumulative percent. ean optional but useful feature for identifying the meanings assigned to each class. This legend identifies the agricultural productivity level associated with each class. ­ 119- ­25.0 -I S.D. X +1 S.D. +2 S.D. Ibs / acre 194 278 362 446 530 SPUB 19 28 36 45 53 I - I YIELD Low Moderate to Low Moderate Moderate to High High - Figure 4.2. Normal distribution of modern bean yield (Ibs/ac) from 46 agricultural soils in Montezuma and Dolores Counties. Colorado. located; associated elevations and soil types for any location can be identified. These images, therefore, depict the agricultural niche, however it may be defined, as it existed in any given year. Thereafter, these depictions may be captured photographically, printed out in color by a color printer, or depicted as shades of gray or symbols by a standard computer printer (Figure 4.3). Comments on GIS Data Processing Given that the above operations were to be repeated 400 times, once for every year in the analysis, it was necessary to automate the processing of the data Here other computer soft. ware was enlisted to facilitate the processing of data. Toward this end, a program was written in interpretive BASIC (GENEPPL.BAS) that generated a PC file in less than five minutes containing 400 sets of EPPL instructions, one for each year in the analysis. After some minor editing, the file (now renamed PRODEPPL.IN to signify that modifications had occurred) was redirected into EPPL and the 400 years of data planes were processed in approximately 12 hours on an IBM/PS2 Model 50z. The batch job produced 400 agricultural productivity data planes, a large disk file (pRODEPPL.OUT) containing statistics for each data plane (a "COUNT' file), EPPL system messages, and possible error messages (of which there were none). GENEPPL.BAS is Appendix F in Van West (1990:561-562). SUPPLEMENTAL ANALYSES WITH SAS An auxiliary system to the GIS was needed to further manipulate the numeric output contained in the EPPL7 count files. A system was needed to link files, read selected data, create new data from various combinations and transformations of extant data, and generate numeric and graphic output of the results. While -120 - Table 4.4. Modem, Nonirrigated Bean Yield for 46 Agricultural Soils in Montezuma and Dolores Counties, Colorado. Number 1 2 3* 4 5** 6** 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Soil Code 8 9 21 28 34 35 36 42 43 47 48 49 50 51 52 53 54 55 56 57 58 60 62 Soil Map Unit A5C A5D C3CD H1CD M2DD M2DD1 M5E R1C R1D R3B R3B1 R3C R3C1 R3D1 R3D1 R4C R4C1 R4D R7C R7D R7D1 RHC ROB Yield Per Soil T e Ib/ac 275 250 400 250 300 250 450 450 400 350 275 350 275 300 250 375 273 300 350 300 250 350 450 Number 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42* 43 44 45 46 Soil Code 63 64 65 66 67 68 69 70 71 73 74 75 76 78 80 81 90 91 92 98 103 104 105 Soil Map Unit ROB1 ROC ROC1 ROD ROHC ROHD ROLC ROLD ROLD1 S5 S51 S53 S5D S6C S7 S8C V1B V1C V1D V4C WOB WOBB woe Yield Per Soil T e Ib/ac 300 450 300 400 350 300 350 300 250 450 450 400 400 423 424 433 500 500 450 500 500 250 500 Notes: Mean yield for 46 soils is 362.02 84.26 pounds/acre. With the exception of M2DD and M2DD1, (the only non-SCS Capability Class 3 and 4 soils, and soils for which yield from only the Pulpit components has been reported), nonirrigated Bean Yield Values per soil type are determined by multiplying the percent of that component by the reported yield value per component, and summing these values, as listed in Table 2.3. For example, soil map unit 16, R4C, is comprised of Cahona, Sharps, and Wrtt Ioams in a relationship 40:35:25 ((.40 x 350 Ibs/ac) + (.35 x 350 Ibslac) + (.25 x 450 Ibslac)) • 140 + 122.5 + 112.5.375 pounds/acre. *Soils not present in study area. "Inadvertently, these two soils were not weighted and the yield from the Pulpit component of each, which is the only agricultural component, was used to represent the entire unit instead of only 45% of that map unit. Had this not been overlooked, the yield for M2DD would have been given as 135 instead of 300, and the yield for M2DD1 would have been given as 112.5 instead of 250. Thus, the correct estimates would have produced soil productivity values that were only 45% of the values used. The overall mean and standard deviation for the 46 soils would have been 355.4076 ± 96.0086 pounds/acre. a number of different programs exist to do this work, including some on powerful microcomputers with large amounts of memory, mainframe systems have some advantages for large datasets requiring complex analyses. Ultimately, the widely used Statistical Analysis System (SAS Institute, Inc. 1985a, 1985b) was selected to do the job. First, the redirected output file that contained the results of the 400 individual analyses of soil moisture conditions and agricultural productivity was uploaded to the WSU mainframe. Next, the EPPL7-redirected output file was modified with a mainframe editing pro- gram, CMS­XEDIT, to remove U1Ulecessary verbiage, add appropriate instructions, and format it for use in SAS data analysis. Thereafter, methods were developed to calculate a single value that represented the total potential production of maize for each year in the 400year series. Further, methods were devised to estimate the number and density of people that could be supported by that annual yield and to identify levels of carrying capacity that would reflect the maximum, sustainable population size over extended periods of time (Appendix C). Subsequently, I wrote SAS programs to concatenate all 400 yearly data sets, convert soil productivity units based on bean yield -121 - AGRICULTURAL PRODUCTIVITY A.D. 902 _LOW • Low­ to­Moderate .Moderate ｄｍｯ､･ｲ｡ｴ･ＭｴｾＭｈ i gh DHigh DNa Data Figure 4.3. EPPL7 dotplot of prehistoric agricultural productivity in the study area as reconstructed for JUly 1, A.D. 902. (SPUBs) into soil productivity units based on maize yield (SPUMs), and calculate the annual supply of maize (Appendix D, Program 1). I also wrote SAS programs to estimate sustainable population given three different levels of demand, calculate a maximum, long­tenn population level and calculate a lower, more realistic, long­term population level for the study area (Appendix D, Program 2). Finally, I wrote SAS programs to graphically portray various aspects of the yield and population information. The procedure described above was conducted first for the entire study area. Thereafter the data for the areas encompassed by the Sand Canyon Survey Locality and the Mockingbird Mesa Survey Locality were processed in the same manner. Finally, data for the areas contained within the radii surrounding the eight tree­ring dated sites, here referred to as "site catchments," were processed. The results of these analyses are described in chapter 5. ­ Estimating Annual Maize Production A single value representing the potential maximum maize production is calculated for each yearly data set in the time series. The value is the total amount of maize that could have been produced from all arable soils in the study area (minus the potential yield that could be derived from the low yield classes for any year). It represents the maximum yield that could have been, but certainly never was, produced in a given year. However, it does provide an upper value for the potential maize supply that could have been produced by human populations, and as such, is the first value required to estimate sustainable human populations. It is computed by summing the individual yield values associated with each productivity class that is considered to produce at least low­tomoderate yields. This is done in the following way (see Appendix D, Program 1). First, each class value that represents the 122- Soil Productivity Unit in bean units (SPUB) is converted to a Soil Productivity Unit in Maize Units (SPUM), as detailed in chapter 3. This is done by means of an equation SPUM = «(SPUB*1O)+4)/0.473)*1.12. Here the SPUB, which represents crop yield in terms of pounds of beans per acre, is converted to a SPUM, which represents crop yield in terms of kg of maize per hectare. Second, each SPUM (kg/ha) is multiplied by the area (or the total number of hectares) included in that class to derive the number of kg of maize produced by those particular soil units or cells (kg of maize per SPUM = SPUM*HA). Finally, those soil productivity units (SPUBs) that are equal to or greater than 19, representing cells where bean yields are greater than or equal to 190 pounds per acre or where maize yields are greater than or equal to 459 kg/ha, are summed. The SPUB 19 was chosen as the lower limit of "low­to­moderate yields" based on the following data. The mean and standard deviation for the 46 of 113 soils in the Montezuma and Dolores County area for which the Soil Conservation Service has assembled mean bean yield data is 362.02 ± 84.26 pounds per acre (Table 4.4). If the yields associated with the 46 soils were distributed normally (Figure 4.2), then a nonnal or "moderate yield" would be any value falling within one standard deviation of the mean. A value falling between +1 and +2 standard deviations would be considered a "moderateto-high yield," and between -1 and -2 standard deviations would be considered a "moderateto-low yield." Any value greater than +2 standard deviations would be considered a "high yield," and any value below -2 standard deviations would be considered a "low yield." Assuming that soils that were capable of producing greater maize yields would have been selected over poor producers, cells containing soils that produced yields less than 19 SPUBs were not allowed to contribute their yield subtotals to the final annual estimate. The use of this minimum yield criterion produces a more conservative estimate of total annual potential maize production (fOTPROD) than would be produced than if all cells were used. Further, it partially compensates for the possibility of having overestimated prehistoric maize pro- ductivity per unit of land made from analogs with modem maize production. Estimatinl: Annual Maximum Human Populations Once an estimate of the potential annual supply of maize was obtained, a method for determining the maximum annual demand for that potential yield had to be devised. Human demand for the supply necessarily includes consideration of land use and cultivation practices, predictable post-harvest crop losses, seed retention rates, storage levels, and human consumption rates. Reasonable assumptions and numeric estimates needed to be established to model these parameters. Thereafter it would be possible to estimate the number of people who could have been supported on that potential food supply, both on an annual basis, and for an extended period of time. Furthennore, the assembled population estimates for the full 400 years would embody a data set that would permit evaluation of the question of whether the study area was capable of supporting large numbers of people in certain periods of time, particularly in the late thirteenth century. The following section details the assumptions used to estimate the yearly maximum population that could be supported by the total annual maize yield (TOTPROD). A specific example of how this is calculated, is found in Appendix C. Equations are developed to estimate a yearly population requiring an annual maize harvest equivalent to one (pOP1YR), two (POP2YR), and three (pOP3YR) year's food supply. They are used in the SAS program presented in Appendix D, Program 2. Assumptions Used in Calculating Population Values • Only a portion of the total land area potentially available to produce crops is used. Some lands remain uncleared and are used for activities and resources not connected with agriculture (e.g., places for hunting, wild plant gathering, or collecting of nonfood resources such as construction or fuel wood). Some cleared places on the landscape are essentially nonagricultural, such as residences and public buildings. 123- In addition, some agricultural lands are deliberately taken out of use or fallowed for one or more seasons to allow moisture to recharge soil profiles. to curtail soil nutrient depletion, and sometimes to reduce loss of topsoil from wind erosion (Beaglehole 1937:36). In this study it is assumed that only 50% of the lands potentially usable for raising crops were cultivated in any year. The remaining half generates the potential gross yield. The estimate of 50% was also used by Kohler et al. (1986:528) in a recent attempt to model agricultural productivity in the vicinity of the Dolores Archaeological Project, and it reflects ethnographic observations of land use by the Hopi (Forde 1931:370), the Tewa of San Juan Pueblo (Ford 1968:157), and maize-growing Mexican peasants (Sanders 1976:141143). Hopi farmland was devoted to maize production. • Not all the maize harvested will be available for consumption. Some will be lost during transportation from the fields to residences and storage facilities, and even after being deposited in storage containers or storage rooms, some additional amount will be lost due to pests and spoilage. In this study, the annual adjusted gross maize yield is reduced by 10% to account for these predictable losses. The resulting value represents the potential net maize yield available to prehistoric farmers. The estimated loss rate of 10% was used by Williams (1989:726) in her ethnohistoric study of fanning populations in the Basin of Mexico and by Hassan (1981:45) in estimating agricultural supply and consumption in Pre--Dynastic Egypt • Of what is harvested. stored. and potentially available for consumption, a certain percent must be reserved for seed. This seed will be used for planting the following season's crop and may also be used for subsequent seasons if the following season is particularly poor. In this study, 10% of the potential net yield must be reserved for planting subsequent crops. The resulting balance represents the actual edible or adjusted net yield. All ethnographers refer to the seed that has been selected and stored with particular care for use in later plantings, but rarely is a percent or weight of grain given. Williams (1989:721) removed 10 kg/ha for her study. Wetterstrom (1976:195) removed 10 kg/ha for each year in storage. Hassan (1981:45) removed 25% in his Egyptian example. While the value used in my study may be high (although not as high as Hassan's value), it may compensate for underestimating or not taking into consideration losses in other parts of the growing-harvesting-and-storing cycle. • While it is not known precisely in what proportions the major prehistoric cultigens of maize, beans, and squash were grown, archaeologists have generally assumed that maize was the major crop of Pueblo Southwest. Maize is more amenable to selection and local adaptation than the other crops, and by virtue of its composition and its regular stOrage on a solid cob, it preserves better than the other two. Further, maize plays a major role in the cultural systems of contemporary and historically-known Pueblo people. Archaeologists have reasoned from various evidence that these traditions have great time depth and were likely established in prehistoric Pueblo societies. In this study it is assumed that approximately 80% of the remaining lands available for agriculture were devoted to maize production and the remainder used for other crops. This value represents the adjusted gross yield or the harvested maize production. The estimate of 80% was used recently by Williams (1989:726) in a study of a sixteenth century Aztec peasant community, Santa Maria Asuncion, in the Basin of Mexico, and has also been used by Sanders (1976:109, 145; 1979:376) in the Basin of Mexico. This value is closely ma1ched by the ethnographic reports of Hack who said that modem Hopi at that time devoted 72% ofall farmland to growing maize (Hack 1942:19), and by Judd (1981:39) who quoted Hough (1930) and claimed that over three-quarters of • While minimum daily caloric requirements vary from person to person and depend on factors such as age, gender, metabolism, and level of activity, it is assumed here that an average of 160 kg of maize per year was required to sustain the average person. This value is equal to .4384 kg per day or 1534 calories per day where 3500 calories are produced from a kg of maize (Cook and Borah 1979:164). It is -124 - accurate in predicting the actual number of people that produced a crop of maize than the calculation for the single year. This may be true not because people stored subsistence foods for this length of time, but rather because they required more yield per person to satisfy the various demands for that supply. assumed that additional calories (and minerals and proteins) will be obtained from foods other than maize. but it is not necessary here to specify a value representing total daily caloric intake. The estimate of 160 kg of maize per year was used by Sanders (1976:145) as an aggregate statistic for maize consumption and is based on values from ethnographic data Williams used a like figure (1989:725). The values used to estimate the rate of consumption vary in archaeological models. The number of calories obtained from a kg of maize varies (e.g., Ford [1968: 157] uses 355 kcal/l00 g. Gross [1985:103] uses 356 kca1/100 g, Hastorf [1980:99] uses 360 kcal/100 g). The number of calories assumed to be necessary to sustain an individual also varies (e.g., Schlanger [1985:44] assumes 2,000 calories a day, and Kohler et al. [1986: 528] assume 2,400 calories a day), and the amount of maize consumed varies (e.g., Schlanger [1985:44] assumes 50% of calories are obtained from maize. and Kohler et al. [1986:528] assume 60% of caloric requirements are obtained from maize). • Population density is computed by dividing the number of people by land area. It is desimble to calculate population density (rather than, or perhaps in addition to, population size) because it represents a standard unit of measure that can be compared from place to place without regard to size or total area. In this study, the entire study area is comprised of 36,759 cells for which complete soils data are available. Each cell is 200 by 200 m (4 hal in area. Thus, the total, analyzable area of the study reton is 147,036 hectares or 1470.36 km . The Sand Canyon Survey Locality contains 652 cells and includes 26.08 km 2. The Mockingbird Mesa Survey Locality contains 449 cells and incorporates 17.96 km 2. Six of the eight tree-ring dated sites (5MT8371,5MT8839,5MT2433, 5MT3834, 5MT1566, and 5M1765) include 197 cells in their 1.6-krn radius, representing 7.88 krn 2. 5MT6970 and 5MT2149 are located near the margins of the swdy area and consequently a portion of the 1.6-km radius fell outside the boundary. Thus, 5MT6970 took in only 108 cells or4.32 km 2 and 5MT2149 included 168 cells or 6.72 krn 2. It is important to note that these area measurements are used for every year in the analysis and do not consider the distribution of individuals or settlements within that area. Thus, the density values produced do not reflect the degree of clustering or aggregation of settlement within the areas identified. Furthermore, the boundaries established for purposes of analysis are arbitrary and do not reflect the acwalbounding of human settlement or activity. However, the estimated values do represent an average population density that could be supported by an area so bounded and are useful numbers for envisioning a upper limit for supportable populations. • In this study, the possibility of storing a portion of edible harvest for one and even two years beyond the needs of the current year is considered (Bums 1983). Here an estimate of population size is made on the full consumption of the adjusted net yield in a single year (POPIYR), on consumption of that yield if a portion is stoced and predictably reduced and used for a second year (POP2YR), and on consumption of that yield if a portion is stored., reduced, and used over two additional years (POP3YR). The goal of attempting to annually grow two years of maize is often cited by ethnographers as the Pueblo ideal (Hough 1915:62; Forde 1931:393; Parsons 1936:954-955; Whiting 1939:13, 15; Titiev 1944:181; Ford 1968:162; Bradfield 1971:21) although it is occasionally said that it is rarely achieved. It seems clear, however, that more maize was grown than was needed to support the immediate producers for a year. Maize was grown also to fulfill reciprocity obligations among kin and was used regularly in ceremonies, weddings (Kavena 1980:14), and trade and exchange transactions (Bradfield 1971:21). Consequently, the calculations for two or even three years of stomge may be more It must be pointed out that the net effect of the second-to-Iast point above, which concerns itself with determining population size at three - 125- contrasting storage/demand levels, is that population estimates will be reduced rather than inflated. In effect, I have assumed that while populations grow and store more than one year's demand for maize, they do this every year without consideration of what "excess" might remain in storage at the beginning of the planting season. For example, if a population in Year 1 managed to raise the equivalent of two year's supply of maize and deposited it in storage after the harvest of Year 1, it would still have the better part of that yield in storage at the beginning of the planting season of Year 2. Given the ever-present risks associated with crop production in the American Southwest, this same population would attempt to plant and harvest an equally large crop in Year 2. Should the goals for the Year 2 maize crop be realized at harvest time, the old grain in storage, which was marked for the second year, will then be consumed (eaten, traded, used as barter tender, or otherwise used) and the fresh grain used to replace the old grain. Should, however, a less than totally successful harvest be realized, some old grain will be kept (i.e., the needed balance preserved) and the amount in reserve from Year 1 and Year 2 harvests will not exceed the storage level modeled for that population. So, while I assume a population may carry over stored reserves, in effect I do not build the carry over into the model; these populations attempt to grow as much as possible, but not more than can be stored or maintained in their storage facilities. Some might argue that this will result in unrealistically low carrying capacity estimates for the storage scenarios. However, because one conclusion of this study will be to argue that short-term climate change is probably insufficient, by itself, to cause the total depopulation of the Four Comers area, some conservatism in population estimation seems defensible. If, in fact, more people could have been supported by their maize production than I estimate, then that strengthens the argument that the thirteenth-century depopulation is not solely attributable to climate change. It is important to emphasize that this model attempts to estimate the maximum values for productive acreage, maize yield, and population size possible at each level of storage or demand for every year in the 4OO-year series. Estimatio2 Lon2-Teun Sustainable Human PQDuiatiQus One Qf the goals of this study is to ascertain whether climatic conditions could have negatively affected the supply Qf agricultural resources to such an extent that they alone could account fQr the major depopulation of the Mesa Verde region in the late thirteenth century. This entailed reconstructing relevant climatic conditions, modeling agricultural productivity, and estimating the size of human populations that could be supported by the available yield. Thanks to the availability of high-resolution tree-ring data, this has been done on an arumal basis. However, these yearly values must be evaluated for patterning through time to incorporate the effects of annual and multiyear fluctuations in climate and crop productivity on sustained population levels. In this study, the concepts of maximum, critical, and optimal carrying capacity have been adapted from Hassan (1981: 166-168) to estimate sustainable pQpulation levels over extended periods of time. These concepts are depicted schematically in Figure 4.4. Maximum carrying capacity is a population size estimate that is equivalent to the long-tenn mean value of the estimated yearly maximum population for the total period of 400 years. This value is the upper limit on population size and would represent a regional population that would frequently experience yield shortfalls when annual production fell noticeably below the mean. Critical carrying capacity is a population size estimate that is equivalent to the minimum annual population value in the 400year period. It is a value below the maximum carrying capacity (the mean) and represents the largest population that would exist throughout the entire period without experiencing significant crop shortages. Thus, it is detennined by the years when yield is least abundant Optimum carrying capacity is a population size estimate that falls within a range of values. Its upper limit is roughly equivalent to the critical carrying capacity value (the minimum value). Its lower limit is not specified, but it is the size of a population operating well below the limit of fluctuating productivity. It is almost always within the range of values equal to 20--60% and is 126 - Maximum Car rying - 1--+---\-f---l-l---\-/--+---1'4-J-++++-++-+--I+--f'r--+1+-t-iY-+I--+Mean Val ue (100%) Capacity 90% 60% 70 % Cr itic al Carrying Capacity Zane of Oplimal Carrying Capacity Minimum Va Iue (60%) 50% ｈＭＺ［Ｎｾｲ｟ＢｔＷｴ 40% 30% 20% 10% Time Figure 4.4. Schematic diagram illustrating the relationship among three concepts of carrying capacity (after Hassan 1981:167). generally between 40-60% of the maximum carrying capacity value (Hassan 1981: 175). To obtain estimates of carrying capacity, the mean value (the maximum carrying capacity), the minimum value (the critical carrying capacity), and other descriptive data were calculated for each 400-year set of maize demand levels (pOPIYR, POP2YR, POP3YR) with SAS Programs (Appendix D, Program 2, Proc Means). This was performed on the data forthe study area as whole, for the two survey Table 4.5. Summary of Model-Building Steps: Operationalizing the Model. I. Paleoenvironmental Data A. AcqUire appropriate tree-ring data (SWOLD7 for years A.D. 901-1970). B. Reconstruct PDSI Values for the time period represented by the tree-ring data for all combinations of weather stations (Bluff, Cortez, Ignacio, Mesa Verde, and Ft. Lewis) and soil moisture groups (11 classes based on available water-holding capacity in the upper and lower soil layers). Product is 55 sets of reconstructions. 1. Calculate PDSI for Instrumented Time Period. 2. Calibrate PDSI and Tree-Ring Data for Instrumented Time Period. a. Initial Calibration. b. Verification. c. Final Calibration. 3. Retrodict PDSI Series, A.D. 901-1970. C. Construct PDSI tables in GIS-compatible fonn (subset using A.D. 901-1300 organized as a set of tables). II. Elevational Data A. AcqUire and process OEMs. B. Produce a single, mosaiced, elevation data plane (DEMWIND.EPP). C. Create new data plane depicting the zones of elevation to be modeled by the weather stations (in this case, five zones modeled by five stations) (5ELEV.EPP). - 127- Table 4.5 (Concluded). III. Soas Data A. Acquire descriptive and locational data on soil map units. B. Produce a single soils data plane depicting all soil map units in the study area (SOILWIND.EPP) C. Create a new data plane depicting all soil map units reclassified into a reduced set of soil classes based on potential water-holding capacity of soil layers (in this case, 11 soil classes subsuming 98 different soil map units) (11AWC.EPP). IV. Soil Productivity Data A. Acquire historic bean and maize production data for study area. B. Acquire data on natural plant productivity for all soils in study area. C. Perform regression-based analyses to determine the relationship between varying levels of agricultural productivity and PDSI. 1. Determine the nature and strength of the relationship between PDSI and historic yield of bean and maize. 2. Estimate bean and maize yields for all soils from known relationship between domestic plant productivity (bean and maize yields) and natural plant productiVity per soil type on a sample of soils. D. Calibrate PDSI values with crop yield values. E. Construct an agricultural productivity table in GIS-eompatible form (a table providing varying values for agricultural yield per soil type given varying PDSI values and climatic conditions). V. Ethnographic Data A. Gather appropriate data useful for modeling prehistoric Pueblo agricultural practices, consumption practices, and per capita demand for yield. VI. GIS-Integrated Final Analysis A. Create new data plane to facilitate the analysis. 1. Create an image that assigns each long-term PSDI reconstruction to its appropriate location (PDSIMAP.EPP). B. Run program that generates annual images of agricultural productivity. 1. Associate each cell on PDSIMAP with its POSt value for each year in the A.D. 901-1300 series, via PDSI tables (product is the annual PDSI reconstruction, TEMPyyyy.EPP). 2. Associate each cell's PDSI value with a PDSI nominal class (1-11). 3. Associate each cell's PDSI nominal class value with a yield value, via agricultural productivity table. 4. Depict the value for each cell's agricultural yield value on a new output image. Image may be displayed and photographed (product is the annual productivity reconstruction, PRODyyyy.EPP). 5. Produce tabular output files (COUNT files) that summarize the total number of cells and their associated area in hectares for each productivity class (prodUct is annual productivity reconstruction count file PRODyyyy.CNT). 6. Store tabular data for subsequent transfer to a statistical analysis program for further processing and analysis. VII. SAS Data Final Analysis A. Calculate total and effective annual agricultural maize yields for study area. B. Calculate the maximum annual population (number of people and density) that could be supported on that effective yield for a population with varying requirements for maize storage for the study area. Product is an annual estimate of the maximum population size, for three levels of storage. C. Calculate the annual maximum, annual minimum, and the long-term mean population (number of people and density) that could be sustained over the period of 400 years (A.D. 901-1300) for the study area. Products are long-term estimates of population size, expressed as the maximum, critical, and optimal carrying capacity, for three levels of storage. ­ 128- localities and for the eight site catchments. THE COMPLETE MODEL The products of the modeling procedure described in chapters 2. 3. and 4 are the annual images (or "maps") of agricultural productivity reconstructed for the A.D. 901-1300 period. the numeric estimates of annual maize productivity. and the numeric estimates of annual and long-term population for the same time period. An example of one of these annual images prepared as a black-and-white EPPL7 DOTPLOT was presented as Figure 4.3. The image is better viewed and interpreted as a colored computer image seen directly on a monitor. captured photographically as a color slide. or observed as a series of colored images prepared as an animated time series. - . The numeric estimates of annual total maize productivity and annual and long-term population size for the study area are presented in abbreviated form in Appendix E. The estimates made for the two subareas. the Sand Canyon Survey Locality and the Mockingbird Mesa Survey Locality are presented in Appendices F and G. The annual estimates for the eight treering dated site catchments are presented in Appendices H through O. The data and analytic tools used to reconstruct the prehistoric dryland farming habitat of the Mesa Verde Region Anasazi may be summarized in outline form. Table 4.5 shows the steps taken to develop the final analytic products. This table may also be seen as a synopsis specifying the methods and sequential products that allow the conceptual model identified in chapter 1 to be operationalized. 129- Page Blank in Original 5 RESULTS This chapter presents and explores data created by the model of prehistoric agricultural productivity and sustainable human population in southwestern Colorado for the tenth through the thirteenth centuries A.D. The primary data are estimated values of annual maize yield and annual potential population for three different levels of demand/storage. Appendix E supplies the annual estimates of maize yield and annual maximum population for the study area as a whole. Appendices F and G present these estimates for two contrasting survey localities within the study area, the Sand Canyon Locality and Mockingbird Mesa. Appendices H-O supply these values for the catchments surrounding eight tree-ring dated sites within the study area. The purpose of this chapter is to examine these data and suggest their usefulness for addressing problems of interest to archaeologists working with horticultural societies. My exploratory studies, however, are neither prescriptive nor exhaustive; they are meant only to demonstrate some of the uses for these data at different scales of analysis. The data generated for the study area as a whole embody an assessment of agricultural productivity and supportable population for a 1470.36 km 2 area. My analysis of these values may be seen as representing one of the studies that can be accomplished with such information on a broad geographic or regional scale. Here I examine certain characteristics of the dataset, determine human carrying capacity levels that would have been sustainable over the full 4OO-year period, and fmally arrive at some tentative conclusions about the effect of known climatic variation on agricultural production and human population in the study area. On a smaller geographic scale, the data produced for the two block survey localities within the study area may be seen as one of the results that can be generated for a subregional level of analysis. At this level it is possible to compare archaeological estimates of population derived from architectural and artifactual remains to the ecologically derived estimates produced by the model. In this way, I compare "observed" values to "expected" values and draw conclusions as to local carrying capacity limits and whether productive limits have been exceeded. This exercise was particularly instructive because the environments of the two localities contrast in some ways that potentially affect agricultural productivity and the number and density of people who may be supported by agriculture. In addition to comparing the two localities to each other, I compare the localities to the larger study area, to gain an appreciation for the degree to which each of these spatial samples represents the whole. The data produced for the eight tree-ring dated sites may be seen as analysis on yet a smaller geographic scale, the site catchment level. Studies like those performed at the regional and subregional levels can, of course, also be performed at this level. In addition, it is possible to characterize the productivity and population characteristics of the catchment during the time that the site was occupied. I use this information to evaluate whether the site occupation period was better than, worse than, or essentially the same as the long-term condition for that place. I also use the information to predict the most favorable times for settlement of that catchment during the 400 years examined. Further, it is possible to compare the catchment of sites that are known to be contemporary and assess their relative productivity. In this way, these data may be used to predict the more favorable places for productive agriculture during a given time. I do this in a limited way for three partially contemporary "central place" sites in my sample that are thought be cOimected to the Chacoan occupation of the study area. I also do this with two small village sites that are nearly identical in their dating. I classify 5MT8371 (DCA site), 5MT8839 (Norton House), 5MT2433 (Aulston Pueblo), and 5MT3834 (Mustoe Ruin) as small village sites on the basis of their limited structural remains and non-public architecture. All have been described as having two or fewer pit structures and 10 or fewer surface rooms. Alternatively, I classify 5MT6970 (Wallace Ruin), 5MT1566 (Lowry Ruin), 5MT2149 (Escalante Ruin), and 5MT765 (Sand Canyon Pueblo) as "central place" sites on the basis of having 25 or more rooms and public architecture such as large or special kivas and other unusual structures such as tri-walled, D-shaped buildings. Finally, it is possible to compare classes of sites to assess the role that differential agricultural potential and sustainable population size may play in site location or duration of occupation. Here I compare potential productivity and population values for the class of four "small village" sites against the class of four "central place" sites. THE STUDY AREA Annual values for total maize yield (TOTPROD) and the maximum number and maximum density of people who could be supported by that yield for a population requiring either one year (pOPIYR), two years (pOP2YR), or three years (pOP3YR) of maize in storage at the end of harvest have been estimated (Appendix E). Total maize productivity for the 400-year series is depicted graphically in Figure 5.1 and trends in mean population levels for the three storage/demand levels are depicted graphically in Figure 5.2. Simultaneous inspection of these two graphs reveals that population size varies (by defmition) directly with maize production. In some years maximum potential yield and maximum potential population are quite high and in some years they are much lower; often this shift happens quickly, one year following another. It is clear, however, that the size of a real population could not vary from year-to-year in this manner. Rather, longer-term trends in productivity control the real size of a population that can be sustained in a given place. Consequently, longterm trends were derived from the 400-year series of values. The 400-year mean value, the 400-year minimum value, and a range of values equivalent to 20% through 60% of the 400-year mean value were taken for POPI YR, POP2YR, and POP3YR to represent the maximum carrying capacity, the critical carrying capacity, and the optimal carrying capacity, respectively (Table 5.l). These data better approximate a sustainable population than do productivity estimates from any individual year. Hassan (1981) has suggested from a worldwide sample that the actual size of populations that are able to maintain their numbers over an extended period of time, usually exists within a range of values between 20% to 60% (and more commonly 40% to 60%) of the longterm mean estimate that can be calculated for their sustaining area. This range is termed the optimal carrying capacity. The maximum number of people who could be supported during the least productive years in the 400year series. including those infamous years of the "Great Drought of 1276-1299" (Douglass 1929, see Burns 1976 for a discussion of this concept), is the critical carrying capacity and is represented by the minimum population value of the 400-year sequence. Using the POP2YR set of values, the low value of 21 persons/krn 2 occurs 15 times with values nearly as low - 22 persons/krn2 and 23 persons/krn2occurring an additional 12 times. Occasionally two (but never three) population lows occur two years in a row (A.D. 906-907; 980-981). Other lows occur at close intervals (e.g., AD. 901 and 906-907; A.D. 972 and 980-981; A.D. 1062 and 1067; AD. 1146, 1150, 1156, and 1161; and AD. 1254 and 1258). By contrast, there are a few. fairly long intervals when the three lowest population values do not occur (A.D. 1020-1061, 1091-1130. 11871216. 1228-1253) and a number of periods of 15 or more years in length when the minimum population supportable was. in fact. reasonably high (Table 5.2). Generally, however, these very low values occur once every 10 to 25 years, a time likely to be recalled by adults in the population. The long-term minimum value. the critical carrying capacity, then, represents the long-term maximum number of people whose demands for maize could be met without suffering, barring other factors that 132 - 1.IE+oa 1.0E+oa 9.0E+07 a.OE+07 ..-. w w ｾ 9 7.0E+07 ｾ 6.0E+07 5.0E+07 ".OE+07 3 . Of +0 7 IIImn,nllllllfttlnIlRtllnlnHlnml.IIII11Il1I1I11MIIIIJlIIIIIIIAIIIIIIIIIIRlllfllllnlDllllflnMI1"nllA111RllllftIlHlmnnnUIIII"""mIIllPlU'KlIIIIIIIIIQlIUlllUlllllmUlIlnllltm"nnm"",,,,"IIUIIIOIIIIIIIIIIIIIII""""'IIII1I11""'"III""""I11"""'""III''"II'''I11'''''IIIII'III'"'''''''''"" 900 1000 1100 "' 1200 YEARS (A.D.) Figure 5.1. Maize production in the study area. A.D. 901-1300. The vertical needles represent annual yields. The undulating line represents weighted mean yields. . 1300 ISO 140 130 120 110 >< t:: 100 CI) ｾ 90 C Z ....0 W ｾ ｾ ..J ::J ｾ 0 ｾ 80 10 60 '0 40 30 20 10 0 , , 900 1000 1100 1200 YEARS (A.D.) Figure 5.2. Supportable population density in the study area, A.D. 901-1300. The vertical needles represent estimated maximum annual population. The undulating lines represent the weighted means for POP1YR (top), POP2YR (middle), and POP3YR (bottom). 1300 Table 5.1. Carrying Capacity Estimates at Three Levels'ofStorage/Demand for A.D. 901-1300 for the 1470.36 km 2 Study Area. . Maximum Carrying Storage Level Capacnr POP1YR 88± 19 (persons per km 2) POP1YR 131,473 ± 28,222 (number of persons) POP2YR 35±7 (persons per km 2) POP2YR 53,246 ± 11 ,429 (number of persons) POP3YR 19±4 (persons per km2) POP3YR ...rumber of persons) 28,752 ± 6172 Mean Value ± Standard Deviation. Critical Carrying Capacityb Optimal Carrying Capacityc 52 (59%) 18 - 35 - 53 77,439 26,295 - 52,589 - 78,884 21 (60%) 7 - 14 - 21 31,363 10,649 - 21,298 - 31,948 11 (58%) 4-8-11 16,936 5,750 - 11,501 - 17,251 bMinimum Value. ｫＢｏｾ - 40%· 60% of Mean Value. affect successful crop production. The maximum carrying capacity, the long-term mean, is theoretically the highest long-term population size that could be sustained most, but not all, of the time. It is the least realistic of the three levels of carrying capacity, but it does provide an upper threshold for estimating a realistic aggregate population for the study area as a whole. The population density values provided in Table 5.1 are high; higher than some researchers might think possible, particularly towards the end of the thirteenth century when the Mesa Verde Area was permanently abandoned by Anasazi populations. The estimates for POPI YR are probably the least accurate; we suspect that prehistoric Puebloans did attempt to store at least two years of maize at the end of harvest in the event that production was insufficient the following season. Consequently, the population estimates for POP2YR are likely to be better estimators of sustainable population, although they suffer from the conservative bias reported in the previous chapter. POP2YR represents the population size that could be supported for two years from the maize produced in the present year, or the population that could be supported if the production in the present year were followed by a year with no production whatsoever. Since this is rarely or never the case, POP2YR (and, all the more, POP3YR estimates) have a conservative bias). - Thus, by placing the maximum inferential weight on the use of the middle storage/demand level (pOP2YR) and the more likely estimates of sustainable population size (the critical and optional carrying capacity values), I calculate that some 21,298-31,363 persons, representing a density of some 14-21 persons/km2 for the 1,470.36 km 2 study area could have been supported in any given year within the A.D. 901-1300 time period. The lower value represents the 40% of mean value from the optimal carrying capacity range, and the upper represents the critical carrying capacity value. This upper value is very similar to the value recently suggested by Rohn (1989) for aggregate population in the Montezuma Valley, an area that includes the study area but extends fully from the eastern slopes of the Abajo Mountains in southeast Utah to the valley bottom below the Mesa Verde in Southwest Colorado. Rohn asserts that 30,000 is a conservative estimate of the number of people who lived in the Montezuma Valley in the thirteenth century (Rohn 1989:166), an estimate derived from his knowledge of archaeological data, particularly survey data in the Montezuma and Dolores County area. Further, this estimate of 30,000 does not include the numbers of people he estimates for nearby Mesa Verde or the Mancos Valley. Therefore, it would seem that the POP2YR estimates are in line with one of the few general estimates of population that has been proposed for the 135- Table 5.2. Periods of Greatest Occupational Attractiveness in the Study Area. Rank by Rank by Minimum Years a Pop. Densityb 4 6 7 7 8 8 1 4 2 2 5 1 3 3 6 5 Beginning Date A.D. 931 955 1020 1091 1091 1100 1228 1259 Ending Date A.D. 952 971 1034 1125 1120 1120 1250 1275 Number of Years 22 17 15 35 30 21 23 17 Minimum Population Density Value (km2) 26 26 26 27 29 30 29 27 aBased primarily on number of years and secondarily on the minimum population density that can be sustained. bBased primarily on minimum population density and secondarily on the number of years that this condition can be sustained. heartland of the Mesa Verde region. In sum, I believe that two conclusions may be drawn from this preliminary analysis of the data generated for the study area as a whole. First, climatic variation alone as it affected agricultural production cannot be cited as the simple and sufficient cause for the major depopulation of the Mesa Verde Region of the Anasazi in the late thirteenth century A.D. Barring crop failures due to plant disease or pests or loss of significant quantities of farmland due to soil erosion or nutrient depletion, there was always enough productive land to support thousands of people in the study area (minimum population size of 31,363 person or 21 persons/km2), even during the difficult times of the middle ll00s, which coincided with the collapse of the Chacoan system, and the so­called Great Drought of the late 1200s. Second, the distribution of productive land changes from year to year, although there are locations in the study area that are more consistently productive and others that are more consistently unproductive than other locations over long periods of time. Inspection of the annual images generated by the GIS makes this point abundantly clear since it is possible to see where the most and least productive lands are located in any year. Examination of the full 400 years reveals those places that are reasonably constant, and therefore predictable, for considerable intervals of time. Therefore, it may be concluded that access to productive resources (Le., the arable land or the products produced from that land) was an issue of extreme importance (Adler 1990; Kohler 1989). In this situation, either people must be allowed to go to where they can grow the food, or the food must be grown and distributed to where the people are located. If mobility and access to the productive lands are restricted, or if redistribution systems are not in place to support dispersed populations or uneven production, then the potential aggregate population figures will overestimate the actual population that could have been supported. These issues fall as much in the sociopolitical realm of human cultural systems as in the environmental realm, and they force us to seriously consider the complex interactions of climatic variation. environment, and human behavioral systems in issues of culture change and stability. LOCALITIES WITHIN THE STUDY AREA: THE BLOCK SURVEY AREAS Sand Canyon Survey Locality The Sand Canyon survey locality, as digitized for this study, is a 26.08 k:m 2 area surrounding two large thirteenth century pueblos, Sand Canyon Pueblo (5MTI65) and Goodman Point Ruin (5MT604), located near the heads of Sand and Goodman Canyons on the highland mesatop setting known as McElmo Dome (Figure 2.4). The locality straddles portions of two adjacent 7.5­minute USGS quadrangle maps, Woods Canyon to the west and Arriola to the east, and is located within the south­central portion of the study area. Elevations in the survey locality range from 1,890­2,164 m (6,2QO­7,100 ft). Precipitation and temperature 136 - patterns affecting local soils were therefore modeled with PDSI reconstructions from the Cortez, Ignacio, and Mesa Verde series. Survey and testing within the locality was undertaken by the Crow Canyon Archaeological Center between 1985 and 1987 (Adler 1988; Van West 1986; Van West et al. 1987); 429 sites were recorded. Of these, 420 single and multiple component prehistoric sites were attributed to the Anasazi culture. Of these 420, 18 habitation components with a mean room count of six rooms per site date to the AD. ＹＳｾＹＸＰ period, 73 habitation components with a mean room count of six rooms per site date to the AD. Ｙ Ｘ ｾ 1060 period, 89 habitation components with a mean room count of eight rooms per site date to the AD. 1ＰＶｾ 1150 period, and 95 habitation components with a mean room count of 20 rooms per site date to the AD. Ｑ ＵｾＱＳＰ period (Adler 1988:27, and revisions, personal communication, March 1989). Room counts and room measurements on these habitation sites, in conjunction with assumptions about rooms per household units, mean number of people per room, and site longevity have permitted Adler to make momentary population estimates based on inferred habitation life spans of 20 years (Le., 20-year momentary population estimates) and habitation life spans of 50 years (Le., 50-year momentary population estimates) for each archaeologically defmed time period (Adler 1988:27). These archaeological data represent a set of "observed" population values that can be compared to the "expected" carrying capacity values derived by the environmental model. Annual values for total maize yield (TOTPROD) and population at three levels of storage/demand (pOP1YR, POP2YR, and POP3YR) for the Sand Canyon survey locality are found in Appendix F. Total maize productivity for each year in the 4()()..year series for the Sand Canyon survey locality is graphed in Figure 5.3 and trends in mean population density for the three levels of storage/demand are displayed in Figure 5.4. Mockingbird Mesa Survey Locality The Mockingbird Mesa survey locality, as digitized and measured in this study, is a 17.96 km 2 area cotenninous with the mesatop surface of Mockingbird Mesa (Figure 2.4). Mockingbird Mesa is a north-northeast to south-southwest oriented mesa situated north and west of Sandstone and Yellowjacket Canyons and south and east of Hovenweep and Negro Canyons. It is depicted on the Negro Canyon 7.5minute USGS quadrangle map and is positioned in the west-central portion of the study area. Elevations in the survey locality range from 1,829-1,975 m ＨＶＬＰｾＴＸ ft). Consequently, reconstructions drawn from the Cortez and Ignacio series were used to model soil moisture conditions in local soils. The Bureau of Land Management conducted archaeological survey in this locality between 1981 and 1984; 684 single and multiple component sites was recorded. Of these, 700 components representing some 550 sites were assigned to the Anasazi culture. Fettennan and Honeycutt (1987:120), who summarize the work, report that 47 habitations representing 66 households can be attributed to the early Pueblo IT period (A.D. 925-10(0), 37 habitations representing 72 households to the late Pueblo II periods (A.D. ＱｯｾＰＩＬ 47 habitations representing 118 households to the early Pueblo III (AD.ll00--12oo), and 40 habitations representing 155 households to the late Pueblo III (AD. 1200-1275). Household counts, in conjunction with assumptions concerning the numbers of people per household, human life expectancy, and site life expectancy, allowed them to make population estimates based on inferred habitation life spans of 12 years and 100 years (or in the case of late Pueblo III, of 75 years) for each of the defined archaeological periods (Fetterman and Honeycutt 1987:119, 120). Sarah Schlanger (1985:194-195) used a sample of these same archaeological data from Mockingbird Mesa and ascribed sites and their components to a different set of archaeological periods based on ceramic assemblages resulting from the Dolores Archaeological Project. She assigned two sites representing 20 people to A.D. ＰＸＹｾＲ period, 30 sites representing period, five 285 people to the A.D. ＵＲＰＱｾＸＹ sites representing 50 people to the A.D. 10251100 period, 14 sites representing 155 people to the A.D. 1100-1175 period, and 49 sites 137 - 2600000 2500000 2400000 2300000 2200000 2100000 ..... w 00 ｾ ...... ｾ 2000000 1900000 1800000 1700000 1600000 1500000 1400000 1300000 1200000 .,. 900 1000 1100 1200 1300 YEARS (A.D.) Figure 5.3. Maize production in the Sand Canyon Survey Locality, A.D. 901-1300. The vertical needles represent annual yields. The undulating line represents weighted mean yields. 200 190 180 170 UiO 15<> 140 b 130 (I) 120 ffi 110 Q 100 S ..... w ｾ \0 ｾ 90 ｾ 80 70 60 5<> 40 30 20 10 0 . 900 , 1000 1100 1200 1300 YEARS (A.D.) Figure 5.4. Supportable population density in the Sand Canyon Survey Locality, A.D. 901-1300. The vertical needles represent estimated maximum annual population. The undulating lines represent the weighted means for POP1YR (top). POP2YR (middle), and POP3YR (bottom). representing 530 people to the A.D. 11751250 period. Household counts, in conjunction with assumptions about the number of the people per household, the frequency of structure reoccupation, and the life expectancy of a living room, permitted her to make population estimates based on an inferred habitation life span of 20 years (i.e., 20-year momentary population estimates) for the each of the archaeologically defined time periods (Schlanger 1985:188, 194, 195, 198, 199). To explore the comparison of archaeological data with the data created by the agricultural productivity model, I have chosen to use Schlanger's population estimates instead of Fetterman and Honeycutt's estimates because Schlanger's data are more comparable to those generated by Adler for Sand Canyon. The archaeological time periods are more alike, both use 20-year momentary population estimates, and both provide density values for population in their respective areas. Schlanger, Fetterman and Honeycutt, and Adler use somewhat different values and assumptions for their estimates, however, so the results are not strictly comparable. The interested reader should refer to their publications for details of method. Armual values for total maize yield (TOTPROD) and population at three levels of storage/demand (pOP1YR, POP2YR, and POP3YR) for the Mockingbird Mesa survey locality are reported in Appendix F. Total maize productivity for each year in the 400-year sequence for the Mockingbird Mesa survey locality is portrayed in Figure 5.5 and trends in mean population density for the three levels of population are portrayed in Figure 5.6. The Localities and Their Relationship to the Study Area Table 5.3 compares the 400-year mean maize yield and population values for the two localities with those of the larger study area. While both the number and density of people able to be supported by the 400-year mean yield are provided, the more useful figure is the population density value as it is standardized by area. The popUlation density values predicted by the productivity model for the - Mockingbird Mesa survey locality are nearly identical with those of the study area as a whole, whereas the density values predicted by the model for the Sand Canyon locality are markedly higher than for the study area or for Mockingbird Mesa. This would seem to indicate that the Mockingbird Mesa Locality is generally representative of conditions in the study area as a whole, although it does not include the lowest and highest elevations or the least and most productive soils in the study area. It also demonstrates that there are places within the study area that are better than others insofar as productive land is concerned. The Sand Canyon locality is a more productive and predictable location in which to farm than is the Mockingbird Mesa locality (The coefficient of variation-the ratio of the standard deviation to the mean multiplied by 100-indicates that there is more overall variation associated with the annual values for Mockingbird Mesa than for the study area, and much less variation associated with the Sand Canyon locality. The coefficient is an inverse measure of overall predictability). Other places within the study area, particularly lands in the north and north-east portions of the study area, appear to be to be equally if not more productive. Table 5.4 provides the maximum, critical, and optimal carrying capacity values for the three areas expressed as persons/km2 for a population requiring two years maize at the end of harvest (pOP2YR). The Mockingbird Mesa survey locality at any time within the A.D. 901 - 1300 period could have supported a population density of at least 18 persons/km2, a value equal to 50% of the long tenn mean value of 36 ± 9 persons/km 2. This minimum value or critical carrying capacity occurs 16 times over the 4OO-year time span, in a similar pattern to that of the study area. By contrast, the data indicate that the Sand Canyon survey locality at any time during the same 4OO-year period could always have supported a density of at least 37 persons/km2, a value equal to 67% of the long-tenn mean value of 55 ± 10 persons/km 2. Interestingly, this 4OD-year minimum value occurs 29 times, or almost twice as often as in the other locality. In absolute tenns, this long-tenn minimum value, or critical carrying capacity, is twice as high as for the Mockingbird Mesa survey locality and 140- 1300000 1200000 II 00000 1000000 .... ｾ ｾ ｾ ...... 900000 800000 700000 600000 500000 400000 .. 900 1000 1100 1200 1300 YEARS (A.D.) Figure 5.5. Maize production in the Mockingbird Mesa Survey Locality, A.D. 901-1300. The vertical needles represent annual yields. The undulating line represents weighted mean yields. 150 140 130 120 110 >- t: 100 ffi 0 90 fI) Z 10 j 70 9 ­ ｾ N :J ll. 0 ll. 60 50 40 30 20 10 0 , , 900 1000 1100 1200 1300 YEARS (A.D.) Figure 5.6. Supportable population density in the Mockingbird Mesa Survey Locality, A.D. 901-1300. The vertical needles represent estimated maximum annual population. The undulating lines represent the weighted means for POP1YR (top), POP2YR (middle), and POP3YR (bottom). Table 5.3. Comparison of Population Values for the Study Area, Sand Canyon Survey Locality, and Mockingbird Mesa Survey Locality A.D. 901-1300. Study Area POPNUM 1,470.36 Sand Canyon Locality POPKM POPNUM 26.08 26.08 Mockingbird Mesa Locality POPKM POPNUM 17.96 17.96 Area (km2) POPKM 1,470.36 TOTPRODa (kg) C.V.b 64,925,217 ± 13,936,845 21.5 64,925,217 ± 13,936,845 21.5 1,779,087 ± 337,524 19.0 1,779,087 ± 337,524 19.0 POP1YRc Mean, S.D. Range 88±19 52....141 131,473 ± 28,222 77,439-207,382 137±26 93-198 3,602 ± 683 89±23 2,437-5,174 46-142 POP2YRd Mean, S.D. Range 35±7 21-57 53,246 ± 11,429 31,363-83,989 55± 10 37-80 1,458±276 987-2095 801,336 ± 204,994 25.6 36±9 18-57 801,336 ± 204,994 25.6 1,622 ± 415 834-2,553 656± 168 337-1,034 POP3YRe Mean, S.D. 19±4 28,752±6,172 29±5 787±149 19±5 354±90 Range 11-30 16,936-45,354 20-43 532-1,131 10-31 182-558 aTOTPROO is the total mean productivity of maize rounded to the nearest whole number; all population values (POPIYR, POP2YR, POP3YR) are truncated integers. bCoefficient of Variation, the ratio of the standard deviation to the mean multiplied by 100, rounded to the nearest tenth. This yields a percentage that is rounded to the nearest integer. cPOP1YR .. «(TOTPROO*.324)/160)/area. dpOP2YR .. «(TOTPROO *.324)*.81 )1320)/area. epOP3YR .. «(TOTPROO*.324)*.6561 )/480)/area. almost twice as high as for the study area (at 21persons/km2). It is a good indicator of the productivity and predictability of the Sand Canyon survey locality for fanning. Table 5.4. Carrying Capacity Values (POPKM) for the Study Area, Sand Canyon Survey Locality, and Mockingbird Mesa Survey Locality, based on POP2YRa Estimates for A.D. 901-1300. Study Area Sand Canyon Locality Mockingbird Mesa 1,470.36 26.08 17.96 Area (J<m2) 64,925,217 1,779,087 801,336 TOTPRODa (kg) ± 13,936,845 ± 337,524 ± 204,994 19.0 25.6 C.V.b 21.5 80 57 Maximum value 57 Maximum CC:c Mean value 35 ± 7 55± 10 36±9 37 (67%) 18 (50%) Critical CC: Minimum value 21 (60%) Optimal CC: 60% of Mean 21 33 22 40% of Mean 14 22 14 20% of Mean 7 11 7 Note: All population values are truncated integers. aTOTPROO is the total mean productivity of maize, rounded to the nearest whole number. bCoefficient of variation (ratio of the standard deviation to the mean multiplied by 100) rounded to the nearest tenth. cMaximum, mean ± standard deviation, and minimum population density values (POPKM) for POP2YR are found in Appendix E (entire study area) F (Sand Canyon locality) and G (Mockingbird Mesa locality). - 143- Comparing Archaeological Estimates with Productivity Estimates of Population Sand Canyon Survey Locality Figure 5.7 plots the 20­year and 50­year momentary population densities (based on inferred habitation life spans of 20 and 50 years, respectively) for each of the archaeological periods established by Adler (1988:27 and revised values, personal communication, March 1989) for the Sand Canyon survey locality. These are overlaid on a plot of the long­term mean value (the maximum carrying capacity), the maximum value, and the minimum value (the critical carrying capacity) calculated for the 400­year period from the annual population density values associated with POP2YR. Also depicted are those intervals when the minimum value (the critical carrying capacity) "lifts" for a period of time from its low of 37 persons/km2 to some greater value. These elevated periods were determined and plotted in the following way. Using population density data contained in Appendix F for POP2YR, the minimum population value for a consecutive series of 10 years was recorded for each 10year period beginning at A.D. 901. For example, the minimum value for A.D. 901­910 is 37 personjkm2. Similarly, the minimum value for each rurming set of lO­year periods (A.D. 902­911, 903­912, 904­­913, and so on) is recorded. When the minimum level changed (e.g., from A.D. 908­917 the minimum value is 46 rather than 37), that new minimum value was plotted at the fmal year of the 1Q­year period (e.g., 46 was plotted at A.D. 917) and was allowed to stay at that level until a new 10­year minimum value was recorded. Here I chose to plot the minimum value at the end of the lO­year period, rather than at the beginning or middle, to simulate an effective lO­year memory of past productivity that may be drawn on to make decisions about current or future agricultural behavior. In this short ­term view of time­a frame of reference more appropriate to a human lifetime and human recollection­this "temporary" lifting of the long­term minimum value may be perceived as stability, depending on the length of time that this condition persists. The actual or perceived reality of a heightened and enduring productive landscape may have significant behavioral ­ ramifications, some of which may be detected archaeologically. Evidence of population in.crease, agriCUltural intensification, or environmental degradation might, for example, be detected during or following an extended period of heightened critical carrying capacity. Examination of Adler's momentary population estimates (based on inferred habitation life spans of 20 years) indicates that local, aggregate population in the Sand Canyon survey locality never approached the critical carrying capacity of 37 persons/km2, despite steady population growth throughout the 400 years. However, if momentary population estimates are based on inferred habitation life spans of 50 years, then local population requirements for maize (under the POP2YR assumptions) would only have been met adequately during the first three rriodS. The minimum value of 37 persons/km ,however. was exceeded at times within the final A.D. 115a-13oo period, an interval when archaeological estimates indicate a population density of 40 persons/km2 was present. This indicates that there would have been years of shortage in meeting the demands of a population requiring two years of maize at the end of harvest if habitation sites were occupied as long as 50 years. These years of shortage in the fmal period are: A.D. 1150, 1156, 1176, 1185, 1191, 1215, 1227, 1254, and 1295. Overall, however, it would appear that the Sand Canyon survey locality was a highly productive and predictable farming location and that potentially large local populations would have been living well below the limits imposed by the productive environment. Mockingbird Mesa Survey Locality Figure 5.8 plots the 2Q­year mean momentary population density estimates (based on inferred habitation life spans of 20 years) for each of the archaeological periods established by Schlanger (1985:199) for the Mockingbird Mesa survey locality. As in Figure 5.7, these are overlaid on a plot of the long­term mean value (the maximum carrying capacity), the maximum value, and the minimum value (the critical carrying capacity) calculated for the 4OQ­year period using POP2YR values. Again, periods of elevated carrying capacity are depicted. 144- SAND CANYON SURVEY LOCALITY POPULATION DENSITY Mo'imum so Value so 70 70 ｾ N e .¥ ...... III c 0 ｾ 60 60 Mean Value (55) -Moximum 50 50 t1> ­ a.. ｾ >. III .f>o. VI Carrying Capacity c ｾ 40 Minimum Value (37)- 7 ｾ t 7 A7 1 7 7R7 j;. 7 j. X77-;=] 14 0 - t1> 0 30 30 C 0 0 20 20 10 'a Critical Carrying Capacity Optimal Carrying Capacity Zone (11-37) ｾ a. 0 a.. at 0 0 I'-"f " ａｏｾ ｏ i' (C cic cr C <, cCf CC, C c,c C 950 I 900-950 I 950-9S0 ｾ I()(X) I C coc ｣ｾ C <icc 1050 980-'060 ｾ CC,t tfCC«<,cC( C <,t 1100 I 1060-1150 ｣ｾ C tVCf CC,t t ( t c,c 1150 I ｣ｾ t < c, ｾ ttl' C ',' ' ( ' " ｾＧｴｻ＼ ＱＲｾ 1200 1050-1300 'to '(' ｉｾｏ I Years (AD) Figure 5.7. Comparison of estimates of carrying capacity for the Sand Canyon Survey Locality generated by the model with estimates derived from archaeological survey in the locality, A.D. 901-1300. N e ｾｭｵ ｩＮ｡ｍ MOCKINGBIRD MESA SURVEY LOCALITY POPULATION DENSITY 60 Value(S7) .>I- "- ｾ III ｾ III a.. ｾ .... ,J:.. 0\ ｾｯ § >- III 40 40 Meon Value (36) 30 30 C III Critical -Carrying 20 Capacity 0 Ｍｾ c C :J C- Minimum 20 Value 08l10 o Maximum Corrying Copocily 10 Optimol Carrying Capacity Zone (7-18) 2 ".,tont/ .. J a.. o AD 900 . Ｙｾ 920-980 1000 1050 1150 '200 0 '300 1250 ＱＲｾＰＭＧＳＰ ＹＸＰＭＱＰＲｾ Years (AD.) Figure 5.8. Comparison of estimates of carrying capacity for the Mockingbird Mesa Survey Locality generated by the model with estimates derived from archaeological survey in the locality, A.D. 901-1300. A comparison of Schlanger's data for Mockingbird Mesa with the long­tenn estimates generated by the model reveals a different scenario than that reconstructed for the Sand Canyon survey locality. While Schlanger provides momentary estimates based only on inferred life spans of 20 years for her five periods. these values reveal that two periods of major population growth are followed by two periods of major population decline. In both cases. the productive capabilities of the Mockingbird Mesa survey locality were significantly exceeded by a population estimated from archaeological data In the growth periods. A.D. 98G­1025 and A.D. II 75­1250. archaeological population estimates are 28.5 pers/km2 and 53 persons/km 2• whereas the 4OO­year minimum value or critical carrying capacity was 18 persons/km2 and the lO­year minimum value is always less than 35 persons/km 2. The earlier period is followed by a period of low population (A.D. 1025­1100. archaeologically estimated to be 5 persons/km 2) and the later period is followed by a time of unknown population density (A.D. 125G­1300). but a time in which it is known that at some point the Mesa is pennanently abandoned. These data suggest a repeated history of population overshoot and collapse when human demand exceeded agricultural supply. In concluding this section. it may be said that locality­based data for which archaeological survey and population estimates are available are potentially very infonnative. They can suggest patterns that would allow a variety of specific hypotheses to be tested in a given location. In the case of the two localities examined here. the productivity data in conjunction with the archaeological data were able to suggest two quite different histories of population in the 400 years examined. TIlis is particularly interesting since these two localities are relatively close to each other (about 15 km or 9.3 mi apart from center to center) and both possess excellent soils with relatively high available water capacities. However. the Sand Canyon survey locality is 124­385 m (2()().­620 ft) higher overall in elevation than the Mockingbird Mesa survey locality. and it is not as circumscribed by canyon topography. This preliminary study of locality­based data is useful in that it clearly demonstrates that there are places within the study area that are more pro- ductive and predictable than others, and even . among the better locations. there are some that are consistently superior. SITE CATCHMENTS WITffiN THE STUDY AREA: TREE­RING DATED SITES Small Village Sites 5MT8371, DCA Site Description. This unnamed site that I have chosen to call the "DCA site"(after the institution that excavated it. the Division of Conservation Archaeology at the San Juan County Museum near Fannington. New Mexico) is reported by Dykeman (1986). It is inferred to be a single component. early Pueblo II habitation site consisting of a single household. Excavations revealed a single, semisubterranean rectangular pit structure with multiple floor features that appeared to be associated with a number of extramural features including storage pits. postholes. shallow pits, an outdoor activity area, and a midden (Dykeman 1986:i). Location. The site is located in the SEI/4. SEI/4. of Section 1. T37N. RI9W. N.M.P.M. on the Negro Canyon 7.5­minute quadrangle map. with UTM coordinates 4.152.000 m N. and 678.800 m E (Zone 12). It is situated on the west edge of Mockingbird Mesa (although not in the Mockingbird Mesa survey locality). on a ridge near the head of Negro Canyon. at an elevation of 1,914 m (6,280 ft). The PDSI reconstructions used to represent soil moisture in this 7.88 km 2 catchment were drawn from the Cortez and Ignacio series. Dating. The assignment of calendrical dates to 5MT8371 by its excavator derived from three sources: tree­ring dates. ceramic studies, and radiocarbon estimates. The finnest dates are associated with the tree­ring dated samples (Table 5.5). All were charred wood from the roof of Pit Structure 1. Five samples were dated; two were cutting dates dating to A.D. 935. The latest date on the site was 940 vv (Dykeman 1986:68). A date of A. D. 952 for the ceramic assemblage was calculated by Dykeman on the basis of ceramic analysis and the application of two modal ceramic date fonnulae (Dykeman 1986:85). Four charcoal 147 - Table 5.5. Tree-Ring Dates from Site 5MT8371, DCA site (TRL# A-666). Structure and Feature Pit Structure 1-- Tr.2, Feat.2 Pit Structure 1--Tr.2, Feat.2 Pit Structure 1--Level 1 Pit Structure 1--Tr.2. Feat.2 Pit Structure 1--Roof fall TRL Cat. Number CVM-114 CVM-115 CVM-121 CVM-113 CVM-119 Field Number 32 33 126 32 59 Tree Species Juniper Pinon Pinon Juniper Juniper Outside Date Symbol 932 w 935 r 935r 936 w 940 w Note: The symbols (Robinson and Harrill 1974:4-5) used with the outside date are as follows: B Bark is present. G Beetle galleries are present on the surface of the specimen. L A characteristic surface patination and smoothness, which develops on beams stripped of bark, is present. c The outermost ring is continuous around the full circumference of the specimen. This symbol is used only if a full section is present. r Less than a full section is present, but the outermost ring is continuous around the available circumference. v A subjective judgment that, although there is no direct evidence of the true outside on the specimen. the date is within a very few years of being a cutting date. vv There is no way of estimating how far the last ring is from the true outside. + One or more rings may be missing near the end of the ring series whose presence or absence cannot be determined because the specimen does not extend far enough to provide an adequate check. ++ A ring count is necessary due to the fact that beyond a certain point the specimen could not be dated. The symbols B. G, L, c, and r indicate cutting dates in order of decreasing confidence, unless a + or ++ is also present. The symbols L, G, and B may be used in any combination with other symbols except v and w. The rand c symbols are mutually exclusive, but may be used with L, G, B, +, and ++. The v and vv are also mutually exclusive and may be used with the + and ++. The + and ++ are mutually exclusive but may be used in combination with all the other symbols. samples from four different features were submitted for 14C analysis. Only one fell within the range of the tree-ring dates and dates associated with ceramics. That uncalibrated date, from Feature 1, was reported as 1000 ± 40 B.P. (A.D. 1950, with a half-life of 5568 years) and assigned an A.D. 950 date by the excavators (Dykeman 1986:69). Given the beginning cutting date of A.D. 935 and the final dates from the structure, the structure was interpreted as likely to have been constructed sometime with the A.D. 935-940 period (Dykeman 1986:68). Since 5MT8371 is believed to have been occupied for 15-20 years, it is assigned an A.D. 935-950 date for this study. Site Catchment and Agricultural Productivity. Table 5.6 summarizes population density data for 5MT8371 using POP2YR levels of demand. Over the entire 4OQ-year period, it appears that values generated for the maximum carrying capacity for 5MT8371 are nearly identical to those for the nearby Mockingbird Mesa survey locality and to the study area as a whole (35 ± 10 persons/km 2 versus 36 ± 9 persons/km2 and 35 ± 7 persons/km2, respectively). However, the critical carrying capacity, - the minimum value, is considerably less for 5MT8371 than either Mockingbird Mesa or the study area (9 persons/km2 versus 18 persons/km2 and 21 persons/km2, respectively). In every respect, 5MT8371 is markedly less productive than the figures estimated for the Sand Canyon survey locality, which has a maximum carrying capacity of 55 ± 10 persons/km2 and a critical carrying capacity of 37 persons/km2 at all times. During the 16-year period when 5MT8371 was occupied, the carrying capacity values for 5MT8371 were higher than they were for the 400-year period, indicating that the small village site was established and inhabited during a time that was slightly better than the long-term average conditions for that catchment. Again, maximum and optimal carrying capacity values for 5MT8371 are very similar to those for the Mockingbird Mesa survey locality and the study area (38 ± 5 persons/km2 versus 38 ± 5 persons/km2 and 37 ± 4 persons/km2, respectively). The critical carrying capacity value for 5MT8371 during its period of occupation is higher than its for either Mockingbird Mesa or the study area as a whole (28 persons/km2 148- Table 5.6. Comparison of Population Density Values (POPKM) for 5MT8371, DCA Site, for A.D. 901-1300 and A.D. 935-950. 5MT8371 Catchment 7.88 350,996 ± 104,956 29.9 C.V.b c 57 Maximum value 35 ± 10 Maximum CC: Mean value c Minimum value c 9 (25%) Critical CC: OptimalCC: 60% of Mean 21 14 40% of Mean 7 20% of Mean Mockingbird Mesa 17.96 801,336 ± 204,994 25.6 57 36±9 18 (50%) 22 14 7 A.D. 901-1300 (400 years) Area (kfn2) TOTPRODa(kg) Sand Canyon Locality 26.08 1,779,087 ±337,524 19.0 80 55± 10 37 (67%) 33 22 11 Study Area 1,470.36 64,925,217 ± 13,936,845 21.5 57 35±7 21 (60%) 21 14 7 A.D. 935-950 (16 years) TOTPROD(kg) 372,413 851,962 1,855,196 67,853,516 ± 57,456 ±127,904 ± 207,518 ± 7,473,915 15.0 11.2 11.0 15.4 C.V. 48 (AD 938, 46 (AD 938) 65 (AD 942, 948) 41 (AD 938, 942, Maximum value 944) 944, 946, 948) Maximum CC: Mean value 38 ± 5 38±5 58±6 37±4 41 (71%) 24 (63%) 26 (70%) Critical CC: Minimum value 28 (74%) (AD 937) (AD 937,939) (AD 937, 944) (AD 937) 35 23 Optimal CC: 60% of Mean 23 22 15 23 40% of Mean 15 15 8 12 7 20% of Mean 8 Note: All population values are truncated integers. aTOTPROD is the total mean productivity of maize, rounded to the nearest whole number ± standard deviation beoefficient of Variation, the ratio of the standard deviation to the mean, multiplied by 100, rounded to the nearest tenth. cMaximum, mean ± standard deviation, and minimum population density values (POPKM) for POP2YR area found in Appendix H (OCA), G (Mockingbird Mesa), F (Sand Canyon locality), and E (entire study area). versus 24 persons/km2 and 26 persons/km2 , respectively), indicating its comparative attractiveness. Nevertheless, the catclunent is still less attractive than other places in the region at exactly the same time, as indicated by the values for the Sand Canyon survey locality, which are much higher (maximum carrying capacity of 58 ± 6 persons/km 2 and a critical carrying capacity of 41 persons/km2 for the entire 400year period). Site Catclunent and Prime Periods for Agricultural Prodyction. Given that the 16-year occupation between A.D. 935-950 was more productive than that over the long-term for the site catclunent of 5MT8371, the data available for 5MT8371 were searched for other equally - long periods that could support a minimum of 28 persons/km2 (Appendix H, Figure 5.9). Table 5.7 summarizes these data for the catchment of 5MT8371. In this study, I assume that the length of time a place can sustain a high carrying capacity is a measure of temporal predictability-the longer the time, the more predictable a place. I also assume that the size of the minimum population density a place can sustain is a measure of spatial productivitythe greater the minimum population, the more productive a place. By these two measures, 5MT8371, in comparison to the other time periods of equal of greater occupational attractiveness, is characterized by moderate predictability and productivity, whereas the A.D. 1098-1120 period is the most productive 149- Table 5.7. Periods of Equal or Greater Occupational Attractiveness in the Catchment of 5MT8371 (A.D. 935-950). Rank by Rank by Minimum Beginning Date A.D. Years a Pop. Densityb 4 4 (5MT8371) 931 8 8 1069 1 2 1091 2 1 1098 5 5 1167 6 6 1187 3 3 1228 7 7 1259 Ending Date A.D. 952 1084 1125 1120 1185 1204 1250 1275 Number of Years 21 16 35 23 19 18 23 17 Minimum Population Density Value (km2) 28 28 28 35 28 28 28 28 aBased primarily on number of years and secondarily on the minimum population density that can be sustained. bBased primarily on minimum population density and secondarily on the number of years that this condition can be sustained. time, and the A.D. 1091­1125 period is the most predictable for the catchment. If agricultural productivity and predictability were key variables for settlement decisions in the catchment of 5Mf8371, then sites dating to this time period should be present. 5Mf8839 Norton House Description. This site is reported by Fuller (1987) and by Kuckelman (1988). It is considered to be a single-component Pueblo II habitation. The site was located in the right-ofway of one of the Dolores River Project canals, the South Canal. Excavation revealed substantial architecture including "one pit structure with an adjoining pit room, a separate pitroom southwest of the pit structure and 11 exterior features" (Kuckelman 1988:373). The main rectangular pit structure was found to have had two occupational surfaces and to have burned on abandonment. No other pit structures or surface structures were defmed during survey and excavation, but it was suggested that modem plowing has likely destroyed surface rooms (Kuckelman 1988:403). Site Location. The site was located in the SWI/4 of the NWI/4 of Section 3, T38N, R18W, N.M.P.M., depicted on the Pleasant View 7.5-minute quad map, at UTM 4161580 m N and 692100 m E (Zone 12) on deep mesatop soil at an elevation of 2,093 m (6,868 ft). The site was in an area that is currently under cultivation, about three miles west of the modem farming community of Pleasant View. The PDSI reconstructions used to represent soil - moisture in the 7.88 km 2 catchment were drawn from the Mesa Verde series. Dating. Three sources of dates were explored to date this site: tree-rings, archaeomagnetism, and ceramics. All 19 tree-ring dated samples came from the burned roof fall of Pit Structure 1; eight of these were cutting dates (Table 5.8). Of these eight, the earliest cutting Table 5.8. Tree-Ring Dates from Site 5MT8839, Norton House (TRL# A-699). TRL Cat. Number CVM-174 CVM-184 CVM-189 CVM-176 CVM-178 CVM-170 CVM-171 CVM-172 CVM-175 CVM-177 CVM-186 CVM-179 CVM-181 CVM-185 CVM-188 CVM-180 CVM-187 CVM-190 CVM-182 Field Number 8 24 34 10B 14 1A 4 5 9 12 27 15 18 26A 31A 17 290 35 20 Outside Date Symbola 934 vv 971 vv 987 vv 1020 +H 1027 vv 1029 r 1029 H 1029 vv 1029 r 1029 v 1035 +rB 1037 r 1037 r 1037 r 1039 r 1040 +vv 1041 r 1041 HB 1048 H Note: All specimens are from Pit structure 1, and all are juniper. aFor explanation of outside date symbols, see Table 5.5. 150- 1$0 140 130 120 llO >- t:CIl ffi 0 -j ｾ .... .... Ul ｾ ｾ 0 ｾ 70 60 50 40 30 20 10 0 I 900 . 1000 lIOO 1200 YEARS (A.D.) Figure 5.9. Population density supportable within the 1.6-km-radius catchment of 5MT8371, DCA Site, A.D. 901-1300. 1300 date is A.D. 1029 r. The fmal date from the site is A.D. 1048 +T. A small cluster of dates at A.D. 1029 and another at A.D. 1037 may represent the major construction episodes at the site with the later dates representing minor repairs. Alternatively, the primary construction occurred in A.D. 1037 with significant reuse of wood previously cut and/or used elsewhere, if the second replastering of the floor was not accompanied by roof repair/construction. Kuckelman concludes that it is likely that "the pit structure was either constructed in A.D. 1029 or in A.D. 1037 and was abandoned after A.D. 1048" (Kuckelman 1988:386). Archaeomagnetic dates from the hearth in Pit Structure 1 proved of little value (Kuckelman 1988:386). Ceramic associations, however, corroborated a middle-to-Iate eleventh century placement, ca A.D. 1000--1075 (Kuckelman 1988:386). For the purposes of this study, the occupational dates assigned to this site are A.D. 1029-1048, a period of 20 years. Site Catchment and Agricultural Productivity. The 400-year maximum carrying capacity for the catchment is 67 ± 10 persons/km2 (Table 5.9). These long-tenn population density values are notably higher than those sustainable in the Mockingbird Mesa survey locality (36 ± 9 person/km2), in the study area as a whole (35 ± 7 persons/km2), or even in the Sand Canyon survey locality (55 ± 10 persons/km2). The critical carrying capacity for 5MT8839 is also higher than the other areas with a 4OO-year minimum value of 49 persons/km 2, whereas the others can only sustain a long-tenn minimum population of 18 persons/km 2, 21 persons/km 2, and 37 persons/km 2, respectively. These high population levels reflect the generally greater potential for Table 5.9. Comparison of Population Density Values (POPKM) for 5MT8839, Norton House, for A.D. 901-1300 and A.D. 1029-1048. 5MT8839 (Norton) 7.88 653,070 ± 99,827 15.3 C.V.b 94 Maximum value c Maxirn.am CC: Mean value c 67 ± 10 Minimum value c 49 (73%) Critical CC: 40 Optimal CC: 60% of Mean 27 40% of Mean 20% of Mean 13 Mockingbird Mesa 17.96 801,336 ± 204,994 25.6 57 36±9 18 (50%) 22 14 7 A.D. 901-1300 (400 years) Area (kfn2) TOTPRODa(kg) A.D. 1029-1048 (20 years) TOTPROD(kg) Sand Canyon Locality 26.08 1,779,087 ± 337,524 19.0 80 55± 10 37 (67%) 33 22 11 Study Area 1,470.36 64,925,217 ± 13,936,845 21.5 57 35±7 21 (60%) 21 14 7 733,352 1,646,677 614,991 59,804,508 ± 76,714 ± 173,871 ± 287,165 ± 10,941,721 12.5 23.7 17.4 18.3 C.V. 77 (AD 1042) 47 (AD 1042) 65 (AD 1042) 44 (AD 1042) Maxiroom value Maximum CC:Mean value 33±6 64 ± 7 33 ± 7 51 ± 9 19 (58%) 37 (73%) 25 (76%) Critical CC: Minimum value 50 (78%) (AD 1036, 1048) (AD 1035) (AD 1032) (AD 1035, 1041) 20 Optimal CC: 60% of Mean 38 20 31 13 40% of Mean 26 13 20 7 20% of Mean 13 7 10 Note: All population values are truncated integers; density km 2 aTOTPROD is the total mean productivity of maize, rounded to the nearest whole number ± standard deviation. bCoefficient of variation, the ratio of the standard deviation to the mean, multiplied by 100, rounded to the nearest tenth. cMaximum, mean ± standard deviation, and minimum population density values (POPKM) for POP2YR are found in Appendix I (Norton), G (Mockingbird Mesa), F (Sand Canyon) and E (entire study area). - 152- agricultural productivity of the nearly uniform, deep, upland mesatop soils located in the northern portion of the study area. During the 20 years when 5MT8839 is thought to have been occupied, the maximum carrying capacity is, in fact, not as good as the average conditions for that catchment over the 400-year span, although the difference is small. In those 20 years, the maximum carrying capacity is 64 ± 7 persons/km2. However, the 20-year minimum value, or critical carrying capacity, is slightly higher than the longterm minimum value at 50 persons/km2. During its period of occupation, 5MT8839 still bears the same superior productivity relationship to both sUIvey localities and the study area as a whole. Site Catelunent and Prime Periods for Al:ricultural Production. The population density reconstructions for 5MT8839 were searched for periods of at least 20 years when the critical carrying capacity never fell below 50 people/km2 (Appendix I, Figure 5.10). These periods are identified in Table 5.10. Using measures of temporal predictability and spatial productivity described for 5MT8371 , the most predictable period of high productivity in the catchment is A.D. 1010-1061, whereas the most productive period of high population density is A.D. 1227-1253. Therefore, it appears that 5MT8839 was occupied in the period of highest temporal predictability. 5MT2433, Aulston Pueblo Description. 5MT2433 is reported by Kane . (1975a) and by Morris (1986a). It was located within and adjacent to the right-of-way for one of the Fairview laterals, a water pipeline associated with the Dolores River Project. Surface evidence consisted of a large sherd and lithic scatter that incorporated two concentrated areas of rubble and eight areas with dark soil and vitrified adobe. Analysis of surface ceramics indicated that the site was multiple-component with occupations dating to BM III or Pueblo I, Pueblo II, and possibly Pueblo III (Morris 1986a:4). Subsurface investigations were restricted to two 5-6-m-wide project corridors that passed thorough portions of the site. One pipeline trench exposed three pits in one area (Features 1-3), but these were not excavated further. A second pipeline trench revealed the profiles of two pit structures, a burned kiva (Pit Structure 1) and a mealing room (Pit Structure 2). These were fully excavated and provide the dates for this analysis. Of the two, only burned Pit Structure 1 provided wood samples suitable for tree-ring dating, and on the combined evidence derived from tree-rings and ceramic and architectural characteristics, a middle Pueblo II date is inferred for this portion of the site. Location. The site was located in the SEl/4 of the SEl/4 of Section 35, T38N, RI8W, N.M.P.M. on the Pleasant View 7.5-minute quadrangle map. Its UTM coordinates are 4,152,600 m N, 695,180 m E (Zone 12). The elevation of the site is 2079 m (6,820 ft) on a broad, relatively flat northeast-southwest trending ridge southeast of Sandstone Canyon and northwest of Woods Canyon. Today this site is totally within cultivated farmland. The PDSI Table 5.10. Periods of Equal or Greater Occupational Attractiveness in the Catchment of 5MT8839 (A.D. 1029-1048). Rank by Years a 2 1 6 3 tie 5 4 3 tie Rank by Minimum Pop. Densityb 4 3 (5MT8839) 6 5 tie 2 1 5 tie Beginning Date A.D. 931 1010 1068 1091 1192 1227 1255 Ending Date A.D. 971 1061 1089 1130 1214 1253 1294 Number of Minimum PopUlation Years Density Value (km2) 42 50 62 50 22 50 40 50 23 60 27 60 40 50 aBased primarily on number of years and secondarily on the minimum population density that can be sustained. bBased primarily on minimum population density and secondarily on the number of years this can be sustained. - 153- 300 >- t:fIl 200 ｾ 0 ,... VI ｾ ｾ Z ｾ ｾ 0 ｾ \00 o ...ｾＡｉＱＧ ｾＧ ＺＧｕＡｉＧＺ ＧＺ Ｇ［ ＢｾＧ ［Ｇ ＢＧＺ ｾＢＺ ｩ 900 ••ＬＡＺｭＬｾ ＬＢＺＡ ｾＬＢ \000 ..ＺＬ ＬＺ Ｌ ｾＬ ＧＨＬ Ａ［Ｌ ＬＺＢＧ［ ＧＢ［ Ｇ ＺＧ Ｇ［ＢＧ ＷＧ 1100 1200 YEARS (A.D.) Figure 5.10. Population density supportable within the 1.6-km-radius catchment of 5MT8839, Norton House, A. D. 901-1300. 1300 Table 5.11. Tree-Ring Dates from Site 5MI2433, Aulston Pueblo (TRL# A-689). TRL Cat. Number CVM-143 CVM-145 CVM-155 CVM-136 CVM-191 CVM-126 CVM-194 CVM-124 CVM-144 CVM-156 CVM-140 CVM·141 CVM·153 CVM-123 CVM-152 CVM-193 CVM-122 CVM-128 CVM-131 CVM-154 CVM-151 CVM-157 CVM-125 CVM-146 CVM-134 Field Number 42 44 54 33 39 13 57 6 43 55 37 38 52 5 51 56 3 14 23 53 50 58 7 45 29 Outside Date Symbol a 966 w 973 w 989 w 1006 +w 1006 +w 1012 w 1025 w 1026 ++v 1027 w 1029 w 1030 v 1030 w 1030 r 1031 +w 1032 w 1034 ++r 1035 +v 1035 +v 1035 ++r 1035 w 1035 rB 1040 v 1041 +v 1041 r 1042 w . tatiOns for the dates, with one assigning initial roof construction in AD. 1035 followed by remodeling shortly after A.D. 1041. The alternative interpretation places construction in the early 1040s and regards the earlier timbers as reused wood. In either case, the kiva was in use during the decade of the 1040s and, according to the author, may have been use as late as the early 1050s (Morris 1986a:29). Based on the above evidence, Aulston Pueblo is assigned an AD. 1030-1050 occupation for this study. Note: All specimens are from Pit Structure 1 (a kiva), and all are juniper. aFor explanation of outside date symbols, see Table 5.5. reconstructions used to represent soil moisture in this 7.88 kIn 2 catchment were drawn from the Mesa Verde and Ignacio series. Dating. Architectural comparisons, ceramic analysis, and tree­ring samples provide the data for assigning calendrical dates to this site. Excavation revealed two floors within the Pueblo II-style kiva, but distinct dates could not be assigned to each surface. Ceramic analysis indicated a terminal occupation dating to approximately A.D. 1050 (Morris 1986a:29). All 25 tree-ring-dated samples came from the burned roof fall of the kiva (Table 5.11). Five of these are cutting dates. The earliest cutting dates are A.D. 1030 v and 1030 r. The latest sample from the site is AD. 1042 vv. Two small clusters are discernible in the tree-ring data, one ca. AD. 1035 and one in the early 10408. Morris suggests two alternative interpre- - Site Cate!unent and Agricultural Productivity. The 400-year maximwn carrying capacity value for MT2433 is 52 ± 11 persons/km2 with a critical carrying capacity value of 32 persons/km 2 (Table 5.12). It is markedly higher than maximum and critical carrying capacity values for both the Mockingbird Mesa survey locality and for the study area as a whole (36 ± 9 persons/km2 with an 18 person/km2 minimum for Mockingbird Mesa, and 35 ± 7 ｾｲｳｯｮＯｫｭＲ with a minimum of 21 persons/km2 for the study area), but not as high as values for the Sand Canyon survey locality (55 ± 10 persons/km2 with a minimum of 37 persons/km2). During the 21-year occupational period assigned to this site, the maximum carrying capacity of the catchment drops to 48 ± 11 persons/km2 with a critical carrying capacity of 32persons/km 2. These estimates resemble those for the Sand Canyon survey locality, which is about 11 kIn to the south of the catchment for Aulston Pueblo, however, the Sand Canyon locality exhibits somewhat greater productivity and potential population densities for the same period of time (52 ± 9 persons/km2 with a critical carrying capacity of 37 persons/km2). Nevertheless, the catchment of 5MT2433 has higher density values than either the Mockingbird Mesa survey locality or the study area as a whole during this 21-year period (33 ± 7 persons/km2 with a critical carrying capacity of 19 persons/km2 for Mockingbird Mesa and 33 ± 6 persons/km2 with a critical carrying capacity of 25 persons/km2 for the study area). 5MT2433, Aulston Pueblo (ca. A.D.10301050) can also be compared to 5MT8839, 155- Table 5.12. Comparison of Population Density Values (POPKM) for 5MT2433. Aulston Pueblo, for A.D. 901-1300 and A.D. 1030-1050. . 5MT2433 (Aulston) 7.88 500,847 ± 115,085 23.0 C.V.b Maximum value 77 Maximum CC: cMean value 51 ± 12 Minimum value 32 (63%) Critical CC: OptimalCC: 60% of Mean 31 40% of Mean 20 20% of Mean 10 A.D. 901-1300 (400 years) Area (krJl2) TOTPRODa(kg) A.D. 1030-1050 (21 years) TOTPROD (kg) 462,421 ± 106,346 23.0 C.V. 64 (AD 1037, Maximum value 1042, 1049) 48 ± 11 Maximum CC:Mean value Critical CC: Minimum value 32 (67%) (AD 1032) Optimal CC: 60% of Mean 29 40% of Mean 19 20% of Mean 10 Mockingbird Mesa 17.96 801,336 ± 204,994 25.6 57 36±9 18 (50%) 22 14 7 Sand Canyon Locality 26.08 1,779,087 ± 337,524 19.0 80 55 ± 10 37 (67%) 33 22 11 735,616 ± 165,411 22.5 47 (AD 1042) 1,667,214 ± 295,324 17.7 65 (AD 1037, 1042, 1049) 52±9 37 (71%) (AD 1032) 31 21 10 33± 7 19 (58%) (AD 1035) 20 13 7 Study Area 1,470.36 64,925,217 ± 13,936,845 21.5 57 35± 7 21 (60%) 21 14 7 60,277,686 ± 10,888,254 18.1 44 (AD 1042) 33±6 25 (76%) (AD 1035,1041) 20 13 7 Note: All population values are truncated integers. aTOTPROD is the total mean productivity of maize, rounded to the nearest whole number. bCoefficient of variation, the ratio of the standard deviation to the mean multiplied by 100, rounded to the nearest tenth. cMaximum, mean ± standard deviation, and minimum population density values (POPKM) for (POP2YR) are found in Appendix J (Aulston Pueblo), G (Mockingbird Mesa), F (Sand Canyon) and E (entire study area). Norton House, with which is likely contemporary (ca. A.D. 1029-1048). Aulston is located approximately 9.4 km (5.9 mi) southsouthwest of Norton House. Norton is only slightly higher than Aulston (14.6 m or 48 ft), but it is located in a larger area of deep mesatop soils than is Aulston. Both exhibit the same trend of diminished agricultural productivity and lower levels of human carrying capacity during the occupational episode. Norton, however, notably outperforms Aulston and can support a greater population density. Whereas Aulston can support a maximum carrying capacity of 48 ± 11 person/km2 and a critical carrying capacity of 32 persons/km2 during the occupational period, Norton could support a maximum carrying capacity of 64 ± 7 persons/krn2 with a critical carrying capacity of 50 persons/km 2. All other things being equal, - Norton would have been a more desirable location than Aulston. Aulston, however, was still much superior to other locations in the study area, and its catchment would have been an attractive location for farming. Site Catchment and Prime Periods for Agricultural Production. It is interesting to note that the 21 years assigned to Aulston's occupation, A.D. ＬＰＵｏＱｾＳ include a year that is as low as any year in the 4OO-year record. Thus, the minimum value for the 21-year occupational episode is as low as that for the 400-year long-term record. Further, given that the period of occupation is not as good as the prevailing long-term conditions, identifying time periods that were equally good or better could not be done. Had it been done, anytime within the 400 years would be equally as good. 156- Instead, I reduced the length of time for my search to 16 years beginning at AD. 1035, the time of the first date cluster, and ending at AD. 1050. In this 16­year period the critical c::?ing capacity is lifted by two people per km , resulting in a sustainable population minimum of 34 persons/km2 rather than 32 persons/km 2. Using 34 persons/km2 as the minimum population size and 16 years as the minimum duration period, nine other time periods of equal or greater potential for occupation can be identified within the 400­year series (Appendix J, Figure 5.11). Table 5.13 summarizes these data for 7.88 krn 2 catchment of 5MT2433. A.D. 1091-1122 is the time of the highest productivity and A.D. 1091-1130 ties with A.D. 1255-1294 as the time of highest temporal predictability within the 400 years considered. It would seem from the ranking that the component dated as A.D.1030-1050 at 5MT2433 existed in the catchment in an interval that was relatively predictable and productive, but not at either extreme. It would be of interest to know if the other site components at 5MT2433 that were not dated by absolute methods do in fact date to these prime periods for agricultural production. 5MT3834, Mustoe Ruin Description. This site is reported by Gould (1982) and its dating is discussed by Gould and later by Ahlstrom (1985:499-501). It is a multiple-component habitation site with an earlier Pueblo II component dating to sometime between A.D. 900-1050 and a later Pueblo III component dating to the AD. Ｑ ＷｾＱＲＳＱ period (Gould 1982:342-345). Excavations recovered evidence of the earlier component in the fonn of three surface rooms, a kiva, a stockade partially enclosing the settlement, an extramural cist, and a trash midden. The later component was represented by a roomblock of at least five rooms, a kiva, a tower, and a trash midden. The later kiva was burned and provided all the samples used for tree-ring dating. Consequently, this is the component used for this study. Location. The site was located within the Sand Canyon survey locality, approximately 1 km (0.6 mi) southwest of Goodman Point Ruin (5MT604) and 4 km (2.5 mi) east of Sand Canyon Pueblo (5MT765). It is located in the NW 1/4 of the SWI/4 of Section 4, T36N, R17W, N.M.P.M. on the Arriola 7.5-minute quadrangle map. Its UTM coordinates are 4,142,050 m N and 700,550 m E (Zone 12). It is situated on relatively level, deep mesatop soils on the highland feature known as Goodman Point, west of Goodman Canyon, at an elevation of 2,054 m (6,740 ft). The PDSI reconstructions used to model soil moisture conditions in this 7.88-km 2 catchment are from the Ignacio and Mesa Verde series. DatinK. The assignment of calendrical dates to 5MT3834 derives primarily from Table 5.13. Periods of Equal or Greater Occupational Attractiveness in the Catchment of 5MT2433 (A.D. 1030-1050). Rank by Years a 5 7 6 1 tie 2 9 8 4 3 1 tie ｾｂ｡ｳ･､ Rank by Minimum Pop. Densityb 3 4 7 (5MT2433) 5 tie 1 9 8 6 2 5 tie Beginning Date A.D. 931 954 1033 1091 1091 1134 1157 1192 1228 1255 Ending Date A.D. 952 971 1054 1130 1122 1149 1174 1214 1253 1294 Number of Minimum PopUlation Years Density Value (km2) 22 39 18 36 22 34 40 34 32 40 34 16 18 34 23 34 26 40 40 34 ｰｲｾｭ｡ ｬｹ on ｲ･｢ｾｮ of years ､ｮｾ seco.ndarily on the ｭｵｩｮｾ population density that can be sustained. Based prlmanly on minimum population density and secondarily on the number of years this can be sustained. - 157- 190 180 170 160 ISO 140 ｾ f::en ｾ 130 120 110 0 100 ｾ ­ til 00 E-< 90 j 80 ::J I:l.. 0 70 I:l.. 60 SO 40 30 20 10 0 I I 900 1000 1100 1200 YEARS (A.D.) Figure 5.11. Population density supportable within the 1.6-km-radius catchment of 5MT2433, Aulston Pueblo, A.D. 901-1300. 1300 tree­ring dating and to a lesser degree on' . artifactual associations. Twenty­nine wood samples were dated (Table 5.14). Nine of the 29 were cutting dates. The earliest cutting date is A.D. 1065 r and is apparently a reused timber from an earlier structure. The next seven cutting dates cluster from A.D. 1173 to 1175 and are interpreted as representing the initial construction of the kiva (Ahlstrom 1985). A second cluster ca. A.D. 1227­1231 ends with the final cutting date and tenninal date for the site. Ahlstrom speculates that this was likely a repair episode due to its weaker representation in the sample (Ahlstrom 1985:499). No evidence of occupational discontinuity occurs in the A.D. 1173­1231 period. Thus, the full 59 years bracketing the Table 5.14. Tree-Ring Dates from Site 5MT3834, Mustoe Ruin (TRL # A-452). TRL Cat. Number CVM-58 CVM-68 CVM-49 CVM-45 CVM-48 CVM-55 CVM-23 CVM-36 CVM-21 CVM·51 CVM·46 CVM-30 CVM-27 CVM-43 CVM-40 CVM-67 CVM-24 CVM-26 CVM-63 CVM-28 CVM-29 CVM-32 CVM-66 CVM-25 CVM·50 CVM-54 CVM-60 CVM-33 CVM-62 Field Number 43 56 34 30 33 40 5 20 2 36 31 13 9 28 24 55 6 8 50 10 11 16 54 7 35 39 46 17 49 Outside Date a Symbol 1016 w 1027 w 1057 +w 1059 +w 1060 ++w 1065 r 1066 +B 1086 +w 1090 ++w 1093 +w 1139 ++B 1150 +v 1160 ++v 1173 r 1174 v 1174 v 1174 r 1174 r 1174 rB 1174 H 1174H 1174 H 1174 HB 1175 r 1229 +v 1229 HB 1230 w 1231 w 1231 rB . A.D. 1173­1231 period is used to date this site. with construction episodes inferred from tree­ring clusters at A.D. 1175 and 1231. Site Catchment and Agricultural Productivity. 5MT3834 is within the Sand Canyon survey locality. The long-tenn population density values for the catchment of Mustoe are, not unexpectedly, most like those for the Sand Canyon survey locality. The 400-year maximum carrying capacity value for 5MT3834 is 56 ± 11 persons/km2 with a critical carrying capacity of 39 persons/km 2, whereas the 400year maximum carrying capacity for the Sand Canyon survey locality is 55 ± 10 persons/km2 with a critical carrying capacity 37 person/km2 (Table 5.15). Both 5MT3834 and the Sand Canyon survey locality are noticeably more productive than either the Mockingbird Mesa survey locality or the study area as a whole (36 ± 9 persons/km2 with a minimum value of 18 persons/km2 for Mockingbird Mesa survey locality and 35 ± 7 persons/km2 with a minimum of 21 persons/km2 for the study area). Given the long period of occupation assigned to this site, it is not surprising that the maximum and critical carrying capacity values for the 59-year occupation period at 5MT3834 are the same as those for the 400 years. Although there is a one person/km2 increase in the maximum carrying capacity values for both survey localities and the study area for the A.D. 1173-1231 period, the critical carrying capacity values for these three areas remain the same as for the 4OO-year period. Site Catchment and Prime Periods for A2ricultural Production. Noting that population density values are not different for the occupation period and that this site is in a portion of the study area that has already been cited as being notably productive, a different method was used to identify those time intervals of high productivity. Figure 5.12 was visually examined for intervals of sustained higher population density. Data contained in Appendix K were consulted and Table 5.16 summarizes the results. The period that exhibits the greatest temporal predictability is A.D. 1091-1122, and the period with the highest productivity is A.D. 1228-1257. The occupation dates assigned to 5MT3834 only partially overlap with Note: All specimens are from the kiva, and all are juniper. aFor explanation of outside date symbols, see Table 5.5. - 159- Table 5.15. Comparison of Population Values (POPKM) for 5MT3834, Mustoe Ruin, for A.D. 901-1300 and A.D. 1173-1231. 5MT3834 (Mustoe) 7.88 553,196 ± 108,069 C.V.b 19.5 c 80 Maximum value Maximum CC: Mean value c 56 ±11 Minimum value c 39 (70%) Critical CC: 34 OptimalCC: 60% of Mean 40% of Mean 22 11 20% of Mean Mockingbird Mesa 17.96 801,336 ± 204,994 25.6 57 36±9 18 (50%) 22 14 7 A.D. 901-1300 (400 years) Area (kfn2) TOTPRODa(kg) Sand Canyon Locality 26.08 1,779,087 ± 337,524 19.0 80 55± 10 37 (67%) 33 22 11 Study Area 1,470.36 64,925,217 ± 13,936,845 21.5 57 35±7 21 (60%) 21 14 7 A.D. 1173-1231 (59 years) POP2YR TOTPROD (kg) C.V. Maximum value Maximum CC: Mean value Critical CC: Minimum value 567,693 1,823,532 817,051 66,314,895 ± 112,686 ± 195,801 ± 339,201 ± 13,901,244 19.9 24.0 18.6 21.0 80 (AD 1184) 57 (AD 1184) 80 (AD 1184) 57 (AD 1184) 37±8 57±10 36±7 59 ± 11 39 (66%) (AD 18 (49%) (AD 37 (65%) (AD 21 (58%) (AD 1175, 1186, 1186,1227) 1175,1186, 1186,1227) 1191,1215, 1191,1215, 1227) 1227) 34 35 22 Optimal CC: 60% of Mean 22 40% of Mean 24 15 23 14 7 11 20% of Mean 12 7 Note: All population values are truncated integers. aTOTPROD is the total mean productivity of maize, rounded to the nearest whole number. bCoefficient of variation, the ratio of the standard deviation to the mean multiplied by 100, rounded to the nearest tenth. cMaximum. mean ± standard deviation. and minimum population density values (POPKM) FOR POP2YR are found in Appendix K (Mustoe Ruin), G (Mockingbird Mesa), F (Sand Canyon) and E (entire study area). the last period of high population density. However, the nonnal minimum population density that can be supported in the catchment area is already so high at 39 persons/km 2 that this may not have been an effective issue here for potential farmers and inhabitants. Table 5.16. Periods of Greatest Occupational Attractiveness in the Catchment of 5MT3834 (A.D. 11731231). Rank by Rank by Minimum Beginning Ending Number of Minimum Population Yearsa Pop. Densityb Date A.D. Date A.D. Years Density Value (km 2) 952 22 46 3 2 931 1 3 1091 1122 32 45 2 1 (5MT3834) 1228 1257 30 47 aBased primarily on number of years and secondarily on the minimum population density that can be sustained. bBased primarily on minimum population density and secondarily on the number of years that this condition can be sustained. ­ 160- 200 190 180 170 160 ISO >- l::CI.l 130 ffi 0 12:0 Z 9ｾ 110 100 j 90 ｾ 10 0 70 0'1 ｾ ｾ 140 60 SO 40 30 2:0 10 0 I , 900 1000 1100 1200 YEARS (A.D.) Figure 5.12. Population density supportable within the 1.6-km-radius catchment of 5MT3834, Mustoe Site, A.D. 901-1300. 1300 Central Place Sites 5MT6970, Wallace Ruin and 5MT4126, Ida Jean Ruin. . masonry over the kiva walls and bench, and built a Chaco style sub-floor ventilator shaft under the original floor. Kiva B produced 24 cutting dates assigned to the year A.D. 1124. Description. The Wallace Ruin is described by Bradley (1974, 1984, 1988b) and Powers et all (1983). The Ida Jean Ruin. sometimes referred to as North McElmo #8, is described by Brisbin and Brisbin (1973) and Powers et al, (1983). These sites apparently are contemporary, are only 0.8 kIn (0.5 miles) apart, and, along with the nearby Haney Ruin, are considered to be Chacoan outliers of the Lake View Group in the Mesa Verde Area (powers et a1.1983: 164-167). Location. 5MT6970. Wallace Ruin. is located some 7 kIn (4.3 mi) northeast of the modem town of Cortez. It is located within the NW 1/4 of the SE1/4 of Section 16, T36N. R15W, N.M.P.M. on the Dolores West 7.5minute quadrangle map. Its UTM coordinates are 4139660 m N and 721000 mE (Zone 12). It is situated on a valley bottom immediately west of Simon Draw, an intermittent stream flowing into McElmo Creek, at an elevation of 1,896 m (6,220 ft). Site 5MT6970, Wallace Ruin, is a multistoried, V-shaped pueblo with an enclosed plaza. It is fronted with gate-like, formalized trash mounds and a road segment south of the plaza wall. Although the entire site has not been excavated, it is estimated to have almost 100 rooms, some of them built in three stories, and five kivas. Four separate building episodes have been identified by its excavator on the basis of architectural characteristics, with a continuous occupation from approximately A.D. 1045-1125 (incorporating the first three phases), followed by a hiatus, and then a final reuse and modification of the structure in the middle to late thirteenth century (Bradley 1988b:29). 5MT4126, Ida Jean Ruin, is 40 m north and 810 m west of Wallace Ruin at VTM coordinates 4,139,700 m N and 720,190 m E (Zone 12). It is located within the SW1I4 of the NW1I4 of Section 16, T36N, R15W, N.M.P.M. on the Dolores West 7.5-minute quadrangle map. It is located on a ridge top west of Simon Draw at an elevation of 1,917 m (6,290 ft). Site 5Mf4126, Ida Jean Ruin, is less wellknown than the Wallace Ruin and has been extensively damaged. It is a two-storied pueblo with approximately 55 rooms built as a Vshaped or E-shaped roomblock with two enclosed kivas and a detached great kiva. Two kivas were adjacent in an upraised plaza enclosed by a plaza wall. The western of these two kivas, Kiva A, is described by Brisbin and Brisbin (1973) who salvaged information from this site. It originally had been connected to the eastern kiva, Kiva B, by a tunnel. They note that it had been originally built as a Mesa Verde-style, keyhole-shaped kiva with six pilasters, a southern recess, and a floor level ventilator system. Later, renovations done in the Chaco style had replaced the Mesa Verde-style pillar pilasters with Chaco-style low pier pilasters, filled and sealed off the southern recess, added a nine-foot-tall veneer of Chaco banded For this study, a location positioned halfway between these two sites was chosen to represent the centerpoint of a single catchment since they are believed to be contemporary and components of the same cultural complex operating out of this location. This created a catchment that was only 4.32 kIn 2 in area since the two sites are adjacent to the eastern margin of the study area. The name and number for the better known 5MT6970, the Wallace Ruin, was retained, and the name and number for the lesser known 5MT4126, Ida Jean Ruin, was dropped. PDSI reconstructions used to represent soil moisture conditions in the catchment were drawn from the Cortez series. Datin2. At the present time, a total of 44 wood samples provenienced to seven separate features is reported for 5MT6970 (Table 5.17a). Feature 25, a three-story room associated with Building Period 1 (A.D. 1000-1075), produced 15 datable samples. Three of the 15 were cutting dates, the earliest being A.D.1037 v, the next AD. 1045 rB, and the last AD. 1052 r. The terminal tree-ring date was A.D. 1071 +vv. Examination of the set of dates from this feature reveals a strong cluster of 11 samples at ca. AD. 1045, which likely 162 - Table 5.17a. Tree-Ring Dates from Site 5MT6970, Wallace Ruin (TRL# A-645). Struct., Feature Feat. 5 Feat. 5 Feat. 6 Feat. 6 Feat. 6 Feat. 6 Feat. 6 Feat. 6 Feat. 7 Feat. 7 Feat. 7 Feat. 7 "rRL Cat. Numb. WLR-13 WLR-14 WLR-11 WLR-4 WLR-5 WLR-10 WLR-9 WLR-6 WLR-19 WLR-18 WLR-31 WLR-34 Feat. 7 WLR-38 Feat. 7 WLR-36 Feat. 7 WLR-35 Feat. 7 WLR-32 Feat. 8 Feat. 8 Feat. 8 Feat. 8 Feat. 8 Feat. 8 Feat. 9 Kiva 3 Kiva 3 Kiva 3 Feat. 25 Feat. 25 Feat. 25 Feat. 25 Feat. 25 Feat. 25 Feat. 25 Feat. 25 Feat. 25 Feat. 25 Feat. 25 Feat. 25 Feat. 25 Feat. 25 Feat. 25 Fill, hearth 4 unknown unknown WLR-22 WLR-28 WLR-25 WLR-27 WLR·23 WLR-20 WLR-29 WLR-17 WLR-16 WLR-30 WLR-61 WLR-62 WLR-51 WLR-52 WLR-53 WLR-54 WLR-55 WLR-56 WLR-58 WLR-59 WLR·60 WLR-64 WLR-67 WLR-63 WLR-65 WLR-68 Field Numb. 27 30 20 27 36 39 33 29 51 50 primary secondary secondary secondary secondary secondary 138 139 134 129 137 132 37 B6 B2 112 113 95 96 97 98 101 103 110 111 109 117 116 119 27175 WLR·48 W2422 WLR-47 W2421 Tree Spec. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Out. Date Symbola 1064 vv 1101 vv 995 vv 1060 ++vv 1073 vv 1073 vv 1074 +vv 1096 vv 1017 vv 1061 vv 1073 vv 1037 vv Jun. 1059 vv Jun. 1061 vv Jun. 1071 +vv Jun. 1084 vv Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Pond. Pond. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. 1048 +vv 1055 vv 1076 vv 1077 vv 1080 vv 1084 vv 1040 +vv 1040 vv 1054 vv 1062 vv 1035 H 1037 v 1045 HB 1045 HB 1045 H 1045 rB 1045 H 1045 HB 1045 HB 1045 HB 1045 HB 1045 +rB 1045 HB 1052 r 1071 +vv 1039 ++B Jun. 988 vv Jun. 1021 vv constrl!ction ｯｾ the room represents the ｩｾｴ｡ｬ and a major bUlldmg event m the ｹｲｯｴｾｩｨ of the multistoried pueblo. No dated tree­nng samples are known to be associated with Build­ ing Period 2 (A.D. 1075­1100). Features (Rooms) 5. 6. 7. 8. 9. and Kiva 3 are associated with Building Period 3 (A.D. 1075­1125) on architectural evidence. but none of these samples provided a ｣ｵｴｩｮｾ ｾ｡ｴ･Ｎ N? tree­ring dates are available for BUlldmg Penod 4 (middle to late 1200s). Thirty­two wood samples taken from three different kivas in Ida Jean Ruin proved datable (Table 5.17b). Two noncutting date samples Table 5.17b. Tree-Ring Dates from Site 5MT4126, Ida Jean Ruin (North McElmo #8, TRL# A-384). ­ Struct., TRLCat. Field Tree Out. Date Feature Numb. Numb. Spec. Symbol a Kiva A MVC-1902 39 Jun. 1070 +vv Kiva A MVC-1904 41 Jun. 1078 ++vv KivaB MVC-1876 13 Pond. 1118 vv Kiva B MVC-1889 26 Pond. 1120vv KivaB MVC-1892 29 Pond. 1122 H KivaB MVC-1891 28 Pond. 1123 +v KivaB MVC-1894 31 Pond. 1123 +r Kiva B MVC-1870 7 Pond. 1124 v KivaB MVC-1866 3 Pond. 1124 r KivaB MVC-1871 Pond. 1124 r 8 KivaB MVC-1874 11 Pond. 1124 r KivaB MVC-1875 12 Pond. 1124 r Kiva B MVC-1877 14 Pond. 1124 r KivaB MVC-1878 15 Pond. 1124r KivaB MVC-1879 16 Pond. 1124r Kiva B MVC-1880 17 Pond. 1124r KivaB MVC-1881 18 Pond. 1124 r Kiva B MVC-1882 19 Pond. 1124 r KivaB MVC-1884 21 Pond. 1124 r KIvaB MVC-1885 22 Pond. 1124 r Kiva B MVC-1886 23 Pond. 1124 r KivaB MVC-1887 24 Pond. 1124r KivaB MVC-1890 27 Pond. 1124 r KivaB MVC-1893 30 Pond. 1124 r Kiva B MVC-1895 32 Pond. 1124 r KivaB MVC-1897 34 Pond. 1124 r KivaB MVC-1898 35 Pond. 1124 r KivaB MVC-1899 36 Pond. 1124r KivaB MVC-1900 37 Pond. 1124 r KivaB MVC-1868 5 Pond. 1124c Kiva B MVC-1873 10 Pond. 1124 c KivaC MVC-1901 38 Jun. 998 vv aFor explanation of outside date symbols, see Table 5.5. 163- Table 5.18. Comparison of Population Values (POPKM) for 5MT6970, Wallace Ruin, for A.D. 901-1300 and A.D. 1045-1125. A.D. 901-1300 400 years Area(kJn2) TOTPRODa (kg) C.V.b Maximum value MaximumCC:MeanvahJe Critical CC: Min. value Optimal CC:60%C of Mean 40% of Mean 20% of Mean ｾ A.D. 1045-1125 81 years TOTPROD (kg) C.V. Maximum value Maximum CC: Mean value Critical CC: Min. value Optimal CC: 60% of Mean 40% of Mean 20% of Mean 5MT6970 (Wallace) 4.32 297,798 ± 53,732 18.0 78 56±10 37 (66%) 34 22 11 Mockingbird Mesa 17.96 801,336 ± 204,994 25.6 57 36±9 18 (50%) 22 14 7 Sand Canyon Locality 26.08 1,779,087 ± 337,524 19.0 80 55±10 37 (67%) 33 22 11 5MT1566 (Lowry) 7.88 409,875 ± 102,360 25.0 68 42±10 26 (62%) 25 17 8 305,680 ± 52,243 17.1 78 (A.D. 1065) 58±9 37 (64%) (A.D. 1067, 1090) 828,844 ± 202,383 24.4 57 (A.D. 1065) 1,831,160 ± 369,007 20.1 79 (A.D. 1116) 433,100 37±9 18 (49%) (A.D. 1067, 1090) 35 23 12 22 15 57 ± 11 37 (65%) (A.D. 1062, 1067, 1081, 1082, 1090) 34 23 11 7 ± 112,453 5MT2149 (Escalante) 6.72 291,075 ± 53,857 18.5 58 34±6 24 (70%) 20 14 7 295,823 ± 60,601 26.0 68 (AD 1065, 1116) 45± 11 26 (58%) (A.D. 1055, 1062, 1067, 1068, 1081, 1082, 1090) 27 18 20.5 58 (AD 1079, 1087, 1116) 36± 37 24 (67%) (A.D. 1048, 1053, 1055, 1062, 1067, 1068, 1081, 1082, 1092, 1104) 22 14 9 7 f'Jofe: Air population values are fruncated integers. atotal mean productivny of maize, rounded to the nearest whole number. bCoefficient of Variation, the ratio of the standard deviation to the mean multiplied by 100, rounded to the nearest tenth. cpercent of mean value rounded to the nearest whole number. Study Area 1470.36 64,925,217 ± 13,936,845 21.5 57 35±7 21 (60%) 21 14 7 67,377,205 ± 14,683,637 21.8 56 (A.D. 1065) 37±8 21 (57%) (A.D.1067, 1090) 22 15 7 came from Kiva A, and one non­cutting date sample came from Kiva C. Neither structure has been described in print. The remaining 29 samples came from Kiva B and 24 of these were cutting dates, all dating to A.D. 1124. Ahlstrom (1985:499­500) describes this as a "strong terminal cluster" and states that it likely represents a major construction event at the site. This large cluster provides strong evidence of kiva roof building activities in A.D. 1124 or 1125. Thus, the Wallace Ruin contributes the initial occupation date of A.D. 1045 and the Ida Jean Ruin provides the terminal date of A.D. 1124. Architectural and artifactual data provide strong support for the initial construction and use of the Wallace Ruin and Lake View Group in general. In this study, the period of A.D. 1045­1125 is used to represent the major occupational period of the combined Wallace­Ida Jean "central place" location. ｾＮ Site Catchment and Agricultural ProducTable 5.18 summarizes population values for 5MT6970. The 400­year maximum carrying capacity is 56 ± 10 persons/km2 with a critical carrying capacity of 37 persons/km 2 . These very high density values are almost identical to the 400­year values for the Sand Canyon survey locality, which has a maximum carrying capacity of 55 ± 10 persons/km2 and a critical carrying capacity of 37 persons/km2. They are both markedly higher than those 400­year values for the Mockingbird Mesa survey locality and for the study area (maximum carrying capacity of 36 ± 9 persons/km2 with a critical carrying capacity value of 18 persons/km2, and a maximum carrying capac. ity of 35 ± 7 persons/km2 with a critical carrying capacity of 21 persons/km2, respectively). Potential mean population values for the 81­year occupation period at 5MT6970 are even higher than the long­term values. The maximum carrying capacity for the A.D. 1045­1125 :r;riOd in this catchment is 58 ±9 persons/km with a critical carrying capacity of 37 persons/km 2. The Sand Canyon survey locality mean values for this period are also a little higher with a maximum carrying capacity of 57 ± 11 persons/km 2 and a critical carrying capacity value of 37 persons/km2. Again, both are much higher than those for the 81­year period for the Mockingbird Mesa survey locality and for the study area (maximum carrying capacity of 37 ± 9 persons/km2 with a critical carrying capacity of 18 persons/km2, and a maximum carrying capacity of 37 ± 8 persons/km2 with a critical carrying capacity of 21persons/km2), but they, too, are slightly higher than the long­term conditions for these places (Figure 5.13). Site Catchment and Prime Periods for Agricultural Production. As noted for several other sites, the minimum population density that can be sustained over the 81­year period between A.D. 1045­1125 is 37 persons/km 2, the same critical carrying capacity value characteristic of the long­term period. Inspection of values contained in Appendix L and Figure 5.13 revealed that there are, however, a number shorter time intervals that support noticeably higher pgpulation densities than 37 persons/km 2. Table 5.19 summaries the results. Table 5.19. Periods of Greatest OCcupational Attractiveness in the Catchment of 5MT6970 (A.D. 1045-1125). Ranka by Yrs. 4 3 6 1 2 5 Rankbby Min. Pop. Density 3 6 5 (5MT6970) 2 (5MT6970) 1 4 Beginning Date A.D. 931 982 1042 1091 1228 1281 Ending Date A.D. 952 1004 1061 1122 1250 1300 Number of Years 22 23 20 31 23 20 Minimum Population Density (km2) 45 41 43 45 46 44 aBased primarily on number of years and secondarily on the minimum population density that can be ｾｵｳｴ｡ｩｮ･､Ｎ Based primarily on minimum population density and secondarily on the number of years that this condition can be sustained. ­165 - 200 190 180 110 160 ISO 140 ｾ t:: 130 CI) ｾ 120 0 ­ 0\ 0\ ｾ 110 100 90 ::J c.. 0 c.. 80 10 60 SO 40 30 20 10 0 , , 900 1000 1100 1200 YEARS (A.D.) Figure 5.13. Population density supportable within the 1.6-km-radius catchment of 5MT6970, Wallace Ruin, A.D. 901-1300. 1300 A.D. 1091­1122 is the period of highest temporal predictability at a minimum population density, while A.D. 1228-1250 is the period of greatest production that is sustained for a considerable interval of time. It would appear that the main occupation and use of the combined Wallace - Ida Jean Ruin occurs during the times of highest predictability and noticeably high productivity. 5MT1566, Lowry Ruin Description. 5MT1566, Lowry Ruin, is a multiple component, multi-story habitation often assigned a role in the Chacoan presence in the Mesa Verde area, although earlier Basketmaker and later, non-Chacoan occupation is present as well. Paul Martin excavated major portions of this site in the 1930s and provides the major report on this site (Martin 1936). Stabilization work. in the 1970s and reanalysis of wood samples in the 1980s has resulted in additional information on Lowry Ruin (Ahlstrom 1985; White and Bretemitz 1976). Excavation revealed a rectangular structure of 40 or more rooms, some up to three stories tall, eight kivas, and a detached great kiva 61 m (200 ft) southeast of the major roomblock. Portions of a road segment have also been identified at the site (powers et al. 1983: 170). Location. 5MT1566 is located in the SWl/4 of the NWl/4 of Section 2, T38N, RI9W, N.M.P.M. on the Ruin Canyon 7.5minute quadrangle map. Its UTM coordinates are 4,161,640 m N and 683,620 m E (Zone 12). It is situated on a high point of land on a mesa top approximately 1 km (.6 mi) west of Cow Canyon at an elevation of 2,049 m (6,720 ft) and southwest of the modem farming community of Cahone. The PDSI reconstructions used to model soil moisture conditions in this 7.88-km 2 catchment are drawn from the Ignacio series. Dating. Contrasting architectural styles and stratigraphic relationships indicated to Martin that Lowry was not a single pre-planned unit built at one point in time. Rather, he identified at least six construction episodes (Martin 1936:37, 194) associated with the masonry pueblo, some occurring after occupational hiatuses. Based on what tree-ring dates he had - av·ailable to. him for portions of the pueblo, . Martin assigned an A.D. 1086-1106+ date for the site (Martin 1936:204). Reassessment of the 53 wood samples from the entire site (Table 5.20) permitted Ahlstrom to identify 33 cutting dates associated with the first three building stages (Ahlstrom 1985:339). He concluded that the first episode associated with the original nucleus of the building (Rooms 10, 15, 19,21, and the great kiva) occurred at ca A.D. 1085-1090, the second episode (Rooms 8, 11,26, Kiva B, and Great Kiva) between A.D. 1106-1120, and the third episode (Kiva A) at ca. A.D. 1120 (Ahlstrom 1985:340). Thus, for the purpose of this study, a main occupation period from A.D. 1086-1120 is assigned to 5MT1566. The first cutting date for the first building episode associated with the construction of the Great Kiva is A.D. 1086. The final cutting date associated with the third period and the remodeling of Kiva A is A.D. 1120. Site Catchment and Agricultural Produc- tiY.i.tt. Table 5.21 summarizes population values for 5MT1566. The 400-year maximum for 5MT1566 is 42 ± 10 carrying ｣｡ｾｩｴｹ persons/km with a critical carrying capacity of 26 persons/km 2. These values are not as high as those for the Sand Canyon survey locality (maximum carrying capacity of 55 ± 10 persons/km 2 and 37 person/km2), but they are higher than for both the Mockingbird Mesa survey locality and the study area as a whole (maximum carrying capacity of 36 ± 9 persons/km2 with critical carrying capacity of 18 persons/km2, and maximum carrying capacity of 35 ± 7 persons/km2 with a critical carrying capacity of 21 persons/km2, respectively). The 35-year occupation period for 5MT1566 has a higher maximum carrying capacity (46 ± 10 persons/km2), but the same critical carrying capaci!y as for the 4OO-year period (26 persons/km 2)for this catchment. This is true for the other three areas during this period, as well. The maximum carrying capacity lifts to 58 ± 10 persons/km2 for the Sand Canyon survey locality, to 38 ± 8 persons/km2 for the Mockingbird Mesa survey locality and to 38 ± 7 persons/km2 for the study area 167- Table 5.20. Tree-Ring Dates from Site 5MT1566, Lowry Ruin (TRL# A-323) TRL Cat. Number Great Kiva GP-586 a,b GP:580-2 Great Kiva GP-588 Great Kiva Great Kiva GP-569 (567, 568,570,571 ) GP-574 (575) Great Kiva GP-587 Great Kiva Great Kiva GP-580 Great Kiva GP-580-1 a,b Great Kiva GP-584 a,b,c Great Kiva GP-577 Great Kiva GP-576 (578,585) Room 10 GP-566 (565) LOW-49 Room 15 Room 15 LOW-21a Room 15 LOW-25 FML 33-2 Room 19 FML 33-10 Room 19 FML 33-15 Room 21 LOW-41 Kiva A LOW-37 Kiva A Kiva A LOW-38 LOW-39 Kiva A Kiva A LOW·40 Kiva A LOW·42 LOW-36 Kiva A FML 33-3 Kiva A LOW-45 KivaB LOW-47 Kiva B LOW-46 Kiva B LOW-32 Kiva 1 GP-559 a,b Kiva 1 LOW-31 Kiva 1 GP-562 Kiva 1 GP-560 Kiva 1 LOW-27 Kiva II orB LOW-26 Kiva II or B LOW-3 Room 8 GP-563 Room 8 LOW-30 Room 8 LOW-28 Room 8 LOW-22 Room 11 LOW-29 Room 11 GP-591 Room 25 FML-33-14 Room 26 FML-33-1 Room 27 FML-33-20 Room 27 FML-33-21 Room 27 GP-582-1 a,b Room 35 Room on east GP-561 side of ruin Field Num. 26 21 6 2 3 4 5 7 1 10 12 11 Tree Outside Species Date/Symbol Jun. 910 w a Jun. 932 ++w Jun. 1043 ++B 1086 rB Jun. Jun. Jun. Pili. Pond. Jun. Jun. Jun. 1086 +v 1086 v 1106 v 1106v 1109 ++v 1119 w 1172 r Pond. Pond. Pond. Pitt Jun. Jun. Pitt Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Pond. Pond. Jun. Jun. Jun. Jun. Jun. Jun. Pin. Pin. Jun. Jun. Pin. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. 968w 988 +w 989w 1088 v 1085 HL 1090 +r 1089 r 985+w 1027 +w 1040 ++w 1053 +w 1066 vv 1098 +B 1108 ++v 1120 rL 1105 vv 1105 vv 1109 +v 988 vv 991 vv 1105 vv 1105 vv 1108 rB 1106 v 1106 r 1016 vv 1055 ++vv 1106 r 1110 r 839 vv 1120 vv 990 vv 1106 cL 1067 w 1103 rL 1103 rL 1097 vv 1011 +r - 168- Site Catchment and Prime Periods for Agricultural Production. The minimum population value for the 35­year period is equivalent to the minimum value for the entire 400­year series for that catchment, 26 persons/km2. Thus, the value is useless for purposes of identifying periods of equal or greater occupational attractiveness. If instead a slightly shorter period of occupation is identified, A.D. 1091­1120, representing 30 years of continuous occupation, then a minimum value of 32 person/km2 can be used to search for like sequences. Inspection of the data contained in Appendix M and Figure 5.14 for 5MT1566 reveals that no other consecutive sequence of years in the 400­year time series produces a minimum of 30 years with a minimum population value as high as 32 persons/km2. The two closest time periods are A.D. 12281253 for 26 years when the minimum population value is never lower than 32 persons/km2 and A.D. 931­952 for 22 years when the minimum population value is 31 persons/km2 (fable 5.22). Therefore, it would seem that interval used to date the main occupation of Lowry Ruin contains within it the period of highest temporal predictability and productivity for the entire 400 years under consideration for this catchment. 5Mf2149, Escalante Ruin Description. This site is most thoroughly described by Hallasi (1979) and is Table 5.20 (Concluded). Struet .• Feature No Provenience No Provenience No Provenience No Provenience TRL Cat. Number FML-33-24 Field Num. Tree Species Jun. Outside Date/Symbol 925 +v FML-33-9 Jun. FML-33-13 Jun. 1084 w LOW-34 Pirt. 1089 w 946+w Note: For explanation of outside date symbols, see Table 5.5. also discussed by Powers et al. (1983) and White and Bretemitz (1979). Escalante is a multiple-component, roughly square, single story pueblo of some 25 rooms enclosing a Chaco-style kiva. A second, smaller, Mesa Verde-style kiva is located immediately south of the structure. Construction and/or use of this site dates to two periods, an earlier Chacoan period and a later Mesa Verdean period, with at least three occupational surfaces being recognized. Location. 5MT2149 is located at the apex of a prominent hill on the west side of the Dolores River and its floodplain, at a place where the west-flowing river takes a great bend to the north. The location is the highest point in the area at 2,195 m (7,200 ft) and takes in a particularly panoramic view of the Mesa Verde region. It is located in the NW1/4 of the SE1/4 of Section 7, T37N R15W, N.M.P.M. on the Dolores West 7.5-minute quadrangle map. Its UfM coordinates are 4,150,500 m N and 717,080 m E (Zone 12). The PDSI reconstructions used to model soil moisture in this 6.72 km 2 catchment are drawn from the Mesa Verde series. (The catchment is a little smaller than the normal 7.88 km 2 formed by creating a 1.6-km radius around the site. This is due to its position close to the eastern margin of the study area where the eastern edge of the catchment is truncated). Dating. Occupation dates have been assigned to this site on the basis of architectural styles, stratigraphic evidence, ceramic associations, and limited tree-ring data. Calendrical assignments are made primarily from ceramics and tree-ring dates. However, limited excava- - tions and few cutting dates result in a tentative and incomplete chronology of site use. Ceramic dates for the earlier Chacoan period place initial construction sometime ca. A.D. 1075-1100. A second occupation at or about A.D. 1150 after a short abandonment is suggested, and a third and final post-A.D. 1200 reoccupation after a second abandonment is inferred (Hallasi 1979:396-397). Thirty-five wood samples were dated for Escalante Ruin (Table 5.23). Of these, nine were cutting dates. Ahlstrom (1985:340--343) reviews dating arguments for this site. He acknowledges the interpretive problems posed by ceramics, stratigraphy, and the tree-ring data. He tentatively suggests that the weak cluster of dates in the 1130s may be remodeling events associated with Kiva A, and the strong cluster at A.D. 1129 for Room 20 may represent a construction episode at the site. Whether major construction and use occurred prior to the late 1120s is not clear from the current list of tree-ring dates from the site. No dated samples are available for the stratigraphically earliest excavated room. The earliest cutting date from 5MT2149 is A.D. 1124 r from Room 9, a room built during the second occupation. The latest cutting date is A.D. 1138 v from Kiva A. The terminal date on the site is A.D. 1138 vv from Kiva A. Thus, the period from A.D. 1124-1138 is used to represent a likely period of occupation at Escalante Ruin for this study. These dates are assigned with full acknowledgment that the initial construction and occupation of the site may have preceded the beginning date by two or more decades and persisted for some years later than the final date. Site Catchment and Airicultural ProducTable 5.24 summarizes population values for 5MT2149. The 400-year maximum carrying capacity for the site is 34 ± 6 persons/km2 with a critical carrying capacity of 24 persons/km 2. These values are most similar to the values for the study area as a whole and the ｾＮ 169- 170 160 1.50 140 130 > 120 t:: 110 ｾ 100 C/) 0 ­j Z 0 .... -.J 0 ::J 0.. 0 90 80 70 60 0.. .50 40 30 20 10 0 I I 900 1000 1100 1200 YEARS (A.D.) Figure 5.14. Population density supportable within the 1.6-km-radius catchment of 5MT1566. Lowry Ruin, A.D. 901-1300. 1300 Table 5.21. Comparison of Population Values (POPKM) for 5MT1566, Lowry Ruin, for A.D. 901-1300 and A.D. 1086-1120. A.D. 901-1300 400 years Area(kffi2) TOTPROoa (kg) C.V.b Maximum value MaxirTlJm CC: Maximum Vaule Critical CC: Min Value Optimal CC: 60%C of Mean 40% of Mean 20% of Mean .... .... A.D. 1086-1120 35 years TOTPROD(kg) .....:J C.V. Maximum value Maximum CC: Mean value Critical CC: Min value Optimal CC: 60% of Mean 40% of Mean 20% of Mean 5MT1566 (Lowry) 7.88 409,875 ± 102,360 25.0 68 42 ± 10 26 (62%) 25 17 8 Mockingbird Mesa 17.96 801,336 ± 204,994 25.6 57 36 ± 9 18 (50%) 22 14 7 Sand Canyon Locality 26.08 1,779,087 ± 337,524 19.0 80 55 ± 10 37 (67%) 33 22 11 5MT6970 (Wallace) 4.32 297,798 ± 53,732 18.0 78 56 ± 16 37 (66%) 34 22 11 5MT2149 (Escalante) 6.72 291,075 ± 53,857 18.5 58 34 ± 6 24 (70%) 20 14 7 442,524 ± 97,810 22.1 68 (AD 1116) 46± 10 26 (57%) (AD 1090) 28 18 9 843,185 ± 184,260 21.9 54 (AD 1116) 38±8 18 (47%) (AD 1090) 23 15 8 1,861,444 ± 335,490 18.0 79 (AD 1116) 58± 10 37 (64%) (AD 1090) 35 23 12 308,549 ± 46,326 15.0 74 (A.D.1116) 58 ±8 37 (64%) (A.D.1090) 35 23 12 301,478 ± 62,644 20.8 58 (AD 1116) 36± 7 24 (67%) (A.D.1090, 1104) 22 14 7 Note: All population values are truncated intergers. a is the total mean productivity of maize rounded to the nearest whole number. bCoefficient of Variation, the ratio of the standard deviation to the mean multiplied by 100, rounded to the nearest tenth. cpercent of mean value rounded to the nearest whole number. Study Area 1470.36 64,925,217 ± 13,936,845 21.5 57 35± 7 21 (60%) 21 14 7 68,542,192 ± 13,413,144 19.6 55 (AD 1116) 38± 7 21 (55%) (AD 1090) 23 15 8 Table 5.22. Periods of Greatest Occupational Attractiveness in the Catchment of 5MT1566 (A.D. 1086-1120). Rank by Yrs. a 3 1 2 Rank by Min. Pop. Densityb 3 1 (5MT1566) 2 Beginning Date A.D. 931 1091 1228 Ending Date A.D. 952 1120 1253 Number of Years 22 30 26 Minimum Population Density Value (km2) 31 32 32 aBased primarily on number of years and secondarily on the minimum population density that can be sustained. bBased primarily on minimum population density and secondarily on the number of years that this condition can be consecutively sustained. Table 5.23. Tree-Ring Dates from Site 5MT2149, Escalante Ruin (TRL# A-354). Struct., Feature Kiva A Kiva A Kiva A Kiva A Kiva A Kiva A Kiva A Kiva A Kiva A Kiva A Kiva A Kiva A Kiva A Kiva A Kiva A Kiva A Kiva A Kiva A Kiva A Room 6 Room 6 Room 6 Room 9 Room 9 Room 9 Room 9 Room 20 Room 20 Room 20 Room 20 Room 20 Room 20 Room 20 Room 20 Room 20 TRL Cat. Number ESC-23 ESC-19 ESC-21 ESC-6 ESC-18 ESC-16 ESC-65 ESC-66 ESC-22 ESC-11 ESC-26 ESC-69 ESC-70 ESC-10 ESC-5 ESC-7 ESC-25 ESC-9 ESC-8 ESC-44 ESC-45 ESC-48 ESC-52 ESC-64 ESC-56 ESC-53 ESC-123 ESC-130 ESC-122 ESC-127 ESC-124 ESC-125 ESC-126 ESC-129 ESC-131 Field Num. 64 46 51 5 45 43 72 73 52 17 69 79 80 13 4 7 68 10 9 32 33 40 53 62 57 54 56 63 54 60 57 58 59 62 64 Tree Species Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Pond. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Pond. Jun. Pond. Jun. Pond. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Jun. Outside Date/Symbol 988 w 1002 w 1027 w 1057 w 1073 +w 1088 w 1100 w 1105 w 1110 w 1116w 1118 w 1127 w 1128 +w 1128 w 1134w 1136 +v 1137 w 1138 w 1138 v 1084 w 1084 w 1098 w 1112 w 1118 ++rB 1124 r 1134 +B 1112w 1128 +r 1129 r 1129 r 1129 rB 1129rB 1129 rB 1129 rB 1129 rB Note: For explanation of outside date symbols, see Table 5.5. - 172- Mockingbird Mesa survey locality (maximum carrying capacity of 35 ± 7 persons/km2 with a critical carrying capacity of 21 persons/km 2, and maximum carrying capacity 36 ± 9 persons/km2 with a critical carrying ｹｴｩ｣｡ｾ of 18 persons/km , respectively). The 400-year maximum and critical carrying capacity values for the Sand Canyon survey locality are much higher at 55 persons/km2 and 37 persons/km 2. The 15 years associated with the occupation for 5MT2149 have only a slightly higher maximum carrying capacity at 35 ± 6 persons/km2, but the critical carrying capacity remains the same at 24 persons/km 2. Maximum carrying capacity values for the study area and the Mockingbird Mesa survey locality remain the same during this period although their critical carrying capacity is lifted a slight amount (22 persons/km2 for both areas). Both maximum and critical carrying capacity values for the Sand Canyon survey locality are elevated slightly from the long-term conditions during the A.D. 1124--1138 period (56 ± 12 per-sons/km2 Table 5.24. Comparison of Population Values (POPKM) for 5MT2149, Escalante Ruin, for A.D. 901-1300 and A.D. 1124-1138. A.D. 901-1300 400 years Area (kffi2) TOTPRODa (kg) C.V.b Maximum value Maximum CC: Mean value Critical CC: Min. value Optimal CC: 60%C of Mean 40% of Mean 20% of Mean ..... ....,J V.l A.D. 1124-1138 15 years TOTPROD (kg) 5MT2149 (Escalantei 6.72 291,075 ± 53,857 18.5 58 34 ± 6 24 (70%) 20 14 7 Mockingbird Mesa 17.96 801,336 ± 204,994 25.6 57 36 ± 9 18 (50%) 22 14 7 Sand Canyon locality 26.08 1,779,087 ± 337,524 19.0 80 55 ± 10 37 (67%) 33 22 11 5MT6970 (Wallace) 4.32 297,798 ± 53,732 18.0 78 56 ± 10 37 (66%) 34 22 11 5MT1566 (lowry) 7.88 409,875 ± 102,360 25.0 68 42 ± 10 26 (62%) 25 17 8 405,238 ± 293,201 ± 1,782,290 ± 808,937 ± 291,715 ± 125,280 400,000 203,427 58,062 51,911 30.9 19.8 22.4 25.2 17.8 C.V. 66 (A.D. 1124) 76 (A.D. 1129) 42 (AD 1124, 55 (AD 1129) 74 (AD 1124) Maximum value 1127,1129) 42 ± 13 55 ± 11 36±9 56 ± 12 35±6 Maximum CC: Mean value 26 (62%) 40 (73%) 22 (61%) 37 (66%) 24 (69%) Critical CC: Min. value (AD 1131, 1133) (A.D. 1126, 1131, (A.D. 1131, (AD 1131, (AD 1131) 1133) 1137) 1133) 25 33 34 21 22 Optimal CC: 60% of Mean 17 22 14 14 22 40% of Mean 8 11 11 20% of Mean 7 7 Note: All population values are truncated intergers. atotal mean productivity of maize rounded to the nearest whole number. bCoefficient of Variation, the ratio of the standard deviation to the mean multiplied by 100, rounded to the nearest tenth. cpercent of mean value rounded to the nearest whole number. Study Area 1470.36 64,925,217 ± 13,936,845 21.5 57 35±7 21 (60%) 21 14 7 64,471,631 ± 15,725,883 24.4 49 (AD 1129) 35±8 22 (63%) (AD 1131) 21 14 7 ISO 140 130 120 110 >t:: 100 ffi 0 90 CI) Z 10 j 70 9 ..... ­J +>- I ｾ 60 ｾ 0 ｾ SO 40 30 20 10 0 900 , , 1000 1100 1200 YEARS (A.D.) Figure 5.15. Population density supportable within the 1.6­km­radius catchment of 5MT2149, Escalante Ruin, A.D. 901­1300. 1300 Table 5.25. Periods of Greatest Occupational Attractiveness in the Catchment of 5MT2149 (A.D.1124­1138). Rank by Yrs. a 5 2 4 3 1 Rank by Min. Pop. Densityb 5 2 (5MT2149) 4 3 1 Beginning Date A.D. 938 11 05 1192 1228 1259 Ending Date A.D. 954 1130 1214 1253 1286 Number of Years 16 26 23 25 28 Minimum Population Density Value (km2) 30 30 30 30 30 aBased primarily on number of years and secondarily on the minimum population density that can be liustained. bBased primarily on minimum population density and secondarily on the number of years that this condition can be sustained. sons/km 2. This value is no different than the long­term minimum population value that can be sustained over the 4OO­year time series. Site Catchment and Prime Periods for Ag­ Instead, Figure 5.15 was inspected for intervals of at least 15 years that sustained higher popuricultural Production. The main period of lation densities. Data contained in Appendix N occupation for 5MT2149 is estimated to be A.D. 1124­­1138 with a minimum population were also reviewed and the results are listed in value for the 15­year period of 24 per­ Table 5.25. Since there is no difference in the minimum population density for any of these intervals, Table 5.26. Tree­Ring Dates From Site 5MT765, Sand Canyon Pueblo they can only be ranked by (TRL# A­856). the length of time that this high level endures. The data Struct., TRL Cat. Field Tree Outside indicate that the A.D. 1259Feature Number Num. Species Date/Symbol 1286 period is the most CCC­210 Jun. 1250 w Struet. 101 53 predictable period of at least 87 927w Struet. 102 CCC­362 a,b Jun. 15 years in length in the Struet. 102 CCC­234 39 Jun. 933 +w entire 4oo­year sequence. 44 Struet. 102 CCC­239 Jun. 958w However, the interval to Struet. 102 CCC­217 22 Jun. 988w which the Escalante Ruin is Struet. 102 CCC­241 a,b 46 Jun. 1005 +w currently dated is almost as Struet. 102 CCC­360 85 Jun. 1029 ++r long. Therefore, it would Struet. 102 CCC­155 40 Jun. 1039 +w Struet. 102 CCC­339 48 Jun. 1075 w appear that during the earlier 51 Struet. 102 CCC­166 Jun. 1076 w Chacoan episode at the Struet. 102 CCC­161 46E Jun. 1090 ++w Escalante Ruin, the site was Struet. 102 CCC­224 a,b,c 29 Jun. 1112 w used and grew during a time Struet. 102 CCC­346a,b 56 Jun. 1115 w of high temporal predict­ Struet. 102 CCC­356 1124 w 66 Jun. ability and high spatial pro­ Struet. 102 CCC­215 20 Jun. 1130 w ductivity. Struet. 102 CCC­355 a,b Jun. 1130 w 65 Struet. 102 CCC­132 Jun. 1137 w 8 5Mf765, Sand Canyon Struet. 102 CCC­131 7 Jun. 1146 w Pueblo Struet. 102 CCC­147 30 Jun. 1160 w Struet. 102 CCC­127 1 Jun. 1165 w 29 Struet. 102 CCC­145 Jun. 1167 +w Description. Sand Can­ CCC­186 Struet. 102 78 Jun. 1172 w yon Pueblo is an extensive, Struet. 102 CCC­370 Jun. 1175 ++w single­component, late CCC­172 Struet. 102 50 Jun. 1180 +w Pueblo III site with some CCC­235 a,b,c Struet. 102 40 Jun. 1181 w 300­400 rooms, 89 kivas, 14 Struet. 102 CCC­348 a,b 58 Jun. 1190 ++w towers, aD­shaped, bi­ or triwith a critical carrying capacity value at 37 persons/km2). ­ 175- Table 5.26 (Continued). Struet., Feature Struct. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struct. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 TRL Cat. Number CCC-144 CCC-231 a,b CCC-207 CCC-350 a,b CCC-367 CCC-185 CCC-154 CCC-179 CCC-174 CCC-160 CCC-176 CCC-133 CCC-136 CCC-342 a,b CCC-232 CCC-208 CCC-345 CCC-162 CCC-163 CCC-135 CCC-204 CCC-363 CCC-219 CCC-148 CCC-178 CCC-343 CCC-347 CCC-184 CCC-364 CCC-202 CCC-369 a,b CCC-349 CCC-341 CCC-343 CCC-222 CCC-225 CCC-227 CCC-230 CCC-236 a,b CCC-238 CCC-228 a,b,c CCC-151 CCC-366 eeC-167 CCC-226 eCC-152 CCC-169 CeC-158 CeC-168 Cee-221 CCC-223 CCC-153 CCC-159 CCC-177 Field Num. 28 36 FS4 60 92 76 38 67 60 46B 61 9 12 52 37 FS 7 55 47 48 10 FS 7 88 24 31 66 53 57 72 89 FS 7 59 51 53 27 30 32 35 41 43 33 34 91 54 31 35 56 44 55 26 28 37 45 62 Tree Outside Species Date/Symbol Jun. 1195 w Jun. 1198 w Jun. 1200 w Jun. 1200 w Jun. 1213 w Jun. 1213 w 1214 w Jun. Jun. 1217 ++w Jun. 1220 w Jun. 1223 w Jun. 1223 w Jun. 1226 ++w Jun. 1226 ++w Jun. 1228 w Jun. 1228w 1229 w Jun. Jun. 1230 ++v 1232 r Jun. 1232 w Jun. Jun. 1233 +v Jun. 1233 ++w Jun. 1234 +v Jun. 1234 +v Jun. 1235 r 1235 r Jun. Jun. 1235 +v Jun. 1235 r 1235 rB Jun. Jun. 1235 r Jun. 1235 r Jun. 1235 r Jun. 1235 r Jun. 1235 v Jun. 1235 +v Jun. 1235 r B Jun. 1235 r Jun. 1235 r Jun. 1235 w Jun. 1235 r 1235 r Jun. Jun. 1236 v Jun. 1242 w Jun. 1242 r Jun. 1242 r Jun. 1242 r Jun. 1247 +w Jun. 1248 +w Jun. 1249 +v Jun. 1249 r Jun. 1249 H Jun. 1249 +w Jun. 1250 HB Jun. 1250 +w Jun. 1250 H - 176- walled structure (containing two kivas), a Great Kiva. and at least one prehistoric reservoir. Its presence was noted by Prudden (1903) and Fewkes (1919), it was recorded in 1967 (Lister 1967), and portions were stabilized in 1978 (Metzger and Breternitz­Goulding 1981). Systematic and intense study began in 1983 by the Crow Canyon Archaeological Center of Cortez, Colorado. The most recent overview of this work is in Bradley (1992). Annual reports of investigations are also available (e.g., Adams 1985; Bradley 1986, 1987, 1988a; Kleidon and Bradley 1989). Location. 5M1765 is situated on the mesa rim and along the upper canyon slopes of Sand Canyon, one of the major tributaries to McElmo Creek, directly north of Sleeping Ute Mountain. It is a large, walled, U-shaped community conforming to the topography of the canyon headlike rincon, incorporating a significant spring within its immediate territory. The site is located in the N1(1 of the NE1/4 of Section 12, T36N, R18W, N.M.P.M. on the Woods Canyon 7.5-minute quadrangle map. Its geographic center is situated at an elevation of 2,073 m (6,800 ft) at UTM coordinates 4,141,200 m N and 696,800 m E (Zone 12). The . PDSI reconstructions used to model soil moisture in soils of this 7.88-km 2 catchment are drawn from the Mesa Verde and Ignacio series. Datin!:. 5M1765 is one Table 5.26 (Continued). Struet., Feature Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 102 Struet. 107 Struet. 107 Struet. 108 Struet. 108 Struet.204 Struet.204 Struet.208 Struet.208 Struet.501 TRL Cat. Number CCC-206 CCC-218 a,b CCC-233 CCC-351 a,b CCC-353 CCC-170 CCC-344 CCC-340 a,b CCC-352 a,b CCC-359 CCC-201 CCC-138 CCC-150 CCC-200 CCC-214 CCC-354 CCC-165 CCC-209 CCC-171 CCC-205 CCC-194 CCC-192 CCC-371 CCC-372 a,b CCC-375 CCC-374 a,b CCC-850 CCC-409 CCC-787 Struet.501 Struet.501 CCC-843 CCC-831 Struet.501 Struet.501 CCC-842 CCC-824 Struet.501 Struet. 501 CCC-758 CCC-784 Struet.501 CCC-795 Struet.501 CCC-792 Struet. 501 CCC-834 Struet.501 CCC-836 Struet.501 CCC-788 Struet. 501 Struet.501 Struet.501 Struet.501 Struet.501 CCC-841 CCC-760 CCC-741 CCC-n2 CCC-791 Field Num. FS 7 23 38 61 63 57 54 49 62 70 FS7 15 34 FS7 19 64 52 FS2 58 FS7 16 11 4 331/ 1 216/2 388/ 114 693/ 10 3881 166 693/9 3881 162 388/83 3881 224 3881 131 3881 113 3881 219 3881 221 3881 115 693/5 388/86 388/2 388/90 3881 119 Tree Outside Species Date/ Symbol Jun. 1250 +v Jun. 1250 HB Jun. 1250 +B Jun. 1250 w Jun. 1250 w 1251 H Jun. Jun. 1251 +w Jun. 1251 +v Jun. 1251 +v 1251 H Jun. 1257w Pitt Pit'!. 1261 w 1264 r Pitt Pit'!. 1264w 1267 rB Pin. Jun. 1270 H 1271 rB Jun. 1271 rB Jun. 1274 r Pitt 1274 r Pili. Jun. 1160 +w Jun. 1166 +w Jun. 1036 ++w Jun. 1229 w Pitt 1235 w Pin. 1267 +w 702w Jun. Pitt 1244 rB Jun. 904+w Jun. Jun. 925w 928w Jun. Jun. 9nw 983 ++w Jun. Jun. 987w 994 ++w Jun. 994+w Jun. 998 +w Jun. 1007w Jun. 1019 +w Jun. 1038 ++w Jun. Jun. Jun. Jun. Jun. 1061 w 1063 w 1084 +w 1088 w 1125 ++w ­ 177- of the best-dated sites in this portion of the Mesa Verde Region. Architectural and artifactual associations place the entire site finnly within the thirteenth century. As of the spring of 1990, a total of 329 wood samples had been dated for 5MT765 (Table 5.26). Of these, 124 were cutting dates. A total of 12 masonry structures, either a room or a kiva, had produced datable wood samples. Five of these structures produced samples with cutting dates and architectural evidence sufficiently strong to suggest a specific year of construction (Bruce Bradley, personal communication, May 1990). Table 5.27 summarizes these data. The earliest cutting date for excavated portions of 5MT765 is A.D. 1200 rB and the latest is A.D. 1274 r, which is also the [mal date for the site. Major clusters of tree­ring dates occur at A.D. 1203, 1225, 1235, 1250, 1251, 1252, 1260, 1262, 1266, and 1274. Based on the [mdings of current investigations in the greater Sand Canyon survey locality (Adler 1988; Huber 1989; Huber and Bloomer 1988; Van West 1986; Van West et at. 1987; Varien 1990) Bradley suspects wholesale movement of kiva roof lumber in the thirteenth century to 5MT765 from previous village locations (Bruce Bradley, personal communication, March 1990). He tentatively concludes that construction and use dates for 5MT765 do not begin as early as the first cutting dates in the 1200s might suggest. Rather, he places the first Table 5.26 (Continued). Struet., Feature Struet.501 TRL Cat. Number CCC-793 Struet.501 CCC-825 Struet.501 Struet.501 CCC-757 CCC-806 Struet.501 CCC-813 Struet.501 Struet. 501 CCC-743 CCC-822 Struet.501 CCC-782 Struet.501 Struet.501 Struet.501 CCC-769 CCC-742 CCC-808 Struet.501 CCC-807 Struet.501 Struet.501 Struet. 501 Struet.501 CCC-838 CCC-754 CCC-766 CCC-780 Struet.501 CCC-802 Struet.501 Struet.501 Struet.501 Struet.501 Struet.501 Struet.501 Struet. 501 Struet.501 CCC-846 CCC-748 CCC-761 CCC·764 CCC-765 ccc-no CCC-n3 CCC-776 Struet.501 Struet.501 CCC-n8 Struet.501 CCC-797 Struet.501 CCC-798 Struet.501 CCC-800 Struet.501 CCC-801 Struet.501 CCC-803 Struet.501 CCC-809 Stnlet.501 CCC-810 Struet.501 CCC-816 CCC-796 Field Tree Outside Num. Species Date/ Symbol 1141 w 388/ Jun. 127 1145 w 388/ Jun. 163 1146 ++w 388/82 Jun. 1149 w 388/ Jun. 124 1152 w 388/ Jun. 149 1170 w Jun. 388/4 1188 w Jun. 388/ 158 1189 w Jun. 388/ 104 1190 w 388/77 Jun. 1194 w 388/ 1 Jun. 1195 ++8 388/ Jun. 138 Douglas 1196 +r 388/ fir 126 1196w 693/6 Jun. 1200 w 388/22 Jun. 1200 r8 388/74 Jun. 1201 v 388/ Jun. 102 1201 w 388/ Jun. 121 1202 r 552/1 Jun. 388/ 14 Jun. 1203 r 1203 r 388/87 Jun. 1203 r 388/92 Jun. 1203 r 388/73 Jun. 1203 r 388/78 Jun. 1203 r 388/93 Jun. 388/ Jun. 1203 r 120 1203 r 388/99 Jun. 1203 r 388/ Jun. 143 1203 r Jun. 388/ 132 Jun. 1203 r 388/ 135 1203 r 388/ Jun. 137 1203 r 388/ Jun. 139 Jun. 1203 r 388/ 142 Jun. 1203 v 388/ 140 1203 r Jun. 388/ 144 1203 r Jun. 388/ 152 - 178- solid evidence for con­ struction ca. A.D. 1252 on the basis of architectural properties in conjunction with tree­ring dates, and he has evidence for more-orless continuous construction and site use until at least A.D. 1274. For the purposes of this study, an occupation of A.D. 1252-1274, a period of 23 years, is assigned to 5MT765, with the knowledge that the initial construction and use of this site may be several decades earlier and it may have been used for a decade later than the last date. Site Catchment and Agricultural Productivity. Table 5.28 summarizes population values for 5MT765. The 400-year maximum ｃ｡ｲｹｩｮｾ capacity within the 7.88 kIn radius of the site is 47 ± 9 persons/km2 with a critical carrying ｣｡ｾｩｴｹ of 10 persons/km . This extremely low population density value happens only once in the 4OO-year sequence, at A.D. 1277. Otherwise, the low is 31 per-sons/km 2, which happens 31 times in the four centuries. Although the maximum carrying capacity for this catchment is higher than the maximum carrying capacity values for either the Mockingbird Mesa survey ± 9 perlocality ｾＳＶ sons/km ) or the study area as a whole (35 ± 8 persons/km2), the critical carrying capacity value of 10 persons/km2 for the catchment is significantly lower than either of these two respectively). Furthermore, the maximum and critical Table 5.26 (Continued). Struet.. Feature Struet.501 TRL Cat. Number CCC-817 Struet.501 CCC-819 Struet.501 CCC-823 Struet. 501 CCC-830 Struet. 501 CCC-835 Struet.501 Struet.501 Struet.501 CCC-840 CCC-845 CCC-815 Struet.501 Struet.501 CCC-774 CCC-789 Struet.501 Struet.501 Struet.501 CCC-740 CCC-744 CCC-805 Struet.501 Struet.501 CCC-775 CCC-786 Struet.501 CCC-804 Struet.501 CCC-818 Struet.501 CCC-832 Struet. 501 Struet.501 Struet.501 Struet.501 CCC-745 CCC-762 CCC-759 CCC-820 Struet.501 CCC-829 Struet.501 Struet.501 Struet.501 Struet. 501 Struet.501 Struet.501 CCC-749 CCC-753 CCC-756 CCC-767 CCC-771 CCC-799 Struet.501 CCC-828 Struet.501 CCC-833 Struet.501 CCC-811 Struet.501 CCC-794 Struet.501 CCC-821 Outside Tree Field Species Date/ Symbol Num. 1203 r Jun. 388/ 225 1203 r Jun. 388/ 155 1203 r Jun. 388/ 161 1203 r Jun. 388/ 164 1203 r 388/ Jun. 220 1203 r Jun. 693/8 1203 r Jun. 548/7 1204 w Jun. 388/ 151 1215 +r 388/94 Jun. Jun. 388/ 1215 ++v 117 Jun. 1220 +r 388/3 Jun. 388/8 1220 +v Jun. 388/ 1220 +w 123 388/96 Jun. 1221 r Jun. 388/ 1221 r 110 Jun. 388/ 1221 w 122 Jun. 1221 r 388/ 154 388/ Jun. 1221 r 171 Jun. 388/6 1222 r 388/88 Jun. 1222 r 388/85 Jun. 1223 w 388/ Jun. 1223 w 159 Jun. 388/ 1223 v 170 388/ 15 Jun. 1224 r 388/21 Jun. 1224 w 388/81 Jun. 1224 r 388/75 Jun. 1224 r 388/79 Jun. 1224 r 388/ Jun. 1224 r 136 388/ Jun. 1224 r 169 388/ Jun. 1224 r 172 Jun. 388/ 1225 r 146 388/ Jun. 1238 r 128 388/ Jun. 1238 r 160 ­ 179- carrying capacity values for the Sand Canyon survey locality, the area in which ＵＶＷｔｾＵ Ｌ､･ｾ｣ｯｬｳｩ are far higher than those for the site (55 ± 10 persons/km2 and 37 persons/km2). This discrepancy can be partially accounted for by noting that ＵｾｔＷＶ is located on the mesatop edge and sloping canyon walls of Sand Canarea yon. The ･ｾｩ､ｭ around the site is characterized by shallow or steeply would sloping sediments ｾｨｴ result in low yields given dry­farming techniques. Therefore, exploitation of arable lands higher on the Ｌｰｯｾｳ･ｭ and thus, at a ｧｲ･ｾ ｲ distance from the site proper, would have been essential to increasing the minimum population supportable at the site itself. The maximum and critical carrying capacity values for the 23­year ｮｯｩｾｰｵ｣ ｯ period assigned to ＵＶｾ indicates that this A.D. 1252­1274 period would have been able to support a mean population density of 48 ± 8 persons/km2 with a minimum population density of 31 persons/km2 ｾ all times. While the 23­year maximum carrying capacity value is only slightly higher than the 400­year average of at 47 ± 9 persons/km2, the critical carrying capacity value is over three times as high ｾ 31 persons/km 2. During the same 23­year interval, the is catchment for Ｕ ｾ ｔ Ｗ Ｖ Ｕ again markedly better than rd ｡ｳ･ｾ the ｩ｢ｧｮｫ｣ｯｾ survey locality or the study area as whole (37 ± 9 persons/km2 with a minimum of Table 5.26 (Continued). Struet., Feature Struet.501 Struet.501 TRL Cat. Number CCC-755 CCC-812 Struet.501 Struet. 501 Struet.501 Struet.501 Struet.501 CCC-746 CCC-751 CCC-763 CCC-768 CCC-n9 Struet.501 CCC-837 Struet.501 CCC-790 Struet.501 Struet.501 CCC-747 CCC-814 Struet.501 Struet. 504 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struct. 1004 Struct. 1004 Struet. 1004 CCC-839 CCC-851 CCC-737 CCC-736 CCC-717 CCC-703 CCC-718 CCC-731 CCC-699 CCC-712 CCC-700 CCC-723 CCC-716 CCC-725 CCC-726 CCC-701 CCC-730 CCC-739 CCC-729 CCC-732 CCC-713 CCC-719 CCC-702 CCC-714 CCC-706 CCC-696 CCC-708 CCC-721 CCC-735 CCC-738 CCC-695 CCC-707 CCC-734 CCC-728 CCC-727 CCC-710 CCC-698 CCC-704 Field Tree Outside Num. Species Date/ Symbol 388/80 Jun. 1246 ++v 388/ Jun. 1249 +r 145 388/7 Jun. 1250 r 723/1 Jun. 1250 r 388/91 Jun. 1250 r 388/76 Jun. 1250 r 388/ Jun. 1250 r 100 388/ Jun. 1250 r 223 388/ Jun. 1251 +r 116 388/ 11 Jun. 1252 +vv 388/ Jun. 1252 +r 150 693/7 Jun. 1252 v 345/1 Jun. 1065 vv 530/1 Jun. 911 VV 584/1 Jun. 924 +vv 530/33 Jun. 994vv 530/ 13 Jun. 1155 vv 530/34 Jun. 1156 ++vv 530/29 Jun. 1170 vv 530/8 Jun. 1183 +vv 530/ 14 Jun. 1184 vv 530/9 Jun. 1189 vv 530/41 Jun. 1190 vv 530/25 Jun. 1200 ++B 530/26 Jun. 1201 ++vv 530/43 Jun. 1205 vv 530/ 10 Jun. 1209 vv 530/27 Jun. 1223 vv 530/3 Jun. 1223 vv 530/28 Jun. 1238 +r 530/30 Jun. 1242 vv 530/ 18 Jun. 1243 r 530/35 Jun. 1244 r 530/ 11 Jun. 1246 vv 530/ 15 Jun. 1247 vv 530/22 Jun. 1248 vv 530/5 Jun. 1249 r 530/ 17 Jun. 1249 +r 530/36 Jun. 1249 r 1249 r 530/2 Jun. 1249 r 588/1 Jun. 586/1 Jun. 1250 r 530/ 19 Jun. 1250 v 530/32 Jun. 1250 r 530/40 Jun. 1253 +vv 530/44 Jun. 1262 vv 530/24 Jun. 1263 +vv 1264 +r 530/7 Jun. 1264 +r 530/20 Jun. ­ 180- 18 persons/km2, and 36 ± 7 persons/km2 with a minimum of 21 persons/km2, respectively). Nevertheless, the catclunent for the site during this interval is not as good as other potential catclunents within the Sand Canyon survey locality, which has a population density potential that is virtually unchanged from long­term conditions (maximum carrying capacity of 55 ± 14 persons/km2 and a critical carrying capacity of 37 persons/km2). Site Catclunent and Prime Periods for AgriculturaJ Production. With the exception of year A.D. 1277 when the maximum arumal population density that could be supported on the potential productivity from the 7.88 km 2 catchment was only 10 persons/km2, the minimum population density value for the 23­year occupation period is the same as it is for the 400­year period at 31 persons/km 2. Consequently, Figure 5.16 and values contained in Appendix 0 were examined for intervals of sustained higher population density. The results of this search are contained in Table 5.29. The period of highest temporal predictability is the A.D. 931­971 period and the period of highest productivity is the A.D. 1228­1253 period. It would seem likely that 5MT765 was established in this period of greatest productivity for the catchment, and continued through another period of relatively high and predictable productivity (A. D. 1255- Table 5.26 (Continued). Struet., Feature Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1004 Struet. 1203 Struet. 1205 Struet. 1205 Struet. 1205 Struet. 1205 Struet. 1205 Struet. 1205 Struet. 1205 Struet. 1205 Struet. 1205 Struet. 1205 Struet. 1205 Struet. 1205 Struet. 1205 Struet. 1205 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. 1206 Struet. Struet. Struet. Struet. Struet. Struet. 1206 1206 1206 1206 1206 1206 TRL Cat. Number CCC­722 CCC­724 CCC­711 CCC­705 CCC­410 CCC­416 CCC­414 CCC­423 CCC­412 CCC­422 CCC­425 CCC­421 CCC­418 CCC­424 CCC­413 CCC­420 CCC­411 CCC­415 CCC­417 CCC­377 CCC­198 CCC­397 CCC­395 CCC­408 CCC­400 CCC­399 CCC­381 CCC­405 CCC­338 a,b CCC­197 CCC­404 CCC­396 CCC­391 CCC­398 CCC­337 a,b CCC­299,300 CCC­388 CCC­379 CCC­383 CCC­386 CCC­387 CCC­389 CCC­390 CCC­332 CCC­328 CCC­302 CCC­306,320 a,b CCC­318 a,b CCC­295 CCC­296 CCC­301,307 CCC­311 a,b CCC­312 1276). Field Tree Outside Num. Species Date/ Symbol 1264 H 530/39 Jun. 1264 H 530/42 Jun. 1265 H 530/ 16 Jun. 1266 r 530/21 Jun. Jun. 242/1 1169 +w Jun. 246/8 1026 +w 1031 w Jun. 246/6 1061 w Jun. 182/8 1181 w 182/7 Jun. 1208 w Jun. 246/4 Jun. 182/9 1229 +w 1231 H 246/ 13 Jun. 1236w 246/ 10 Jun. Jun. 182/9 1248 w Jun. 1249 HB 246/5 246/ 12 Jun. 1251 +rB 246/3 Jun. 1252 ++r 246/7 Jun. 1260 rB Jun. 246/9 1260 rB Jun. 166/2 934w 74 Jun. 972+w Jun. 1031 +w 231/6 231/ 14 Jun. 1036 ++w 224/3 Jun. 1036 +w 231/7 Jun. 1049 ++w 231/ 21 Jun. 1058 +w Jun. 166/3 1059 w 174/ 1 Jun. 1090 +w Jun. 78 1138 +w 82 Jun. 1171 ++w 143/27 Jun. 1176 ++w 231/ 12 Jun. 1181 +w Jun. 143/1 1193 w 231/4 Jun. 1202 ++w 77 Jun. 1224w 27 Jun. 1241 HB 167/2 Jun. 1241 w 166/3 Jun. 1242 rB 167/ 1 Jun. 1242 r 167/2 Jun. 1242 v 167/2 Jun. 1242 r 167/2 Jun. 1242 r 126/22 Jun. 1242 r Jun. 71 1242 r 67 Jun. 1242 r 32 Jun. 1242 r 36 Jun. 1242 r 52 23 24 31 41 43 Jun. Jun. Jun. Jun. Jun. Jun. ­ 1242 1242 1242 1242 1242 1242 181 - r rB rB rB rB rB Concluding Remarks Table 5.30 presents the long­term population density values at the POP2YR level of demand for the catchments of the eight tree­ring dated sites. While it is unwise to suggest general trends in behavior on the basis of only eight sites, it is instructive to conclude this examination of catchment data with a ranking of these different places on the productive landscape. Of the four small village sites, 5MT8839 (Norton House) is clearly located in the potentially most productive and most predictable catchment. This is indicated by the relatively high long­ term mean value, the maximum carrying capacity, and relatively low coefficient of variation. Based on the mean population density value for each catchment, the ranking for small village sites is 5MT8839 (Norton) is highest, followed by 5MT3834 (Mustoe), 5MT2433 (Aulston), and finally 5MT8371 (DCA Site). Of the four central place sites, 5MT6970 (Wallace) has the highest long­term mean population density, followed by 5MT765 (Sand Canyon), 5MTl566 (Lowry), and 5MT2149 (Escalante). It is interesting to note that of the two contemporary small village sites, 5MT8839 (Norton) and 5MT2433 (Aulston), the former is by far the more attractive as a locale for high potential Table 5.26 (Concluded). Struct., TRL Cat. Feature Number Struct. 1206 CCC-317 Struct. 1206 CCC-323,322 Struct. 1206 CCC-326a,b Struct. 1206 CCC-331 Struct. 1206 CCC-335 a,b Struct. 1206 CCC-336 Struct. 1206 CCC-315 a,b Struct. 1206 CCC-327a,b,c Struct. 1206 CCC-324 Struct. 1206 CCC-393 Struct. 1206 CCC-333 Struct. 1206 CCC-385 Struct. 1206 CCC-378 Struct. 1206 CCC-334 Struct. 1206 CCC-382 Struct. 1206 CCC-380 Struct. 1206 CCC-406 CCC-407 Struct. 1206 Struct. 1206 CCC-305 CCC-308 Struct. 1206 Struct. 1206 CCC-314 a,b Struct. 1206 CCC-321 Struct. 1206 CCC-325 a,b Struct. 1206 CCC-329 Struct. 1206 CCC-297,310 Struct. 1206 CCC-303 Struct. 1206 CCC-294 Struct. 1206 CCC-196 Struct. 1206 CCC-313 a,b Struct. 1206 CCC-330 Struct. 1206 CCC-319 a,b Struct. 1206 CCC-394 Struct. 1206 CCC-373 a,b Feature 2 Note: For explanation of outside date Field Num. 51 59 65 70 75 76 48 66 62 231/ 1 72 167/ 1 166/3 73 166/3 166/3 174/ 1 174/ 1 35 38 45 55 64 68 25 79 22 32 44 69 53 231/2 Tree Outside Species Date/ Symbol Jun. 1242 rB Jun. 1242 rB Jun. 1242 rB Jun. 1242 rB Jun. 1242 rB 1242 rB Jun. Jun. 1242 v 1242 v Jun. 1242 w Jun. Jun. 1245 ++rB Jun. 1259 w 1260 w Jun. Jun. 1261 rB Jun. 1261 rB Jun. 1261 B Jun. 1262 r Jun. 1262 r Jun. 1262 r Jun. 1262 r Jun. 1262 r Jun. 1262 r Jun. 1262 +r Jun. 1262 w Jun. 1262+r Jun. 1262 +r Jun. 1262 +r Jun. 1262 +rB Jun. 1262 +rB Jun. 1262 +rB Jun. 1262 +rB Jun. 1262 w Jun. 1265 +rB Jun. 1242 rB 2 symbols, see Table 5.5. maize yields and high population densities. Similarly, of the three partially con­ temporary central place sites, 5MT6970 (Wallace) is po­ tentially more productive than 5MT1566 (Lowry) and that both are markedly more capable of sustaining higher population densities than 5MT2149 (Escalante). If the dates used to bracket the major Chacoan components present at these sites are reasonably correct, then Wallace's productive pri­ macy may be a factor in its early establishment and enduring presence in the region. The mean population value (the maximum carry­ ing capacity) for the four small villages sites as a group is 53.0 ± 12.9 persons/km2 (C.Y. = 24.3) whereas the mean for the four central place sites is 45.0 ± 8.8 persons/km 2 (C.Y.= 19.6). Likewise, the mean mini­ mum popUlation density (the critical carrying capacity) for all the small village sites is 32.35 ± 17.0 (C.Y.= 52.7) persons/km2 and the mean Table 5.27. Suggested Construction Dates (A.D.) for Tree-Ring Dated Structures at 5MT765. Block No. 100 200 500 1000 1200 Struct. No. 101 102 107 108 204 208 501 504 1004 1203 1205 1206 Function room kiva roomlkiva kiva room kiva kiva room kiva room room kiva Last Date or Range of Dates 1250w 1232 r-1274 r 1166 +w 1229w 1267 +w 1244 rB 1200 rB-1252 +w 1065w 1243 r-1266 r 1169 +w 1260 rB 1242rB-1265+rB ­ 182- Total Samples 1 99 2 2 2 2 98 1 40 1 14 67 Cutting Dates 0 26 0 0 0 1 51 0 10 0 2 34 Suggested Constr. Date ? 1274 ? ? ? ? 1252 ? 1266 ? 1260 1262 Table 5.28. Comparison of Population Values (POPKM) for 5MT765, Sand Canyon Pueblo, for A.D. 901-1300 and A.D. 1252-1274. A.D. 901-1300 400 ｾ･｡ｲＡＩ｟ Area (km2) TOTPRODa(kg) C.V.b Maximum value Maximum CC: Mean value CriticalCC: Minimum value Optimal CC: 60%C of Mean 40% of Mean 20% of Mean .... 00 u.J 5MT765 (Sand Canyon) 7.88 459,114 ± 92,902 20.2 71 47 ± 9 10 (21%) 28 19 9 Mockingbird Mesa 17.96 801,336 ± 204,994 25.6 57 36 ± 9 18 (50%) 22 14 7 Sand Canyon Locality 26.08 1,779,087 ± 337,524 19.0 80 55 ± 10 37 (67%) 33 22 11 Study Area 1470.36 64,925,217 ± 13,936,845 21.5 57 35±8 21 (60%) 21 14 7 A.D. 1252-1274 23 years TOTPROD (kg) 466,172 ± 81,809 65,843,368 ± 12,921,943 815,258 ± 202,795 1,750,494 ± 468,955 C.V. 17.6 24.9 19.6 26.8 46 (AD 1259, 1271) Maximum value 59 (AD 1255, 1271) 54 (AD 1259) 72 (AD 1271) Maximum CC: Mean value 48 ± 8 36± 7 37±9 55± 14 21 (58%) (AD 1254) Critical CC: Minimum value 31 (65%) (AD 1254) 18 (49%) (AD 1254) 37 (67%) (AD 1254) 22 22 33 29 Optimal CC:60% of Mean 14 22 19 15 40% of Mean 7 7 10 11 20% of Mean Note: All population values are truncated integers. atotal mean productivity of maize rounded to the nearest whole number. bCoefficient of Variation, the ratio of the standard deviation to the mean multiplied by 100. This value is rounded to the nearest tenth. cpercent of mean value rounded to the nearest whole number. 180 170 160 ISO 140 130 >- l:: 120 ｾ 110 tI) 0 Z 0 .... 90 -( ..J 110 I-- ..- 00 ｾ 100 :J 00 0- 70 60 SO 40 30 20 10 0 I 900 1000 1100 1200 YEARS (A.D.) Figure 5.16. Population density supportable within the 1.6-km-radius catchment of 5MT765, Sand Canyon Pueblo, A.D. 901-1300. 1300 Table 5.29. Periods of Greatest Occupational Attractiveness in the Catchment of 5MT765 (A.D. 1252-1274). . Rank by Yrs. a 1 6 tie 2 4 5 3 6 tie Rank by Min. Pop. Densityb 5 6tie 4 2 3 1 (5MT765) 6 tie(SMT765) Beginning Date A.D. 931 1033 1091 1105 1192 1228 1255 Ending Date A.D. 971 1054 1130 1130 1214 1253 1276 Number of Years 42 22 40 26 23 26 22 Min. Population Density (km2) 32 32 33 38 38 39 32 aBased primarily on number of years and secondarily on the minimum population density that can be ｾｵｳｴ｡ｩｮ･､Ｎ Based primarily on minimum population density and secondarily on the number of years that this condition can be sustained. minimum population density for the central place sites is 24.25 ± 11.1 persons/km2 (C.V.= 45.8) These statistics would seem to indicate that the small villages were located in catchments that were more productive and more capable of sustaining higher minimum population levels than those for the central place sites. Although this is an admittedly small sample, these data seem to indicate that the agricultural potential of a residential location and its immediate environs was a stronger factor in site selection for inhabitants of small villages than it was for inhabitants of central place sites. Only two of the eight sites were occupied in time intervals that could support a higher minimum population density (the critical car - rying capacity) than in the 4OO­year interval. Seven of eight (the exception being 5MT3834, Mustoe), however, were dated to a time period considered among the best for their particular catchments, seemingly indicating an awareness of and selection for locations and intervals of high agricultural productivity. Finally, in closing, it must be said that it is not possible to know how accurate are the absolute population density values presented in this study. Nevertheless, they provide at least a relative ordering of the productive capabilities for the eight site catchments, as well as the two survey localities and the study area. As such, they are useful for making comparisons and suggesting research questions. 185- Table 5.30. Summary of Population Values (POPKM) Using POP2YR Estimates for Eight Tree-Ring Dated Sites, A.D. 901-1300. 00 0\ Occupation Dates Area (km2) TOTPRODa (kg) C.V.b Maximum value Maximum CC: Mean Value Critical CC: Minimum Value Optimal CC: 60%C of Mean 40% of Mean 20% of Mean 5MT8371 (DCA) A.D. 935-950 7.88 350,669 ± 104,956 29.9 57 35 ± 10 9 (26%) 21 14 7 5MT8839 (Norton) A.D. 1029-1048 7.88 653,070 ± 99,827 15.3 94 67 ± 10 49 (73%) 40 26 13 5MT2433 (Aulston) A.D. 1030-1050 7.88 500,847 ± 115,085 23.0 77 51 ± 12 32 (63%) 31 20 10 5MT3834 (Mustoe) A.D. 1173-1231 7.88 553,196 ± 108,069 19.5 80 56 ± 11 39 (70%) 34 22 11 Occupation Dates Area (km2) TOTPROD (kg) C.V. Maximum value Maximum CC: Mean Value Critical CC: Minimum Value Optimal CC: 60% of Mean 40% of Mean 20% of Mean 5MT6970 (Wallace) A. D. 1045-1125 4.32 297,798 ± 53,732 18.0 78 56 ± 10 37 (66%) 34 22 11 5MT1566 (Lowry) A.D. 1086-1120 7.88 409,875 ± 102,360 25.0 68 42 ± 10 26 (62%) 25 17 8 5MT2149 (Escalante) A. D. 1124-11 38 6.72 291,075 ± 53,857 18.5 58 34±6 24 (70%) 20 14 5MT765 (Sand Canyon Pueblo) A.D. 1252-1274 7.88 459,114 ± 92,902 20.2 71 47±9 10 (21%) 28 19 9 7 Note: All population values are truncated integers. atotal mean productivity of maize rounded to the nearest whole number. bCoefficient of Variation, the ratio of the standard deviation to the mean multiplied by 100 and rounded to the nearest tenth. cpercent given is percent of mean value rounded to the nearest whole number. 6 SUMMARY AND EVALUATIONS This chapter briefly summarizes the methods and basic conclusions of the study. It also describes the strengths and weaknesses of the methods and data used to reconstruct paleoclimate and agricultural productivity. I conclude with suggestions for future research that would improve the model and with suggestions for additional studies that might be accomplished with the resulting databases. POP2YR, POP3YR). From these annual estimates, longer time period estimates of a sustainable population-carrying capacity estimates-are made. Three different approximations of carrying capacity are defmed: a maximum carrying capacity equal to the longterm (400-year) mean; a critical carrying capacity equal to the long-term (400-year) minimum annual value; and an optimal carrying capacity range equal to some value between 20% and 60% of the long-term mean value. SUMMARY The preceding chapters describe a method of reconstructing paleoclimate for the late Anasazi occupation period of southwestern Colorado (A.D. 901-13(0) and the predicted influence of climatic variation on dryland agricultural potential. Tree-ring data, strongly reflecting the regional climate of the southeastern portion of the Colorado Plateaus are used to retrodict 1070 years (A.D. 901-1970) of Palmer Drought Severity Indices (PDSI) for the month of June for local soils and local elevational settings. Reconstructed PDSI values associated with each soil and each 4-ha unit of space in the 1470.36-k.m2 study area are reexpressed as potential maize yields. The integration, quantification, and visual display of these productivity values are coordinated through geographic information system (GIS) technology. The method results in the production of 1) annual maps depicting the variable character of the potential agricultural environment and 2) annual values for total maize productivity (TOTPROD) that can be translated into the number (POPNUM) and density (pOPKM) of people that can be supported on that yield. Three different estimates for population are provided, representing a population that annually demands the equivalent of either one, two, or three years of maize in storage at the end of harvest (pOP1YR, The conservative POP2YR value, representing a population requiring two years of maize to be placed in storage at the end of harvest, and the critical carrying capacity value, representing the maximum population size/density that can be sustained during the lowest producing year(s) of the 400-year record, are used explore archaeological questions at three spatiallevels. First, they are used to question whether maize production in the study area was sufficient to sustain a large number of people during the 400 years examined, and to examine the proposition that climatic fluctuation resulted in the collapse of Anasazi agricultural systems and directly promoted the thirteenth century abandonment. Second, they are used to assess the productivity and predictability of maize agriculture in two localities within the study area, and to compare these model-derived estimates of carrying capacity with archaeological estimates of population in order to determine whether the productive limits of more localized areas had been reached at different times and different places. Third, they are used to evaluate and compare the productive catchments of eight tree-ring dated sites and to suggest the climatic and productive conditions under which habitation sites are established and maintained. While each of these explorations of the re- suiting database was used to address different archaeological questions, it is clear from all of them that productivity varied considerably from place­to­place and from year­to­year within the study area, but also that there was always enough productive land to produce enough maize to support a very large population (for example, 31,363 persons at a density of 21 persons/km2 in the 1470.36 km 2 study area over the 400­year period), even in the relatively dry times of the middle twelfth and late thirteenth centuries. If mobility and access to productive land were not restricted, or if redistribution systems were in place to support dispersed populations or uneven production, then the productive environment always could have sustained many people, even during the so­called Great Drought of A.D. 1276­1299. If, however, mobility and access to productive resources were severely restricted, and extensive extra­community food sharing was not regularly practiced, then there would have been times when some populations who were confined to living in certain places were, by virtue of their size, characterized by a demand for maize that was not met by local annual supply. Nevertheless, it is important to emphasize that there were always locations somewhere within the study area that could produce adequate maize crops, and at no time was the "potential dry farming belt" (petersen 1988) completely pinched out by climatic fluctuations. In other words, climatic fluctuations as they affected crop production in and of themselves cannot be used as the sole and sufficient cause for the total abandonment of Northern San Juan Region at the end of the thirteenth century. . tern of annual growth discerned in tree­ring data. Further examination of the patterning is necessary before this correlation can be ascertained, but if this were true it may imply that the prehistoric Pueblo farmers of the Mesa Verde area were not willing to work hard for only very minimal returns, even though they could have met their annual requirements by working more land of lower productive potential. Instead of turning to less productive soils, it is possible that they moved to places where high productivity was more predictable, which occasioned both local abandonments and relocations within the region, and ultimately the regional abandonment of the Northern San Juan District. Should this type of cultural selection for only the best lands have occurred, then estimates of total maize productivity should only be drawn from the highest yield categories. This would, of course, result in lower population estimates and carrying capacity values than the preceding study has indicated. An alternative to this possibility is that environmental resources other than agricultural productivity were the limiting factors in sustaining a large population in the study area Critical environmental resources that have been suggested elsewhere as potentially limiting include potable water shortages (Herold 1961), wood­resource depletion (Kohler and Matthews 1988), soil nutrient depletion in pinyon­juniper woodland zones (Matson et al. 1988, Kohler 1992), and animal protein deficiency (Speth and Scott 1989). Whereas the shortage of drinking water can be linked to meteorological conditions, shortages of lumber and fuel, shortages of animal protein, and depletion of soil nutrients are probably more closely linked to human overuse and poor management practices than limits imposed by the natural environment. Should one or several of these factors be the limiting conditions for sustaining large number of people, then the population estimates currently generated by the model will exceed the true maximum value. The implications of the results are several, given the validity of the model for reconstructing prehistoric agricultural productivity. If Anasazi populations were aware of the differential productivity of places on the landscape and selected for those locations that would consistently produce good yields of maize, as the site catcmnent studies show, then it might indicate that populations only considered the most arable soils (Le., soils that would be classified as moderate­to­high and high yield by the model) as worth farming. Tantalizing hints of this possibility exist in the data generated by the model; the annual contribution of moderate­to­high yield soils seems to mimic the pat­ Yet another possibility is that non­environmental factors (i.e., social or cultural factors) are responsible for the ultimate abandonment of the region. It is beyond the scope of this study to explore what some of these non­environmental factors might be (see, e.g., Adams 1991). It is likely, however, that some 188 combination of factors is responsible for the depopulation of the region in the late thirteenth century. Although this study does not identify what factor or combination of factors led to the final abandonment, it does suggest that some previous assumptions about maize production cannot be used to account for this event. Second, the model is quantifiable, which allows other potential users to judge the appropriateness of the variables and values used and permits modification where these values are found to be inappropriate or inaccurate. In this regard, the model gains strength from using the most current data on the distribution and quality of soils in the study area, newly available digital elevation data, and newly compiled information on soil productivity. STRENGTHS OF THE MODEL The tree-ring-based model of agricultural productivity and sustainable population described in the preceding chapters derives strength from at least four sources. First and foremost, it is a high resolution model. Treering data provide annual information about regional patterns of precipitation and temperature that are made locally relevant by application of Palmer Drought Severity Indices (PDSI), which have been calculated for local soils and local weather stations. Further, it is recognized that tree-ring data provide the highest temporal resolution reconstructions of past environmental conditions (Dean 1988b), that PDSI are integrative measures of precipitation and temperature, which incorporate the effects of previous values on current soil moisture, and that retrodictions of PDSI produce dendroclimatological reconstructions with the highest explained variance of all reconstructed variables (Rose et al. 1982). My use of multiple local weather stations to model the local influences of elevation on precipitation and temperature patterns, and my use of multiple classes of soil moisture rather than just one as suggested in the original method (palmer 1965), is new. As such, it helps to translate regional climate into local terms. A cell size of 4 ha (200 x 200 m) contributes to the model's high spatial resolution, particularly given the size of the study area (1,816 km 2 or 701 mi 2). which required recording values for a minimum of 45,400 cells (200 rows x 227 columns). While other researchers in the Mesa Verde Region have conducted productivity studies at this scale of spatial resolution before. albeit not for such an extensive study area (e.g., Cordell 1975; Darsie 1982; Kohler et al. 1986), no one before has created such a high resolution. locally pertinent record of climatic variability as it affects agricultural production. Third, the implementation of the model by means of geographic information system technology permits the manipulation of large data sets and the visual display of results in a form that can be easily grasped and also numerically expressed. Thus. GIS serves as a management device, as well as an analytic tool, and is ideally suited to this kind of research. Conceptually, the maps produced by the GIS are similar to those produced by Dean and Robinson (1977, 1979) by simpler computer cartographic techniques, in that they both depict changing dendroclimatic values across a study area for a lengthy period of time. The maps produced in this study, however, are more useful for local analysis than the contour maps of the northern Southwest produced by Dean and Robinson because they are annual, not decadal, and because they translate regional climatic patterns .into a measure of local agricultural productivity that is more interpretable from a human behavioral point of view. Fourth, the population estimates are deliberately conservative; I consciously tried to underestimate rather than overestimate the number of people and their average per km density that could be supported in any place and at any time during the 400 years. In part this is due to my attempt to represent the behavior of a population that annually grows and stores two years of maize. This level of agricultural demand is repeated ever year, regardless of how much food may be left in storage facilities. Thus, the demand is maintained constantly and results in lower population estimates than if carryover in storage from year to year had been allOWed. Furthermore, the population density estimates are based on the total number of persons divided by the total area of land (e.g., the full 1,470.36 km 2 study area or the full area included in the localities or site 189 - catchments) after the yields from the lowest producing lands have been removed from consideration. This results in lower population density values than would have been derived if only the actual but varying area of arable land (some value less than 1470.36 km 2 in the study area, for example) had been used to calculate total arumal productivity. LIMITATIONS OF THE MODEL At least three sources of potential weakness can be identified. One pertains to the method of paleoclimatic reconstruction selected. The second pertains to data quality (inherent error and operator error), and the final pertains to potentially important variables not selected for inclusion in the model. Just as the model gains strength by using tree-ring data to reconstruct high frequency paleoclimatic variation, it is limited by the same factors that limit all tree-ring-based reconstructions. Dean (l988b:132-135) reviews these advantages and shortcomings. Perhaps the most unfortunate and unavoidable of these is that there are years when growing conditions are so limiting than very narrow rings or even missing rings (indicating no growth) are produced. This can result in the underestimation of the true severity of the climate (especially the accurate estimate of precipitation) for that year. This limitation is partially mitigated by using chronologies where the specimen depth (the number of tree-ring samples) is adequate and by using multiple chronologies, as in this study (Le., SWOLD7), where unevenness in the frequency of samples associated with each individual chronology is partially evened out by using indices from other chronologies. Similarly, tree-ring records are better at predicting low precipitation values than high precipitation values because more growth is attributed to non-climatic variables when the tree is not water-stressed. This means that reconstructions at lower elevations, which generally receive less precipitation and are warmer, are generally better than reconstructions for the highest elevations where more precipitation and cooler temperatures are typical. This dendroclimatological fact was observed in the varying r 2 values associated with the final full-period calibration equations for the 55 different re- constructions (Table 3.3). Fortunately, however, only 2.4% of the land in the study area fell into the range of elevations modeled by the highest weather station (Ft Lewis), minimizing the impact of this inherent problem. Nevertheless, it is true that no dendroclimatic reconstruction, whether from relatively low elevations on the Colorado Plateau or not, ever retrodicts 100% of the variance observed in the reconstructed variable, because a perfect correlation between tree-ring growth and climate does not exist. Finally, it must be said, that treering chronologies are poor estimators of low frequency climatic variation due to the methods used to standardize ring-width values, which is the kind of longperiodicity variation that alluvial geomorphology and palynology can detect. This method-dependent deficiency to monitor low frequency variability is not particularly problematic for this study, which instead focuses on the high frequency variation in potential annual dry-farming productivity. However, it would be a greater problem if the focus were on floodplain agriculture, for example. Second, the accuracy of the model and its results are limited by data quality. The accuracy of the reconstructions of PDSI, the estimates of bean and com yield, and the estimates .of population are dependent on the quality of the primary data described in chapter 2. If the raw data amassed by other agencies or individuals are inherently inaccurate (e.g., if soils were mapped incorrectly by Soil Conservation Service), estimates for bean yields were markedly underestimated or overestimated, the weather stations selected to represent the local expression of regional climate were not truly representative, or the Digital Elevation Models (the base maps used for this study) were inaccurate, then the resulting manipulations of these data would also be inaccurate. Similarly, if I recorded these data improperly (e.g., errors in judgment, transcription errors, encoding errors, or miscalculations) or overgeneralized certain variables by virtue of the 200 x 200 m cell size resolution (e.g., dominant soil type), then the resulting estimates will be equally compromised. While care to record and calculate values was exercised throughout this study (and on occasion monitored, as with the recording of soil map units from aerial imagery), it was not possible to control for inherent error 190 - in the primary data sources, other than acquiring official and presumably valid data kept by federal agencies. For the present, these data, particularly the soil distribution data, are taken as given, with the acknowledgment that more refined mapping may be necessary, especially for canyon topography within the study area While this is not seen as a major drawback in the search for patterning at the larger geographic scales of the study area and survey localities, it potentially is a limitation for the smaller­scale site catchment investigations. This is so because small patches of arable land may have been missed by soil scientists or ignored by me in the process of determining the dominant soil map unit per 4­ha cell if they covered less than half that area. Third, the ability of the model to accurately depict past envirorunental conditions and their influence on agricultural production may be reduced by non­inclusion of all the important variables. Certain envirorunental phenomenon that are known or have been suspected by some to be relevant to successful prehistoric maize harvests have not been incorporated in this model. The infestation of crops by insects, molds, fungi, and other parasites that can decrease yields or even totally destroy crops has not been included in the model, per se, although a percentage of crop loss is withdrawn from the potential annual yield to partially simulate this problem. Regular depletion of soil nutrients and large­scale erosion of sediments that remove soils and farmland have not been built into the model, although these are variables that might be considered in future versions of the model. However, at this time there is no evidence to indicate that nutrient depletion or major soil loss in the study area was a critical factor in diminishing production. Further, while degree of slope has been incorporated in the soil data, aspect, or the direction of slope, has not. This is a variable that could easily be incorporated into future versions of the model, particularly because of the GIS capability to calculate aspect from elevational data, but only if it was found to significantly affect yield figures. Given the dominant south and southwest aspect of the terrain in the study area as a whole, it was not considered as a major discriminating variable. Finally, the pooling effects of cold air . drainage in canyons and within the troughs of low hills and small upland basins (Adams 1979) has only partially been considered in the model. The significance of cold air drainage into topographic lows is that it reduces temperature and shortens the length of the growing season for plants. While this phenomenon was emphasized in the studies by the Dolores Archaeological Project, given the centrality of the Dolores River and its tributaries, the present study is primarily upland and not canyon or lowland topography. In addition, the reduced productivity of shallow soils, soils on steep slopes, and soils within canyons, is reflected in reduced natural potential productivity values, reduced crop yield values or low available water­holding capacity values associated with such places. These locations ultimately were included in the lowest yield category and effectively were removed from contributing (or only occasionally contributed) to the estimate for total annual productivity (TOTPROD). Therefore, the effect of cold air drainage is indirectly taken into consideration in the model. RECOMMENDATIONS FOR RJTURE RESEARCH Future promising work should include tests . for the sensitivity of the current model to changes in values, thereby assessing its robustness; refinement of the model through improved methods of estimation and more detailed (or additional) data; and continued exploration of the archaeological record with the extant database. Perhaps the most obvious study is a thorough examination of the values used in the current model to assess the strength or robustness of the results. Through repeated runs of the model, where values used to represent critical variables are changed one at a time, it will be possible to determine how great or how small a change is required to substantially alter the basic conclusion of the study. Given that the agricultural productivity of the study area is always high enough to support a minimum of 21 persons/km2 and that this is the more conservative POP2YR value, I believe that the model is quite robust In other words, the model produces results that are solid and will not wildly fluctuate given small changes in 191 - critical variables. Nevertheless, sensitivity analysis should be done to add confidence to the conclusions reached in this research. Refinements to the basic model, once tested and confirmed acceptable, could result from experimentation with the use of different values, which might lead to higher levels of variance explained in the reconstructed environmental variables. For example, instead of using the instrumented June PDSI values, which depict the state of soil moisture conditions as they exist on July 1, July PDSI values, which depict conditions on August 1, might produce higher correlations between tree-ring data and PDSI values and better reflect water stress to maize plants during a critical stage of their development. My reconstructions are based on June PDSI values. The variance (r2 ) in June PDSI values explained by the tree-ring data is very similar to variance explained for July PDSI values derived by Rose et al. 1982 for the southwestern Colorado and by Rose for the Zuni area (personal communication, April 1989). Further, my June PDSI reconstructions closely mirror qualitative commentary drawn from historic sources for the study area Nevertheless, it would be useful to quantify the different results attributed to the use of one month or the other, although I suspect the variance explained will be quite similar given high autocorrelation from one month to the next. Experimentation with different values related to agricultural demand, as well as to agricultural supply, could lead to an appreciation for the tolerances of the resource-topopulation relationship. For example, it would be instructive to reduce the portion of the total land area potentially available to produce crops, perhaps incrementally through time, to simulate what might be the carrying capacity ramifications of an increasingly deteriorating environment for agriculture. Alternatively, crop yield values could be modified to simulate lower prehistoric yields than those currently estimated, particularly if experimental garden research using native seed could demonstrate markedly different production values. Improvements to the database are always warranted, and that is particularly true for the mapping of soils. A more detailed survey to identify and chart the location of distinct soil . types and measure natural and agriCUltural productivity as a check on the values used in this study is recommended. This would be useful in locations known to be characterized by significant archaeological remains so as to compare the distribution of soils of variable quality to the distribution of permanent habitations and seasonal field houses. The recalculation of PDSI values should be accomplished at some future date when treering chronologies have been extended to incorporate more years beyond the current upper limit of A.D. 1970. It would be ideal if extremely dry years in southwestern Colorado (such as 1989, a year drier than recorded historically by any of the five weather stations used in this study) could be incorporated into the instrumented PDSI calculations. If more extremes in climatic fluctuation could be incorporated into the calibration data set, better long-term reconstructions would result. Finally, additional explorations of the archaeological record using these data can be envisioned. If archaeological sites dated by surface ceramic assemblages can be better dated than is currently the practice and assigned calendrical dates rather than portions of the Pecos Classification system (e.g., A.D. . 1050-1075 versus late "Pueblo II"), then it should be possible to closely examine the distribution of settlements of various time periods with respect to the distribution of productive land. This would be done in much the same way that I examined the tree-ring-dated archaeological sites. Should the pattern of site establishment, site use, and site abandonment be found to reflect the temporal and spatial characteristics of productive agricultural land, as suggested by the locality and site catchment studies reported herein, then it will be possible to predict similar patterns for other areas and other places in the Anasazi world. Further, when clusters of settlements can be associated as components of prehistoric communities and the spatial distribution of various prehistoric communities can be identified, then it may be possible to infer which communities coalesced when and where on the basis of access to and control of productive lands. Kohler (1989) and Adler (1990) have suggested recently that competition for and control of productive land may be the key in the formation of 192 - communities and in aggregation. The data contained in this model may provide the backgroWld infonnation to test this idea. In this way, it may be possible to reconstruct the territories of contemporary commWlities and gain insight into the dynamics of intra- and intercommunity cooperation or strife. The maps of annual productivity should be examined closely to identify the best patches of agricultural land in those years or periods when productivity is low or below the longterm conditions for the study area. By this method, the most optimal areas for settlement in stressful times can be established and archaeological survey can be conducted to test the proposal that productive and predictable land was indeed a significant draw in stressful times. In order to further explore the utility of the data, the model could be applied to the well studied and nearby Wetherill Mesa, for which extensive survey has been accomplished - . (Hayes 1964) and where simulations of site establishment and abandonment have already been conducted (Cordell 1975). Comparison of the estimates of population and patterns of settlement should be interesting and informative. It may also be possible to refine the dating of poorly dated sites by suggesting the most likely occupational intervals based on the reconstruction of the temporal variability in agricultural productivity in the vicinity of those sites. In conclusion, it is probably fair to say that this study provides a defensible approach to modeling prehistoric agricultural productivity, one that can be tested, refined, and augmented by future research. As such, it offers a solid base upon which subsequent work may be built and may bring a sense of methodological order to studies of prehistoric settlements and human communities. 193- Page Blank in Original BIBLIOGRAPHY Adams,E. C. 1979 Cold Air Drainage and Length of Growing Season in the Hopi Mesas Area In The Kiva 44:285­296. 1984 Preliminary Report of Work Accomplished During 1983 Under State Permit Nos. 83-8 and 83­9. Crow Canyon Archaeological Center. Submitted to the Office of the State Archaeologist. Colorado Historical Society. Denver. 1985 Annual Report of Test Excavations at 5M1765, Sand Canyon Pueblo, and Archaeological Survey in T36N, R18W, Sections 12, and 24, and T36N, R16W, Sections 29 and 30. Crow Canyon Archaeological Center. Submitted to the Office of the State Archaeologist. Colorado Historical Society, Denver. 1991 The Origin and Development of the Pueblo Katsina Cult. The University of Arizona Press, Tucson. Adams, K. R. 1980 Relative Numbers of Native Macrofossils in Strata of Poor Preservation, with Emphasis on Flotation. In Pan 7: Salmon Ruin Ethnobotanical Report, by V. L. Bohrer. In investigations at the Salmon Site: the Structure of Chacoan Society in the Northern Southwest ill, edited by C. Irwin­Williams and P. H. Shelley, pp. 251­301. Eastern New Mexico University, Portales. Adler, M. A. 1988 Archaeological Survey and Testing in the Sand Canyon PueblolGoodmon Point Ruin Locality, Montezuma County, Colorado, 1987 Field Season. Crow Canyon Archaeological Center. Submitted to the Bureau of Land Management. San Juan Resource Area, Durango and Office of the State Archaeologist, Colorado Historical Society, Denver. 1990 Population Aggregation and the Anasazi Social Landscape: The View From the Four Comers. Paper presented at the 1990 Southwest Symposium, Albuquerque. Adler, M. A., and M. Metcalf 1991 Draft Report on Archaeological Survey of Lower East Rock and Sand Canyons, Montezuma County, Colorado. Crow Canyon Archaeological Center, Cortez, Colorado. Report submitted to the Bureau of Land Management, San Juan Resource Area Office, Durango, Colorado. Ahlstrom. R. V. N. 1985 The Interpretation ofArchaeological TreeRing Dates. Ph.D. dissertation, University of Arizona, University Microfllms, Ann Arbor. Allen, K. M. S., S. W. Green, and E. B. W. Zubrow 1990 Interpreting Space: GIS in Archeology. Taylor and Francis, London. Arrhenius, G., and E. Bonatti 1965 The Mesa Verde Loess. In Contributions of the Wetherill Mesa Archaeological Project. assembled by D. Osborne, pp. 92-101. Memoirs of the Society for American Archaeology No. 19. Washington, D.C. Aronoff, S. 1989 Geographic Information Systems: A Management Perspective. WDL Publications, Ottawa Avery, T. E., and G. L. Berlin 1985 Interpretation ofAerial Photographs. 4th ed. Burgess, Mirmeapolis. Bailey, R., S. Hackenberger, D. Howes, and D. Wherry 1985 Geographic Information Processing in Land Use Modeling and Testing in the Columbia River Basin. Paper presented at the 50th Annual Meeting of the Society for American Archaeology, Denver. Beaglehole, E. 1937 Notes on Hopi Economic Life. Yale University Publications in Anthropology 15. New Haven. Berry,J. K. 1989a Beyond Mapping. GIS World 2(3):8, 3839. 1989b Beyond Mapping. GiS World 2(4):34, 47-48. Betancourt, J. L., and O. K. Davis 1984 Packrat Middens from Canyon de Chelly, Northeastern Arizona: Paleoecological and Archaeological Implications. Quaternary Research 21:56-64. Ones. 2d ed., edited by J. H. Berger and E. F. Berger, pp. 122-127. Interdisciplinary Supplemental Educational Programs, Cortez, Colorado. Betancourt, J. L., and T. R. Van Devender 1981 Holocene Vegetation in Chaco Canyon, New Mexico. Science 214:656-658. 1986 1985 Annual Report of Test Excavations al Sand Canyon Pueblo (5M1765). Crow Canyon Archaeological Center, Cortez, Colorado. Submitted to Bureau of Land Management, Durango, Colorado. Blasing, T. J. 1975 Methods for Analyzing Climatic Variations in the Nonh Pacific Sector and Western North America for the Last Few Centuries. Ph. D. dissertation, University of Wisconsin, University Microfilms, Ann Arbor. 1987 Annual Report of Excavations at Sand Canyon Pueblo (5M1765). Montezuma County. Colorado. 1986 Field Season. Crow Canyon Archaeological Center, Cortez Colorado. Submitted to Bureau of Land Management, Durango, Colorado. Bohrer, V. L. 1980 Part 7: Salmon Ruin Ethnobotanical Report. In Investigations at the Salmon Site: The Structure of Chacoan Society in the Northern Southwest Jl/. edited by C. IrwinWilliams and P. H. Shelley, pp. 163-535. Eastern New Mexico University, Portales. 1988a Annual Report on the Excavations at Sand Canyon Pueblo, 1987 Field Season. Crow Canyon Archaeological Center, Cortez, Colorado. Submitted to Bureau of Land Management, Durango, Colorado. 1981 Methods of Recognizing Cultural Activity From Pollen in Archaeological sites. The Kiva 46:135-142. 1988b Wallace Ruin Interim Report. Southwestern Lore 54(2):8-33. Bohrer, V. L., and K. R. Adams 1977 Ethnobotanical Techniques and Approaches at Salmon Ruin. New Mexico. San Juan Valley Archaeological Project, Technical Series No.2. Contributions in Anthropology, Vol. 8, No.1. Eastern New Mexico University, Portales. Bolton, H. E. 1972 Pageant in the Wilderness: The Story of the Escalante Expedition to the Interior Basin. 1776. Utah Slate Historical Society, Salt Lake City. Bracken, I., G. Higgs, D. Martin, and C. Webster 1989 A Classification of Geographical Systems Literature and Applications. Environmental Publications, University of East Anglia, Norwich. Bradfield, M 1971 The Changing Pattern ofHopi AgriCUlture. Royal Anthropological Institute Occasional Paper 30. London. Bradley, B. A. 1974 Preliminary Report of Excavations at Wallace Ruin, 1969-1974. Southwestern Lore 40(3 & 4):63-71. 1984 The Wallace Ruin and the Chaco Phenomenon. In 1nsighJs 1nto the Ancient - 1992 Excavations at Sand Canyon Pueblo. In The Sand Canyon Archaeological Project: A Progress Report, edited by W. D. Lipe, pp. 79-99. Occasional Paper No.2. Crow Canyon Archaeological Cemer, Cortez, Colorado. . Brandegee, T. S. 1976 The Flora of Southwestern Colorado. The United States Geological and Geographical Survey Bulletin 3(3):227-248. Breternitz, D. A. 1986 Notes on Early Basketmaker ill sites in Mancos Canyon, Colorado. The Kiva 51:263264. Breternitz, D. A., and D. W. Martin 1973 Report of the Dolores River Project Archaeological Reconnaissance, 1972-1973. Ms. on file, Midwest Archaeological Center, National Pm Service, Lincoln, Nebraska Bretemitz, D. A., C. K. Robinson, and E. T. Gross (compilers) 1986 Dolores Archaeological Program: Final Synthetic Report. Bureau of Reclamation, Engineering and Research Center, Denver. Brew,J. O. 1946 The Archaeology of Alkali Ridge, Southeastern Utah. Papers of the Peabody 196- Museum of American Archaeology and Ethnology No. 21. Harvard University, Cambridge, Massachusetts. Brisbin, J. M., and C. J. Brisbin 1973 Contribution to the Sites Being Excavated in the North McElmo Drainage Area of Southwestern Colorado. Ms. on file, Crow Canyon Archaeological Center, Cortez, Colorado. Briuer, F. L. 1988 Geographic Information Systems and Automated Archaeological Information: Investments and Dividends for Archaeological Resource Management Paper presented at the 53rd Annual Meeting of the Society for American Archaeology, Phoenix. Brown, J. 1975 Preliminary Report on the 1975 Excavations at Site 5MT3 Near Yellowjacket, Colorado. Southwestern Lore 41(4):47-50. Brown,L. A. 1938 A Basisfor Rating the Productivity of Soils on the Plains ofEastern Colorado. Colorado Experimental Station Technical Bulletin 25. Colorado State College, Ft. Collins. Bullard, W. R. 1962 The Cerro Colorado Site and Pithouse Architecture in the Southwestern United States Prior to A.D. 900. Papers of the Peabody Museum of Archaeology and Ethnology No. 44. Harvard University, Cambridge, Massachusetts. Bums,B. T. 1976 The "Abandonment Controversy" and a Proposal for its Partial Resolution. Preliminary Exam Paper, Department of Anthropology, University of Arizona. Ms. on file, Arizona State Museum Library, University of Arizona, Tucson. 1983 Simulated Anasazi Storage Behavior Using Crop Yields Reconstructed From Tree-Rings: A.D. 652-1968. 2 vols. Ph.D. dissertation, University of Arizona, University Microfilms, Ann Arbor. Burrough, P. A. 1986 Principles of Geographic Information Systems for Land Resource Assessment. Oxford University Press, Oxford. Calamia, M. A. 1986 Geographic Information System Applicationsfor Cultural Resource Management. Bureau of Land Management, Denver. Brown, P. E., and B. H. Rubin 1982 Patterns of Desert Resource Use: An Integrated Approach to Settlement Analysis. In Granite Reef" A Study in Desert Archaeology, edited by P. E. Brown and C. L. Stone, pp. 267­305. Arizona State University Anthropological Research Papers No. 28. Tempe. Bryan,K. 1925 Dates of Channel Trenching (ArroyoCutting) in the Arid Southwest Science 62:338-344. 1941 Pre-Columbian Agriculture in the Southwest, as Conditioned by Periods of Alluviation. Annals of the Association of American Geographers 31:219--242. 1954 The Geology of Chaco Canyon. New Mexico. in Relation to the Life and Remains of the Prehistoric Peoples ofPueblo Bonito. Smithsonian Miscellaneous Collections, Vol. 122, No.7. Smithsonian Institution, Washington, D.C. - Carmichael, D. L., and D. Christenson 1988 Predictive Modeling of Prehistoric Site Distribution in Central Montana. Paper presented at the 53rd Annual Meeting of the Society for American Archaeology, Denver. CASA 1981 Cultural Resource Inventory of the Shell Mobile C02 Wet Gathering System, Montezuma and Dolores Counties. Colorado. Complete Archaeological Services Associates, Cortez, Colorado. Submitted to Bureau of Land Management, San Juan Resource Area. Durango. Cattanach, G. S., Jr. 1980 Long House. Mesa Verde National Park. Colorado. Publications in Archeology NO.7-H. National Park Service, Washington, D.C. Chandler, S. M., A. D. Reed, and P. R. Nickens 1980 Ecological Variability and Archaeological Site Location in Southwestern Colorado: The Class II Cultural Resource Inventory of the Bureau ofLand Management's Sacred Mountain Planning Unit. Centuries Research, 197- Montrose, Colorado. Submitted to the Bureau of Land Management, Montrose District Office, Montrose, Colorado. Clarke, K. C. 1986 Advances in Geographic Information Systems. Computers, Environment and Urban Systems 10:175-184. Collins, S. H., and G. C. Moon 1981 Algorithms for Dense Digital Terrain Models. Photogrammetric Engineering and Remote Sensing 47:71-76. Connolly, M. 1990 A Historic Land Use Study of Goodman Point. Ms. on file, Crow Canyon Archaeological Center, Cortez, Colorado. Creamer, W. 1985 The Upper Klethla Valley: Computer Generated Maps of Site Location. Paper presented at the 50th Annual Meeting of the Society for American Archaeology, Denver. Darsie, R F., III 1982 Prehistoric Agricultural Potential and Site Location in the Dolores River Valley, Southwest Colorado. Unpublished M.A. thesis, Department of Anthropology, Washington State University, Pullman. Dean, J. S. 1988a A Model of Anasazi Behavioral Adaptation. In The Anasazi in a Chonging Environment, edited by G. J. Gumerman, pp. 25-44. Cambridge University Press, Cambridge. 1988b Dendrochronological and Paleoenvironmental Reconstruction on the Colorado Plateaus. In The Anasazi in a Chonging Environment, edited by G. J. Gumerman, pp. 119-167. Cambridge University Press, Cambridge. Cook, E. R 1982 A Long-Term Drought Sequence for the Hudson Valley, New York. In Climate from Tree Rings, edited by M. K. Hughes, P. M. Kelly, J. R Pilcher, and V. C. LaMarche, Jr., pp. 163-165. Cambridge University Press, Cambridge. Cook, E. R, and G. C. Jacoby 1977 Tree Ring Drought Relationships in the Hudson Valley, New York. Science 198:399401. Cook, S. F., and W. Borah 1979 Essays in Population History: Mexico and California. University of California Press, Berkeley. Dean, J. S., and W. J. Robinson 1977 Dendroclimatic Variability in the American Southwest A.D. 680 to 1970. Laboratory of Tree-Ring Research, University of Arizona Submitted to the National Park Service, Interagency Archeological Services, Denver. U.S. Department of Commerce, National Technical Information Service PB-266 340. 1978 Expanded Tree Ring Chronologiesfor the Southwestern United States. Chronology Series 3. Laboratory of Tree-Ring Research, The University of Arizona, Tucson. Cordell. L. S. 1975 Predicting Site Abandonment at Wetherill Mesa. The Kiva 40:189-202. 1979 Computer Cartography and the Reconstruction of Dendroclimatic Variability in the American Southwest, A.D. 680 to 1970. In Statistical Cartographic Applications to Spatial Analyses in Archoeological Contexts, edited by S. Upham, pp. 79-94. Anthropological Research Papers No. 15. Arizona State University, Tempe. 1984 Prehistory of the Southwest. Academic Press, New York. Cole, S. J. 1989 Iconography and Symbolism in Basketmaker Rock Art. In Rock Art of the Western Canyons, edited by J. S. Day, P. D. Friedman, and M J. Tate, pp. 59--85. Colorado Archaeological Society Memoir No.3. Denver. Cowen, D. J. 1988 GIS versus CAD versus DBMS: What Are the Differences? Photogrammetric Engineering and Remote Sensing 54:1551-1555. Dean, J. S., R C. Euler, G. J., Gumerman, F. Plog, R H. Hevly and T. N. V. Karlstrom 1985 Human Behavior, Demography, and Paleoenvironment on the Colorado Plateaus. American Antiquity 50:537-554. Debloois, E. I., and D. F. Green 1978 SARG Research on the Elk Ridge Project, Manti-La Sal National Forest, Utah. In - 198- Investigations of the Southwestern Anthropological Research Group: The Proceedings of the 1976 Conference, edited by R. C. Euler and G. J. Gumennan, pp. 13­24. Museum of Northern Arizona, Flagstaff. Douglass, A. E. 1929 The Secret of the Southwest Solved by Talkative Tree­Rings. National Geographic Magazine 54:737-770. Douthit, M. L. 1984 Excavations at Site 5MT3777 and Site 5MT3778, Casa de Suenos. Technical Report, August 21,1984. Bureau of Land Management, Anasazi Heritage Center, Dolores, Colorado. Dueker, K. J. 1987 Geographic Infonnation Systems and Computer-Aided Mapping. Journal of the American Planning Association 53:383-390. Duff, U. Francis 1904 Some Exploded Theories Concerning Southwestern Archeology. American Anthropologist [n.s.] 6:303-306. U.S.G.S. National Cartographic Infonnation Center, Reston, Virginia. Erdman, J. A., C. L. Douglas, and J. W. Marr 1969 Environment of the Mesa Verde, Colorado. Archeological Research Series No. 7-B. National Park Service, Washington, D. C. Euler, R. C., G. J. Gumennan, T. N. V. Karlstrom, J. S. Dean, and R. H. Hevly 1979 The Colorado Plateaus: Cultural Dynamics and Paleoenvironment. Science 205:10891101. Farley, J., W. F. Limp, and J. Lockhart 1990 Theoretician's Workbench: Integrating GIS, Remote Sensing, EDA, and Database Management In Interpreting Space: GIS and Archaeology, pp. 141-164. Taylor and Francis, London. Fettennan, J., and L. Honeycutt 1980 Cultural Resource Inventory, MAPCO's Liquid Hydrocarbons Pipeline. Southwestern Colorado. Bmeau of Anthropological Research, Dove Creek, Colorado. Submitted to Woodward-Clyde Consultants, San Francisco. 1982a Cultural Resources Inventory, Shell C02 Laterals and Shell Oil Flowlines. Bureau of Anthropological Research. Submitted to Woodward-Clyde Consultants, San Francisco. Dykeman, D. D. 1986 Excavations at 5MT8371 , An Isolated Pueblo II Pithouse Structure in Montezuma County, Colorado. Studies in Archaeology NO.2. Division of Conservation Archaeology, San Juan County Museum Association, Fannington, New Mexico. 1982b Testing and Excavation Report, MAPCO's Rocky Mountain Liquid Hydrocarbons Pipeline, Southwest Colorado. Bureau of Anthropological Research, Dove Creek, Colorado. Submitted to Woodward-Clyde Consultants, San Francisco. EddY,F. W. 1966 Prehistory of the Navajo Reservoir District, Northwestern New Mexico. Museum of New Mexico Papers in Anthropology Vo1.15, Nos. 1 and 2. Santa Fe. Eddy, F. W., A. E. Kane, and P. R. Nickens 1984 Southwest Colorado Prehistoric Context: Archaeological Background and Research Directions. Office of Archaeology and Historic Preservation, Colorado Historical Society, Denver. Ekren, E. B., and F. N. Houser 1965 Geology and Petrology of the Ute Mountains Area Colorado. Geological Survey Professional Paper 481. U.S. Government Printing Office, Washington, D.C. Elassal, A. A., and V. M. Caruso 1983 Digital Elevation Models. U.S. Geological Survey Circular 895-B. User Services Section, - 1987 The Mockingbird Mesa Survey. Cultural Resource Series No. 22. Bureau of Land Management, Denver, Colorado. Fewkes, J. W. 1919 Prehistoric Villages, Castles and Towers of Southwestern Colorado. Bmeau of American Ethnology Bulletin No. 70. Washington, D.C. Flint, P. R., and S. W. Neusius 1987 Cottontail Procurement among Dolores Anasazi. In Dolores Archaeological Program: Supporting Studies: Settlement and Environment, compiled by K. L. Petersen and J. D. Orcutt, pp. 257-273. Bureau of Reclamation, Engineering and Research Center, Denver. 199- Ford, R. I. 1968 An Ecological Analysis Involving the Population of San Juan Pueblo. New Mexico. Ph.D. dissertation, University of Michigan. University Microfilms, Ann Arbor. 1987 Cultural Resource Inventoriesfor the Dolores Project: The Dove Creek Canal Distribution System and Dawson Draw Reservoir. Four Comers Archaeological Project Report NO.7. Complete Archaeological Service Associates, Cortez, Colorado. Submitted to Bureau of Reclamation, Upper Colorado Region, Salt Lake City. Forde, C. D. 1931 Hopi Agriculture and Land Ownership. Journal of Royal Anthropological Institute 61:357­406. 1988 Cultural Resource Inventoriesfor the Dolores Project: The Ute Irrigated Lands Survey. Four Comers Archaeological Project Report No. 13. Complete Archaeological Service Associates, Cortez, Colorado. Submitted to Bureau of Reclamation, Upper Colorado Region, Salt Lake City. Forrest, E., G. E. Montgomery, and G. M. Juhl 1990 Intelligent Infrastructure Workbook: A Management-Level Primer on GIS. A­E­C Automation Newsletter, Fountain Hills, Arizona. Freeman, I. S. 1958 A History of Montezuma County Colorado. Johnson Publishing, Boulder, Colorado. Fritts, H. C. 1965 Tree­Ring Evidence for Climatic Changes in Western North America. Monthly Weather Review 93:421-443. 1976 Tree Rings and Climate. Academic Press, London. Fritts, H. C., T. J. Blasing, B. P. Hayden, and J. E. KUlZbach 1971 Multivariate Techniques for Specifying Tree Growth and Climatic Relations and for Reconstructing Anomalies in Paleoclimate. Journal of Applied Meteorology 10:84--864. Gill, S. J., and D. Howes 1985 A GIS Approach to the Use of Surface Samples in Intra-Site Distributional Analysis. Paper presented at the International Symposiwn on Data Management and Mathematical Methods in Archaeology and International Union of Pre- and Protohistoric Sciences, Denver. Gillespie, W. 1976 Culture Change at the Ute Canyon Site: A Study of the Pithouse-Kiva Transition in the Mesa Verde Region. Unpublished M.A. thesis, Department of Anthropology, University of Colorado, Boulder. GIS World 1990 The 1990 GIS Sourcebook. 1st edition, GIS World, Fort Collins, Colorado. Fritts, H. C., G. R. Lofgren, and G. A. Gordon 1981 Past Climate Reconstructed from Tree Rings. In Climate and History: Studies in Interdisciplinary History. edited by R. I. Rotberg and T. K. Rabb. pp. 193­213. Princeton University Press, Princeton, New Jersey. Gleichrnan, C., and P. Gleichman 1992 The Lower Sand Canyon Survey. In The Sand Canyon Archaeological Project: A Progress Report, edited by W. D. Lipe, pp. 25-31. Occasional Paper No.2. Crow Canyon Archaeological Center, Cortez. Fuller, S. L. 1985a Dove Creek Canal, Reach 2 Research Design. Four Comers Archaeological Project Report No.1. Complete Archaeological Service Associates, Cortez, Colorado. Submitted to Bureau of Reclamation, Upper Colorado Region, Salt Lake City. Gould, R. R. 1982 The Mustoe Site: The Application of Neutron Activation Analysis in the Interpretation of a Multi-Component Archaeological Site. Unpublished Ph. D. dissertation, Department of Anthropology, University of Texas, Austin. 1985b South Canal Research Design. Four Comers Archaeological Project Report No.3. Complete Archaeological Service Associates. Cortez, Colorado. Submitted to Bureau of Reclamation, Upper Colorado Region, Salt Lake City. - Graybill, D. A. 1989 The Reconstruction of Prehistoric Stream Flow. In The 1982-1984 Excavations at Las Colinas. Environment and Subsistence. by D.A. Graybill, D. A. Gregory, F. L. Nials, S. K. Fish, C. H. Miksicek, R. E. Gasser, and 200- C. R. Szuter, pp. 25­38. Archaeological Series 162, Vol. 5, Pt. 1. Cultural Resource Management Division, Arizona State Museum, University of Arizona, Tucson. Hallisy, S. J. 1974 Salvage Excavations at Sites 5MTUMR2344 and 5MTIJMR.2347 in Mancos Canyon, Southwest Colorado. Submitted to Bureau of Indian Affairs, Albuquerque Area Office, Albuquerque. Gross, G. T. 1985 Anasazi Storage Facilities in the Dolores Region. Colorado. A.D. 600-920. Ph.D. dissertation, Deparunent of Anthropology, Washington State University, Pullman. Hammack, L. C. 1983 Cultural Resource Inventory: Shell Mobil C02 Project 25 KV Electric Distribution System. Yellowjacket and Hovenweep Areas. Complete Archaeological Services Associates, Cortez, Colorado. Submitted to Bureau of Land Management, San Juan Area, Durango, Colorado. Gross, G. T., and D. Howes 1986 Cougar Springs Cave (Site 5MT4797). In Dolores Archaeological Program: Anasazi Communities at Dolores: Early Small Settlements in the Dolores River Canyon and Western Sagehen Flats Area, compiled by T. A. Kohler, W. D. Lipe, and A. E. Kane, pp. 80--1 DO. Bureau of Reclamation, Engineering and Research Center, Denver. Hammack, N. S. 1984 Archaeological Testing and Monitoring. ShelllMobile C02 Yellow Jacket and Hovenweep Systems. Montezuma County, Colorado. Complete Archaeological Service Associates, Cortez, Colorado. Submitted to Bureau of Land Management, San Juan Area, Durango, Colorado. Gumennan, G. J. (editor) 1988 The Anasazi in a Changing Environment. Cambridge University Press, Cambridge. Hammett, J. E., and N. H. Olsen (editors) 1984 Hovenweep 1977. Archeological Report No.4. Office of Sponsored Programs, San Jose State University, San Jose, California. Haase, W. R., IV 1983 Pueblo II and Pueblo JJJ Settlement Patterns on Cedar Mesa, Southwestern Utah. Unpublished MA. thesis, Department of Anthropology, Washington State University, Pullman. Hamon, C. E. 1983 Detennining Capability Class of Soils. SCS Technical Notes. Agronomy Note No. 56, 3rd rev., July. Haase, W. R. IV 1985 Domestic Water Conservation Among the Northern San Jan Anasazi. Southwestern Lore 51(2):15-27. Harris, A. H. 1970 Past Climate of the Navajo Reservoir District American Antiquity 35:374-377. Hack, J. T. 1942 The Changing Physical Environment of the Hopi Indians ofArizona. Papers of the Peabody Museum of American Archaeology and Ethnology, Vol. 35, No.1. Cambridge, Massachusetts. Hart, J. A., and D. B. Wherry 1984 VICAR/IBIS User Reference Manual. Washington State University Digital Image Analysis Laboratory, Computing Service Center, Pullman. Hall, S. A. 1977 Late Quaternary Sedimentation and Paleoecologic History of Chaco Canyon, New Mexico. Geologic Society of America Bulletin 88: 1593-1618. Hal1asi, J. A. 1979 Part II: Archaeological Excavation at the Escalante Site, Dolores, Colorado, 1975 and 1976. In The Archaeology and Stabilization of the Dominguez and Escalante Ruins, pp. 197425. Cultural Resource Series No.7. Colorado State Office, Bureau of Land Management, Denver. - Hasenstab, R. J. 1983a A Preliminary Cultural Resource Sensitivity Analysis for the Proposed Flood Control Facilities Construction in the Passaic River Basin ofNew Jersey. Report submitted for the Passaic River Basin Special Studies Branch, Department of the Anny, New York District, Anny Corps of Engineers. 201- 1983b The Application of Geographic Information Systems to the Analysis of Archaeological Site Distribution. Paper presented at the 48th Annual Meeting of the Society for American Archaeology, Pittsburgh. Summers of 1875 and 1876. In Tenth Annual Report of the U.S. Geological and Geographical Survey of the Territories, Embracing Colorado and Parts ofAdjacent Territories, Being a Report of Progress of the Exploration for the Year 1876 (compiled) by F.V. Hayden, pp. 383-407. U. S. Government Printing Office, Washington, D.C. Hassan, F. A. 1981 Demographic Archaeology. Academic Press, New York. Hastorf, C. 1980 Changing Resource Use in Subsistence Agricultural Groups in the Prehistoric Mimbres River Valley, New Mexico. In Modeling Changes in Prehistoric Subsistence Economies, edited by T. Earle and A. Christensen, pp. 79­120. Academic Press, New York. Honeycutt, L., and J. E. Fetterman 1982 Cultural Resource Inventory of the Empire Electric 115 KV Transmission Line. Bureau of Anthropological Research, Dove Creek, Colorado. Submitted to Empire Electric Association, Cortez, Colorado. 1985 Letter Report on the Cultural Resource Inventory ofParcels Associated with the Dove Creek Canal-Reach I. Bureau of Anthropological Research, Dove Creek, Colorado. Submitted to Bureau of Reclamation, Cortez Projects Office, Cortez. Hayes, A. C. 1964 The Archeological Survey of Wetherill Mesa, Mesa Verde National Park, Colorado. Archaeological Research Series No.7­A. National Park Service, Washington, D. C. Hayes, A. C., and J. A. Lancaster 1975 Badger House Community, Mesa Verde National Park, Colorado. Publications in Archeology 7­E. National Park Service, Washington, D.C. Haynes, D. D., J. D. Vogel, and D. G. Wyant (compilers) 1972 Geology, Structure, and Uranium Deposits of the Cortez Quadrangle, Colorado and Utah. Miscellaneous Investigations Series, Map 1629. U.S. Geological Survey, Washington, D.C. Hevly, R. H. 1988 Prehistoric Vegetation and Paleoclimates on the Colorado Plateaus. In The Anasazi in a Changing Environment, edited by G. J. Gumerman, pp. 92-118. Cambridge University Press, Cambridge. Herold,J. 1961 Prehistoric Settlement and Physical Environment in the Mesa Verde Area. University of Utah Anthropological Papers No. 53. Salt Lake City. Hogan, P. F. 1987 Prehistoric Agricultural Strategies in West-Central New Mexico. Ph. D. dissertation, Washington State University, University Microf1lrns, Ann Arbor. Holmes, W. H. 1878 Report of the Ancient Ruins of Southwestern Colorado, Examined During the - Hough, W. 1915 The Hopi Indians. The Torch Press, Cedar Rapids, Iowa. 1930 Ancient Pueblo Subsistence. Proceedings of the 23rd International Congress ofAmericanists, 1928, pp. 67-69. New York. Huber, E. K. 1989 Preliminary Report of Excavations, 1987 and 1988 At Green Lizard (5MT390l), Montezuma County, Colorado. Crow Canyon Archaeological Center, Cortez, Colorado, and Washington State University, Pullman. Submitted to Bureau of Land Management, San Juan Resource Area. Durango. Huber, E. K., and W. W. Bloomer 1988 Annual Report of Investigations at Green Lizard (5MT390l), Montezuma County, Colorado. Crow Canyon Archaeological Center, Cortez, Colorado, and Washington State University, Pullman. Submitted to the Bureau of Land Management, San Juan Resomce Area, Durango, Colorado. Hughes, M K., P. M. Kelly, J. R. Pilcher, and V. C. LaMarche, Jr. (editors) 1982 Climate from Tree Rings. Cambridge University Press, Cambridge. Hunt, C. B. 1956 Cenozoic Geology ofthe Colorado Plateau. Geological Survey Professional Paper 279. 202- U.S. Government Printing Office, Washington, D.C. Management, Anasazi Heritage Center, Dolores, Colorado. Hurst, C. T., and V. F. Lotrich 1932 An Unusual Mug from Yellow Jacket Canyon. El Palacio 33(21­22): 195­198. Jackson, W. H. 1878 Report on the Ancient Ruins Examined in 1875 and 1877. In Tenth Annual Report of the U.S. Geological and Geographical Survey of the Territoriesfor 1876, pp. 411-450. Washington, D.C. 1933 Another Unusual Bowl from Yellow Jacket Canyon. El Palacio 34(15-16):111-115. Huxhold, W. E. 1990 An Introduction to Urban Geographic Information Systems. Oxford University Press, Oxford. Ives, G. A., and M. Errickson 1983 Dove Creek Canal Reach II, Fairview Pumping Plant, Cahone Pumping Plant, and a Portion of Dove Creek Canal­Reach I. Ms. on rue, Bureau of Land Management, Anasazi Heritage Center, Dolores, Colorado. Ives, G. A., and J. D. Orcutt 1981 The Dove Creek Canal, Reach I, Survey 1981. Ms. on file, Bureau of Land Management, Anasazi Heritage Center, Dolores, Colorado. Ives, J. C. 1971a Preliminary Report of Archaeological Excavations by Fort Lewis College Under Pennit #7O-C0--022 on Public Lands Under the Supervision of the Bureau of Land Management, 1970-1971. Ms. on file, Bureau of Land Management, Anasazi Heritage Center, Dolores, Colorado. 1971b Preliminary Report of Archaeological Excavations by Fort Lewis College Under Pennit #71--C0--021 on Public Lands Under the Supervision of the Bureau of Land Management, 1971. Ms. on rue, Bureau of Land Management, Anasazi Heritage Center, Dolores, Colorado. 1972 Preliminary Report of Archaeological Excavations by Fort Lewis College Under Pennit #72--C0-022 and a Survey Under Pennit #7O-C0--021 on Public Lands Under the Supervision of the Bureau of Land Management, 1972. Ms. on rue, Bureau of Land Management, Anasazi Heritage Center, Dolores, Colorado. 1973 Preliminary Report of Excavations Conducted by Fort Lewis College Under Pennit #73--C0-034. Ms. on rue, Bureau of Land - Judd, N. M. 1981 The Material Culture of Pueblo Bonito. Reprinted. Reprints in Anthropology, vol. 23. J.L. Reprints, Lincoln, Nebraska. Originally published 1954, Smithsonian Miscellaneous Collections, vol. 24. Smithsonian Institution, Washington, D.C. Kane, A. E. 1975a Archaeological Resources in the Great Cut-Dove Creek Area. Dolores Project Report of the 1974 Season. Ms. on rue, Bureau of Land Management, Anasazi Heritage Center, Dolores, Colorado. 1975b Archaeological Resources of the Dolores River Project: Report of the 1975 Field Season. Ms. on rue, Bureau of Land Management, Anasazi Heritage Center, Dolores, Colorado. Karlsb'om, T. N. V. 1988 Alluvial Chronology and Hydrologic Change of Black Mesa and Nearby Regions. In The Anasazi in a Changing Environment, edited by G. J. Gumerman, pp. 45-91. Cambridge University Press, Cambridge. Kavena, J. T. 1980 Hopi Cookery. The University of Arizona Press, Tucson. Kidder, A. V. 1927 Southwestern Archaeological Conference. Science 66:489-491. Kleidon, J., and B. Bradley 1989 Annual Report of the 1988 Excavations at Sand Canyon Pueblo (5MT765). Crow Canyon Archaeological Center, Cortez, Colorado. KoWer, T. A. 1989 Chapter 1: Introduction. In Bandelier Archaeological Excavation Project: Research Design and Summer 1988 Sampling, edited by T. A. KoWer, pp. 1-12. Reports of Investiga- 203- tions 61. Department of Anthropology, Washington Slate University, Pullman. 1992 Prehistoric Human Impact on the Environment in the Upland North American Southwest Population and Environment: A Journal ofInterdisciplinary Studies 13:255268. Kohler, T. A., and M. H. Matthews 1988 Long­tenn Anasazi Land Use and Forest Reduction: A Case Study from Southwest Colorado. American Antiquity 53:537-564. Project Report No. 11. Complete Archaeological Service Associates. Submitted to Bureau of Reclamation, Upper Colorado Region, Salt Lake City. Kvamme, K. L. 1983a Computer Processing Techniques for Regional Modeling of Archaeological Site Locations. Advances in Computer Archaeology ＱＺＲｾＵ Ｎ 1983b New Methods of Investigating the Environmental Basis of Prehistoric Site Location. Unpublished Ph.D. dissertation, Department of Anthropology, University of California, Santa Barbara Kohler, T. A., J. D. Orcutt. E. Blinman, and K. L. Petersen 1986 Anasazi Spreadsheets: The Cost of Doing Agricultural Business in Prehistoric Dolores. In The Dolores Archaeological Program: Final Synthetic Report. compiled by D. A Bretemitz, C. K. Robinson, and G. T. Gross, pp. 525-538. Bureau of Reclamation, Engineering and Research Center, Denver. 1984 Models of Prehistoric Site Location near Pinyon Canyon, Colorado. In Papers ofthe Philmont Conference on the Archaeology of Northeastern New Mexico. edited by C. J. Condie, pp. 34 -370. New Mexico Archaeological Council Proceedings No.6, Albu- querque. Kohler, T. A, and C. R. Van West 1994 The Calculus of Self Interest in the Development of Cooperation: Sociopolitical Development and Risk among the Northern Anasazi. To appear in Resources Stress, Economic Uncertainty, and Human Response in the Prehistoric Southwest, edited by J. A Tainter. Santa Fe Institute, Studies in the Science of Complexity, Addison-Wesley, Reading MA, 1994. In press. 1985 Detennining Empirical Relationships Between Natural Environmental and Prehistoric Site Locations: A Hunter-Gatherer Example. In For Concordance in Archaeological Analysis: Bridging Data Structure, Quantitative Technique. and Theory, edited by Christopher Carr, pp. 208-238. Westport Publishing, Kansas City. 1986a An Overview of Geographic Infonnation Systems for Archaeological Research and Data Management In Proceedings ofthe National Workshop on Microcomputers in Archaeology. Society for American Archaeology and Society for Professional Archaeologists, Washington, D. C. Kuckelman, K. A. 1986 Cultural Resource Inventory for the Towaoc Canal Project: Highline Ditch, Lower Hermana Lateral. Rocky Ford Laterals, and Portions of the Upper Hermana and Lone Pine Laterals. Four Comers Archaeological Project Report No.8. Complete Archaeological Service Associates. Submitted to Bureau of Reclamation, Upper Colorado Region, Salt Lake City. 1986b The Use of Geographic Infonnation Systems for Modeling Archaeological Site Distributions. In Geographic Information Systems in Government, vol. 1, edited by B. K. Opitz, pp. 345-362. Deepak Publishing, Hampton, Virginia. 1988 Excavations at Norton House (Site 5MT8839), A Pueblo IT Habitation. In Archaeological Investigations on Douth Canal, Vol. 2, pp. 373-403. Four Comers Archaeological Project Report No. 11. Complete Archaeological Service Associates. Submitted to Bureau of Reclamation, Upper Colorado Region, Salt Lake City 1988 Development and Testing of Quantitative Models. In Quantifying the Present and Predicting the Past: Theory, Method, and Application of Archaeological Predictive Modeling, edited by W. J. Judge and L. Sebastian. pp. 325-428. U.S. Government Printing Office, Washington, D.C. Kuckelrnan, K. A, and J. N. Morris 1988 Archaeological Investigations on South Canal, 2 vols. Four Comers Archaeological - 204- 1989 Geographic Infonnation Systems in Regional Archaeological Research and Data Management In Archaeological Method and Theory, Vol. 1. edited by M. B. Schiffer, pp. 139-203. The University of Arizona Press, Tucson. Department of Anthropology, Washington State University, Pullman. Lightfoot, R. R., and C. R. Van West 1986 Excavation and Testing Conducted on Private Land Under State Permit 85-22. Crow Canyon Archaeological Center, Cortez, Colorado. Submitted to the Office of the State Archaeologist, Colorado Historical Society, Denver. Kvamme, K. L., and M. A. Jochim 1989 The Environmental Basis of Mesolithic Settlement. In The Mesolithic in Europe: Papers Presented at the Third International Symposium, edited by C. Bonsall, pp. 1-12. John Donald Publishers, Edinburgh, Scotland. Lightfoot, R. R., and M. D. Varien 1988 Report of 1987 Investigations at the Duckfoot Site (5Mf3868), Montezuma County, Colorado. Crow Canyon Archaeological Center, Cortez, Colorado. Submitted to the Office of the State Archaeologist, Colorado Historical Society, Denver. Kvamme, K. L., and T. A. Kohler 1988 Geographic Infonnation Systems: Technical Aids for Data Collection, Analysis, and Display. In Quantifying the Present and Predicting the Past: Theory, Method, and Application ofArchaeological Predictive Modeling, edited by W. J. Judge and L. Sebastian, pp. 493-547. U.S. Government Printing Office, Washington, D.C. Lipe, W. D. 1989 Social Scale of Mesa Verde Anasazi Kivas. In The Architecture of Social Integration in Prehistoric Pueblos, edited by W. D. Lipe and M. Hegmon, pp. 53-71. Occasional Papers No.1. Crow Canyon Archaeological Center, Cortez, Colorado. Larson, D.O., and J. Michaelsen 1990 Impacts of Climatic Variability and Population Growth on Virgin Branch Anasazi Cultural Developments. American Antiquity 55:227-249. Lipe, W. D., and R. G. Matson 1971 Human Settlement and Resources in the Cedar Mesa Area, Southeast Utah. In The Distribution ofPrehistoric Population Aggregates, edited by G. J. Gumerman, pp. 126-151. Prescott College Anthropological Reports No.1. Prescott, Arizona Lange, F., N. Mahaney, J. B. Wheat, and M. Chenault 1986 Yellowjacket Jacket, a Four Corners Anasazi Ceremonial Center. Johnson Books, Boulder. Lipe, W. D. (editor) 1992 The Sand Canyon Archaeological Project: A Progress Report. Occasional Paper No.2. Crow Canyon Archaeological Center, Cortez, Colorado. Lightfoot, R. R. 1985 Report of 1984 Investigations Conducted Under State Permit 84-15. Crow Canyon Archaeological Center, Cortez, Colorado. Submitted to the Office of the State Archaeologist, Colorado Historical Society, Denver. 1987 Annual Report of Investigations at the Duclifoot Site (5MT3868J, Montezuma County, Colorado. Crow Canyon Archaeological Center, Cortez, Colorado. Submitted to the Office of the State Archaeologist, Colorado Historical Society, Denver. 1988 Roofing an Early Anasazi Great Kiva: Analysis of an Architectural Model. The Kiva 53:253-272. 1992 Archaeology of the House and Household: A Case Study ofAssemblage Formation and Household Organization in the American Southwest. Unpublished Ph.D. dissertation, - Lister, R. H. 1967 Progress Report on the 1967 Inventory of Archaeological Remains in Lands Controlled by the Bureau of Land Management in the Cortez - Dove Creek Area of Southwestern Colorado. Ms. on ftle, Bureau of Land Management, Denver. Lowe, C. H. 1964 Arizona's Natural Environment: Landscapes and Habitats. The University of Arizona Press, Tucson. Mackey, J. c., and S. J. Holbrook 1978 Environmental Reconstruction and the Abandonment of the Largo-Gallina Area, New 205- Martin, D. W., and J. D. Scurry 1988 The Role of MOSS of Archaeological Predictive Modeling. Paper presented at the 5th Annual MOSS User's Workshop, New Orleans. Mexico. Journal ofField Archaeology 5:2949. McDonald, C. L., and 1. K. Cain 1985 Applied Computer Graphics in a Geographic Information System: Problems and Successes. IEEE Computer Graphics and Applications 5:34-39. Martin, P. S. 1929 The 1928 Archaeological Expedition of the State Historical Society of Colorado. Colorado Magazine 6(1):1-35. McEwen, R. B., H. W. Calkins, and B. S. Ramey 1983 Overview and USGS Activities. U.S. Geological SlDVey Circular 895-A. User Services Section, U.S.G.S. National Cartographic Information Center, Reston, Virginia 1930 Archaeological Expedition of the State Historical Society of Colorado in Cooperation with the Smithsonian Institution. Colorado Magazine 7(1):1-40. Maffini, G. 1987 Raster versus Vector Data Encoding and Handling: A Commentary. Photogrammetric Engineering and Remote Sensing 53: 13971398. 1936 Lowry Ruin in Southwestern Colorado. Anthropological Series Vol. 23, No.1. Field Museum of Natural History, Chicago. 1938 Archaeological Work in the AckmenLowry Area, Southwestern Colorado, 1937. Anthropological Series, Vol. 23, No.2. Field Museum of Natural History, Chicago. Macomb, J. N. 1876 Report of the Exploring Expedition from Santa Fe. New Mexico to the Junction of the Grand and Green Rivers of the West in 1859. Government Printing Office, Washington, D.C. 1939 Modified Basket Maker Sites: AckmenLowry Area, Southwestern Colorado. 1938. Anthropological Series Vol. 23, No.3. Field Museum of Natural History, Chicago. Marble, D. F. 1984 Geographic Information Systems: An Overview. PECORA 9 Proceedings, Spatial Information Technologies For Remote Sensing Today and Tomorrow. pp. 1&-24. Sioux Falls, South Dakota. Martin, P. S., and W. Byers 1965 Pollen and Archaeology at Wetherill Mesa. Memoirs of the Society for American Archaeology 19:122-135. Marozas, B. A., and J. A. Zack 1990 GIS and Archaeological Site Location. In Interpreting Space: GIS and Archaeology, edited by K. M. S. Allen, S. W. Green, and E. B. W. Zubrow, pp. 165-172, Taylor and Francis, London. Mather, J. R. 1977 Workbook in Applied Climatology. Rev. ed. Publications in Climatology 30(1). Laboratory of Climatology, Elmer, New Jersey. Matson, R. G. 1991 The Origins of Southwestern Agriculture. The University of Ariwna Press, Tucson. Martin, C. W. 1976 Archaeological Inventory of the Sand Canyon Cliff Dwelling Area. Montezuma County. Colorado. Mesa Verde Archaeological Research Center. Submitted to Bureau of Land Management. Montrose District Office, Montrose, Colorado. Martin, D. W. 1971 Evaluation of the Archaeological Resources on Public Lands of the Bureau of Land Management in the Cortez-Dove Creek (McElmo District) of Southern Colorado. Ms. on fUe, Bureau of Land Management. Anasazi Heritage Center, Dolores, Colorado. - Matson, R. G., and W. D. Lipe 1978 Settlement Patterns on Cedar Mesa: Boom and Bust on the Northern Periphery. In Investigations of the Southwestern Anthropological Research Group: The Proceedings of the 1976 Conference. edited by R. C. Euler and G. J. Gumerman, pp. 1-12. Museum of Northern Ariwna., Flagstaff. Matson, R. G., W. D. Lipe, and W. R. Haase 1988 Adaptational Continuities and Occupational Discontinuities: The Cedar Mesa Anasazi. Journal ofField Archaeology 15:245-264. 206- 1986b Archaeological Excavations on Reach II of the Dove Creek Canal. Four Comers Archaeological Project No.9. Complete Archaeological Service Associates. Submined to Bureau of Reclamation, Upper Colorado Region, Salt Lake City. Meko, D. M., C. W. Stockton, and W. R. Boggess 1980 A Tree­Ring Reconstruction of Drought in Southern California. Water Resources Bulletin ＱＶＺＵＹＴｾＮ Metzger, T. R., and S. L. Breternitz­Goulding 1981 Ruin Stabilization Activities at Sand Canyon Pueblo Cliff Dwelling Area. Montezuma County, Colorado. Mesa Verde Research Center. Ms on file, Bureau of Land Management, Anasazi Heritage Center, Dolores, Colorado. Neily, R. B. 1983 The Prehistoric Community on the Colorado Plateau: An Approach to the Study of Change and Survival in the Northern San Juan Area of the American Southwest. Ph.D. dissertation, Southern lllinois University. University Microfilms, Ann Arbor. Mitchell, J. M., Jr., C. W. Stockton, and D. M. Meko 1979 Evidence of a 22­Year Rhythm of Drought in the Western United States Related to the Hale Solar Cycle Since the 17th Century. In Solar Terrestrial Influences in Weather and Climate, edited by B. M. McConnac and T. A. Seliga. pp. 125­143. Reidel Publishing, Dordrecht, Holland. Newberry, J. S. 1876 Geological Report. In Report of the Exploring Expeditionfrom Santa Fe, New Mexico to the Junction of the Grand and Green Rivers of the West in 1859, by J. N. Macomb. U.S. Government Printing Office. Nickens, P. R. 1977 Cultural Resources of the Dolores Project Area, Southwestern Colorado. Submitted to the Bureau of Reclamation, Upper Colorado Region, Salt Lake City. Monmonier, M. S. 1982 Computer Assisted Cartography: Principles and Prospects. Prentice­Hall, Englewood Cliffs, New Jersey. 1981 Pueblo III Communities in Transition: Environment and Adaptation in Johnson Canyon. Memoirs of the Colorado Historical Society NO.2. Boulder. Morley, S. G. 1908 The Excavation of the Cannonball Ruins in Southwestern Colorado. American Anthropologist, n.s., 10:596­­610. Nickens, P. R., and D. A. Hull 1982 Archaeological Resources of Southwestern Colorado: An Overview of the Bureau of Land Management's San Juan Resource Area. In Archaeological Resources of Southwestern Colorado, edited by S. Eininger, P. J. Gleichman, D. A. Hull, P. R. Nickens, A. R. Reed, and D. D. Scott, pp. 1-308. Cultural Resource Series No. 22. Bureau of Land Management, Denver. Morley, S. G., and A. V. Kidder 1917 The Archaeology of McElmo Canyon Colorado. EI Palacio 4:41-70. Morris, E. H. 1939 Archaeological Studies in the La Plata District, Southwestern Colorado and Northwestern New Mexico. Carnegie Institute of Washington Publication No. 519. Washington, D.C. Morris, E. H., and R. F. Burgh 1954 Basketmaker II Sites Near Durango, Colorado. Carnegie Institute of Washington Publication No. 604. Washington, D.C. Morris, J. N. 19800 Monitoring and Excavation at Aulston Pueblo (Site 5MT2433), a Pueblo II Habitation Site. Four Corners Archaeological Project Report No.6. Complete Archaeological Service Associates. Submitted to Bureau of Reclamation, Upper Colorado Region, Salt Lake City. - Orcutt, J. D., and D. A. Goulding 1986 Archaeological Survey of the McPhee Reservoir. In Dolores Archaeological Program: Research Designs and Initial Survey Results, compiled by A. E. Kane, W. D. Lipe, T. A. Kohler, and C. K. Robinson, pp. 185-375. U.S. Department of the Interior, Bureau of Reclamation, Research and Engineering Center, Denver. Palmer, W. C. 1965 Meteorological DroughJ. Research Paper No. 45. U.S. Department of Commerce, 207- Office of Climatology, U.S. Weather Bureau, Washington, D. C. Parker, H. D. 1988 The Unique Qualities of a Geographic Information System: A Commentary. Photogrammetric and Engineering and Remote Sensing 54:1547-1549. Parker, S. 1985 Predictive Modeling of Site Settlement Systems Using Multivariate Logistics. In For Concordance in Archaeological Analysis: Bridging Data Structure, Quantitative Techniques, and Theory, edited by C. Carr, pp. 173­207. Westport, Kansas City. 1986 The Role of Geographic Information Systems in Cultural Resource Management. In Geographic Information Systems in Government, Vol. I, edited by B. K. Opitz, pp. 133­140. Deepak Publishing, Hampton, Virginia. Parker, S., W. F. Limp, J. A. Farley, and I. Johnson 1986 Database Management Geographic Information Systems and Predictive Modeling in the Forest Service. Arkansas Archaeological Survey. Draft report submined to the U.S. Department of Agriculture, Forest Service, Region 8, Atlanta. Parsons, E. C. 1936 Hopi Journal of Alexander M. Stephen. Columbia University Contributions to Anthropology 23 and 24. Columbia University Press, New York. Petersen, K. L. 1981 Ten Thousand Years of Climate Change ReconstructedfromFossil Pollen. Ph.D. dissertation, Washington State University. University Microfilms, Ann Arbor. 1987 Tree-Ring Transfer Functions For Estimating Com Production. In Dolores Archaeological Program: Supporting Studies: Settlement and Environment, compiled by K. L. Petersen and J. D. Orcutt, pp. 217-231. Bureau of Reclamation, Engineering and Research Center, Denver. Petersen. K. L., and V. L. Clay 1987 Characteristics and Archaeological Implications of Cold Air Drainage in the Dolores Project Area In Dolores Archaeological Program: Supporting Studies: Settlement and Environment, compiled by K. L. Petersen and J. D. Orcutt, pp. 187-216. Bureau of Reclarnation, Engineering and Research Center, Denver. Petersen, K. L., and P. J. Mehringer, Jr. 1976 Postglacial Timberline Fluctuations, La Plata Mountains, Southwestern Colorado. Arctic and Alpine Research 8:275-288. Petersen. K. L., and J. D. Orcutt (compilers) 1987 Dolores Archaeological Program: Supporting Studies: Settlement and Environment. Bureau of Reclamation, Engineering and Research Center, Denver. Petersen, K. L., V. L. Clay, M. H. Matthews, and S. W. Neusius (compilers) 1985 Dolores Archaeological Program: Studies in Environmental Archaeology. Bureau of Reclamation, Engineering and Research Center, Denver. Powers, R. P., W. B. Gillespie, and S. H. Lekson 1983 The Outlier Survey: A Regional View of Settlement in the San Juan Basin. Reports of the Chaco Center No.3. Division of Cultural Research, National Park Service, Albuquerque. Price, A. B., W. D. Nettleton, G. A. Bowman, and V. L. Clay 1988 Selected Properties, Distribution, Source, and Age of Eolian Deposits and Soils of Southwestern Colorado. Soil Science Society ofAmerica Journal 52:450-455. Prudden, T. M. 1903 The Prehistoric Ruins of the San Juan Watershed in Utah, Arizona., Colorado and New Mexico. American Anthropologist, n.S., 5:224-288. 1914 The Circular Kivas of Small Ruins in the San Juan Watershed. American Anthropologist, n.s., 16:33-58. 1918 A Further Study of the Prehistoric Small House Ruins in the San Juan Watershed. Memoirs of the American Anthropological Association, vol. 5, No.1. 1988 Climate and the Dolores River Anasazi. University of Utah Anthropological Papers No. 113. Salt Lake City. - 208- 1972 Social Implications of Pueblo Water Management in the Northern San Juan. Sonderdruck: Zeitschrift Fur Ethnologie, Band 97, Heft 2, Braurschweig, pp. 212-219. Reed, A. D. 1979 Part I: The Dominguez Ruin: A McElmo Phase Pueblo in Southwestern Colorado. In The Archeology and Stabilization of the Dominguez and Escalante Ruins, pp. 1­196. Cultural Resource Series NO.7. Colorado State Office, Bureau of Land Management, Denver. 1974 Payne Site Investigations. Southwestern Lore 40(3 & 4):40--52. 1975 A Stockaded Basketmaker III Village at Yellowjacket, Colorado. The Kiva 40:113120. Riley, C. L. 1948 An Archeological Survey at Hovenweep National Monument. Ms. on fLle, Mesa Verde Research Center, Mesa Verde National Park, Colorado. 1977 Culture Change and Continuity on Chapin Mesa. Regents Press of Kansas, Lawrence. 1989 Northern San Juan Prehistory. In Dynamics of Southwest Prehistory, edited by L. S. Cordell and G. J. Gumerman, pp. 149-177. Smithsonian Institution Press, Washington, D.C. 1950 "Defensive" Structures in the Hovenweep Monument. El Palacio 57: 339­344. Rinaldo, J. 1950 An Analysis of Culture Change in the Ackmen­Lowry Area. Fieldiana 36(5):93-106. Rose, M. R., J. S. Dean, and W. J. Robinson 1981 The Past Climate ofArroyo Hondo, New Mexico, Reconstructed From Tree-Rings. Arroyo Hondo Archaeological Series Vol. 4. School of American Research Press, Santa Fe. Ripple, W. J. 1987 GIS for Resource Management: A Compendium. American Society of Photogrammetry and Remote Sensing and American Congress on Surveying and Mapping, Bethesda, Maryland. Rose, M. R., W. J. Robinson, and J. S. Dean 1982 Dendroclimatic Reconstructionfor the Southeastern Colorado Plateau. Final Report to Dolores Archaeological Project and Division of Chaco Research. Laboratory of TreeRing Research, The University of Ariwna, Tucson. 1989 Fundamental of GIS: A CompendiUm. American Society of Photogrammetry and Remote Sensing and American Congress on Surveying and Mapping, Bethesda, Maryland. Roberts, F. H. H., Jr. 1925 Report on Archaeological Reconnaissance in Southwestern Colorado in the Summer of 1923. Colorado Magazine 2(2):1-80. Roys, L. 1936 Lowry Ruin as an Introduction to the Study of Southwestern Masonry. In Lowry Ruin in Southwestern Colorado, by P. S. Maron, pp. 115-142. Anthropological Series Vol. 23, No.1. Field Museum of Natural History, Chicago. 1930 Early Pueblo Ruins in the Piedra District, Southwestern Colorado. Bureau of American Ethnology Bulletin 96. Washington, D. C. Robinson, W. J., and B. G. Harrill 1974 Tree-Ring Datesfrom Colorado V, Mesa Verde Area. Laboratory of Tree-Ring Research, The University of Ariwna, Tucson. Rohn, A. H. 1963 Prehistoric Soil and Water Conservation on Chapin Mesa, Southwestern Colorado. American Antiquity 28:441-455. 1971 Mug House, Mesa Verde National Park, Colorado. Archeological Series No.7-D. National Park Service, Washington, D. C. - Sanders, W. T. 1976 The Agricultural History of the Basin of Mexico. In The Valley of Mexico: Studies in Pre-Hispanic Ecology and Society, edited by E. R. Wolf, pp. 101-159. University of New Mexico Press, Albuquerque. 1979 The Basin ofMexico: Ecological Processes in the Evolution of a Civilization. Academic Press, New York. SAS Institute 1985a SAS User's Guide: Basics. 5th ed. SAS Institute, Cary, North Carolina. 209- 1985b SAS User's Guide: Statistics. 5th ed. SAS Institute, Cary, North Carolina R. D. Jones. pp. 207-222. Arkansas Archaeological Survey Research Series No. 14. Fayetteville. Schaafsma, P. 1980 Indian Rock Art of the Southwest. School of American Research and University of New Mexico Press. Santa Fe and Albuquerque. Schulman. A. 1950 Pre-Columbian Towers in the Southwest. American Antiquity 15:288-297. Schlanger. S. H. 1985 Prehistoric Population Dynamics in the Dolores Area, Southwestern Colorado. Ph.D. dissertation, Washington State University. University Microfilms. Ann Arbor. Sebastian, L. 1988 Leadership, Power, and Productive Potential: A Political Model of the Chaco System. Ph.D. dissertation, University of New Mexico. University Microfilms, Ann Arbor. 1986a 1982 Probabilistic Sampling Survey of Windy Ruin and Yellowjacket Crest Locality. In Dolores Archaeological Program: Research Designs and Initial Survey Results. compiled by A. E. Kane, W. D. Lipe, T. A. Kohler. and Bureau of RecC. K. Robinson, pp. ＴＷｾＰＮ lamation. Research and Engineering Center, Denver. 1992 The Chaco Anasazi: Sociopolitical development in the Prehistoric Southwest. Cambridge University Press, Cambridge. Soil Conservation Service 1951 Soil Survey Manual. U.S. Department of Agriculture Handbook 18. U.S. Government Printing Office, Washington. D. C. 1986b Population Studies. In Dolores Archaeological Program: Final Synthetic Report. compiled by D. A. Breternitz. C. K. Robinson, and G. T. Gross, pp. 493-524. Bureau of Reclamation, Research and Engineering Center, Denver. Speth, J. D., and S. L. Scott 1989 Horticulture and Large-Mammal Hunting: The Role of Resource Depletion and Constraints of Time and Labor. In Farmers as Hunters: the Implications of Sedentism, edited by S. Kent, pp. 71-79. Cambridge University Press, Cambridge. Schlanger, S. H., and P. L. Harden 1986 1979 and 1980 Probability Survey of Cline Crest, Grass Mesa, Beaver Point, Trimble Point, and Hoppe Point Localities. In Dolores Archaeological Program: Research Designs and Initial Survey Results, compiled by A. E. Kane, W. D. Lipe, T. A. Kohler, and C. K. Robinson, pp. 379-446. Bureau of Reclamation, Research and Engineering Center, Denver. Schoenwetter. J. 1970 Archaeological Pollen Studies of the Colorado Plateau. American Antiquity ＳＵＺｾＸＮ Stiger. M. A. 1979 Mesa Verde Subsistence Patterns from Basketrnaker to Pueblo m. The Kiva 44: 133144. Stockton, C. W. 1975 Long-Term Streamflow Records Reconstructed from Tree Rings. Papers of the . Laboratory of Tree Ring Research 5. University of Arizona, Tucson. Schoenwetter, J.• and F. W. Eddy 1964 Alluvial and Palynological Reconstruction of Environments, Navajo Reservoir District. Museum of New Mexico Papers in Anthropology, No. 13. Santa Fe. Scholtz, S. C. 1981 Location Choice Model in Sparta. In Settlement Predictions in Sparta: A Locational Analysis and Cultural Resource Assessment on the Uplands of Calhoun County. Arkansas. by R H. Lafferty m, J. L. Otinger, S. C. Scholtz, W. F. Limp, B. Watkins and - Star, J., and J. E. Estes 1989 Geographic Information Systems: An Introduction. Prentice-Hall, Englewood Cliffs, New Jersey. Stockton, C. W., and W. R. Boggess 1980 Augmentation of Hydrological Records Using Tree Rings. In Proceedings of the Engineering Foundation Conference "Improved Hydrological Forecasting, Why and How." Asilomar Conference Grounds, Pacific Grove. California. March 25-30, 1979, pp. 239-265. American Society of Civil Engineers and American Meteorological Society, New York. 210- Stokes, M. A. and T. L. Smiley . . 1968 An Introduction to Tree-Ring Dating. The U.S. Bureau of the Budget 1979 U.S. National Map Accuracy Standards. In University of Chicago Press, Chicago. Mapsfor America: Cartographic Products of the U.S. Geological Survey and Others, by M. Thompson, p. 104. U.S. Geological Survey, Reston, Virginia. Swannack, J. D., Jr. 1%9 Big Juniper House, Mesa Verde National Park, Colorado. Archeological Research Series Van Devender, T. R., and W. G. Spaulding 1979 Development of Vegetation and Climate in the Southwestern United States. Science 204:701-710. No. 7­C. National Park Service, Washington, D.C. Swedlund, A. C. 1969 Human Skeletal material from the Yellow Jacket Canyon Area, Southwestern Colorado. Van West, C. R. 1986 Cultural Resource Inventory for the 1985 Field Season in the Crow Canyon and Sand Canyon Areas, Montezuma County, Colorado. Crow Canyon Archaeological Center, Cortez. Submitted to the Office of the State Archaeologist, Colorado Historical Society, Denver. Unpublished M.A. thesis, Department of Anthropology, University of Colorado, Boulder. Thompson, C., S. A. Baker, J. R. Allison, and K. L. Winteh 1986 The Nancy Patterson Village Archaeological Research Project, Field Year 1985. Preliminary Report No.3. Brigham Young 1990 Modeling Prehistoric Climatic Variability and Agricultural Productiopn in Southwestern Colorado: A GIS Approach. Unpublished University, Museum of Peoples and Cultures, Technical Series No. 86­27. Provo, Utah. Ph.D. dissertation, Deparnnent of Anthropology, Washington State University, Pullman. Tipps, B. L. 1978 Stabilization at Three Prehistoric Ruins in the San Juan Basin, 1977. Mesa Verde Research Center. Ms. on file, Bureau of Land Management, Anasazi Heritage Center, Dolores, Colorado. 1994a Reconstructing Prehistoric Climatic Variability and Agricultural Production in Southwestern Colorado, A.D. 901-1300: A GIS Approach. Paper presented at the Anasazi Symposium, Mesa Verde National Park. Oct. 3-4, 1991, in press. Ms. 1991. Titiev, M. 1944 Old Oraibi, A Study of the Hopi Indians of Third Mesa. Papers of the Peabody Museum of 1994b The Heuristic Value of Estimates of Prehistoric Agricultural Production: A Case Study from Southwestern Colorado. Paper presented at the Third Annual Southwestern Symposium, January 17-18, 1992, Tucson, Ariwna. To be published in Interpreting Archaeology and Ethnology Vol. 22, No.1. Cambridge, Massachusetts. Toll, H. W., III 1977 Dolores River Archeology: Canyon Adaptations as Seen Through Survey. Cultural Southwestern Diversity: Underlying Principles and Overarching Patterns. Arizona State Uni- Resource Series No.4. Bureau of Land Management, Denver. versity Anthropological Papers, Tempe, in press. Ms. 1992. Tomlin, C. D. 1995 Agricultural Potential and Carrying Capacity in Southwestern Colorado, AD 901-1300. Paper presented at the conference, "Pueblo Cultures in Transition: AD 1150-1350 in the American Southwest." March 25-28, 1990, Crow Canyon Archaeological Center, Cortez, Colorado, in press. Ms. 1992. 1990 Geographic Information Systems and Cartographic Modeling. Prentice-Hall, Engle- wood Cliffs, New Jersey. Tomlinson, R. F., H. W. Calkins, and D. F. MMble 1976 Computer Handling of Geographic Data: An Examination ofSelected Geographic Information Systems. Natural Resources Re- search xm. UNESCO, Paris. Trewartha. G. T. 1954 An Introduction to Climate. McGraw-Hill, New Yorlc. - Van West, C. R., and T. A. Kohler 1994 A Time to Rend, A Time to Sew: A New Perspective on Northern Anasazi Social Development in Later Prehistory. Paper presented at the Conference, "The Anthropology 211- ofHwnan Behavior Through Geographic Information and Analysis," February 1-2, 1992, University of California, Santa Barbara Oxford University Press, Oxford, in press. Ms. 1992. Wansleeben, M. 1988 Geographic Information Systems in Archaeological Research. In Computer and Quantitative Methods in Archaeology 1988, edited by S. P. Q. RaIltz, pp. 435-451. BAR International Series 446 (ii), British Archaeological Reports, Oxford Van West, C.R., and W. D. Lipe 1992 Modeling Prehistoric Climate and Agriculture in Southwestern Colorado. In The Sand Canyon Archaeological Project: A Progress Report, edited by W.D. Lipe, pp. 105-119. Occasional Paper No.2. Crow Canyon Archaeological Center, Cortez, Colorado. Warren, R. E. 1988 Predictive Model of Archaeological Site Location: Two Case Studies from the Midwest. Paper presented at the 53rd Annual Meeting of the Society for American Archaeology, Phoenix. Van West, C. R., M. A. Adler, and E. K. A. Huber 1987 Archaeological Survey and Testing in the Vicinity of Sand Canyon Pueblo, Montezuma County, Colorado, 1986 Field Season. Crow Canyon Archaeological Center. Submitted to the Office of the State Archaeologist, Colorado Historical Society, Denver, and the Bureau of Land Management, San Juan Resource Area, Durango, Colorado. Warren, R. E., S. G. Oliver, J. A. Ferguson, and R. E. Druhot 1987 A Predictive Model of Archaeological Site Location in the Western Shawnee National Forest. Ms. on fIle. Illinois State Museum Contract Archaeological Program, Springfield. Wehde,M. 1982 Grid Cell Size in Relation to Errors in Maps and Inventories Produced by Computerized Map Processing. Phatogrammetric Engineering and Remote Sensing 48:12891298. Varien, M. D. 1990 Preliminary Descriptive Report on the Excavations at Lillian's Site (5MT3936), Roy's Ruin (5MT3930) , Sharlene's Site (5MT3918), and Troy's Tower (5MT3951). Crow Canyon Archaeological Center, Cortez, Colorado. Wetterstrom, W. E. 1976 The Effect ofNutrition on Population Size at Pueblo Arroyo Hondo, New Mexico. 2 vols. Ph.D. dissertation, University of Michigan, University Microfilms, Ann Arbor. Varien, M. D., and R. R. Lightfoot 1989 Ritual and Nonritual Activities in Mesa Verde Region Pit Struetmes. In The Architecture of Social Integration in Prehistoric Pueblos, edited by W. D. Lipe and M. Hegmon, pp. 73-87. Occasional Papers No.1. Crow Canyon Archaeological Center, Cortez, Colorado. Wheat, J. B. 1955 MT-1 Basketmaker ill Site Near Yellowjacket, Colorado. Southwestern Lore 21(2):1826. 1980 Yellowjacket Canyon Archaeology. In Insights Into the Ancient Ones, 2d ed., edited by J. H. Berger and E. F. Berger, pp. 60-66. Interdisciplinary Supplemental Educational Programs, Cortez, Colorado. Varien, M., W. Lipe, B. Bradley, M Adler, and I. Thompson 1990 Mesa Verde Region Settlement, A. D. 1100 to 1300. Ms. on file, Crow Canyon Archaeological Center, Cortez., Colorado. Walkenhorst, R. N. 1983 Cahone Laterals Letter Report. Ms. on file, Bureau of Land Management, Anasazi Heritage Center, Dolores, Colorado. White, A. S. 1976 Emergency Stabilization of 5MT263, East Rock Canyon, Colorado. Mesa Verde Research Center. Ms. on file, Bureau of Land Management, Anasazi Heritage Center, Dolores, Colo- mOO. Walkenhorst, R. N., D. Goulding, M. Errickson, and G.lves 1983 Fairview Lateral Letter Report. Ms. on file, Bureau of Land Management, Anasazi Heritage Center, Dolores, Colorado. - White, A. S., and D. A. Bretemitz 1976 Stabilization of Lowry Ruins. Cultural Resources Series No.1. Bureau of Land Management, Denver. 212- 1977 Hovenweep 1976. Department of Anthropology, San Jose State University Archaeological Report No.3. San Jose, California. 1979 Stabilization of Escalante Ruin 5MT2149 and Dominguez Ruin 5MT2148, Dolores, Colorado. In The Archeology and Stabilization of the Dominguez and Escalante Ruins. Cultural Resource Series NO.7. Bureau of Land Management, Denver. Whitten, P., T. Keams, and M. Swift 1986 A Report on the Archaeological Survey and Testing of a C02 Pipeline Right-of-Way on Cajon Mesa, San Juan County, Utah and Montezuma County, Colorado. Studies in Archaeology No. 1. Division of Conservation Archaeology, San Juan County Museum Association, Fannington, New Mexico. Whiting, A. F. 1939 Ethnobotany of the Hopi. Museum of Northern Arizona Bulletin 15. Flagstaff. Wilshusen, R. H. 1989 Unstuffmg the Estufa: Ritual Floor Features in Anasazi Pit Sttuctures and Pueblo Kivas. In The Architecture of Sociallntegration in Prehistoric Pueblos, edited by W. D. Lipe and M. Hegmon, pp. 89-111. Occasional Papers No. 1. Crow Canyon Archaeological Center, Cortez, Colorado. Williams, B. J. 1989 Contact Period Rural Overpopulation in the Basin of Mexico: Carrying Capacity Models Tested with Documentary Data. American Antiquity 54:715-732. Williamson, R. A. 1984 Living the Sky: The Cosmos of the American Indian. University of Oklahoma Press, Norman. Winter, J. C. 1975 Hovenweep 1974. Department of Anthropology, San Jose Stale University, Archaeological Report No. 1. San Jose, California. 1976 Hovenweep 1975. Department of Anthropology, San Jose State University, Archaeological Report No.2. San Jose, California. - 1984 The Towers of Hovenweep: Theories of Use and A History of Tower Research. In Insights Into the Ancient Ones, 2d ed., edited by J. H. Berger and E. F. Berger, pp. 73-80. Interdisciplinary Supplemental Educational Programs, Cortez, Colorado. Woods Canyon Archaeological Consultants 1985 Archaeological Testing Program of Four Sites in the Goodman Point Area. Montezuma County, Colorado. Woods Canyon Archaeological Consultants. Submitted to Shell Pipeline Corporation. Ms. on file, Bureau of Land Management, Anasazi Heritage Center, Dolores, Colorado. Woosley, A. 1. 1977 Hovenweep 1977, Preliminary Report. San Jose State University. Ms. on fIle, Bureau of Land Management, Anasazi Heritage Center, Dolores, Colorado. Wycoff, D. G. 1977 Secondary Forest Succession Following Abandonment of Mesa Verde. The Kiva 42:215-231. Zahner, R., J. R. Saucier, and R. K. Myers 1989 Tree-Ring Model Interprets Growth Decline in Natural Stands of Loblolly Pine in the Southeastern United States. Canadian Journal for Resources 19:612-621. Zubrow, E. B. W. 1990 Modelling and Prediction with Geographic Information Systems: A Demographic Example from Prehistoric and Historic New York. In Interpreting Space: GIS and Archaeology, edited by K. M. S. Allen, S. W. Green, and E. B. W. Zubrow, pp. 307-318. Taylor and Francis, London. 213- Page Blank in Original ApPENDIX A EIGENVECTOR AMPLITUDES GENERATED BY A PRINCIPAL COMPONENTS ANALYSIS ON SEVEN EXPANDED TREE­RING CHRONOLOGIES (SWOLD7) Year AD Axis 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920. 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 ­.394 .540 .427 .259 ­.954 ­.220 ­.144 ­.361 ­.176 .023 .080 ­.136 .042 .201 ­.233 .299 ­.181 .165 .178 .051 .252 ­.132 .029 .133 ­.110 ­.279 ­.059 ­.616 .024 .410 .033 .378 ­ .140 ­.002 ­ .148 .184 .843 ­.377 ­.115 .315 .545 .044 ­.219 ­.039 .204 .329 7 AX1S 6 ­.128 ­1.114 ­.299 ­.290 .352 ­.163 .167 .510 ­.636 ­.715 ­.223 ­.854 .350 ­ .109 .804 ­.414 .533 .074 ­.161 ­.456 1.156 ­.237 .683 ­.679 .325 ­.378 .075 .631 .482 ­.055 .433 .386 ­.550 ­.719 ­.436 ­.648 ­.110 .224 ­.725 ­.383 ­.077 ­.280 ­.246 .556 .257 .007 Axis AX1S Axis AX1S Axis 4 5 3 2 1 .117 ­.308 1 .510 .545 4.179 .095 .153 ­.481 ­1 . 142 .288 ­.092 1 .651 ­.444 .670 1 .910 .486 .237 ­.487 .913 1 .135 ­.402 .261 ­.016 .124 ­2.920 ­.479 ­.376 .845 .208 5.392 .280 ­.118 .376 ­.089 4.438 .106 ­.571 .202 .600 ­3.303 .368 ­.947 ­1.313 ­.737 .278 ­.639 ­1 .382 ­.424 ­.256 .055 .173 .879 ­.779 ­1 .396 ­3.029 .214 ­1 .505 ­.109 ­1.411 1 .987 .125 .239 .299 ­.765 ­.343 ­.166 .486 ­1 .077 ­.910 .013 ­.285 ­.743 ­.681 1 .566 ­.153 .276 .445 1.409 ­1.096 .299 ­.826 .855 ­.277 ­.491 ­2.646 .857 .773 .489 ­.175 ­2.193 ­.675 .206 .663 ­.493 ""2.625 ­.447 ­.318 ­.469 .769 1.022 .476 ­.842 .606 ­.756 ­.289 ­.021 ­.367 ­.055 .631 4.273 .492 1 .713 1 .411 ­ .104 .283 .135 ­.459 1 .193 ­.851 3.235 .558 .908 ­.031 .113 ­2.157 .087 ­.864 .387 .221 ­.227 .225 ­ .199 .024 ­.180 1.695 .054 .957 .298 ­.391 ­3.109 ­.645 ­.890 .183 ­.708 ­.104 ­.069 ­.868 1 .359 .811 2.296 .163 .712 ­.842 ­.502 ­ .186 .064 ­ .129 .583 ­.224 ­.039 .615 ­ .817 ­.005 ­.046 .586 .129 ­.350 ­2.177 1 .810 .398 .636 ­1.147 .993 .276 .168 .557 .216 ­.214 .267 ­1.050 .044 .734 2.889 .052 ­.546 .444 .074 ­2.127 ­.819 .090 .942 ­.163 1 .417 1.376 ­.275 .346 ­1.047 .636 ­.098 1 .079 .182 .935 ­1.699 .662 ­.492 .826 ­.298 ­.636 ­.615 ­.287 1.104 .720 ­1 .201 ­.201 ­.128 .291 ­.044 .416 ­.392 ­1.325 1 .120 .218 ­.201 .340 ­.189 .343 ­;660 ­.059 .337 ­2.117 Year AD 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 Axis Axis Axis Axis Axis Axis Axis 7 5 4 6 3 1 2 .725 .165 .351 .066 ­.895 ­.395 · 111 ­.035 .387 ­.516 ­.724 ­.212 ­.160 ­.850 .333 .690 .021 .244 .867 .704 ­2.843 .175 .225 ­.496 1. 214 ­1 .628 ­.981 1 .009 .622 ­.219 ­.981 .290 ­.735 ­.368 2.972 .643 ­.180 ­.185 ­.051 ­.462 ­.464 · 198 ­.647 .098 ­.668 ­.066 1 .025 ­.439 2.094 .245 ­.003 .024 ­.480 ­.546 .013 4.221 .547 ­.688 1 .077 .847 .044 ­.282 · 101 ­.383 ­.530 ­.047 ­1 . 109 .037 .393 ­3.366 .481 .691 ­.242 ­.241 ­.768 .908 2.260 .012 .310 ­.325 .736 1 .012 ­1. 538 2.533 ­.208 ­.834 ­.169 ­.068 ­.159 1 .080 ­2.459 ­.689 ­.064 .025 ­.747 1 .597 ­.586 ­3.202 ­.369 ­.267 ­.253 .886 ­. 182 ­1.370 .844 ­.190 .547 ­1.129 1 .465 .863 .523 ­3.067 ­.101 ­.281 ­.482 .167 .240 ­.331 ­.826 ­.098 .099 .797 ­1.117 ­.733 .674 1 .087 .056 ­1 .086 ­.696 .770 2.274 ­1 .382 ­1 .047 ­.484 .100 .304 ­.594 .298 .251 ­4.081 .293 ­.531 .549 ­.246 . 191 .740 ­2.343 1 .014 ­.008 .498 ­.272 ­.167 1.236 ­.212 .044 . 1 71 ­.585 1 .296 ­.721 ­.672 2.885 .269 ­.331 .567 1 .075 .422 .269 ­3.683 .744 .217 ­.216 ­.714 ­1 .004 1 .088 .592 .338 .557 ­ .144 ­.173 .096 .958 4.724 .229 .396 .468 .401 ­.980 ­2.315 ­3.298 ­.243 ­.095 .122 ­.597 ­1 .498 ­1. 530 .817 .212 .604 ­.525 ­.588 ­.514 ­.392 5.149 .006 1 .025 1 . 141 ­.182 .092 1 .407 .301 ­.030 .175 ­.156 .233 ­.279 ­.835 ­2.384 .534 .419 ­.609 ­.429 1 .055 ­.379 2.588 ­.222 .199 • 1 19 ­.325 1 .072 .280 ­.507 .835 ­.215 .011 ­.206 .180 .639 ｾＮ 143 .636 .700 .175 .441 .034 .312 3.542 .499 .428 ­.501 1.038 ­1.275 ­.825 · 141 .330 ­.014 ­.142 .923 ­.586 .022 1 .750 '.480 ­.304 ­.070 ­.232 .448 ­.521 1 .822 .489 .068 ­.309 .415 ­.152 ­.075 ­1. 924 .561 ­.293 .822 .791 .359 . 118 ­2.678 ­.212 .176 ­.855 ­.783 .035 .297 ­4.004 ­.985 ­.868 .024 ­.419 .037 ­.222 ­4.761 ­.593 .029 ­.102 ­1.992 ­.004 .183 ­4.990 .086 ­.548 ­1 .361 .439 ­.457 .024 .221 ­.515 ­.637 ­.017 .803 ­.559 1 .410 3.643 ­.043 .335 .005 .412 ­.370 ­.508 1 .994 .004 .857 .636 .292 ­.796 ­.329 2.388 .080 ­.315 .073 .792 .909 ­2.552 .085 ­.456 ­.306 .170 ­1. 556 1.690 1 .461 ­1. 957 ­ 216- Year AD 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 101 1 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 Axis Axis Axis Axis AX1S Axis Axis 7 6 5 4 3 2 1 .818 .397 ­2.433 1 .223 ­1. 984 .398 ­1 .752 1 .834 1 .044 ­1 .995 ­.029 ­.845 ­1.010 ­.507 .738 .470 ­1 .520 ­.228 ­1 .033 .832 ­.800 .508 ­.986 ­, .029 3.986 .197 .049 ­.148 . 176 1 .300 .286 ­.431 .136 ­2.353 ­.538 .054 ­.188 .292 .079 ­.435 2.926 ­.333 .413 . 107 .535 .336 .193 1 .277 .039 .245 ­.295 ­1 .067 .389 ­.840 .864 ­.298 .017 .226 ­.053 .818 .493 .282 ­.063 .476 1 .402 4.516 ­.070 ­.076 ­1 .091 ­.211 .235 .638 ­2.022 ­.956 . 713 ­.202 ­.332 ­.763 .773 .681 ­2.671 ­.745 .023 · 181 . 171 .709 ­.561 ­1 .843 .448 ­.594 ­.677 .886 ­.640 .354 3.595 ­.096 ­.547 · 161 .261 .142 ­1 .369 .669 ..829 .627 ­.380 .411 ­.454 ­1 .239 ­.709 ­.016 1.509 · 151 .498 ­1.418 .402 .397 ­.589 .195 ­.306 .113 ­1.311 1. 719 .301 1 .694 ­.242 ­.129 .548 .290 ­.491 ­1.411 ­.561 3.620 ­ . 101 ­.131 .674 .321 ­1. 719 .688 ­.610 .474 ­.301 ­.092 .329 ­.464 .217 .309 ­2.107 ­.415 .960 ­.456 ­1 .380 ­.512 .366 .125 ­.143 ­.543 ­.187 .469 ­.727 1 .295 .363 .302 ­.601 .106 .728 .185 .234 3.320 .026 ­.906 .275 1 .016 ­1.359 ­2.519 · 140 ­.208 ­.125 ­.314 ­.505 ­.354 .122 ­.242 .212 ­.379 ­.468 ­.027 ­1 .260 .390 1 .041 ­.173 .434 .146 ­2.151 .877 ­1. 329 ­1 .683 .422 ­.373 ­.784 ­.018 .299 ­.049 ­3.947 .477 ­.224 ­.956 .685 ­.932 .653 ­2.371 .215 ­. 141 1 . 187 .790 ­1.070 2.294 ­1 .969 .476 ­.581 ­.157 1 .504 ­1.419 ­.946 .145 ­.052 .529 .042 .394 ­.474 .966 ­1.276 ­.192 .087 ­.255 .519 .172 .699 ­3.146 ­.477 .305 .885 1 .958 ­.203 ­.660 1 .496 .052 .409 .396 ­.142 ­.631 1 .106 .986 ­.282 .306 ­.069 .266 .139 .325 1 .609 .258 ­.058 ­.267 .944 ­.334 1 .007 ­.649 ­.002 ­.227 .129 .568 ­1 .376 ­.790 .708 .340 .147 .062 ­.474 ­.328 ­.820 4.816 .455 . 181 .787 ­.461 ­.300 1 .018 . 1 .882 ­.490 .665 ­.362 ­.291 ­.393 ­.454 ­1 .098 .114 .484 ­.166 .363 .402 .509 ­.618 ­.106 .099 ­.653 .275 .261 ­.469 .075 ­.264 .551 ­.393 ­.336 ­.728 .229 .566 ­.337 .335 ­.296 ­.139 .874 ­.390 3.274 .487 ­.103 ­.599 ­.146 ­.653 .054 ­2.410 ­.650 .207 1.580 .154 ­.033 .473 ­.153 .096 ­.103 .285 .386 ­.143 ­.786 2.941 ­ 217- Year AD 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 Axis 7 ­.946 .035 .028 ­.416 ­.052 ­.475 ­.334 ­1 . 180 .221 .046 .079 .367 .000 .003 .762 .214 .537 .059 .533 .008 ­.397 .256 ­.035 ­.432 .048 ­.359 .266 ­.133 .231 ­.295 .225 .175 ­.411 .104 ­.205 ­.334 ­.137 .210 . 161 ­.039 ­.319 .193 ­.037 .456 .430 .145 .295 .471 .107 Axis 6 .258 .415 ­.157 .397 .584 ­.075 ­.138 . 165· .697 .679 .242 .710 .225 .245 ­.144 ­.480 .323 ­.216 .490 .199 .045 .425 ­.315 ­.446 .282 ­.350 .002 .254 ­.858 .355 .050 .023 .409 . 151 ­.424 ­.436 ­.721 .175 .021 ­.537 .224 ­.371 .264 .169 .213 ­.251 .324 ­.333 ­.039 Axis Axis 5 4 ­.003 ­.239 1 .225 ­.154 ­ . 115 ­.773 ­.140 ­.340 .600 ­.961 .218 ­.760 ­.903 .094 .196 ­.861 .078 ­.594 .675 ­1.244 ­.492 .380 ­.270 ­.094 .063 .343 ­.830 ­.484 ­.837 ­.125 ­.332 .61 7 ­.252 .906 .105 ­.039 ­.360 ­.912 ­.087 .151 ­.865 ­.550 ­.474 ­.707 .167 ­.256 .493 ­.145 ­.017 ­.151 ­.123 ­.921 . 169 .676 .267 ­.121 ­1.073 ­.601 ­.606 ­.118 ­.114 .425 .545 .207 .613 ­.099 .499 ­.041 ­.287 .833 ­.362 ­1.697 .496 .389 ­.054 .631 .381 ­.352 .039 ­.252 .344 1 .011 .551 ­1 .330 .368 ­.067 ­.346 ­.742 ­.491 ­1 .043 ­.073 .462 ­.076 .200 ­.195 .085 .931 ­.186 ­ 218- Axis 3 1 .510 .400 ­1 .038 ­.824 .626 .308 ­.880 .430 ­.557 ­.166 .265 ­·1 .237 .152 .396 ­.384 .788 ­ . 121 .803 ­1.310 ­1.273 .098 ­.868 ­.375 .137 .285 ­.031 .137 ­.267 .445 ­.421 ­1.226 ­.183 .230 ­1.078 . 147 ­1.176 ­.324 .633 ­.649 .153 1 .124 .077 1 .469 1 .309 ­.098 .444 .045 1 .268 .243 Axis 2 ­.686 .863 ＭＮＵＱｾ .160 ­.647 .331 .461 1 .066 1 .480 ­.720 1 . 165 ­.233 .052 ­1 .583 ­.545 .766 ­.856 .685 ­.271 ­.257 ­.193 1 .035 .666 .273 ­.651 .470 .316 .087 ­.212 ­.095 ­1.513 . 168 ­.163 .614 ­1.703 ­.949 .577 2.590 ­.002 ­.600 ­.308 .818 ­2.220 ­1 .290 .967 ­.367 ­.500 ­.691 .267 Axis 1 ­.075 1 .413 .832 4.295 ­.792 ­1.376 2.067 ­4. 133 .623 ­.998 1 . 192 ­.332 ­.990 .041 1.123 ­2.188 ­1 . 160 2.626 ­3.964 ­2.430 ­4.924 ­.744 4.172 2.965 ­.519 ­.457 .546 ­1.265 .676 ­.274 2.003 ­2.272 ­1 .184 .315 ­2.105 ­1.729 2.677 ­1 .887 3.066 ­.857 2.780 .885 ­2.047 ­1 .438 ­.099 5.669 2.877 ­2.310 2.596 Year AD 1094 1095 1096 1097 1098 1099 1 100 1101 1102 1103 1104 1105 1 106 1 107 1108 1109 1 110 1 111 1 1 12 1113 1114 1 1 15 1 1 16 111 7 1118 1 1 19 1120 1 121 1122 1123 1124 1125 1126 1127 1128 1129 1130 11 31 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 Axis 7 · 107 ­.403 ­ 1 .756 .483 ­.949 ­.094 ­.155 .083 · 141 .177 .817 ­.140 ­.020 ­.098 .449 .123 ­.147 .144 ­.782 .368 ­.349 .267 ­.136 ­.354 .444 ­.073 · 1 13 ­.381 ­.043 ­.185 ­.382 .599 .609 .182 ­.199 ­.181 ­.467 ­.583 ­.242 ­.291 .242 .196 .333 ­.456 ­ . 117 ­1.101 .088 ­.182 ­.172 AX1S Axis Axis 4 5 6 3 ­.172 ­.939 ­. 175 .631 ­.109 ­.261 1 .081 1 .220 ­.300 1 .076 ­.262 ­ . 1 10 .415 ­.485 .217 ­1.117 .448 .608 ­.299 ­.210 .574 ­.081 ­.360 ­ . 511 .182 .406 .631 .108 .019 ­.888 2.135 .651 .005 ­.356 .724 ­.192 .346 .248 · 1 19 ­.350 ­.413 ­.424 ­ 1 .085 ­.769 .582 .339 .076 ­.214 · 1 1 7 ­.324 1 .438 ­.512 .250 .854 ­.686 ­.639 .594 ­.562 .505 ­.844 .798 ­.058 .942 .729 .136 ­.663 ­.213 .183 .095 .317 ­. 161 ­.335 .546 .043 .272 ­1. 398 .416 ­.018 .147 ­.422 . 141 .449 ­.857 · 157 ­.473 .816 ­.265 ­.522 ­.959 .426 ­.300 ­.007 .437 .304 ­.395 ­1 .421 ­.514 .861 .090 ­.776 .224 .249 .489 ­.859 ­ . 118 .237 ­.867 ­.485 ­.353 ­.486 ­.403 ­.019 ­.919 .215 ­1. 025 ­.292 ­.575 .345 .176 1 .035 .361 ­.320 ­.637 ­.508 ­.154 ­.219 ­.065 ­ .183 .001 ­.201 .305 1 .641 ­.090 ­.072 ­.636 ­.340 .194 .005 ­.428 .096 ­.850 ­. 781 .627 ­.214 ­.344 ­.248 .489 .040 .007 .580 .120 .816 .208 .154 ­.026 ­.548 ｾＮＰＳＱ .129 .886 ­.106 .354 ­1 .056 ­.300 ­.800 ­.017 ­.470 .436 .154 ­.357 ­.246 .374 ­.042 ­.052 ­.367 . 114 1 .140 ­.856 ­1. 196 ­.309 .063 .296 .349 .829 ­.344 ­.159 . 117 .754 ­.133 ­.929 ­.801 .256 .897 ­.648 ­.113 ­.090 1 .516 2 ­.780 .874 1 .004 ­.551 . '90 ­.900 ­.684 .836 ­.619 ­1 .687 2.134 ­1 .279 1 .222 ­.486 ­.542 .501 ­.193 .930 ­.323 .474 ­1.628 ­.649 ­1.227 ­.456 .474 .169 .779 . 119 ­ . 161 .938 ­.909 ­.251 .627 ­1 .241 ­.807 .010 .234 .956 ­.786 1 .399 ­1 .528 ­.675 .485 ­.342 ­.418 .329 .103 .082 .092 1 .756 ­2.675 ­1. 622 3.138 .599 3.019 ­.707 ­.744 ­2.252 ­.993 ­.082 .459 ­.317 .739 1 . 160 ­3.119 ­1 .260 .045 ­2.246 .587 ­.510 ­.628 ­2.957 ­2.121 ­1.479 ­1 . 185 .160 3.197 ­3.240 1.160 ­1.917 .256 1 .313 .071 ­1.817 ­4.266 .508 3.101 .028 .617 2.797 1 .601 ­.346 1 . 769 .670 ­1 .134 2.970 ­2.656 ­2.224 Axis AX1S ­ AX1S 219- Year AD 1143 1144 1 145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1 161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 Axis AX1S 7 6 ­.596 .041 ­.442 ­.372 .210 .399 .368 ­.:230 .434 .71 2 .439 1 .149 . 1 13 ­.306 ­ . 119 .302 .338 ­.028 ­.337 ­1 .345 .273 .968 .645 .379 .209 .361 .158 .445 ­.304 .566 .164 ­.172 ­.539 ­.615 .479 ­.466 1 .359 .157 ­.125 ­1.010 .611 .407 ­.836 .258 .309 .481 .241 .365 .053 ­.533 ­.080 .302 .308 ­.037 .069 ­.444 ­.322 ­.230 .109 ­.799 ­.218 ­.924 .161 ­.114 .144 .050 ­.211 ­.400 .390 .122 .008 .350 ­.150 ­.278 .639 .350 .786 .293 ­.528 .404 . 114 .281 .727 ­.477 ­.185 ­.092 .046 .266 .439 ­.132 .089 ­.138 .076 .094 ­.083 ­.748 .072 .596 Axis Axis Axis AX1S AX1S 5 4 1 3 2 .185 ­.051 ­.015 ­.103 1 .546 ­.:220 ­.384 ­.102 .488 ­.122 .741 .074 ­1.201 .995 ­.716 ­.430 ­.001 ­.420 ­.185 4.160 .412 .794 .385 .01 1 .538 .041 ­.733 .880 .31 2 1 .768 ­.193 ­. 157 ­.526 .457 .325 .024 .042 ­.097 .360 4.472 .107 .029 ­.437 .261 3.381 .494 .107 .217 ­4.588 ­.203 .550 ­.238 ­.422 .927 ­.562 .511 ­.061 1 .037 ­.426 .277 ­.037 ­.357 .653 ­.497 ­1.118 ­.442 ­.330 ­.346 .755 3.914 .004 .628 ­.797 ­1 . 133 1 .620 .495 ­.177 .813 ­. 114 2.239 .366 ­.624 ­.192 ­.061 ­3.180 ­.160 ­.179 .465 .290 ­1. 193 ­.047 .684 .698 ­.259 3.411 .438 .009 .485 1 .949 ­5.430 ­.529 ­.112 .756 .020 ­2.244 .493 .787 ":".008 1 .427 ­.681 ­.186 ­.265 .624 .527 ­.461 .144 1 .502 1.089 ­.623 1 .950 1 .381 ­1.967 .447 .590 ­.605 1 .058 2.461 ­.303 ­.565 ­.615 ­.132 1 .086 2.217 .335 ­.461 . 131 .474 ­.714 ­.087 .034 ·.904 1 .071 ­3.389 .637 .230 ­.416 1 .014 .017 ­.741 ­.318 ­.100 .733 1. 110 .195 ­2.290 .270 .405 ­.071 ­.556 2.511 ­.676 ­1.210 .490 1 . 792 1.298 . 177 .572 .709 ­.444 .095 ­.146 ­.462 ­.549 .069 3.068 .548 .100 .272 .036 ­.989 ­.008 .441 .510 .648 .022 .403 .364 ­.685 . 191 ­.703 ­.554 ­1.294 ­1 .399 ­1.099 ­1 .426 .192 ­.666 ­.345 ­.385 2.861 .317 ­.296 ­.103 ­.190 .202 ­.221 ­.612 ­.659 ­.984 ­3.863 .214 ­.093 ­.400 .237 ­1 . 186 . 121 ­.006 .738 .120 4.696 .510 ­.252 .456 ­1 .493 ­.689 .134 .450 ­.222 ­.456 ­.235 ­.169 ­.262 ­.731 ­1 .068 1.589 .066 .042 ­.918 ­.032 ­1 .109 .612 2.618 ­.255 ­.075 ­.062 ­ 220- Year AQ 1192 1193 1 194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 Axis Axis 7 6 .346 ­.327 ­.914 .342 .341 ­.219 ­.641 ­.379 .042 ­.232 ­.683 ­.035 ­.269 ­.257 .154 ­.354 ­.022 .516 .461 ­.825 .316 ­.531 .303 ­.263 .056 ­.170 .699 .150 .410 · 158 ­.370 ­.843 ­.340 ­.051 .010 ­.060 ­.217 · 111 .005 ­.907 .081 .346 ­.659 ­.923 ­.101 ­.142 .174 ­.112 ­.107 .034 .354 .412 ­.260 .624 ­.255 ­.611 ­.426 .580 ­.307 ­.069 ­.240 ­.430 ­.215 · 110 .059 ­.048 .108 .971 ­.195 ­.206 .157 ­ .108 .254 ­.327 ­.248 1 .317 .627 .149 ­.068 .157 ­.286 .434 .511 ­.525 .662 ­.583 1 .233 ­.769 .769 .525 ­.281 ­.886 .318 .563 ­.421 ­1.278 .091 .217 Axis Axis Axis Axis AXis 4 1 3 5 2 2.791 .098 ­.445 ­.854 ­.621 .032 ­.640 ­.657 .004 ­.624 ­.454 ­.646 ­.329 ­.592 ­1.215 ­.248 ­.497 ­.416 ­2.436 ­.551 ­.586 ­.482 ­.415 ­.789 ­2.171 ­.315 ­.750 ­.703 ­1.271 ­2.095 .593 1. 713 ­.061 .005 ­.090 .616 .040 .252 1 .334 .922 .134 ­.031 ­3.549 ­.596 ­1 . 128 .444 ­2.597 ­.226 ­.766 ­.084 .210 ­1.145 ­.798 1 .029 2.249 .405 ­1 .528 .430 ­.698 ­.523 .376 ­.098 ­.066 ­.925 ­.104 ­.304 . 119 1 .017 ­.298 2.901 ­.357 ­.296 1 .233 ­.288 ­.421 .822 ­2.687 .143 .656 ­.215 .077 ­.128 .526 . 114 .046 .323 ­.772 ­2.488 ­.504 ­3.005 .029 ­.194 ­.323 ­.094 ­.727 ­.737 .473 ­1 .568 .661 .035 .750 .010 .362 ­.292 ­.211 .810 ­1 .039 .506 ­1. 1 52 .059 .083 ­.464 ­.294 ­.239 2.525 .992 3.044 ­.375 .537 ­.258 ­.033 .898 .066 2.578 .356 ­.086 ­.371 ­.343 ­.453 4.795 .135 ­.160 .728 .220 ­1 .331 ­.205 . 121 .647 ­.170 ­2.334 .328 ­.023 .336 ­.467 1 .607 ­.502 .097 ­.323 ­.625 2.237 .360 ­.678 ­.261 ­.143 ­.058 ­.865 ­.040 .225 ­.536 ­.782 .147 .350 .095 ­1 .465 1 .224 .713 ­.850 .878 .026 ­.535 ­.352 1 .074 ­.011 ­.168 ­.617 .347 .348 .406 .252 4.392 ­.177 .491 ­2.383 .532 .660 .220 ­.684 .709 .497 ­1.778 ­.772 .066 .280 .681 ­1 .489 ­.417 .537 ­.127 .066 ­.660 ­.243 .079 .757 .642 ­1.925 .197 ­.096 ­.523 ­.262 .930 .065 .945 1 .064 .067 ­.567 .196 .107 .477 1.169 ­2.155 ­.224 .386 .013 .233 1.503 ­.215 .644 .882 .149 ­2.034 ­.795 ­.546 .084 ­.221 ­2. 137 .450 .323 ­1 .138 ­1.029 ­1.056 .434 ­.119 ­.331 ­.116 2.473 ­ 221- Year AD 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 Axis 7 .251 .484 ­.149 .438 .055 ­.308 ­.902 .662 ­.758 ­.037 ­.104 .635 ­.521 .335 .069 ­.753 ­.643 ­.243 ­1.066 ­.291 ­.339 ­.534 .305 .418 ­.362 ­.339 ­.039 ­.133 .503 ­1.542 . 161 ­.613 .350 .466 ­.465 .210 ­.577 .108 ­.606 .768 ­.357 .198 ­.284 .308 .502 .089 ­.487 .720 .311 Axis 6 ­.178 ­.177 ­.370 ­.034 .505 .445 .902 ­.358 .234 ­.804 .312 ­.189 ­.231 ­.180 .073 ­.493 .170 .567 .412 ­.532 1 .493 1 .180 ­.555 .795 .074 1 .158 ­.796 .736 ­.941 .762 ­.556 .280 .389 ­.434 ­.385 .492 ­.637 .278 .169 ­.658 . 116 .430 ­.128 ­.552 .148 .078 ­.177 ­.718 .001 Axis Axis Axis 5 4 3 ­.319 .630 ­.320 .089 ­.053 ­.140 .891 ­.182 .095 ­.125 .430 .859 .476 ­.718 .136 . 161 ­.457 .065 .004 ­.572 .379 .328 ­.274 ­.981 .021 ­.518 .568 .216 ­.037 ­.967 .159 ­.561 ­.352 .677 .007 ­.050 .088 ­.191 ­1.232 ­.156 .189 .549 .922 .555 ­ . 151 .338 ­.094 ­.551 ­.280 .285 ­.325 ­.310 ­1 .050 .162 .265 .975 1 .352 ­.222 .268 ­.807 1 .353 .317 .941 .035 .043 ­.967 ­.143 1 .062 .521 .390 ­.702 ­.284 ­.469 ­.367 ­.644 .594 .972 .658 .012 . 117 ­.378 . 135 ­.140 .373 ­.515 ­1. 209 ｾ Ｑ Ｎ Ｖ Ｒ Ｗ .576 .018 ­.588 ­.242 ­1 .042 ­.535 ­.480 ­.323 ­.893 .466 1 .269 .415 .335 .574 .237 ­.841 1.129 1 .255 .376 ­.405 .195 ­.617 1.106 ­.092 .708 ­.175 .700 ­.747 .216 1 .183 ­.468 ­.412 ­.181 1 .078 .382 ­.248 ­.411 . 166 ­.218 ­.828 ­.155 .697 .265 .130 ­.750 ­.373 .102 ­.805 ­.331 ­.197 ­.354 .067 ­.162 ­1.370 ­.193 .245 ­1.168 .419 ­.613 ­.707 ­ 222- Axis 2 1 .364 ­.356 .998 .588 1 . 176 ­.107 ­.001 ­2.634 ­.393 .500 ­.872 ­1.399 .035 ­.053 ­.786 ­.741 .574 ­.310 . 141 .775 ­1 .500 .481 ­.666 ­.460 ­.518 .641 .227 1 .514 ­.545 .653 ­.825 .765 ­.147 .300 ­.297 ­.507 .683 ­.429 .389 ­.590 .298 ­.818 1 .575 ­.317 .894 ­1.425 2.323 ­.517 .506 Axis 1 ­3.654 ­.871 .040 ­2.043 ­.717 3.137 ­.342 .362 ­1 .830 1 .584 4. 141 1 .747 ­1.339 6.078 ­1 .946 .875 ­1 .954 4.493 ­4.559 1 .044 .380 ­2.636 2.257 .992 ­.655 ­1. 398 ­1 .551 ­3.014 .086 ­.126 ­1 .184 .015 1 .969 .899 ­3.821 4.534 ­1.052 3.490 ­2.051 4.473 ­1.221 1.352 ­.066 1.680 .864 2.244 ­.706 3.600 ­.612 Year AD 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1 311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 Axis Axis 6 7 ­.470 ­.009 .543 ­.414 ­.215 ­.876 ­.275 .409 .486 .106 .109 ­.060 .412 .319 .192 1.106 1 .074 ­.219 ­.358 ­.308 .417 .707 .655 ­.792 .734 .154 .596 ­.160 .323 ­.783 ­.098 1 .230 ­.210 ­.503 ­.096 ­.263 .542 ­.288 .262 1 .087 ­.192 1 .029 ­.342 ­.091 .040 ­.266 ­.531 .453 .008 ­.640 .048 ­1. 939 .077 ­.277 ­.330 .189 . 151 ­.641 .592 ­.396 .072 ­1 .261 .107 ­.182 .346 ­ . 11 1 ­.001 .220 .342 ­.493 .526 .323 ­.416 .302 .285 ­.038 .633 . 128 .173 ­.330 .107 .421 ­.209 ­.094 .537 ­.371 .063 .308 .201 .287 .243 .045 ­.142 .305 ­.093 .324 ­.326 ­.173 Axis 5 ­.803 .435 .755 ­.461 ­.012 ­.200 .268 ­.373 ­.339 ­.150 ­.335 ­.275 .149 .002 . 175 ­.553 ­ . 211 ­. 119 ­.075 ­.087 ­.174 . 185 .414 ­.039 .210 .762 .545 .549 .340 ­.033 .310 .416 .505 ­.203 ­.230 ­.323 .169 ­.086 ­.024 ­.386 .441 .353 ­.686 ­.739 .060 .059 ­.506 ­.373 ­.327 ­ Axis 4 .053 ­.017 ­.380 .715 ­.244 .134 .421 ­.681 ­.108 ­.674 ­.017 ­.504 ­.867 ­.035 .018 ­.675 .388 .426 .656 ­.927 .012 ­.572 ­.857 ­1 .031 .549 ­.859 1.136 ­.050 .425 ­.467 ­.096 ­.380 . 113 .248 .179 ­.535 ­.389 ­.351 ­.257 .601 ­.125 ­.407 ­.248 ­1 .468 ­ 1 . 764 ­.180 ­.519 ­.079 ｾＸＵＹ 223- Axis Axis Axis 1 2 3 ­2.153 1. 118 ­2.810 .687 ­1 .924 1 .607 ­.249 .955 1 .308 ­.223 1 .293 ­2.566 .241 ­1 .252 1 .415 ­.695 2.195 2.427 1 .300 ­1 .441 .301 ­.548 1 .467 ­.863 .258 .564 ­3.113 ­1 .657 2.100 2.314 ­1 . 108 1 .381 ­.533 ｾＮＷＵ .093 ­1.463 ­.020 .266 ­2.039 1 .334 ­1.810 ­1. 137 .250 ­1.008 1 .697 ­.189 ­.781 ­3.775 .003 ­.262 1 .388 ­.446 1 .619 1 .413 .993 ­ . 111 1 .038 .120 ­.598 ­3.514 1 .580 ­1.059 ­3.003 .515 ­.474 .244 .734 ­1 .964 ­.518 ­.422 .150 ­4.432 .545 1 .061 ­1.013 ­1.212 ­.978 3.368 1 .548 .476 3.229 .743 .826 ­1.596 .432 .414 ­2.352 ­.698 ­.331 ­1 .261 ­.681 ­.655 ­.642 .263 ­.412 ­2.125 1 .375 ­.385 ­.443 .695 ­.640 .625 ­.582 ­.059 1 .921 ­1 .068 ­.316 ­4.279 .827 ­1 .343 ­.281 .721 ­.208 ­.327 ­.109 ­.260 2.154 ­.933 .959 ­.180 1.190 ­.933 ­3.199 .894 ­1 .060 ­.326 .372 ­.024 ­2.891 .845 ­.623 ­2.910 ­.659 ­1 .044 ­.504 ­.440 .385 2.855 3.133 ­2.636 ­2. 159 ­.005 ­1.634 ­1 .029 ­.150 .842 3.885 Year AD 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 Axis 7 ­.158 .289 .425 .065 .096 ­.330 . 11 7 .256 ­.205 .260 ­.109 .505 .027 ­.074 .139 .376 ­.536 ­.256 ­.014 .005 . 182 .382 ­.052 .041 .090 .468 .126 .060 .236 .305 ­.745 ­.254 ­.164 ­.112 .096 ­.072 .346 ­.446 ­.062 ­.267 .104 ­.089 .016 .609 .069 .496 ­.004 .226 .292 Axis Axis Axis AX1S 6 4 5 3 ­.013 ­.506 .869 .077 ­.106 ­.272 .202 ­.426 ­.446 .280 ­.186 ­.905 .245 ­.103 .212 .240 .637 ­.341 ­.816 · 135 .033 ­.155 .082 .395 .036 ­.066 ­.629 · 167 . .473 ­.953 .409 ­.366 .076 .312 ­.091 .255 .663 ­.360 .098 ­.294 .166 ­.399 .895 ­1.127 .702 .595 .204 ­.078 .494 ­.312 .885 ­.015 .289 ­.323 ­.039 ­.262 .262 ­. 196 .285 ­1 .743 ­.793 .127 ­.455 .301 .437 ­.323 ­.270 ­.988 ­.423 ­.158 .362 ­.578 .722 ­.355 .865 .936 .933 .041 .468 .558 .577 ­.252 ­.272 .740 ­.650 ­.122 ­.178 .015 .767 ­.675 ­.154 .445 .466 1.007 .484 .226 ­.380 ­.181 ­.516 .466 .028 ­.023 .416 .540 ­.221 ­.602 ­.514 ­.346 ­.247 ­.363 ­.442 .126 ­.400 ­.209 .015 .519 ­.730 ­.674 .751 ­.583 .311 ­.808 ­.018 · 196 .365 ­.479 ­.744 ­.774 .823 .138 ­.083 .288 .725 ­.800 ­.004 .015 .590 .013 .821 .289 .350 ­.387 .479 ­.836 .565 .080 1 .232 .062 ­.378 .954 ­.128 ­.373 .340 ­.751 1 .252 .098 .175 ­.239 ­.253 ­1 .063 ­.038 ­.499 .494 ­.773 .329 ­.167 ­.773 ­1.205 .244 .069 .618 ­.143 ­.583 .375 ­1.147 ­1.514 .453 ­.734 ­.554 ­1 .839 .155 .228 ­.379 · 11 3 ­.612 .376 .241 .310 .374 ­.349 ­ .135 .206 .327 .635 ­.052 .370 Axis 2 ­.494 .208 ­.193 ­.616 .477 .911 .333 1 .297 ­1 .078 .055 .227 .741 1 .603 ­1 .522 .370 ­.375 .503 .21 7 ­1.375 .074 ­.661 .079 1 . 147 ­ .190 ­.202 .505 1.048 ­1 .098 . 125 .902 ­.510 .284 .095 1 .018 ­.281 ­.108 .256 .055 ­.600 .898 .017 ­1 .887 .044 ­2.481 .013 ­.450 .930 .019 .184 1 ­1 .834 2.740 1 .919 .925 .893 ­.846 ­.392 ­.375 4.198 .258 ­1 .567 1 .446 .458 2.940 ­3.461 ­1. 895 1 .621 ­2. 114 ­1 .055 ­3.610 ­2.475 1 .821 ­.397 2.123 2.658 1.158 ­.386 .065 .045 ­2.577 3.051 ­2.897 .581 ­1.632 ­.298 ­2.048 2.089 2.805 . 113 ­1 .060 .749 ­2.148 ­.207 .484 ­3.207 ­.187 ­1.072 ­1 .069 ­.349 ­ Axis 224- Year AD 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 . 1433 1434 1435 1436 AX1S 7 ­.280 ­.280 .172 ­.757 ­ . 111 ­.029 .261 ­.379 ­.292 .352 ­.328 ­.200 ­.460 ­.219 ­.203 ­.117 ­1.043 ­.051 ­.594 ­.253 ­.038 ­.148 .203 ­.013 ­.471 ­.767 ­.307 ­.404 .058 ­.136 ­.561 ­.787 ­.344 ­.402 ­.401 . 199 ­.086 ­.040 .007 ­.623 ­.424 .562 .011 .308 ­.092 .516 .064 ­.424 ­.078 Axis Axis 5 6 ­.904 .812 .001 ­.596 .823 .470 .194 ­.886 .468 ­.627 ­ . 113 1 .168 .102 ­.839 .790 .51 6 ­ . 151 ­.756 .095 ­.456 .027 ­.494 .892 .309 .525 ­.172 ­.364 ­.206 ­.189 ­.243 .028 .109 .434 ­.185 .380 .466 .339 ­.408 .074 .235 .011 .468 ­.554 ­.750 .588 ­1 .330 .384 .467 ­.703 ­.632 .437 ­. 151 ­.407 ­.957 .020 ­.020 ­.084 ­.259 .441 ­.744 .392 ­.427 ­.164 ­.130 ­.088 ­ .144 .713 .198 ­.389 ­.380 . 71 1 ­.425 ­.139 .065 .040 ­.242 .348 .406 .028 ­.520 ­.990 ­.248 ­.247 .575 .063 ­.015 .334 .863 .510 . 113 .356 .719 ­.008 ­.576 .720 ­.031 ­.009 .521 ­ AX1S Axis Axis Axis 1 4 3 ... ­.438 .136 .025 .318 .417 .339 .292 1 .358 .421 1 .051 .034 3.597 ­.019 .315 ­.006 ­.045 ­.420 .737 .646 ­1.479 .369 .223 .176 ­.887 ­.102 ­.167 ­.200 ­.365 ­.231 .167 1 .241 ­1 .446 .284 .574 ­.036 1 . 164 .079 ­.664 ­.297 1 .909 .874 .175 1 .691 ­.937 ­.178 ­.170 ­.710 3.814 .;.. . 179 .378 .211 2.527 .139 .981 .152 .122 ­.083 .770 .294 .997 ­.159 .834 1 .519 .168 .858 ­.33"3 1 .020 ­2.619 .071 1 .008 .065 ­.196 .307 ­.475 .379 ­1.876 ­.532 ­.461 ­.667 2.906 .829 ­1 .546 .098 ­.393 .155 ­.239 ­.907 ­3.097 ­.585 ­.683 .898 ­1.215 .357 2.850 .578 .392 .274 ­.183 .537 ­3.298 .174 .077 1 .419 4.222 .073 ­1 .426 1 . 105 ­3.770 ­.004 .886 ­.839 3.460 .061 .645 .557 ­.381 .547 ­.762 ­.280 ­.829 ­.621 .593 ­.296 2.608 ­.141 .990 1 .053 ­.762 .529 1 .454 ­1 .367 ­.160 ­.175 ­1 .052 1 .981 1 .745 .447 .704 .803 ­1 . 701 ­.141 ­.255 ­.478 3.417 ­.246 1 .272 ­1 .924 1 .051 .514 .291 ­1.378 .867 .490 .991 ­1.171 ­2.860 .223 .051 .358 ­1 .824 .915 .677 ­4.557 .322 .546 ­.375 ­.960 .383 ­.015 .234 ­.157 ­.244 .519 ­.138 .242 ­1 .459 ­.207 .282 ­1 .598 .235 .487 ｾ Ｎ Ｐ Ｑ Ｘ .319 ­3.363 .663 .381 1 .076 ­1 . 186 ­.947 .574 1 .047 .337 ­.090 ­1.272 ­.011 .956 ') 225- Year AD 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 Axis 7 .581 ­.495 ­.467 ­ . 118 ­.178 ­.018 ­.201 ­.045 .327 .392 ­.168 .176 ­.227 .105 .459 .224 .597 .072 .176 .571 ­.395 .521 ­.246 ­.293 ­.020 .060 .494 ­.449 .134 .446 ­ .147 .013 ­. 127 ­.041 ­.246 ­.143 ­.215 .437 ­.289 .724 .030 .833 ­.259 ­.054 .615 .421 ­.669 .531 .360 Axis 6 .125 .169 .537 ­.056 .145 ­.326 .125 ­.891 .376 ­.045 ­.227 .703 .572 .205 ­ . 113 .407 ­.302 .071 ­.434 .338 ­.308 ­.452 ­.356 .583 ­.343 ­.989 ­.499 ­.152 ­.675 ­.353 ­.127 .404 ­.332 .147 .186 ­.870 .616 .136 .504 .067 ­.007 ­.899 ­.059 .178 ­.408 ­.568 .642 ­.690 ­.490 Axis 5 .652 .680 .199 ­.134 .550 .556 ­.749 .148 .582 ­.324 ­.951 ­.327 .311 .023 ­.329 .007 .476 .344 ­.366 ­.364 .631 ­.425 ­.050 ­.827 .452 .236 .457 ­.302 ­.044 . 161 .763 1 . 147 .826 ­. 711 .328 ­.120 .345 .412 ­.194 .335 .318 .143 ­.171 .396 ­.519 .469 .437 .004 ­.589 ­ Axis 4 ­.581 .089 .821 ­.186 .236 ­.348 .050 .032 ­.247 .431 ­.119 ­.283 ­.333 .266 ­.345 . 151 ­.245 ­.719 .282 .221 ­.613 ­.040 .436 ­.522 .344 ­.499 ­.558 .005 ­.430 ­.127 ­.454 ­.187 .418 .432 .395 .280 .285 ­.279 .303 ­.725 ­.058 ­.182 .002 ­.048 1 .334 ­.423 .006 .452 .107 226- Axis AX1S Axis 3 1 2 .339 .21 6 ­.004 .109 .572 3.282 ­.051 .712 ­2.285 .345 .160 ­.502 ­.200 ­.269 ­2.536 ­.297 1 .881 1. 736 ­1 .064 1 .535 ­.924 .445 ­.683 .281 .122 ­.609 2.998 ­.586 2.159 ­.980 ­.176 .923 ­.408 ­.843 .728 .996 .444 .039 2.844 .406 ­.084 2.558 ­.753 ­.483 ­1 .843 ­.129 ­.129 .593 .463 ­.408 ­.782 ­.337 .165 .307 .354 .973 3.961 ­.695 ­.524 .404 .272 1.199 3.278 ­.923 ­.945 ­3.337 .248 ­.612 1.086 1 .268 ­.617 . 186 1 .406 ­1. 179 2.022 .034 ­1. 379 ­2.244 ­.092 .085 1 .997 .820 1 .052 3.183 .. 369 1.150 .449 .387 . 161 ­3.081 ­.016 ­.691 ­2.181 ­.643 .832 .611 ­.306 ­.322 ­1.723 .013 ­.536 .848 .198 ­. 197 4.219 ­.484 .878 ­.528 .597 .529 1 .895 ­.397 ­.763 1 .255 ­.315 ­.542 3.479 ­.378 .325 ­.542 ­.176 ­1.760 .602 ­.045 .707 ­1.599 ­.343 1 .256 1.699 .259 ­.276 4.609 ­.061 ­.105 ­1. 929 1.189 ­.340 ­.982 ­.193 ­.431 .762 .625 ­.370 ­5.364 1 .083 .564 ­2.560 Year AD 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 Axis Axis Axis Axis AX1S Axis Axis 7 4 6 2 1 5 3 ­.570 ­1 .459 1.304 .524 ­.786 .543 ­2.439 .402 ­.441 ­.185 ­.253 ­.573 .282 2.807 .920 ­1 .091 .971 .276 .059 ­.771 ­1. 782 ­.415 ­.440 ­.261 ­.377 .429 ­.051 ­1 .522 ­.358 ­.324 .518 .815 ­.070 .103 ­.948 ­.289 ­.075 ­.153 ­.147 ­.269 ­2.002 ­2.830 .382 .178 ­.234 ­. 137 .189 .867 ­.954 ­.017 .174 .616 ­.688 ­1 .007 .642 .033 .551 ­.792 .579 ­.125 .634 .537 ­1 .071 .703 .145 .543 6.018 .256 · 110 ­.425 .298 1 .003 .868 ­.044 .866 ­.427 ­.226 .108 ­.278 .076 ­.089 ­.269 ­.165 1 .362 .542 .259 ­.439 ­3.829 1 .028 ­.383 ­.657 ­.406 .143 ­.773 .334 1 .072 ­.976 .695 ­.471 1 .327 .070 .751 3.997 .070 .688 .349 ­.509 .241 .062 .405 ­.286 .165 ­. 161 .018 .052 ­.964 ­.720 .254 ­.136 .168 ­.217 .145 ­.750 .104 ­.062 1 .082 ­.535 ­.138 .904 ­.746 ­.606 .724 ­.022 ­.310 ­1.007 ­.204 .056 .807 1 .080 1 . 1 70 .127 .632 .109 .125 .439 4.541 .169 .054 .039 ­1 .467 ­.208 1 .214 · 112 ­.031 .688 ­.083 1 .483 ­.624 .714 .654 ­.133 ­.164 ­.221 ­.853 ­.993 ­1 .006 .202 ­.241 ­.082 .157 .514 1 .350 1 .650 .076 .599 .778 ­.830 ­.866 ­1 .047 1 .660 ­.206 ­3. 110 .387 .572 ­.791 ­.037 .610 ­.699 ­1.078 .247 .581 ­1 .460 .084 .614 ­3.292 .618 ­.335 .573 ­.143 .104 ­1 .094 .399 ­1 .606 .061 .078 ­.695 1 .475 .941 .474 ­2.556 .050 .011 .439 .317 .930 ­.828 5.238 .084 ­.667 ­.402 .080 ­.593 .211 2.299 .032 .870 .988 1 .344 ­.773 · 113 ­.807 .807 ­.466 .700 ­.626 .157 ­.194 ­1.663 .333 .536 ­.337 ­.072 ­.265 .337 ­1 .704 ­.095 ­.615 ­.178 .073 .908 1 .921 ­1.671 ­1 .088 .041 .456 .791 ­.034 .282 3.674 ­.534 ­.858 ­.844 .348 ­.566 ­.300 1.259 .316 .209 .389 .562 ­.250 ­.904 2.115 ­.584 ­.633 . 115 .274 .057 ­.043 . 11 5 .436 ­.366 .252 .690 ­.321 ­.853 ­'1.261 ­.098 .300 ­.418 ­.795 .231 ­.274 .096 ­1 .032 ­.407 ­.119 1 .043 ­.098 .208 .228 ­.667 ­.706 ­.911 .731 ­.319 .972 ­1.751 ­.326 .498 .252 ­.231 ­.123 ­.024 ­.058 ­.139 .372 ­.160 .705 .134 .348 ­1 .106 ­.245 .124 ­.075 .109 1 .367 .768 2.862 .498 .043 .762 ­.109 ­1 . 133 .361 .273 ­.252 .124 ­.051 ­.218 .245 .544 ­.557 ­ 227- Year AD 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 Axis 7 .259 .202 .103 ­.387 1 . 103 .927 .877 ­.192 .956 ­.384 .364 .626 .145 ­.443 ­.125 . 110 .097 ­.338 ­.692 ­.725 ­.517 ­.199 .251 .314 ­.406 ­.355 .232 .569 ­.636 .270 .436 .569 ­.363 .299 1 .047 .606 ­.143 1 .029 ­.084 .534 .224 ­1 • 149 1 .736 ­.071 ­.381 ­.051 ­.083 ­.073 ­.402 Axis 6 .207 ­.319 ­.006 .087 ­.291 .501 ­. 199 . ­.180 . 141 ­1.113 .094 ­.429 ­.682 ­.748 .230 .357 .325 .220 .059 .487 ­1 .099 ­.140 .082 .097 ­1 .288 .584 ­.323 ­.429 ­.201 ­.521 ­.252 ­.249 ­.261 ­.395 ­.330 .098 ­.720 ­.267 .005 ­.567 .220 .666 ­.419 ­.087 .436 ­.159 ­.131 .506 .564 Axis 5 .207 .998 .185 .466 ­.001 .364 .409 ­.099 .225 .840 ­.738 .379 .384 ­.169 ­.265 ­.736 .446 .088 ­.011 · 114 ­.926 · 122 ­.032 .679 .037 .414 ­.093 .580 .074 ­.117 .429 · 115 .781 .515 ­.223 ­.960 ­.675 ­.802 · 166 ­ ..320 .422 .135 .590 .284 ­.002 ­.233 .072 ­.027 ­ .168 ­ Axis Axis Axis AX1S 4 1 3 2 ­.472 .656 ­.832 1 .748 ­.455 ­.150 .318 ­1.714 .603 .523 ­.351 ­1.761 ­.552 .479 .776 2.335 ­.019 .496 .769 ­1. 961 ­.557 ­.444 1 .463 ­2.752 ­.090 ­.243 1 . 109 ­.897 .044 1 .31 1 ­.124 4.586 ­.169 ­.172 .879 ­2.909 .747 .583 ­.246 .480 ­1 .028 .583 ­.801 1 .880 .693 .,...304 ­.813 .840 1 . 136 ­.499 .679 .431 ­.298 ­. 161 ­.559 .767 1 .007 ­.000 ­.787 .382 ­.522 ­1.010 .190 ­2.041 .207 .154 ­.692 1 .699 .626 ­.992 . 111 .249 .344 .440 ­1.800 ­1 .301 ­.143 ­. 131 1 .394 1 .318 .801 .242 .708 ­1 .495 ­1 . 103 ­.823 ­1 .244 ­1 .931 .504 ­.025 ­.362 ­1 .780 .407 .633 .299 ­.331 ­.403 .710 .358 .127 .066 ­.825 ­1.667 2.315 .376 ­.681 .997 1 . 701 .251 .355 ­.750 2.812 ­.369 .052 ­.989 ­.306 ­.309 1.164 ­.496 ­.360 .007 .764 ­.395 ­2.133 ­.146 .668 .521 1 .318 .055 .196 2.094 .220 ­1 .083 ­.261 ­.280 ­.129 .766 1.150 ­1.417 ­.282 ­.391 ­.372 ­.245 ­1 .831 ­.218 1 .071 ­.836 .290 .496 ­1 .525 ­.442 .874 ­.254 ­.201 ­.288 4.757 .487 1.100 .449 .254 ­.347 ­.420 .400 .739 ­.417 ­.909 ­.014 2.065 1 .597 ­2.252 .034 .241 . 11 1 1.278 ­.509 .483 ­.166 .427 ­.613 4.307 ­.774 3.362 ­.358 ­.901 .334 ­.074 .061 1 .324 .412 2.154 .038 .056 .335 ­ .134 ­.385 3.664 228- Year AD 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 .1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 Axis Axis 7 6 .314 ­.389 ­.424 .421 ­.027 ­.156 ­.219 .521 ­.302 ­.076 ­.779 .358 ­1.012 .120 ­.243 .339 ­1 .253 .862 ­.923 ­.135 .514 ­.041 ­.347 .202 1 .349 ­.057 ­.274 .398 .064 .305 .075 1 .905 .345 ­.982 .651 ­.298 .372 .239 ­.245 ­.712 ­.618 ­.361 ­.043 ­.265 1 .066 ­.304 ­.242 .245 .511 ­.134 .436 ­.326 1 .040 .422 .588 .274 ­.463 .145 ­.066 .142 ­.656 .555 .075 ­.363 .012 .644 .369 .153 ­.648 ­.023 .342 .115 ­.651 .726 .014 ­.303 . 119 .532 ­.540 .346 .081 ­.859 .311 ­.024 ­.469 .507 1 .062 ­1. 192 .169 . 180 ­.028 ­.013 .124 ­.259 ­.278 .235 ­1 .138 ­.246 Axis Axis Axis Axis Axis 4 5 3 2 .370 ­ . 177 .265 .930 .184 ­.095 .041 ­.343 .292 ­.806 ­.902 .005 .057 .086 .583 .422 ­.514 ­.142 ­.393 ­.330 ­.666 ".312 ­.740 1 . 155 .91 1 ­.052 ­.273 .229 ­.213 .027 ­.522 ­.655 ­1.077 ­.073 1 .021 ­ .145 .368 ­.350 .352 ­1.176 1 .088 ­.069 ­.436 ­.683 .087 .064 ­.372 .899 .264 2.213 .453 .009 ­.755 ­.950 ­.381 2. 135 .104 ­.328 .208 ­.496 .224 ­.274 .536 .434 .734 ­.819 ­.007 .158 ­.155 .001 ­.518 ­1 .495 ­.346 ­.439 ­.391 .588 ­.240 .022 .800 .628 ­.783 .058 .508 .298 ­.220 ­.077 ­.104 ­.465 ­.278 ­.304 ­.033 .062 .575 ­.009 .284 .667 ­.707 .260 ­.782 ­.899 .140 .256 ­.245 ­ .198 .404 ­.966 ­.238 .836 .041 1 .015 ­.477 1 .267 ­.055 .495 .298 .882 ­.581 .904 ­1.911 .511 .287 .382 .051 ­.214 1 .393 ­.629 ­.124 ­.363 ­.024 ­.224 ­.599 ­1. 243 .703 .782 ­.743 ­.778 .528 . 196 ­.909 ­.200 . 151 .266 ­.391 .969 .644 ­.134 ­.740 .650 1 .512 .105 ­.020 ­.058 .494 ­.060 ­.269 .474 .774 .317 ­.012 1 .855 ­.013 .828 ­.445 ­.381 .304 ­.330 .506 ­.534 ­.345 .471 .231 .580 ­.238 .318 .082 ­.193 ­.148 .299 ­.137 ­.780 ­.008 ­.345 ­.284 .691 .019 .722 ­.555 ­.075 .325 .006 ­1. 135 .073 .158 .328 ­.042 .937 1 2.551 4.020 ­.662 1 .903 ­.892 . 110 3.484 ­ 229- ｾＮＶＷＸ 1 .994 3.143 ­3.804 ­1 .060 ­2.637 ­1.137 2.566 ­2.327 3.394 .846 .026 ­1 .091 ­.043 ­1.871 .136 .844 ­1.627 ­1.788 ­4.584 ­2.507 ­.690 ­1.679 .400 ­.602 .730 ­2.964 ­2.825 1 .520 ­2.908 ­3.191 .543 .989 2.379 2.430 1.423 ­2.472 1 .542 ­2.369 ­1.208 1 . 181 2.448 Year AD 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 Axis 7 .636 ­.312 ­.942 .005 ­.946 .006 ­.393 ­.350 .609 ­.200 .401 .342 .294 ­.166 · 1 19 ­.776 .214 .247 .462 .795 ­ . 151 .454 .636 .053 ­.184 · 174 .318 · 171 .506 ­.822 ­.285 .073 .420 .350 ­.008 .068 .021 ­.301 .213 · 162 · 169 · 192 ­.068 ­.246 ­.159 .705 ­.317 ­.109 ­.843 Axis 6 ­.849 ­.281 ­.405 .668 ­.713 ­.003 ­.096 .222 ­.310 .448 .371 ­.216 .685 .110 .995 .280 .477 .220 .514 1 .391 ­.720 .662 .440 .780 .471 .166 ­.676 ­.538 ­.050 .430 .400 .610 ­.087 ­.095 .205 ­.404 ­.589 ­.027 ­.638 ­.525 ­.205 ­.864 .772 ­.274 ­.096 ­.749 ­. 101 ­.859 .089 Axis Axis Ax­i s Axis Axis 5 3 4 1 2 1 . 1 14 .169 ­.510 ­.541 ­1 .709 .066 ­.064 .049 .372 ­1 . 144 .258 ­.626 ­1 .056 ­.345 ­1 .360 .872 .159 1 .898 .153 .492 . 761 ­.122 ­.176 1 .006 .674 .580 ­.295 ­.357 ­.823 2.322 ­.459 .993 .050 .747 ­1 .522 .390 ­.162 ­1 .676 ­.443 ­.247 ­.012 ­1 .481 .478 ­.341 .617 .634 ­.268 ­.164 ­.186 1 .025 .195 .103 .053 .453 ­.774 .015 1 .616 ­.218 ­.327 ­.263 .517 ­.609 2.025 .613 ­.388 . ·148 ­1 .954 ­.074 .204 ­2.677 ­1 .402 .687 .243 .100 ­1. 839 ­1 .294 .787 .507 ­.094 2.179 .104 ­.204 ­1 . 105 .314 ­1 .407 ­.417 ­.246 .216 ­.814 ­.419 ­.202 .996 .086 .010 ­3.629 .547 .288 1 .273 1.204 ­.700 ­1.578 ­.717 ­.025 ­.810 ­.378 .527 .678 1 .280 2.390 .258 ­.871 ­.012 ­.291 ­.346 ­3.206 .398 ­.402 .243 ­.895 ­.314 .105 ­.619 ­.053 .610 2.556 ­.143 ­.168 .221 .449 1 .846 .303 .153 ­.000 ­.719 .848 ­.308 .464 ­.531 ­.048 ­1. 660 '.048 1 .537 ­.556 .450 .530 ­.290 ­.437 .780 ­.448 .991 .210 .510 .297 .271 .322 .557 ­.384 .350 ­.330 3.619 .420 ­.478 .027 ­.727 ­1.749 ­.299 .189 .629 ­.170 1 .929 ­. 175 . 114 .059 .218 2.860 ­.058 ­.248 ­.583 .332 2.384 ­.378 ­.274 .886 ­.474 1 .239 ­.609 .251 ­1 .073 .086 1 . 779 .480 ­.142 ­.352 ­.363 ­.029 .190 . 116 .107 ­.369 . ­.258 ­.230 .089 ­.166 ­.461 ­1.747 .541 ­.407 ­.700 ­.209 ­.760 .543 ­.651 ­.267 ­.230 ­.978 .229 .130 .427 ­.044 3.046 . 101 ­.231 ­.454 .294 .106 ­.434 ­1 .441 ­.366 ­.608 .233 .463 .218 .927 .553 .534 .541 ­.285 .334 . 181 ­2.740 .572 .073 ­.098 ­1 .271 .613 ­ 230- Year AD 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 Axis 7 .425 .349 .304 .162 ­.440 ­ . 11 7 ­.114 ­.119 .010 .086 ­.104 ­.061 ­.158 .077 ­.055 .106 .059 .235 ­.151 .217 ­.187 ­.221 ­.431 .562 .088 .144 .378 ­.136 ­.079 ­.583 .359 ­.305 .046 .467 .203 ­.201 . 110 ­.001 .144 .103 .570 .029 ­.124 .035 .013 ­.118 .012 .545 .076 Axis Axis Axis 4 6 5 3 ­.209 .246 ­.474 ­.304 ­.092 .132 .963 ­.236 ­.712 ­.540 .080 .021 ­.162 ­.045 .743 . 191 ­.257 ­.424 1 .018 ­.640 .015 ­.472 .193 ­.227 .412 ­.493 .955 ­.357 ­.754 ­.264 ­.013 ­.242 .438 ­.005 ­.379 ­.928 .036 ­.255 ­.107 ­.756 ­.684 1 .127 ­.175 ­.147 .496 .015 .375 .104 .738 ­.428 .660 .195 .782 .105 .143 .045 ­.344 .049 ­.004 .075 .341 ­.039 .823 ­.095 .376 ­.002 ­.556 ­.415 .179 .651 ­.308 1 .170 ­.712 ­.228 ­1 .304 .405 .545 .501 .720 ­.313 .347 ­.400 .612 .172 .885 ­.160 ­.680 ­.032 ­.682 .188 .396 ­.232 ­.186 .055 ­.186 ­.414 .860 .275 ­.934 .139 .511 .237 ­.059 ­.066 .061 ­.323 ­.555 ­.323 ­.104 1 .014 ­.170 .510 1 .304 ­.575 .378 ­.526 .401 ­.012 .129 ­.469 .463 ­.299 .054 .237 ­ .132 .060 .208 ­.698 .312 .997 .353 1 .021 ­.412 ­.624 .257 ­.522 ­.377 ­.127 ­.651 .737 .094 ­.092 .237 ­.124 ­.077 .590 ­.448 . 111 .350 .100 ­1.296 ­.593 .046 .811 .142 ­.601 .215 ­1.675 ­.112 ­.687 .062 ­ .911 ­.252 .422 .565 .249 .258 ­.750 ­.786 ­.393 .407 ­.214 .468 .269 .060 ­. 181 ­.008 .368 .401 .243 ­.489 ­.190 ­.477 ­1 .442 ­.814 . 116 ­.000 ­.635 ­.436 ­.106 .579 .666 .048 ­.381 .549 ­.319 2 ­.423 ­.361 .616 .037 ­.192 .148 ­.579 .743 ­.725 ­.235 ­.003 ­.391 .527 .251 ­.327 ­.320 ­1 .003 ­1.178 ­.396 ­.815 ­.007 .413 .020 ­1. 717 ­.673 .936 .807 ­.994 ­.155 ­.970 ­.025 1 .341 ­.956 ­.774 ­.164 .669 ­.616 ­1 .435 ­1 .449 1 .364 .899 ­.121 1 .682 ­1 .960 .046 . 126 .515 .319 ­.739 ­.680 ­1.811 3.198 5.023 ­1 .335 ­.377 ­.262 ­3.007 ­1 .067 .366 ­3.461 ­.672 .728 ­.742 3.349 ­1 . 190 .782 ­3.265 .863 ­3.134 1.488 ­.363 2.196 ­.192 ­3. 161 2.588 ­.649 .650 ­3.524 .565 ­1.825 .544 1 .835 .519 2.890 ­.180 ­1 .357 1.529 ­5. 114 ­.259 .516 ­1.985 1.688 ­2.471 ­2.985 ­.247 1 .760 4.854 ­.045 Axis Axis ­ Axis 231- 1 Year AD 17 31 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 17541755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 Axis 7 . 152 ­.012 ­.828 .332 ­.433 .325 .269 .209 .359 .609 .001 ­.037 .055 .047 ­.095 ­.491 ­1.033 ­.251 .265 ­.414 ­.407 ­.685 .004 -.818 -.754 -.222 .103 .562 -.016 .447 -. 131 -.334 .258 -.388 .054 . 1 12 .377 -.438 -.640 -.259 -.740 -.883 -.054 -.013 -.460 -.762 -.202 -.012 -.399 Axis 6 ­.157 ­.500 ­.759 ­.412 ­.099 ­.360 ­.025 ­.780 ­.785 ­.083 ­.450 ­.373 ­.653 ­ . 116 ­.504 .500 1 .745 ­.283 .325 .195 ­.493 ­.771 ­.758 -.587 .009 -.093 -.447 .139 .072 -.104 -.427 .053 -.625 .160 .087 .637 -.637 1 .217 .914 .050 1.438 .917 -.650 -.129 .046 .183 -.194 -.112 -.192 Axis 5 .061 .920 ­.479 .299 ­1. 057 ­1.113 .397 .665 ­1.131 .417 ­.449 .162 .527 ­.569 .206 ­.045 ­.360 ­.274 .742 .524 ­1.024 .153 ­.440 .776 .169 .395 -.504 -.106 -.418 .517 -.843 -.007 -.364 .445 -1 .998 1 .065 .265 .361 .984 -.834 -.627 .584 -.696 .814 .332 -.299 .103 -.850 .545 - Axis 4 ­.036 .564 ­.233 ­ 1 . 134 ­.669 ­.325 .039 ­.074 .066 .366 ­.412 .21 7 .170 ­1 .454 .203 .132 ­.338 .164 .083 .292 ­1 .607 .388 ­.784 1 .003 -.274 -1.243 .366 -.068 .008 1 .311 -.267 .347 -.692 .737 .513 .045 -.548 .242 -.433 .705 .889 1 .248 .557 .428 -.032 -.020 -.712 .167 -.260 232- Axis AX1S 3 2 .081 .258 ­.912 ­.355 ­ . 161 . 1 14 ­.497 ­.222 .919 .491 ­.081 ­.196 ­.006 ­1. 189 .097 ­.937 .399 ­.897 ­.142 .025 .537 ­.250 ­.654 .062 ­.150 ­.505 .312 .133 .252 ­.487 .505 .006 ­.887 .292 ­.558 .494 .201 ­.425 ­.777 ­.163 1 .441 1 .453 ­.299 ­.144 .557 1 .445 -.364 1 .058 -.580 1 .293 -.227 .148 1 .471 -.856 .666 .232 .869 -.379 -.747 -.373 -.069 1 .311 -.307 .701 -.395 -1.276 -.401 .400 -.258 .565 -.178 -.552 1 .073 1.575 -.207 .125 -.808 .069 -.023 1 .025 -.371 .023 -.851 -.269 .516 -.047 -.264 -.378 .531 -.052 -.486 -.167 -.837 .448 1 .362 .499 -.896 -.502 Axis 1 .092 ­1.032 2.461 ­1 .408 3.506 ­.609 2.603 .829 2.662 .736 .201 .851 ­2.629 .921 ­3.227 ­4.501 ­4.275 5.993 ­4.676 1 .654 ­.875 2.933 ­.192 -2.074 1 .549 1 .913 1 .986 -.832 -1 .096 .869 -1.181 -1. 694 2.665 -2.436 .863 -2.544 .663 -2.065 -.630 -1 .640 -4.675 -.693 4.609 -.131 .574 .044 1 .946 -.102 1 .870 Year AD 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 Axis 7 . 161 ­.713 ­.038 ­.025 .691 .842 .303 ­.532 1 .013 ­.320 1 . 164 ­.035 ­.204 .398 ­.402 1 .214 ­ . 111 .344 .897 .732 .090 .084 ­.010 .409 .055 .334 ­.774 .549 ­.060 .016 .824 .646 .214 ­.264 .493 .149 ­.015 1 .076 ­.334 1 .019 .761 .407 ­.622 ­.531 ­.253 .215 .513 ­1 . 190 .396 Axis 6 ­.083 .139 ­.422 .516 ­.057 ­.510 ­.671 .526 ­.173 .309 ­.479 .025 .425 .333 .290 .237 ­.871 .098 .225 ­.710 ­.553 ­.380 ­.376 ­.098 .486 .242 ­.280 .203 ­.255 ­.061 ­.850 ­. 131 . 110 ­.019 ­.250 .406 ­.108 .428 ­.356 ­.112 ­.114 .323 ­.456 ­.089 .800 ­.165 .296 ­.075 ­.209 Axis Axis Axis Axis 4 5 3 2 ­.644 .01 1 ­.712 .176 .169 ­.018 ­.068 ­.584 .008 ­.092 ­.602 .927 ­.359 ­.998 .521 ­.728 .487 ­.021 .286 .466 ­1.068 ­1 . 129 ­.830 .134 .237 .428 .031 ­.413 .574 .056 .014 .400 ­.581 ­.523 ­.013 .785 .145 1 .226 ­.270 ­.555 .368 .567 1 . 511 .862 .276 ­.609 ­.126 ­.489 .586 ­.287 ­.536 .022 .543 .776 ­. 118 1 .027 .519 ­.284 .425 ­.209 .704 .354 .424 ­.440 .147 .578 ­.119 ­.346 ­1. 293 ­.277 ­.459 ­.607 ­.318 ­.583 .301 .801 .51 1 ­.245 .076 ­.021 ­1.693 ­.604 . 191 .553 ­.601 .204 .321 ­.622 ­.201 ­.489 ­.919 ­ . 311 ­.770 ­.405 .027 .886 .653 ­.314 ­.162 ­.180 ­.109 ­.157 .044 .907 ­.048 ­.589 ­.086 ­.543 .455 .013 .233 ­.048 .414 .326 .800 .329 ­.479 ­.385 1 .034 ­.639 .641 ­ . 172 ­.821 .509 .180 ­.605 .342 ­.173 .143 ­.440 ­.750 ­.500 ­.063 .537 .626 .623 ­.241 .145 .787 ­1 . 152 .156 .168 .898 ­.679 ­.144 ­.137 ­.070 ­1 . 128 ­.489 ­.563 ­1.173 ­.634 ­.090 .701 .280 .873 ­.087 .236 .724 .341 ­.653 .962 .017 1 .412 ­.212 ­.370 ­1.549 ­.463 ­.732 ­.468 ­.507 ­.921 ­1.111 .866 .222 .261 .208 .342 ­.441 .256 ­.390 ­.154 ­.597 .215 ­.762 .128 ­1.226 ­.295 ­.146 ­.537 ­.096 .889 .010 .806 .104 .021 ­ 233- Axis 1 3.350 .050 2.447 ­2.483 ­3.099 .848 1 .476 ­2.288 2.044 1 .267 ­ .011 ­2.341 ­1.913 ­4.809 .845 ­.099 1 .094 .81 1 ­.643 ­.482 ­.337 1 .527 ­.430 .729 ­1 .039 2.058 3.710 ­2. 152 .912 .511 .517 ­.681 1.146 1 .001 .020 ­2.732 ­5.305 ­1.770 3.886 3.492 .736 ­2.110 4.350 2.374 1. 117 ­.535 .408 1 .091 ­3.423 Year AD 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 ｾＸＴＶ 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 Axis 7 ­.604 .927 ­.542 .446 .573 .424 1 .222 1 .481 .130 .208 .503 .022 .980 .613 .213 -.027 ­.155 .276 -.320 .777 .034 .367 .196 1 .207 .153 ­.582 -.079 ­.749 ­.766 ­.848 ­.233 .819 ­.127 ­.773 .782 ­1 .420 .269 ­.241 ­.571 .262 .067 .368 1 .303 .603 .952 ­.353 1 .010 ­.282 -.110 Axis Axis Axis 6 4 5 ­.432 .934 ­.778 ­ . 611 ­.516 ­.043 ­.769 ­.464 .682 ­.747 ­.472 ­.329 ­.227 .182 ­.102 .016 ­.108 ­.520 .81 1 .209 .232 .265 .353 .597 .046 .480 .156 .580 1 .092 1 .008 .401 1.004 .396 .108 .240 .935 .746 ­.135 .364 ­.369 ­.194 .846 .403 .304 ­1.277 .549 .696 1.984 -.706 ­1 .450 .488 -.170 .911 ­1 .485 ­.597 -.270 .820 .145 ­.313 .202 ­.537 1 .911 ­.482 -.065 .143 .165 .682 -.085 .265 ­.389 .066 .549 ­.397 .692 ­.637 -.890 ­.135 .173 .620 -.301 .643 .226 ­.261 .559 .192 ­.192 .190 .646 .716 .447 ­.357 ­ . 181 .517 -.307 -.007 .461 ­.468 -.031 ­.193 .773 .208 .807 ­.163 ­.198 ­.419 -.180 ­.511 .075 .093 .128 -.205 ­.392 1.189 -.304 .196 ­.228 -.406 1 .075 .953 .653 .762 ­.997 .079 .935 .354 .388 .163 ­.453 ­.187 -.047 -.081 ­.112 1 .319 ­.419 -.400 .679 .396 .024 .398 .566 ­.284 .789 .572 -.800 .802 ­.499 ­1 .840 ­ 234- Axis Axis Axis 3 2 1 ­1 .082 .906 1 . 71 2 1 .126 ­.178 ­1. 107 ­.621 ­.244 .445 ­1 .489 ­.424 ­.593 ­.232 . 171 ­, .995 .936 1 .263 ­1.127 ­1 .044 ­.674 ­.889 ­.368 ­1.411 ­.037 ­.400 .689 ­1 . 779 ­.304 ­.517 ­2.901 .359 .338 ­4.655 ­.308 .018 ­4.454 .885 .958 ­2.256 ­.461 .028 2.613 ­.356 .266 ­.331 ­.587 ­.398 ­1. 862 .586 .720 2.003 .146 .086 ­.447 ­.339 .842 4.698 -.506 ­1.118 ­1 .321 ­.477 1.083 -3.034 ­.745 ­.839 ­.795 .532 ­. 149 3.830 ­.688 ­.991 ­2.965 ­ . 511 ­.333 .162 ­.432 .421 -.076 ­ . 191 . 131 ­1 .859 ­1.255 .620 ­1.314 ';".510 1 .382 -.079 -.082 ­.329 ­1 .833 ­.558 -.310 2.589 1 .124 -.520 ­1 .094 .280 .725 5. 177 ­1.787 ­1.716 ­.874 ­.646 .308 .632 .459 .655 3.247 ­.495 .145 ­1 .085 .600 .614 -2.250 ­.759 ­.118 -2.550 1 . 191 ­.195 -3.650 -.208 .341 -2.880 .688 .870 2.232 ­.653 1 .597 1 .687 .418 1 .641 .431 -.501 .137 1 .457 .078 -.089 -.410 -.220 .353 .702 .382 1 .055 2.299 .744 -.503 ­2.753 Year Ap 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 191 1 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 Axis Axis 7 ­.388 · 116 ­.393 .213 · 1 19 ­.214 ­.124 ­.372 ­.381 ­.202 ­.571 1 .416 .390 ­.675 ­.306 1 .993 ­1.922 .445 .843 ­.261 .639 1 .555 ­1 .080 .532 .485 ­1 .503 2.040 ­.710 ­2.099 .208 ­.307 ­.989 ­.547 .565 ­.741 ­.805 · 110 ­.278 ­.750 ­.028 ­1 .086 ­1 .186 ­1.050 ­1.215 ­1 .467 ­.938 .425 ­. 113 ­1.017 6 ­.724 ­.126 ­.743 ­.107 1 . 132 1 .230 1 .391 1 .001 .458 .206 .822 .970 .662 .075 .705 .692 ­.433 .419 .432 ­.686 .315 .524 ­.918 .310 .018 ­1. 169 .813 ­2.021 ­1.705 ­.701 ­.368 ­1 .245 .681 ­1.373 .190 ­.788 ­.806 .665 .180 ­.246 .342 ­.516 ­.509 ­.371 .252 ­.796 .471 ­.608 ­ .196 Axis Axis Axis Axis 5 4 3 ­ . 113 .057 .316 .246 ­.671 ­.795 ­.200 ­.377 ­.413 ­.034 ­.661 .064 1 .255 ­.093 ­.041 .304 ­.122 .108 .478 ­.146 ­.423 .364 ­.396 ­.002 .084 ­.069 ­.967 ­.358 ­.883 1 .654 .810 ­.954 .323 ­.560 ­.073 ­.397 ­.366 .387 ­.301 ­.677 .352 .313 ­.433 .231 ­1 .427 .559 ­.124 ­.178 .046 ­.658 .516 ­.461 .915 .339 .713 .550 ­.345 ­.827 ­.678 ­.853 ­.544 ­.293 .015 .501 ­.210 ­.001 ­.027 ­.639 ­.307 ­.347" ­.945 ­.008 .557 ­.681 .877 ­.645 .315 ­1. 225 .085 ­.244 ­.020 ­.012 ­.204 .595 ­.495 .334 .703 ­.395 ­.376 .142 ­1 .083 ­.177 2.154 .087 ­.652 ­.061 .007 .276 ­.396 1 .287 .492 1 .079 ­1.436 ­1.276 ­.150 1 .074 ­.762 .470 ­.069 . 161 ­.288 .21 1 .025 ­.388 1 .824 .630 .540 ­.526 ­.332 .274 1 .247 1 .642 .094 .550 .685 .662 ­1 .066 ­.154 ­.206 .708 ­.193 .200 ­1.016 .294 ­.923 .580 .401 ­.168 ­.617 .849 1 .035 .718 1 .048 ­1.279 ­.697 ­.916 .895, 2 ­.600 1 .019 ­.957 .500 .780 ­.339 .109 .072 .230 1 .802 ­1.016 ­.646 ­1 .609 ­.760 ­.026 .279 .295 ­.362 .756 .264 ­.398 .838 .284 ­.320 1. 713 ­1 .661 1 . 1 10 ­.646 ­1.763 ­.367 ­1 .029 ­2.526 ­1 .689 ­1 .229 . 112 .489 ­.873 ­1 .326 ­.875 ­1 .847 .167 .293 ­.300 .086 ­1.326 .266 ­1 .254 ­1.151 ­.492 1 1 .800 1 .537 3.664 1 .104 ­.824 1 .966 ­1 .061 ­1 .482 .166 .250 ­2.463 .445 1.208 ­1 .968 1 .325 2.579 2.438 ­1 . 145 3.968 ­2.347 .268 4.918 2.233 ­.307 3.340 .059 4.679 ­3.244 ­1 .255 ­3.542 ­1. 658 .882 1 .291 ­3.034 ­2.013 1 .903 ­3.357 ­2.859 ­3.762 ­.418 1 .819 ­4.058 ­4.550 ­1 .068 .503 1 .410 ­1 .457 3.499 ­3.617 ­ Axis 235- Yea,.. AD 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 Axis Axis Axis AX1S AX1S 4 7 6 .5 3 ­.466 ­1. 226 ­.081 .474 ­.048 ­.932 ­.348 .270 ­.560 ­.803 ­1 .430 . 150 1 .488 ­.168 .134 ­.493 . 131 ­ . 131 .310 .340 .340 .354 .257 ­1. 636 .205 ­.852 .047 ­.770 ­.353 .271 ­. 715 ­.290 ­.294 .040 .893 1 .256 .189 ­.406 1 .029 .158 .071 .052 ­1 .228 ­.640 .290 .448 1 .384 ­.013 ­.373 .021 ­.538 ­.130 ­.645 .551 .273 .700 ­1.131 .589 ­.839 .132 ­.226 .469 ­.008 ­.567 ­.009 ­.790 ­ . 911 ­.231 ­.216 .433 .034 ­.060 .502 ­.695 . 1 16 . 11 7 1 .050 ­.183 .487 ­.929 . 114 ­.807 ­ .417 .130 .529 ­.331 .575 ­1.191 ­.086 .602 .580 .235 .539 .700 .212 ­1 .944 .377 .472 .498 ­.321 .807 .914 . 145 ­.022 .639 ­.042 1 .051 ­.393 ­.155 .220 ­.367 1.123 1 .342 .170 ­.248 ­.181 .431 ­.363 .061 ­.154 ­.064 .011 ­.837 .465 ­.322 .561 ­.251 .782 .415 .152 .389 .245 ­.645 .170 .401 .301 ­.081 .577 ­.170 .351 ';".717 ­.452 ­.637 ­.086 ­.471 ­.974 ­.042 .065 ­.778 .185 ­.460 ­.168 .656 .287 ­.558 ­.576 ­.229 .095 ­1.274 ­.106 ­.467 ­.482 .482 ­.168 .534 ­.521 .010 1 .164 ­.760 ­.118 1 • 164 ­.421 ­.337 ­.154 ­.836 ­.616 ­.229 .746 1 . 101 .202 ­.288 ­.274 .437 .284 ­.419 ­.346 ­.244 ­.157 ­.978 .831 .830 ­.622 .797 .842 .102 . 196 ­.088 ­.005 ­.103 ­1.386 ­.976 .182 ­1.062 .510 .093 . 711 1.133 1 .038 ­.247 .005 .781 .274 .489 ­.246 .055 .050 .833 ­.092 ­.887 .531 ­ 236- AX1S 2 ­ 1. 067 ­.136 ­.096 .295 1 .275 .715 .398 .312 .031 .072 ­.075 ­1.101 .041 .555 ­.842 .496 1 .200 .264 ­ . 171 .434 ­1 . 138 ­.251 ­.433 ­.907 .318 ­1.939 ­.803 ­.183 ­.849 ­.235 ­1 . 794 .506 1 .455 .245 .060 ­.891 .055 ­.451 ­1 .041 ­.736 ­.845 .501 .807 .678 , Axis ­.109 .386 .555 ­.774 .563 ­3.635 .801 2.382 ­2.154 .647 ­2.686 .324 .871 ­.633 ­4.352 ­2.221 ­.005 ­.891 ­.687 3.209 ­.391 ­1.621 ­2.921 4.176 4.651 ­1. 363 2.996 1 .325 2.666 3.121 .573 ­1 .053 2.723 ­1.410 1.444 .260 1 .138 1 .535 ­2.807 1 .315 2.481 ­1.555 ­.615 ­1 .148 ApPENDIX B HISTORIC COMMENTARY ON CLIMATE AND AGRICULTURE IN MONTEZUMA COUNTY, A.D. 1894-1970 1894 1896 1899 1902 1904 1905 1906 1908 "The season of 1894 was one of water shortage allover the county and there is much talk of doing something about it" (Freeman 1958:248). "The season of 1896 was another very dry season. Grain crop is short and there is only one crop of hay. This set the people to thinking about a reserve water supply" (Freeman 1958:249). "The summer of 1899 [in Mancos] was very dry, no feed on winter range. Severe winter follows and there is heavy loss of stock from starving" (Freeman 1958:211). "The Cortez town board has a well drilled for artesian water for a town water supply. They drill to 594 ft., 8 in., but get not water. The board pays the bill and quits.... Cortez is still having water trouble.... This was in the spring of 1902.... The dry season brought about a movement in Cortez to incorporate the town so they can bond and build a water system as the only solution of the water problem. Sentiment is divided. The town is small and valuation low for such an undertaking" (Freeman 1958:252). Montezuma Valley Journal, September 5, 1902 refers to the drought of the past three months and says it is "probably the severest year known in this section of the county ... [yet] one tract produced 33 bushels of wheat per acre...[and a] ranch a few mile southwest of Cortez averaged 25 bushels weighing 72 pounds per bushel...[and] a good cutting of alfalfa has been harvested south of Cortez this season.... The fact of the case is many farmers of this valley have been so accustomed to growing 40­50 bushels of wheat and 6­10 tons of alfalfa per acre that they think there is almost a failure of crops when these figures were lowered by the drought of the present season (Thompson, personal communication, January 1990). ''The summer of 1902 very dry. Mancos river ceases to flow in August Town wells go dry. Montezuma Valley and Indian reservation very dry. Water for stock very scarce. Navajo Indians hard hit They are leaving the reservation and begging for food" (Freeman 1958: 213). "Winter of 1903­1904 was very dry and there was almost no snow, even on the mountains. Some farmers planted no crops in the spring because they figured there would be not water. Most farmers planted their land and raised pretty good crops. It was cold and ice accumulated in the mountains, melted and made lots of water for a time. There were some good summer showers" (Freeman 1958:253). "As a result of the dry winter and spring of 1904 the spring at Navajo Springs, always freely flowing, all but went. Plans were made to move the [Ute Mountain Ute] agency if the water failed entirely" (Freeman 1958:254). Montezuma Journal, July 29, 1904 records that a statewide weekly crop bulletin reported that "rainfall has been light and more is needed, especially in the extreme southern counties where severe drought prevails (Thompson, personal communication, January 1990). "On May 15, 1905, the Mancos river was the highest it had been for many years. It floods the lower street of town and the water was within six inches of the bridge. The Dolores river was also on a rampage" (Freeman 1958:214). "In 1906 the county produced the largest grain crop in its history up to that time" (Freeman 1956:255). "The years of 1906 and 1907 were years of excessive rain in Montezuma Co. The water commissioner at Mancos was not called out these two years" (Freeman 1958:217). "There were big crop yields in 1908. Wheat made from 25 to 50 bushels per acre, oats 40 to 80 bushels threshermen report There was also a good hay crop. Range cattle got very fat and an unusual amount of good cattle were shipped" (Freeman 1958:257). 1909 1910 1911 1913 1917 1918 1920 1922 1924 'There was more rain than in any year in recent history, and there was a big crop of everything grown on the farms: a big fruit and potato crop, good grain and hay crops, thousands of fat cattle and sheep were marketed, and the market for every product was the best in recent years" (Freeman 1958:257­258). ''Rains of early September, 1909, wiped out bridges, roads and railroads; crop damage is widespread. Never in the history of this part of the state has the country undergone such a persistent bombardment form the elements, and continuous rains and recurring floods wrought disaster to every industry. Farms, mine, mills, the railroads and highways-all have had their share of loss. Since July 18 it has frequently rained two days and nights without ceasing. All streams were higher than they had ever been known" (Freeman 1958:218). "During the rainy seasons of 1909 and 1911 flood waters came down the [McElmo] canon in such volume and with such terrific force that it washed away houses and large parts of orchards and farms causing damage that ran into the thousands of dollars. One farmer lost 40 acres of grain and several acres of his farm went with it. Two and three acres were a common loss. It was at this time that the channel widened and deepened to what it is today" (Freeman 1956: 132). "In 1910 Montezuma County farmers are harvesting a bountiful grain crop, 60 bushes of wheat and 80 to 90 bushels of oats per acre is not an uncommon yield" (Freeman 1958:261). "A rain historic in intensity and in its disastrous consequences fell the week of July 10, 1911, in Montezuma Valley, and well over the county. A number of important flumes, in the irrigating system, were broken in a dozen or more places, and it was reported that McElmo Creek was the highest it had ever before been known. The flood worked havoc and the repair work was a big expense to the district, already burdened with financial difficulties. The rain was already heavy in the mountains and all streams were on a rampage and much damage was done to roads, bridges and crops" (Freeman 1958:264). 'The year 1911, with its abundance of rain, produced good crops and many record yields of wheat, oats, barley and hay were reported from every part of the county. There was an unusually good crop of corn in McElmo, and in other localities where com is grown. There are report of yields of 100 bushels of oats per acre, 75 of barley, 40 to 60 bushels of wheat and com up to 100 bushels per acre" (Freeman 1958:265). "After a season of unprecedented rainfall, a number of minor floods and one major one, the wet season climaxes by a general downpour on the evening and night of Oct. 6, 1911; resulting in the greatest flood in this southwestern section since it has been known to the white man (Freeman 1958:221)." In this wet year, Freeman notes that historic sources report that the rivers were raging. These rivers included the San Miguel, Dolores, Mancos, La Plata, Animas, Pine, Piedra, San Juan, Navajo, Chama, and Rio Grande (Freeman 1958:221-222). 'The summer of 1913 very dry. No winter range in the lower country. Some canlemen bought hay. Others shipped, and shipping was heavy. Plans to restock in the spring" (Freeman 1958:267). "The year 1917 brought a dry summer. There are good crops, but feed on the winter range is very poor and many more cattle must be fed through the winter than usual" (Freeman 1958:271). "On June 21 and June 23 the county had two good rains, the second assuming flood proportions. Very unusual rains for June, but they were welcome as the previous winter and spring had been dry" (Freeman 1958:273). "The spring floods of 1920 were the most destructive in several years. The railroad along the Dolores river was washed out in a dozen places, and there was heavy damage along several other stretches of the road in the Basin. Water was so high no repairs could even be started for days. It was weeks before through service could be resumed" (Freeman 1958:313). Winter of 1921-1922 was a winter of heavy snowfall at Mesa Verde National Park (Freeman 1958:173). "There is a good grain crop in the county and dealers report much more binding twine than ever before is being used. Albert Roessler reports 60 bushes of wheat per acre. He has been on his ranch nine years, has had a better crop every year, and this year the best ever" (Freeman 1958:233). "Another good farm, the W.A. Bay place, reported good yields: 98.5 bushels of oats per acre off of 8 acres, 40 lbs. per bushel; 6 acres of Marquis wheat yields 50 bushels per acre, the result of good farming for some years. Owner says he is just getting the land in shape again to produce good crops" (Freeman 1958:233). - 238- 1926 1927 1934 1935 "It was a rainy season and the growers estimate that 50 per cent of the seed was lost in harvesting" (Freeman 1958:283). . "Alfalfa seed growers made a little history in November, 1926, when they shipped 100,000 pounds of alfalfa seed­­two full car loads­from Mancos. The seed brought 14 cents a pound or $14,000 for the shipment. Thomas Coppinger and Talcott and Carpenter were the main producers, with many small producers. In some cases, the seed grown on an acre of land was worth nearly as much as the land was worth" (Freeman 1958:237). "In the month of April, 1926, the rain gauge at Spruce Tree Camp registered a rainfall of 7.31 inches, and in May following 1.36 inches was recorded" (Freeman 1958:175). "Heavy rains of flood proportions, struck the county last days of June. Washed out roads, railroad and bridges, two railroad bridges on Dolores river above town and badly damaged one below town. This rain was notable in that is was the heaviest rain ever known to fall in this county in the month of June, uniformly a dry month with almost no rain" (Freeman 1958:284). "In the fall of 1927 beans definitely came to the county's farming industry as a main crop. Many yields of 800 pounds per acre reported and the contract price to the growers was $5 per hundred. Many fields of 60 to 120 acres were planted to beans. Highest yields were reported from the dry farming territory near Lewis. Beans were shipped by the car load and many cars were shipped" (Freeman 1958:306). "The Ackmen section of the dry lands has more wheat than can be harvested" (Freeman 1958:306). "A record drought struck Montezuma County in 1934 and was probably the driest year the county has experienced since it occupancy by the white man. With a very light winter snowfall, ahnost no spring storms and very scant summer rains, crop production was at the lowest level the county has ever known and distressing conditions were bound to ensue. These conditions being generally prevalent in the southwest, the government took measures to relieve the situation. Funds were made available for buying cattle off the rancher thus reducing the nwnber to what could be taken care of. Buyers came to every ranch and bought cattle paying $12.50 a head for average cattle of every kind and young stock in proportion. The best of these were shipped into southern areas where cattle were less plentiful and there was plenty of feed. Hundreds of head of the poorest stuff in the county were simply driven off to some isolated spot and shot In cases where stock were mortgaged half of the proceeds were given to the mortgage holder and half to the owners. There was no flood water from snow and little water in the reservoirs. A Thompson Park reporter said they had the only drouth the Park had ever experienced" (Freeman 1958:286287). "Grasshoppers, prairie dogs, bean beetles, potato psyllids added to the dry weather calamities in Montezuma County. The government furnished tons of grasshopper poison and prairie dog poison and bean growers and potato growers fought with sprays, and with the constant help and advise of County Agent Barr, any great calamity from these plagues was averted" (Freeman 1958:267). "Rain fell during the week of July 20, 1934, some rain in every part of the county and a good rain in the vicinity of Ackmen and Yellow Jacket. Bean and corn had survived the dry weather and the rain brought on a fair crop in may places in the dry farming area It was something to have a good crop in 1934" (Freeman 1958:287). 1934 is mentioned frequently by present and former Goodman Point residents as a drought year that resulted in low crop yields (Connolly, personal communication, January 1990). "The dry year of 1934 was followed by one of plenty of water and good crops were reported from every part of the county. Thirty to forty bushels of wheat per acre, 60 to 90 bushes of oats, 40 to 70 bushels of barley were common yields where a good job of farming was done. Considerable certified pure seed was used this year and plots planted with this seed invariably out­yielded crops from other seed.... Almost all seed grain and potatoes planted this year were treated before planting to ward off disease and grain smut" (Freeman 1958:288). "The long drouth of 1934, affecting the entire county, was effectively broken in the winter of 1934­1935. Although the soil was dry to a great depth the fall of rain and snow was sufficient to restore moisture all the way down" (Freeman 1958:239). ­ 239- 1937 1946 1947 1951 1954 1956 1959 1960s1970s "The snowfall of the winter of 1936­1937 on Mesa Verde National Park was the greatest on record according to the report made public in April, 1937. The total snowfall, according to the record, as reported by Paul R. Frank, Acting Superintendent, was 145.5 inches. It was a winter of deep snows allover the San Juan Basin and deep mud the following spring" (Freeman 1958:240). Montezuma Valley Journal, August I, 1946 headline states "wheat harvest now exceeds previous dry year estimates" and records that "some of the non­irrigated wheatfields are now yielding 15 bushels to the acre and less fortunate growers who have as low as 3­5 bushel returns are nevertheless showing a profit due to the fact that the local price is very favorable" (Thompson, personal communication, January 1990). 1947 recalled as being an example of a year with good crops by Goodman Point area residents (Connolly, personal communication, January 1990). 1951 is recalled as a year of severe drought resulting in almost total crop failure by former and present Goodman Point area residents (Connolly, personal communication, January 1990). Montezuma Valley Journal, July 12, 1951 reports that "precipitation records compiled by Montezuma Valley Irrigation show only 0.02 inches of rainfall in June with a sparse 3.26 inches recorded since January.... Few vigorous stands of wheat are reported left unparched on proximity farms with pinto beans planting destined to suffer the same fate if no rainfall is forthcoming in the next two or three weeks" (Thompson, personal communication, January 1990). Montezuma Valley Journal, October 25, 1951, records the following: "With inspection reports coming in on the insured bean acreages the 1951 county yield of bean crops is showing up as the lightest per acre the county has ever known, according to EL. Murphy, the County FMA chairman. Mr. Murphy has been raising beans in the Pleasant View area for 30 years and has seen some light crops and some extra heavy ones, but never a complete failure. Even this year, with moisture for the last two years at only 30% of the one year average, the crop is not a complete failure. The yield on a county average, will be about 20% of normal" (Thompson, personal communication, January 1990). "Only 65 per cent of normal snowfall was recorded in the winter of 1953­1954, and a shortage of irrigation water ensued" (Freeman 1958:299). "This is one of the shortest crop years the [Yellowjacket] locality has ever experienced.... A 700,000 bushel wheat crop is the largest handled to date. There is probably a 200,000 bushel bean crop this season, one of the lightest ever produced and a 300,000 bushel wheat crop" (Freeman 1958: 146). "In the summer of 1956, an extremely dry period of weather, and on account of SO much water being taken out above [of the Dolores river], the river quit running at the pump station and the town [of Dove Creek] was again desperately short of water for a while" (Freeman 1956:153). 'The past few years [1953­1956] have been rather drier than the average season, but there has been a fair crop right along, nothing approaching a complete failure, and during the summer of 1956, good bean and wheat land was exchanging right along at $100 to $125 an acre [in the Dove Creek area]" (Freeman 1958:155). Montezuma Valley Journal, September 24, 1959 records that "About 20% of the bean crop has been harvest but the yield has been reported as the worst in years. Etter said the yield is running from one half to five bags with an average of two sacks per acre" (Thompson, personal communication, January 1990). These two decades remembered by Goodman Point area residents as being generally good and relatively predictable years for crop production (Connolly, personal communication, January 1990). ­ 240- ApPENDIX C CALCULATING ANNUAL POPULATION SIZE: AN EXAMPLE FOR A.D. 902 The criteria specified in chapter 4 were used to detennine the total potential productivity (TOTPROD) for the region as a whole during a single year, A. D. 902. This was a very moderate year which exhibited soils that were typified by PDSI classes 6, 7, and 8­nearly nonnal, incipient wet, and slightly wet In A.D. 902, a total of 73,033,490 kg of maize was estimated to be the maximum potential yield that could be produced from the 1470.36­km 2 study area. The following example details how estimates for POP1YR, POP2YR. and POP3YR were calculated. Eｾ｡ｵｴｬｏｮ 1. TOTPROD * .5 Prod uct Potential gross yield = 2. Potential gross yield*.8 = Adjusted gross yield 3. Adjusted gross yield*.9 = Potential net yield (yrl) 4. Potential net yield*.9 = Adjusted net yield (yrl) 5. Adj. net yield (yrl) 1 POP1YR consumption rate = 6. POPIYR 1 total area = POPIYR POP1YR (density) 7. Adj. net yield (yrl)*.9 = Potential net yield (yr2) Pot. net yield (yr2) *.9 = Adj. net yield (yr2) 8. Adj. net yield (yr2) 1 POP2YR consumption rate = 9. POP2YR 1 total area = POP2YR POP2YR (density) 10. Adj. net yield (yr2) *.9 = Potential net yield (yr3) 11. Pot. net yield (yr3) *.9 = Adj. net yield yield (yr3) 12. Adj. net yield (yr3) 1 POP3YR POP3YR consumotion rate = POP3YR (density) 13. POP3YR 1 total area = ExampJe 73,033,490 kg *.5 =36 516,745 ｫｾ 36,516,745 kg *.8 =29,213,396 kg 29,213,396 kg *.9 =26,292 056 kl! 26,292,056 kg *.9 =23 662,851 kl! 23,662,851 kg 1160 kg per person = 147,893 oeoDle 147,893 people 11470.36 km 2 = 100.6 people/km2 23,662,851 kg*.9 =21,296,566 ｫｾ 21,296,566 kg *.9 = 19,166,909 ｫｾ 19,166,909 kg 1320 kg per person = 59 897 people 59,897 people 1 1470.36 k:m 2 =40.7 people/km2 19,166,909 kg.*.9 = 17,250,218 kl:! 17,250,218 kg*.9 15,525,196 kl:! 15,525,196 kg 1480 kg per person = 32 344 peoDle 32,344 people 1 1470.36 km 2 21.9 oeoDIe/km2 = = The above can be simplified as: POPIYR = ««TOTPROD*.5)*.8)*.9)*.9)/160 or a POPIYR =(TOTPROD*.324)/I60 and POP1YR (density) = ««(TOTPROD*.5)*.8)*.9)* .9)/160)/1470.36 or a POP1YR (density) =«TOTPROD* .324)/160)/1470.36 a POP2YR =«TOTPROD*.324)*.81)/320 and a POP2YR (density) =«(TOTPROD*.324)*.8l)/320)/1470.36 a POP3YR = «TOlPROD*.324)*.6561)/480 and a POP3YR (density) = (TOlPROD*.324)* .6561)/480)/1470.36 « aThese are the equations that were used in SAS programs to calculate population size for the entire study area. ­ 242- SAS ApPENDIX D PROGRAMS FOR CALCULATING TOTPROD AND POPKM The frrst program takes the edited set of count fJ1es generated for each year in the study and does the following. It concatenates all 400 data sets, frrst as sets of 50 years and then as a single set Next, it creates a new variable SPUM by an equation that converts SPURs to SPUMs. It also creates a new variable called PRODMKG (productivity of maize in kilograms). The data set CONCAT is sorted by year and a listing file is created that reports the results for each variable (SPUR HA YEAR SPUM and PRODMKG) in the form of 5 columns and as many rows as there are SPURs for each of the 400 years. Finally, it calculates a TOTPROD (total maize productivity) value for each year by summing the SPUM values when the SPUR values are greater than or equal to 19. A disk file contains the results of this step and is a list with one column (TOTPROD) and 400 rows, one for each year. This disk me is renamed and used as data for programs designed to calculate population size and population density. Program 1 generates TOTPROD: 1 2 3 4 38 39 40 41 85 86 87 88 OPTIONS TMSG=NOTES; DATA Y901; INPUT SPUB 5-6 HA 34-39; YEAR=901; CARDS4; (EDITED COUNT FILE DATA FOR A.D.901 HERE) 16486 16487 16488 16530 16531 16532 16533 16534 16535 16536 16537 16538 16539 16540 16541 16542 16543 16544 16545 16546 16547 16548 16549 16550 16551 ;;;; DATA Y1300; INPUT SPUB 5·6 HA 34-39; YEAR=902; CARDS4; (EDITED COUNT FILE DATA FOR A.D. 1300 HERE) ;;;; DATA CONCAT1; SET Y901 Y902 Y903 Y904 Y905 Y906 Y907 Y908 Y909 Y910 Y911 Y912 Y913 Y914 Y915 Y916 Y917Y918 Y919 Y920 Y921 Y922 Y923 Y924 Y925 Y926 Y927 Y928 Y929 Y930 Y931 Y932 Y933 Y934 Y935 Y936 Y937 Y938 Y939 Y940 Y941 Y942 Y943 Y944 Y945 Y946 Y947 Y948 Y949 Y950; DATA CONCAT2; SET Y951 Y952 Y953 Y954 Y955 Y956 Y957 Y958 Y959 Y960 Y961 Y962 Y963 Y964 Y965 Y966 Y967 Y968 Y969 Y970 Y971 Y972 Y973 Y974 Y975 Y976 Y9n Y978 Y979 Y980 Y981 Y982 Y983 Y984 Y985 Y986 Y987 Y988 Y989 Y990 Y991 Y992 Y993 Y994 Y995 Y996 Y997 Y998 Y999 Y1 000; DATA CONCAT3; SET Y1 001 Y1002 Y1003 Y1 004 Y1 005 Y1 006 Y1 007 Y1 008 Y1009 Y1010 Y1011 Y1012 Y1013 Y1014 Y1015 Y1016 Y1017 Y1018 Y1019 Y1020 Y1 021 Y1 022 Y1 023 Y1024 Y1 025 Y1 026 Y1027 Y1 028 Y1 029 Y1 030 Y1031 Y1032 Y1033 Y1034 Y1035 Y1036 Y1037 Y1038 Y1039 Y1040 Y1041 Y1042 Y1043 Y1044 Y1 045 Y1046 Y1047 Y1 048 Y1 049 Y1 050; DATA CONCAT4; SETY1051 Y1052 Y1053 Y1054 Y1055 Y1056 Y1057 Y1058 Y1059 Y1 060 Y1 061 Y1062 Y1063 Y1 064 Y1 065 Y1 066 Y1 067 Y1 068 Y1 069 Y1070 Y1071 Y1072 Y1073 Y1074 Y1075 Y1076 Y01n Y0178 Y0179 Y1080 Y1081 Y1082 Y1083 Y1084 Y0185 Y1086 Y1087 Y1088 Y0189 Y1090 Y1091 Y1092 Y1093 Y1094 Y0195 Y1096 Y1097 Y1098 Y1099 Y1100; DATA CONCAT5; SET Y1101 Y1102 Y1103 Y1104 Y1105 Y1106 Y1107 Y1108 Y1109 Y111 0 Y1111 Y1112 Y1113 Y1114 Y1115 Y1116 Y1117 Y11118 Y1119 Y1120 Y1121 Y1122 Y1123 Y1124 Y1125 Y1126 Y1127 Y1128 Y1129 Y1130 Y1131 Y1132 Y1133 Y1134 Y1135 Y1136 Y1137 Y1138 Y1139 Y1140 Y1141 ;;;; DATA Y902; INPUT SPUB 5-6 HA 34-39; YEAR=902; CARDS4; (EDITED COUNT FILE DATA FOR A.D. 902 HERE) ;;; DATA Y903; INPUT SPUB 5-6 HA 34-39; YEAR=902; CARDS4; (EDITED COUNT FILE DATA FOR A.D. 903 HERE) 16552 16553 16554 16555 16556 16557 16558 16559 16560 16561 16562 16563 16564 16565 16566 16567 16568 16569 16570 16571 16572 16573 16574 16575 16576 16577 16578 16579 16580 16581 16582 16583 16584 Y1142 Y1143 Y1144 Y1145 Y1146 Y1147 Y1148 Y1149 Y1150; DATACONCAT6; SETY1151 Y1152Y1153Y1154Y1155Y1156Y1157Y1158 Y1159 Y1160 Y1161 Y1162 Y1163 Y1164 Y1165 Y1166 Y1167 Y1168 Y1169 Y1170 Y1171 Y1172 Y1173 Y1174 Y1175 Y1176 Y1177 Y1178 Y1179 Y1180 Y1181 Y1182 Y1183 Y1184 Y1185 Y1186 Y1187 Y1188 Y1189 Y1190 Y1191 Y1192 Y1193 Y1194 Y1195 Y1196 Y1197 Y1198 Y1199 Y1200; DATA CONCAT7; SET Y1201 Y1202 Y1203 Y1204 Y1205 Y1206 Y1207 Y1208 Y1209 Y1210 Y1211 Y1212 Y1213 Y1214 Y1215 Y1216 Y1217 Y1218 Y1219 Y1220 Y1221 Y1222 Y1223 Y1224 Y1225 Y1226 Y1227 Y1228 Y1229 Y1230 Y1231 Y1232 Y1233 Y1234 Y1235 Y1236 Y1237 Y1238 Y1239 Y1240 Y1241 Y1242 Y1243 Y1244 Y1245 Y1246 Y1247 Y1248 Y1249 Y1250; DATA CONCAT8; SET Y1251 Y1252 Y1253 Y1254 Y1255 Y1256 Y1257 Y1258 Y1259 Y1260 Y1261 Y1262 Y1263 Y1264 Y1265 Y1266 Y1267 Y1268 Y1269 Y1270 Y1271 Y1272 Y1273 Y1274 Y1275 Y1276 Y1277 Y1278 Y1279 Y1280 Y1281 Y1282 Y1283 Y1284 Y1285 Y1286 Y1287 Y1288 Y1289 Y1290 Y1291 Y1292 Y1293 Y1294 Y1295 Y1296 Y1297 Y1298 Y1299 Y1300; DATA CONCAT; SET CONCAr1 CONCAT2 CONCAT3 CONCAT4 CONCAT5 CONCAT6 CONCAT7 CONCAT8; SPUM=(((SPUB *10)+4)/.473r1.12; PRODMKG=HA*SPUM; PROC SORT; BY YEAR; PROC PRINT; DATA_NULL_; SET CONCAT; BY YEAR; FILE PUNCH; IF FIRST.YEAR THEN DO; TOTPROD=O; PUT YEAR @; IF SPUB GE 19 THEN TOTPROD+PRODMKG; RETURN; END; IF SPUB GE 19 THEN TOTPROD +PRODMKG; IF LAST.YEAR ; PUT TOTPROD; RETURN; The second program calculates population density for a population intending to place one year, two years, and three years of maize in storage. A variant of this program produces the maximum number of people who could be sustained on TOlPROD rather than producing the density of people per km 2. ｔｨｾ only difference between these two is the deletion of the division by the total area in the equation that calculates population size for each storage level. The program does the following. First it specifies the input two variables, YEAR and TOlPROD. The data set that contains this information is contained within the renamed punch file generated by the first program. Next, it creates three new variables, POPlYR, POP2YR, and POP3YR by the equations given. Next, it sorts the data by year and creates an output file with five colums (YEAR TOTPROD POPl YR POP2YR POP3YR) and 400 rows, one for each year in the study. Finally, it calculates the mean, standard deviation, minumum value, and the maximum value, among other useful descriptive statistics, for the variables TOlPROD, POPl YR, POPl YR, and POP3YR, and then appends this to the output file. Program 2 generates POPKM: 1 2 3 4 5 404 405 OPTIONS TMSG=NOTES; DATA POP; INPUT YEAR TOTPROD; CARDS; (DATA FROM RENAMED PUNCH FILE THAT REPORTS ANNUAL TOTPROD HERE) ­ 244- 406 407 408 409 410 411 412 413 414 DATA POP1; SET POP; POP1YR = INT(((TOTPROD*.324)/160)/1470.36); POP2YR= INT((((TOTPROD*.324)*.81)/320)/1470.36); POP3YR= INT ((((TOTPROD*.324)*.6561)/480)/1470.36); PROC SORT DATA= POP1; BY YEAR; TITLE1 'REGIONAL POPULATIONAL DENSITY (PERSONS/KM2)'; PROC PRINTDATA=POP1; PROCMEANS DATA = POP1; VAR TOTPROD POP1YR POP2YR POP3YR; - 245- ApPENDIX E Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Study Area, A.D. 901­1300a Year A.D. 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 Maize Yield TOTPROD (kg) 38351244 73033490 49252912 52740208 80344869 38241754 38647208 74453111 76407925 72693730 97327555 70538178 68766106 72502842 59249961 55794185 80986788 77865620 80097380 55078313 68494597 38332301 49468899 49738722 74199938 68448963 53010315 78762226 Number of People in 1470.36 kfTl2 Population Density per km2 POP1YR POP2YR POP3YR POPIYR POP2YR POP3YR 16984 31452 52 77661 21 11 59896 32344 100 147892 21 40 40393 21812 67 99737 27 14 43253 23356 72 106798 29 15 65892 35582 162698 110 44 24 16936 31363 52 77439 21 11 17115 31695 53 21 11 78260 61060 32972 102 41 150767 22 154726 62664 33838 105 42 23 32193 21 147204 59617 100 40 79820 43103 134 197088 54 29 57850 31239 97 142839 39 21 30454 94 139251 56396 38 20 59461 32109 146818 99 40 21 48592 26239 81 119981 33 17 24709 31 16 112983 45758 76 111 66419 35866 45 163998 24 157677 63859 34484 107 43 23 162197 35472 110 24 65689 44 111533 45171 24392 75 30 16 56174 94 138701 30334 38 20 31437 77622 16976 52 21 11 100174 40570 21908 68 27 14 100720 40791 22027 27 14 68 150254 32860 102 41 60853 22 138609 56136 30313 94 38 20 107345 43475 23476 73 29 15 159493 64594 34881 108 43 23 121133 59819157 49059 26491 1300 82 33 Mean (n­400) 131473 53246 28752 64925217 88 35 19 Standard 7 28222 11429 13936845 6172 Deviation (N­400) Minimum Value 21 52 38241754 77439 31363 16936 (N=400) Maxiroom Value 102410868 57 30 83989 141 45354 N=400 ate: All Dvalues are rounded to the nearest whole number. All population values rOP1 YR. POP2YR. ｾｒｙＳｐｏ are truncated Ｎｾｴｮｩ For a fuller presentation, see an West 1990:5 79. 18 19 4 11 30 F Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Sand Canyon Survey Locality, A.D. 901­1300a ApPENDIX Year A.D. 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 Maize Yield TOTPROD (kg) 1203519 1927934 1341613 1471751 2174287 1203519 1203519 1917705 2185747 2067733 2483208 2067733 1916569 2039129 1579631 1472888 2186221 2051726 2127782 1472888 1916569 1203519 1340477 1465670 1917705 1915716 1472888 2064797 Number of People in 26.08 km 2 Population Density per km2 POP1YR POP2YR POP3YR POPIYR POP2YR POP3YR 2437 987 532 93 37 20 3904 1581 853 149 60 32 2716 1100 594 104 42 22 2980 1207 651 114 46 24 4402 1783 962 168 68 36 2437 987 532 93 37 20 2437 987 532 93 37 20 3883 1572 849 148 60 32 4426 1792 967 169 68 37 4187 1695 915 160 65 35 5028 2036 1099 192 78 42 4187 1695 915 160 65 35 3881 1571 848 148 60 32 4129 1672 903 158 64 34 3198 1295 699 122 49 26 114 46 25 2982 1207 652 4427 1792 968 169 68 37 4154 1682 908 159 64 34 4308 1745 942 165 66 36 2982 1207 652 114 46 25 3881 1571 848 148 60 32 2437 987 532 93 37 20 2714 1099 593 104 42 22 2967 1202 649 113 46 24 148 60 32 3883 1572 849 148 60 32 3879 1571 848 2982 1207 652 114 46 25 914 160 64 35 4181 1693 1300 3244 1313 1601984 709 Mean (n=400) 3602 1458 787 1779087 Standard 276 149 337524 683 Deviation (N=400) Minimum Value 2437 1203519 987 532 (N=400) MaximJm Value 5174 2555172 2095 1131 N=400 D values are round to the nearest whole number. All Note: IT TP rOP1YR, POP2YR, POP3YR) are truncated integ:rs. For a fuller presentation, see Van West 1990:58 591. 124 137 26 50 55 10 27 29 5 93 37 20 198 80 43 population values G Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Mockingbird Mesa Survey Locality, A.D. 901­1300a ApPENDIX Year A.D. 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 Maize Yield TOTPROD (kg) 412084 958569 573175 590413 1036234 412084 412084 1023922 893405 872947 1196681 792818 860728 875788 695925 686075 1036234 1023922 1029226 670541 855330 412084 648568 586170 1023922 855330 590413 1023922 Population Density per km2 Number of People in 17.96 km 2 POP1YR POP2YR POP3YR POPIYR POP2YR POP3YR 834 337 182 46 18 10 1941 786 424 108 43 23 1160 470 253 64 26 14 1195 484 261 66 26 14 2098 849 458 116 47 25 834 337 182 46 18 10 834 337 182 46 18 10 2073 839 453 115 46 25 1809 732 395 100 40 22 386 98 39 21 1767 715 134 54 29 2423 981 529 89 36 19 1605 650 351 1742 705 381 97 39 21 1773 718 387 98 39 21 1409 570 308 78 31 17 303 77 31 16 1389 562 2098 849 458 116 47 25 115 46 25 2073 839 453 2084 844 455 116 46 25 1357 549 296 75 30 16 1732 701 378 96 39 21 834 337 182 46 18 10 1313 531 287 73 29 15 1186 480 259 66 26 14 453 115 46 25 2073 839 96 39 21 1732 701 378 484 261 66 26 14 1195 2073 839 453 115 46 25 1300 695925 1409 570 308 78 31 Mean (n=400) 1622 801336 656 354 46 18 9 23 Standard 204994 415 168 90 Deviation (N=400) 46 Minimum Value 18 834 412084 337 182 (N=400) 57 Maxirrum Value 1034 1261125 2553 142 558 N=400 Note: All T TPR D values are rounded to the nearest whole number. All population values rOP1 YR, POP2YR, POP3YR) are truncated Ｎｳｲ･ｾｴｮｩ For a fuller presentation. see Van West 1990:59 -603. 17 10 5 10 31 H Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Catchment of 5MT8371 (DCA Site), A.D. 901­1300a ApPENDIX Year A.D. 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 Maize Yield TOTPROD (kg) 93426 406041 278234 278234 468553 93426 93426 468553 365883 361336 498407 300529 361336 365883 343587 339040 468553 468553 468553 339040 361336 93426 327675 231861 468553 361336 278234 468553 Number of People in 7.88 km2 Population Density per kfTl2 POP1 YR POP2YR POP3YR POPIYR POP2YR POP3YR 189 76 41 24 9 5 822 333 179 104 42 22 71 28 15 563 228 123 71 28 15 563 228 123 48 26 948 384 207 120 189 72 41 24 9 5 76 41 24 9 5 189 48 26 948 384 207 120 740 300 162 94 38 20 731 296 160 92 37 20 51 28 1009 408 220 128 608 246 133 77 31 16 37 20 731 296 160 92 94 38 20 740 300 162 695 281 152 88 35 19 686 278 150 87 35 19 48 26 948 384 207 120 948 384 207 120 48 26 948 384 207 120 48 26 686 278 150 87 35 19 731 296 160 92 37 20 189 76 41 24 9 5 663 268 145 84 34 18 469 190 102 59 24 13 120 48 26 948 384 207 731 296 160 92 37 20 71 28 15 563 228 123 948 384 207 120 48 26 1300 343587 695 281 152 Mean (n=400) 350669 709 287 154 Standard 104956 212 46 86 Deviation (N=400) Minimum Value 189 93426 76 41 (N=400) Maximum Value 556372 1126 456 246 N=400 to the nearest whole number. TPR D values are rou Note: All rOP1YR, POP2YR, POP3YR) are truncated integers. For a fuller presentation, see Van West 1990:604-615. 88 89 26 35 35 10 19 19 5 24 9 5 142 57 31 I population values I Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Catchment of 5MT8839 (Norton House), A.D. 901­1300 a ApPENDIX Year A.D. 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 Maize Yield TOTPROD (kg) 478574 737789 576793 579445 789503 478574 478574 695641 684843 682949 903350 682949 681055 682949 582097 579445 789503 787609 787609 579445 681055 478574 576793 576793 681055 681055 579445 787609 Number of People in 7.88 km2 Population Density per km2 POP1YR POP2YR POP3YR POPIYR POP2YR POP3YR 969 392 211 122 49 26 1494 605 326 189 76 41 1168 473 255 148 60 32 1173 475 256 148 60 32 1598 647 349 202 82 44 969 392 211 122 49 26 969 392 211 122 49 26 1408 570 308 178 72 39 71 38 1386 561 303 175 1382 560 302 175 71 38 94 50 1829 740 400 232 1382 560 302 175 71 38 1379 558 301 175 70 38 71 38 1382 560 302 175 1178 477 257 149 60 32 1173 475 256 148 60 32 82 44 1598 647 349 202 81 44 1594 645 348 202 202 81 44 1594 645 348 1173 475 256 148 60 32 70 38 1379 558 301 175 392 211 122 49 26 969 32 1168 473 255 148 60 1168 473 255 148 60 32 70 38 1379 558 301 175 1379 558 301 175 70 38 60 32 1173 475 256 148 81 44 1594 645 348 202 582097 653070 99827 1178 1321 202 477 535 81 257 288 44 149 167 25 60 67 10 32 36 5 478574 969 392 211 122 49 26 909146 1841 745 402 233 94 51 population values ApPENDIX J Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Catchment of 5MT2433 (Aulston Publo), A.D. 901­1300a Year A.D. 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 Maize Yield TOTPROD (kg) 314812 556164 328830 393274 612425 314812 314812 554175 661581 617539 689806 617539 554175 598312 453607 393274 615740 568288 588556 393274 554175 314812 328830 419491 554175 554175 393274 571603 Number of People in 7.88 kn12 Population Density per km2 POP1YR POP2YR POP3YR POPIYR POP2YR POP3YR 637 258 139 80 32 17 1126 456 246 142 57 31 665 269 145 84 34 18 796 322 174 101 40 22 34 1240 502 271 157 63 17 637 258 139 80 32 17 637 258 139 80 32 1122 454 245 142 57 31 1339 542 292 170 68 37 64 34 1250 506 273 158 71 38 1396 565 305 177 1250 506 273 158 64 34 1122 454 245 142 57 31 1211 490 264 153 62 33 918 372 200 116 47 25 796 322 174 101 40 22 64 34 1246 504 272 158 146 59 31 1150 466 251 61 33 1191 482 260 151 796 322 174 101 40 22 1122 454 245 142 57 31 258 139 80 32 17 637 84 34 18 665 269 145 107 43 23 849 344 185 454 245 142 57 31 1122 1122 454 245 142 57 31 796 322 174 101 40 22 1157 468 253 146 59 32 1300 454460 920 372 201 116 47 Mean (n=400) 500847 1013 410 221 128 51 Standard 115085 233 94 50 29 12 Deviation (N=400) Minimum Value 314812 637 258 139 80 32 (N=400) Maximum Value 739948 1498 606 327 190 77 _CN=400) Note: All TOTPROD values are rounded to the nearest whole number. All population values (POP1YR, POP2YR, POP3YR) are truncated Ｎｾ･ｴｮｩ aFor a fuller presentation, see Van West 1990:62 39. 25 27 6 17 41 K Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Catchment of 5MT3834 (Mustoe Ruin), A.D. 901­1300a ApPENDIX Year A.D. 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 Maize Yield TOTPROD (kg) 377040 592250 388311 460010 672000 377040 377040 593387 732902 667359 760407 667359 592250 659782 526121 461146 672000 604469 634304 461146 592250 377040 387174 441654 593387 591398 461146 605605 Number of People in 7.88 kl1l2 Population Density per km2 POP1YR POP2YR POP3YR POPIYR POP2YR POP3YR 763 309 166 96 39 21 61 33 1199 485 262 152 786 318 171 99 40 21 203 118 47 25 931 377 1360 551 297 172 69 37 763 309 166 96 39 21 763 309 166 96 39 21 1201 486 262 152 61 33 1484 601 324 188 76 41 37 1351 547 295 171 69 1539 623 336 195 79 42 1351 547 295 171 69 37 1199 485 262 152 61 33 1336 541 292 169 68 37 1065 431 233 135 54 29 204 118 47 25 933 378 1360 551 297 172 69 37 1224 495 267 155 62 33 1284 520 280 163 66 35 933 378 204 118 47 25 1199 485 262 152 61 33 763 309 166 96 39 21 784 317 171 99 40 21 894 362 195 113 45 24 1201 486 262 152 61 33 1197 485 261 151 61 33 204 118 47 25 933 378 1226 496 268 155 63 34 1300 536539 1086 440 237 137 55 Mean (n=400) 1119 553196 453 244 141 56 Standard 27 11 108069 218 88 47 Deviation (N=400) Minimum Value 96 377040 763 39 309 166 (N-400) Maxirrom Value 80 776849 1573 637 344 199 N=400 Note: All T TPR D values are rounded to the nearest whole number. All population values ＮｒｙＱｐｏｾ POP2YR, POP3YR) are truncated integers. For a fuller presentation, see Van West 1990:64Q-651. 30 30 6 21 43 L Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Catchment of 5MT6970 (Wallace Ruin), A.D. 901­1300a ApPENDIX Year A.D. 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 Maize Yield TOTPROD (kg) 198673 329891 236351 243549 361620 198673 198673 357832 330611 310152 402935 280886 302196 323981 300094 278309 364462 355559 360673 275847 295945 198673 248948 216953 357832 301249 243549 360673 Number of People in 4.32 km2 Population Density per km2 POP1YR POP2YR POP3YR POPIYR POP2YR POP3YR 402 162 87 93 37 20 154 62 33 668 270 146 478 193 104 11 0 44 24 493 199 107 114 46 24 732 296 160 169 68 37 402 162 87 93 37 20 402 162 87 93 37 20 724 293 158 167 67 36 669 271 146 154 62 33 628 254 137 145 58 31 815 330 178 188 76 41 124 131 53 28 568 230 611 247 133 141 57 30 656 265 143 151 61 33 607 246 132 140 56 30 563 228 123 130 52 28 738 298 161 170 69 37 720 291 157 166 67 36 730 295 159 169 68 36 558 226 122 129 52 28 599 242 131 138 56 30 402 162 87 93 37 20 116 47 25 504 204 110 439 177 96 101 41 22 724 293 158 167 67 36 57 30 610 247 133 141 493 199 107 114 46 24 730 295 159 169 68 36 132 140 56 1300 300094 607 246 Mean (n=400) 297798 602 243 131 139 56 Standard 53732 108 44 23 25 10 Deviation (N=400) 87 93 37 Minimum Value 198673 402 162 (N=400) 183 194 78 MaxiroomVaJue 414774 839 340 (N=400) Note: All TOTPROb values are rounded to the nearest whole number. All population values (POP1YR, POP2YR. POP3YR) are truncated integers. aFar a fuller presentation, see Van West 1990:652-663. 30 30 5 20 42 M Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Catchment of 5MT1566 (Lowry Ruin), A.D. 901­1300a ApPENDIX Year A.D. 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 Maize Yield TOTPROD (kg) 256677 450955 261696 314642 501817 256677 256677 451713 555520 511478 568212 511478 450955 495850 363799 315400 503806 456922 483632 315400 450955 256677 260939 344212 451713 450387 315400 459669 Number of People in 7.88 knl2 Population Density per km2 POP1YR POP2YR POP3YR POPIYR POP2YR POP3YR 519 210 113 65 26 14 199 115 46 25 913 369 529 214 115 67 27 14 637 258 139 80 32 17 1016 411 222 128 52 28 210 113 65 26 14 519 113 65 26 14 519 210 914 370 200 116 47 25 1124 455 246 142 57 31 419 226 131 53 28 1035 1150 466 251 146 59 31 1035 419 226 131 53 28 913 369 199 115 46 25 1004 406 219 127 51 27 736 298 161 93 37 20 638 258 139 81 32 17 1020 413 223 129 52 28 925 374 202 117 47 25 979 . 396 214 124 50 27 139 81 32 17 638 258 913 369 199 115 46 25 519 210 113 65 26 14 528 214 115 67 27 14 697 282 152 88 35 19 914 370 200 116 47 25 912 369 199 115 46 25 638 258 139 81 32 17 930 376 203 118 47 25 1300 369292 747 302 163 94 38 Mean (n=400) 409875 829 335 180 104 42 Standard 26 10 102360 207 83 45 Deviation (N=400) Minimum Value 65 26 256677 519 210 113 (N=400) MaxilTXJm Value 657925 1332 68 539 291 169 N=400 Note: All T TPR D values are rounded to the nearest whole nu er. All population values rOP1YR, POP2YR, POP3YR) are truncated integers. For a fuller presentation, see Van West 1990:664-675. 20 22 5 14 36 ApPENDIX N Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Catchment of 5MT2149 (Escalante Ruin), A.D. 901­1300 a Year A.D. 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 Maize Yield TOTPROD (kg) 202822 316176 249706 249706 351031 202822 202822 310588 310588 310588 477949 310588 310588 310588 249706 249706 351031 351031 351031 249706 310588 202822 249706 249706 310588 310588 249706 351031 Number of People in 6.72 km2 Population Density per kl1l2 POP1YR POP2YR POP3YR POPIYR POP2YR POP3YR 410 166 89 61 24 13 640 259 140 95 38 20 505 204 110 75 30 16 505 204 110 75 30 16 710 287 155 105 42 23 410 166 89 61 24 13 410 166 89 61 24 13 628 254 137 93 37 20 628 254 137 93 37 20 628 254 137 93 37 20 967 391 211 144 58 31 628 254 137 93 37 20 628 254 137 93 37 20 628 254 137 93 37 20 505 204 110 75 30 16 204 110 75 30 16 505 710 287 155 105 42 23 710 287 155 105 42 23 155 105 42 23 710 . 287 110 75 30 16 505 204 628 254 137 93 37 20 410 166 89 61 24 13 505 204 110 75 30 16 505 204 110 75 30 16 628 254 137 93 37 20 628 254 137 93 37 20 505 204 110 75 30 16 710 287 155 105 42 23 1300 204 249706 505 110 75 30 Mean (0=400) 291075 588 238 128 87 34 Standard 16 6 108 44 53857 23 Deviation (N=400) Minimum Value 61 24 410 202822 166 89 (N=400) Maximum Value 967 58 477949 391 144 211 N=400 Note: All TPR D values are rounded to the nearest whole number. All popUlation va ues rOP1YR, POP2YR, POP3YR) are truncated integers. For a fuller presentation. see Van West 1990:676-687. 16 18 3 13 31 0 Total Annual Maize Productivity and Maximum Potential Population for Three Levels of Storage in the Catchment of 5MT765 (Sand Canyon Pueblo), A.D. 901­1300a ApPENDIX Year A.D. 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 Maize Yield TOTPROD (kg) 303731 507235 370220 378176 566526 303731 303731 501646 516895 510455 652432 510455 501646 508182 385469 378176 566526 559991 563590 378176 501646 303731 370220 383045 501646 501646 378176 559991 Number of People in 7.88 krJ12 Population Density per km2 POP1YR POP2YR POP3YR POPIYR POP2YR POP3YR 615 249 134 78 31 17 1027 415 224 130 52 28 749 303 163 95 38 20 765 310 167 97 39 21 1147 464 250 145 58 31 615 249 134 78 31 17 615 249 134 78 31 17 1015 411 222 128 52 28 132 53 29 1046 423 228 1033 418 226 131 53 28 167 67 36 1321 535 288 1033 418 226 131 53 28 128 52 28 1015 411 222 416 225 130 52 28 1029 780 316 170 99 40 21 167 97 39 21 765 310 1147 464 250 145 58 31 143 58 31 1133 459 248 .462 249 144 58 31 1141 167 97 39 21 765 310 1015 411 222 128 52 28 134 78 31 17 615 249 163 95 38 20 749 303 314 169 98 39 21 775 1015 411 222 128 52 28 1015 411 222 128 52 28 167 97 39 21 765 310 1133 459 248 143 58 31 1300 386227 782 316 171 99 40 Mean (0=400) 459114 929 376 202 117 47 Standard 92902 188 76 41 23 9 Deviation (N-400) 204 82 44 25 10 Minimum Value 101079 (N.400) Maxiroom Value 683688 1384 560 302 175 71 -.1N=400, Note: A110TPROD values are rounded to the nearest whole number. All population values (POP1YR, POP2YR, POP3YR) are truncated integers. aFor a fuller presentation, see Van West 1990:688-699. 21 25 5 5· 38 Washington State University Department of Anthropology REPORTS OF INVESTIGATIONS 1957 Stallard, Bruce, An Archaeological Survey in the Wells Reservoir in the State of Washington. (Out of Print). 2 1958 Stallard, Bruce, Preliminary Surveys for Highway Salvage Archaeology in the State of Washington: A 3 1959 Daugherty, Richard D., Archaeological Excavations in the Ice Harbor Reservoir, 1959 Progress Report. 4 1959 Newman, Thomas Stell, Toleak Point-An Archaeological Site on the North Central Washington 5 6 1959 Osborne, Douglas, Archaeological Tests in the Lower Grand Coulee, Washington. (Out of Print). 1960 Sprague, Roderick, Archaeology in the Sun Lakes Area of Central Washington. Appendix by Roald 7 1961 Clinehens, Stephen S., Further Archaeological Excavations in the Lower Grand Coulee of Central 9 1%1 Fryxell, Roald, Geologic Field Examination of the Park Lake Housepit Site (45-GR-90), Lower Grand 10 1961 Daugherty. Richard D., Excavations in the lee Harbor Reservoir, 1957-1960: A Preliminary Report. 12 14 1961 Mallory, Oscar L., An Archaeological Survey of Pacific Gas Transmission Company's Alberta to California Pipeline System: MP 108.0 to MP 722.0. (Out of Print). 1961 Combes, John D., An Archaeological Survey of Pacific Gas Transmission Company's Alberta to California Pipeline System: MP 108.0 to MP 722.0. Phase JI. (Out of Print). 1962 Mallory, Oscar L., Continued Archaeological Appraisal of the Lower Coulee, Central Washington. 15 16 1962 Ice, Dannie M., Archaeology of the Lava Bulle Site, Deschutes County, Oregon. (Out of Print). 1962 Guinn, Stanley J., White Rock Village Archaeological Site: A Preliminary Report of Investigations. 18­1 1963 Combes, John D., A Report on the Excavations of a Late Indian Burial Site in the Ice Harbor Reservoir 18­11 1962 Fryxell, Roald, Geologic Examination of the Ford Island Archaeological Site (45-FR-47), Washington. 19 1962 Daugherty, Richard D., Archaeological Research in the Boundary Dam Reservoir Area. Final Report. 20 21 1962 Combes, John D., Excavations at Fort Spokane, Summer of 1962: A Preliminary Report. (Out of Print). 1962 Fryxell, Roald (in cooperation with Richard D. Daugherty), Interim Report: Archaeological Salvage in 22 1963 Guinn, Stanley J., A Maritime Village on the Olympic Peninsula of Washington. (Out of Print). 1963 Fryxell, Roald (in cooperation with Richard D. Daugherty), Late Glacial and Post Glacial Geological and Final Report. (Out of Print). (Out of Print). Coast. (Out of Print). Fryxell. (Out of Print). Washington. (Out of Print). Coulee, Washington. (Out of Print). (Out of Print). 13 (Out of Print). (Out of Print). Region, Washington. Section 1. (Out of Print). Section JI. (Out of Print). (Out of Print). the Lower Monumental Reservoir, Washington. (Out of Print). 23 24 1963 25 26 1964 1964 27 1964 28 1964 Archaeological Chronology of the Columbia Plateau, Washington: An Interim Report to the National Science Foundation, 1962-1963. (Out of Print). Steen, Virginia, A Study of the Chemical and Physical Characteristics of Pumice from Glacier Peak, Washington and Mt . Mazama (Crater Lake), Oregon: Report of Progress, June 16 to September 7,1963. (Out of Print). Barnes, Paul L., Archaeology of the Dean Site: Twin Falls County, Idaho. (Out of Print). Ackerman. Robert E., Prehistory in the Kuskokwim-Bristol Bay Region, Southwestern Alaska. (Out of Print). Fryxell, Roald, and Earl F. Cook, editors, A Field Guide to the Loess Deposits and Channeled Scablands of the Palouse Area, Eastern Washington. (Out of Print). Ackerman, Robert E., Archaeological Survey, Glacier Bay National Monument, Southeastern Alaska, Part 1. (Out of Print). Continued inside back cover . . . Continued from back cover . . . 29 1964 Combes, John D., Excavations at Spokane House-Fort Spokane Historic Site, 1962-1963. (Out of Print). 30 1964 Combes, John D., A Preliminary Investigation at Old Military Fort Spokane, Washington. (Out of Print). 3I 1964 Fryxell, Roald, and Richard D. Daugherty, Demonstration of Techniques for Preserving Archaeological Stratigraphy: A Report to the Wenner-Gren Foundation. (Out of Print). 32 1965 Sprague, Roderick, The Descriptive Archaeology of the Palus Burial Site, Lyons Ferry, Washington. (Out of Print). 1965 Rice, David G., Archaeological Test Excavation in Fryingpan Rockshelter, Mount Rainier National 33 Park. 34 1965 Nelson, Charles M., Archaeological Reconnaissance in the Lower Monumental and Little Goose Dam Reservoir Areas, 1964. (Out of Print). 1966 Kenaston, Monte Ray, The Archaeology of the Harder Site, Franklin County, Washington. (Out of 35 Print). 1965 Ackerman, Robert E., Archaeological Survey: Glacier Bay National Monument, Southeastern Alaska. 36 Part 11. (Out of Print). 37 1966 Sprague, Roderick, and John D. Combes. Excavations in the Little Goose and Lower Granite Dam Reservoir, 1965. (Out of Print). 1966 Larabee, Edward McM., and Susan Kardas, Archaeological Survey of Grand Coulee Dam National 38 Recreation Area. Part 1: Lincoln County Above Normal Pool. 39­1 1966 Nelson, Charles M., A Preliminary Report on 45COI, A Stratified Open Site in the Southern Columbia Plateau. (Out of Print). 39­II 1966 Rice, David G., An Archaeological Reconnaissance of the Wynochee Dam Reservoir, Greys Harbor County, Washington. (Out of Print). 1967 Daugherty, Richard, Barbara A. Purdy, and Roald Fryxell, The Descriptive Archaeology and Geo40 chronology of the Three Springs Bar Archaeological Site, Washington. (Out of Print). 41 1967 Walker, Deward E., Jr., Mutual Cross-Utilization of Economic Resources in the Plateau: An Example from Aboriginal Nez Perce Fishing Practices. (Out of Print). 42 1967 Chance, David H., Archaeological Survey of Coulee Dam National Recreation Area. Part 11: Spring Draw-Down of 1967. (Out of Print). 43 1967 Sprague, Roderick, A Preliminary Bibliography of Washington Archaeology. (Out of Print). 44 1968 Ackerman, Robert E., The Archaeology of the Glacier Bay Region, Southeastern Alaska, Part 111. (Out of Print). 45 1968 Rice, David G., Archaeological Investigations in the Coulee Dam National Recreation Area, Spring 1968. 1969 Bernard, Russell, editor, Los Otomies: Papers from the lxmiquilpan Field School. 46 47 1969 Nelson, Charles M., The Sunset Creek Site (45-KI-28) and Its Place in Plateau Prehistory. (Out of Print). 48­1 1971 Bernard, H. Russell, and Michael Kenny, editors, The Inland Empire: Papers from the Spokane Area Field School in Cultural Anthropology and Linguistics. (Out of Print). 48­II 1971 Irwin, H. T., D.J. Hurd, and R. M. LaJeunesse, Description and Measurement in Anthropology. Computer Series 1. (Out of Print). 49 1971 Leonhardy, Frank C., Gerald Schroedl, Judith Bense, and Seth Beckerman, Wexpusnime (45-GA-61): Preliminary Report. 50 1972 Rice, David G., The Windust Phase in Lower Snake River Region Prehistory. (Out of print). 5I 1973 Schroedl, Gerald F., The Archaeological Occurrence of Bison in the Southern Plateau. 52 1974 Croes, Dale. Jonathan Davis, and H. T. Irwin, The Use of Computer Graphics in Archaeology: A case Study from the Ozette Site, Washington. 1975 Adams, William H., Linda Gaw, and Frank C. Leonhardy, Archaeological Excavations at Silcott, 53 Washington: The Data Inventory. (Out of Print). 54 1977 Adams, William H.• Silcott, Washington: Ethnoarchaeology of a Rural American Community. 1978 Bodley, John H., Preliminary Ethnobotany of the Peruvian Amazon. 55 1979 Bodley. John H., and Foley C. Benson, Cultural Ecology of Amazonian Palms. 56 1979 Brown, Christopher L., and Carl E. Gustafson, A Key to Postcranial Skeletal Ref/wins of Cattle/Bison, 57 Elk and Horse (reprinted 1990). Continued inside front cover . . .