an NSF supported program

Village Ecodynamics Project


Research Plan

The very solid foundation provided by the work described in the prior results—which itself builds on generations of scholarship in this region—and the substantial research insights to which it is leading, compel us to expand this research temporally, spatially, topically, and methodologically. Whereas our earlier work focused on households, we will now focus on groups. Within SW Colorado, we will expand our research area to encompass MVNP (Fig. 6), and we will extend our research to include portions of the Northern Rio Grande (NRG) region in New Mexico. For our original research area, the period from AD 600-1300 included the entire sweep of intensive farmer settlement, but study of thehistories of these same populations after they leave SW Colorado ca. AD 1280 requires us to broaden our temporal focus to the AD 600-1600-year period. Finally, we will expand the ABM to study processes such as the domestication of (or adoption of domesticated) turkey and the increase in size of the co-resident social unit from the household to (putatively) the lineage/clan, villages, and towns. Our focus on groups requires that we pay close attention to the processes generating solidarity within these ever-larger groups as well as the nature of the competitive or cooperative processes among groups. We also propose to bring new techniques in the analysis of ancient DNA to bear on classic questions of population expansion and migration, and we will use a variety of new methods to assist in the difficult process of building the productivity reconstructions on which our ABMs depend.


Our modeling strategy so far has been to concentrate on households. Even in our ABM framework, households are not entirely atomistic, since they have relations of exchange of several types with other households. Further, we have enriched our modeling by use of systems-level models (or what White [2007] calls “aggregate (‘sufficient unit’) equation modeling”) to study the dynamics of specific processes, such as warfare, that are not represented in the ABMs.

These studies make it clear that our agent-based models severely under-represent the sorts of social and cultural processes we can either see in the archaeological record, or identify by means of comparing the archaeological record with the output from our models. There is a long tradition in the northern Southwest of identifying nested social units in architectural arrangements. At the Pueblo III Mesa Verde alcove site of Mug House, for example, Rohn (1971:31-41) recognized household units or suites, courtyard units composed of several suites, dual divisions organizing the suites, and finally communities (see also Rohn 1977:265-288) typically composed of households from several nearby sites. More recently, several archaeologists have re-emphasized that the community is a potentially significant unit of social identity, resource management, and settlement decision making (e.g., Adler 2002; Varien 2002; Varien and Potter 2008).

Moreover, some of our research to date strongly suggests the importance of modeling culture. For example, in the comparisons between our archaeological data and our ABMs, we think we see throughout the AD 600-900 period the persistence of cultural models dictating locational strategies biased towards easy access to big game, even as these cultural models became increasingly untenable with local and regional population growth and depletion of deer populations. This strategy experienced a crisis that resulted in increasing violence and the rapid depopulation of our original study area in the early AD 900s, and was replaced by a completely different settlement strategy (Kohler and Reed 2008).

Our decision to focus on groups is based not only on our perceptions of their importance in this record, but also on recent development of analytical methods that make their modeling practical, of theory that makes their representation desirable. The social sciences, including archaeology, seem to be on the cusp of a revolution that will allow us to escape from old qualitative dichotomies that pit concepts provided first by Durkheim against those favored more recently by methodological individualists to understand the relations between people and the groups in which we live. This revolution recognizes that there are deep phylogenetic reasons for the existence in humans of both tendencies towards cohesion within groups, and resistance of highly centralized power structures or invasion by other groups (Henrich and Gil-White 2001; Turner 2002; White 2007).

These new perspectives emphasize that competitive processes—and perhaps other considerations—can sometimes drive group sizes to be very large, even as people maintain nested subsets of relationships at smaller spatial and social distances. These provide “reservoirs” of social structure onto which the system can relax in periods of diminished competition. These scalable and dynamic social relationships are difficult to model and analyze, particularly in their ecological settings. But recent developments in ABM (e.g., in Kohler and van der Leeuw 2007) and network modeling (e.g., Barábasi 2002; Batty 2005; Watts 2002) demonstrate that it is possible to model and begin to understand systems with very complex and distributed synchronous or lagged social and environmental causal processes. Culture change in small-scale societies is such a process. By “culture” we mean not only the “cognitive universes constructed and maintained by human groups” (Watson 2008:29) but, more visibly, sets of strategies for dealing with resources, for constructing societies, and for maintaining communities, that may persist, at least for awhile, even if they become maladaptive.

For several decades “group selection” processes have been viewed as untenable for most organisms. But this is not—or should not be—the case for humans, since culture is a group-level property that presents special opportunities for cultural group selection that potentially explains the high levels of within-group altruistic behaviors that set humans apart from other primates (Sober and Wilson 1998). Bowles (2006; see also Soltis et al. 1995) has recently demonstrated theoretical and empirical feasibility for a model in which some reproductive leveling, such as monogamy or food sharing within groups, combined with lethal intergroup competition, can lead to high levels of (within-group) altruism.

Our next generation of models will still represent households, but they will form (from the bottom up) groups that may have slightly different strategies for resource management, for approaching conflicts with other groups, and for exchange within their group and between groups. This will be accomplished by extending the reach of the “cultural algorithms” (Kobti et al. 2006a,b; Reynolds 1994) in which certain aspects of exchange evolve in our previous models, and by partitioning the “belief space” in these algorithms so that groups can develop their own cultural traditions with respect to key sets of strategies that we wish to study (especially the evolution of turkey domestication, warfare, and exchange processes). This framework will give us the possibility of studying cultural group selection processes as well. Limited resources stress the survival of the social networks and even the social group itself (Kobti 2004).

One approach to building groups computationally is to implement individual household-level strategies enabling specialization in harnessing resources. This would enable the rise of task differentiation within a socio-economic system allowing for various types of cooperation and defection strategies within and across groups. Segregation of groups’ “belief spaces” will facilitate modeling the ability of a group to leave a region in which social or resource stressors are present. New social operators responsible for group generation, combination and splitting will be examined in the model and their outcomes compared with what we see in the local archaeological records. As we design and interpret our models, key consultants will be Samuel Bowles of SFI, an economist with a broad portfolio of modeling expertise and particular interests in evolution-of-cooperation problems, and Steven LeBlanc of Harvard’s Peabody Museum, an expert on aggregation, warfare, and the phylogeography of the prehispanic Southwest.

Exploring group composition through ancient genetic evidence

Human coprolites are an important source of ancient genetic material, especially when human remains are unavailable for study, and advancements in ancient DNA (aDNA) methods make it possible to use these specimens to investigate ancient population genetics (Kemp et al. 2006; Poinar et al. 2001). Following consultation with Pueblo groups, aDNA extracted from coprolite samples will be analyzed from existing collections held at MVNP (≈240 samples) and the AHC (≈105 samples), with additional collections currently being sought. We propose here to perform replicated extractions of aDNA and determine haplogroups and haplotypes from ≈100 coprolites. These data, sampled appropriately and combined with previous genetic studies of contemporary Southwest indigenous populations (Kemp 2006; Malhi et al. 2003) have the potential to address:

  • The biological identities of the populations in each of the two population cycles recorded in the archeological record of SW Colorado. We hope to be able to examine the suggestion, for example, that PI populations on different sides of the Dolores River may have been distinct (Wilshusen and Ortman 1999);
  • The extent to which the introduction of maize into SW Colorado was made by expanding populations of farmers or whether maize was adopted by local populations;
  • The extent to which Pueblo populations of the northern Rio Grande descended directly from populations that left SW Colorado, as we assume here on other grounds;

On a broader scale, these data will add detail to the record of population expansion recorded in the genomes of all Southwest populations studied to date. Furthermore, these data hold promise for resolving debates about the rate of mitochondrial DNA evolution (Kemp et al. 2007), which will add precision to molecular estimates, the goal of which is to find tight correspondence between the archaeological and genetic records (Kohler et al. 2008).

Geochemical Studies to Improve Estimates of Maize Paleoproductivity

Three major factors exert primary control on production of maize: heat (freeze-free days), water, and nitrogen (N). In the existing ABM, tree-ring-based reconstructions of climate (precipitation and temperature) are used to drive potential agricultural production. Nitrogen depletion in the model is subsumed within a “soil degradation” method and treated as an adjustable parameter. In this expansion of our research we seek to improve the paleoproductivity estimates by treating N cycling and depletion quantitatively.

Growing maize extracts a substantial amount of N from the soil; i.e., there is about 5 g of N in the above-ground plant. If maize is planted at a density of 7400 plants/acre (the highest density observed today among Zuni farmers), then about 37 kg of N is lost from the soil zone each year. Plots of soil pedon N concentrations across the San Juan Basin indicate that all of the San Juan Basin and much of the Four Corners area is nitrogen depleted relative to Zuni fields. In addition, almost all the N is in the organic form (N-org) and only about 3% of it is annually converted to the inorganic NO3- form that is biologically available. The timing of that conversion and transport of soluble N through the root zone is crucial to the productivity of maize. If, e.g., 1/3rd of the 3% N released were lost to the soil zone prior to root-ball establishment, a San Juan Basin field would be N-depleted in 3 years. In addition, N losses can occur via the growth of “weeds” within fields. Therefore, much of the NO3- may not be available to the plant during its growing season and rapid depletion of the N reservoir could occur. Also, N-org in the lower part of the soil zone may decompose an order of magnitude more slowly than the N-org in the upper soil zone, limiting the production of NO3- at depth (Fontaine et al. 2007). Lastly, optimum use of N assumes that it functions as a well-mixed reservoir, which implies that the maize must be planted so as to access the entire soil volume, a process which would seldom be achieved.

Nitrogen can be replenished in a variety of ways. There is however no southwestern documentation of manuring or adding vegetational debris with the exception of early historical observations of the Zuni, who built dams that concentrated organic matter on the distal ends of alluvial fans before planting maize. Moreover, a calculation shows that if all of a settlement’s human waste (urine and solid) were added (without subsequent loss) to a field during a 120-day growing season, it would only replace the N consumed by <2 bushels of maize.

Another way to increase N concentration is to take advantage of natural N-fixing plants such as mountain mahogany and antelope bitterbrush. During the early historic period, both plants thrived in the Mesa Verde-McElmo Dome area. A recent study by Smith (2004) has shown that bitterbrush releases 8-65 kg N/ha/yr to Mesa Verde soils with the release being proportional to the density of bitterbrush; each plant releases 27 g N/yr.

In quantifying the availability and persistence of N in the study areas, we seek to answer several questions:

  1. How much N (N-org, NH4+, and NO3-) exists in the soil zone?
  2. How much N do bitterbrush and mountain mahogany contribute to the soil zone?
  3. When does N-org transform to NO3- and what is the efficiency of that process?
  4. Does bitterbrush recycle deep soil-zone N to the surface?
  5. What is the residence time of soil organic matter (SOM) as a function of depth?
  6. How much N would escape the root zone of maize during infiltration of winter snowmelt and spring precipitation and how much N has been sequestered below the “active” soil zone?

To answer these questions we propose the following studies:

  1. Perform an inventory of the forms of N in the soil zone of the study area. Use soil probe or auger to sample each of 14 “productively lumped” soil groups in pristine or long abandoned settings. Each sediment core would consist of 20-cm increments that would be analyzed for N-org, NH4+, and NO3-. In addition, pH, conductivity, K, P, Mo, Fe, and S would be measured. Probe transects would be run between nitrogen-fixing plants, and areas containing different densities of bitterbrush (including no bitterbrush) would be sampled.
  2. Consolidate bitterbrush and mountain mahogany density data collected primarily by M. Lisa Floyd of Prescott College.
  3. At 4 sites in both the Mesa Verde and Rio Grande areas, lysimeter sets will be installed at 25-cm intervals to monitor NH4+, NO3- fluxes. Lysimeters will be pumped and samples taken at least monthly intervals and at times of snowmelt runoff and precipitation. Other metals, pH and conductivity will also be measured.
  4. Smith (2004) has shown using 15N that 90% of foliar bitterbrush N in the piñon-juniper woodland of Mesa Verde was derived by atmospheric fixation; therefore, it appears that 10% of bitterbrush’s N comes from the deep soil zone. We will use this value in concert with bitterbrush densities and soil-zone N concentrations to calculate the amount of N that is cycled from the deep soil zone to the soil surface.
  5. To determine SOM residence times, 14C analyses of soil organic carbon (SOC) will be determined at approximately 6 sites. Samples will be taken from 3 depths: 0-10 cm, 40-50 cm, and 90-100 cm. This determination will assist in determining the refractory nature of N-org with depth.
  6. The lysimeter study (item 3) will assist in determining how much N escapes the root zone of maize. With regard to N sequestration at depth, deep soil samples will be taken with the auger to determine the presence or absence of NO3- and Cl maxima (see, e.g., Walvoord et al. 2003). Such an accumulation may only occur in the valleys that incise the mesas in the study area depending on the thickness of the soil cap which is primarily composed of well-drained loess.

These studies will allow us to build more accurate estimates of nutrient depletion rates into the ABM. They will also allow us to improve estimates of base-rate productivity (without human impact) across the larger areas that we will be working with in this round of research.

Expansion of Research in SW Colorado

Within SW Colorado we propose to extend our research area to include 11 additional USGS 7.5˚ topographical quadrangle maps that encompass the Mesa Verde landform and MVNP. This approximately doubles the area in SW Colorado that we currently model.

MVNP is a National Register and World Heritage Site visited by more than 500,000 people/year. Following recent catastrophic wildfires that burned 2/3rds of the park area, NPS archaeologists re-surveyed most of the park, identifying hundreds of new sites and updating the documentation for existing sites. As a result it is reasonable to consider the entire park as a full-coverage block survey area encompassing 210 km2 where all the ancestral Pueblo habitation sites have been identified (see Fish and Kowalewski 1990 on the general advantages of full-coverage survey). Although the spatial extent of this survey, and its location within the heartland of a major prehispanic society, are unequaled in North America, much of the research potential in these data has not yet been tapped. This project will help MVNP archaeologists organize and synthesize these data, place them into a regional context, and analyze them to provide the first synthetic and quantitative demographic reconstruction for MVNP, thus providing a foundation for the additional research that this dataset deserves.The unique nature of these data will also allow us to make progress on many issues raised by our VEP research, conducted in areas just northwest of MVNP:

  • To what extent was the dynamic population history we observed a product of movement back and forth between our original research area and adjacent areas? Initial assessments of the settlement data from MVNP suggest somewhat complementary population histories in these two areas, leading us to ask why one area or the other might have been more attractive under certain socionatural conditions;
  • The MVNP data will allow us to further validate our agent-based models, a process that was limited in our original project by the relatively small size of its full-coverage survey areas. This made it difficult to assess their goodness-of-fit as “continuous-value maps” (Visser and de Nijs 2006). The much larger full-coverage survey presented by MVNP, in combination with the smaller block survey areas in our original study area, will provide a context within which it is possible to use these more powerful methods to assess model goodness-of-fit and hence to improve model performance;
  • What are the relative contributions of resource distributions and productivity, population density, physiography, and previous investment in socioeconomic infrastructure (e.g., Janssen et al. 2003) in producing social group territories? We have the opportunity to investigate these issues in unrivaled detail on Mesa Verde because of its very large full-coverage survey. The MVNP surveys encompass the local sustaining areas of multiple communities, as well as the hinterlands between them. This provides us a unique opportunity to investigate the social and environmental factors underlying the development of group territories and boundaries;
  • MVNP and the adjacent Ute Mountain Ute Tribal Park contain the highest concentration of cliff-dwellings in the U.S. These sites have been sheltered from the elements since the time of their construction, and contain large numbers of well-preserved roof timbers that can be dated to the year of harvest. Such data provide an unrivaled opportunity to investigate the long-term impacts of timber harvesting on forest structure with no new excavations, and will help us to calibrate our dataset of surface ceramics to time of manufacture.

Our presentation of results to a group of regional resource managers in a day-long workshop at the AHC in 2004 prompted MVNP cultural resource staff to invited us to extend our work to the park. In 2007 Glowacki and Johnson, members of the original team, evaluated the data sources available in the park, using funding secured by CCAC, and from this created a workplan for bringing these data into VEP formats. This report (Glowacki 2007) is the basis of our workplan for this component of our research.

Modeling Considerations. Extending our maize paleoproduction model to Mesa Verde and areas to its south and north gives us an opportunity to attempt to model the effects of cold-air drainage on agricultural production, since it has long been believed that the generally south-sloping mesa tops in the MVNP provide growing seasons that are unusually long, given their elevation, due to beneficial cold-air drainage (Erdman et al. 1969). We are examining these effects using strategically placed thermochrons, plus data from existing MVNP temperature records, and are drawing on USGS expertise coordinated by Benson to help us develop a model for cold-air drainage for Mesa Verde, the original VEP study area, and the new NRG study area.

Addition of MVNP and a discontinuous portion of the NRG will also require us to rethink how we handle household movement in the ABM. Except for those households that we seed at the beginning of each run, the current simulation structure does not allow households to either enter the simulated area (except by being born into it) or to leave it (except by dieing). We propose to build a border around our simulated areas in Southwest Colorado, and in the NRG, in which we do not model productivity, but into which agents can move if they cannot find an acceptable situation within their current area. They will be suspended in this “netherworld” and will age but not have children. From there they will costlessly monitor conditions in the closest portions of each simulated area and move to whichever offers the best opportunities, if any do, or remain in place in the “unmodeled” area. In this way we will see whether we can generate migration streams between these areas that resemble those we reconstruct from the archaeological record.

Finally, as mentioned above, given the importance of turkey in Pueblo II and later times (Rawlings 2006), raising domesticated turkey will become a household economic option, with its own cost structure and benefits drawn in part from experience of domestic-scale turkey farmers in SW Colorado.

Several important aspects of our modeling routines, developed earlier in this research, will be retained or only slightly refined. We will continue to use approaches pioneered by Johnson (2006; see also Kohler et al. 2007:73-78) for “growing” the grazed and browsed vegetation, bringing in annual climatic variability from tree-ring records, which in turn allow us to model the growth and density of rabbits, hare, and deer on this landscape. Likewise Johnson’s similar approaches to growing wood for fuels will be retained. By contrast, we have not found that attempting to build paleohydrological models of spring flows that are climatically sensitive improves the fit of our ABMs to the archaeological record better than simply using the locations of those water sources. We economize in the present research by simply locating relevant domestic water sources on the landscape, and assigning them average approximate flow rates from any available records, rather than building elaborate hydrological models.

Expansion of Research into the Northern Rio Grande

 Perhaps the greatest shortcoming of archaeological research in the Mesa Verde region is the fact that the sequence ends ca. A.D. 1280, and archaeologists have not had a clear understanding of which subsequent peoples, if any, should be viewed as the most direct heirs to the Mesa Verde tradition.

Recent research is beginning to clarify this problem, however, and leads us to view the Northern Rio Grande as an area where it is reasonable to speak of in-migrating Mesa Verde people continuing their tradition in some sense. We base this conclusion on:

  • Population reconstructions which clearly show that population growth in the NRG was complementary to population declines in the Mesa Verde region (Anschuetz and Scheick 2006; Crown et al. 1996; Duff 1998; Snead et al. 2004);
  • Bioarchaeological research using skeletal morphology suggests rather strong genetic continuity between Mesa Verde and the NRG populations, and further suggests that earlier populations of the Rio Grande were either swamped genetically or displaced by in-migrating Mesa Verde people (Kuckelman in press; Ortman 2007);
  • Although many details of pottery and architecture in the NRG exhibit stronger continuities with indigenous forms than with Mesa Verde forms during and immediately after the migration period (but see, e.g., Cordell 1995), landscape-oriented research shows striking continuity in place-making practices, such as the construction and placement of shrines, between the Mesa Verde and Rio Grande regions (Anschuetz 1998; Douglass 1917; Marshall and Walt 2007; Ortman in press);
  • Recent work sourcing obsidian artifacts from the original VEP study area (Arakawa 2006) shows that nearly all the obsidian found in Pueblo III period contexts derives from the Jemez Mountains in New Mexico. This finding supports the inference of long-distance contacts between the two areas prior to the final migrations.

While the above research indicates that 14th-century inhabitants of the Northern Rio Grande include at least some inheritors of the Mesa Verde tradition, other studies have examined how the northern San Juan’s ancestral Pueblo people contributed aspects of their knowledge and understanding in forging new forms of economy and social organization in the NRG landscape (Anschuetz 2007; Anschuetz and Wilshusen 2007). This cultural-historical continuity between Mesa Verde and the NRG creates several opportunities. First, we have an opportunity to hold cultural tradition constant to some extent in comparing land use and environmental impacts in two somewhat different, sequentially occupied environments.

Second, in the centuries following migration to the Rio Grande ancestral Pueblo economies developed in significant ways as a result of specialized production (Spielmann 1998, 2002) and the emergence of trade fairs (Ford 1972). Thus, linking the Mesa Verde and Rio Grande sequences brings into focus a significant economic transition, away from a kin-based system grounded in reciprocity to a more “economic” (market-based) system focused on exchange of goods. We therefore have the opportunity to examine the socionatural factors behind this transformation, as well as the effects of it. Was there something about the distribution of resources in the Rio Grande that encouraged economic development, or something about resource distributions in the Mesa Verde region that inhibited such development?

Finally, linking Mesa Verde and Rio Grande archaeology allows us to make use of the rich ethnohistoric records of the latter area (e.g., Harrington 1916; Marshall and Walt 2007) to document social and ethnic boundaries that were in place at the onset of the historic period. This in turn provides an opportunity for us to investigate the formation of such boundaries in the past as the Rio Grande landscape filled with in-migrating peoples from the Mesa Verde region and elsewhere.

We therefore propose to extend archaeological and modeling efforts to a 2600-km2 area of the NRG, bounded roughly by the city of Santa Fe on the south, the Jemez Mountains on the west, the Sangre de Cristo Range on the east, and Rio del Oso in the north (Fig. 6). This incorporates BNM, Los Alamos National Laboratory, the tribal lands of the six contemporary Tewa pueblos (Nambe, Ohkay Owingeh [San Juan], Pojoaque, San Ildefonso, Santa Clara, and Tesuque) and significant portions of the SFNF. This also incorporates the study areas of several recent projects that provide significant data for this project, including the Rio Sarco Project (Snead 1995), the Bandelier Archaeological Survey (Powers and Orcutt 1999), the Bandelier Archaeological Excavation Project (Kohler 2004), the Pajarito Archaeological Research Project (PARP) (Hill and Trierweiler 1986), the Rio del Oso Survey (Anschuetz 1998), and the Los Alamos Land Transfer and Conveyance Project (Vierra 2007).

Modeling Considerations. Adding the NRG will require us to rethink some portions of our maize production model, which has assumed that maize is dry farmed. Abundant research, much of it by one of our consultants, Kurt Anschuetz, has demonstrated the importance here of various water-management strategies—including the diversion and conveyance of seasonal flows in canals and culturally modified watercourses on the alluvial fans bordering the NRG’s permanent streams—from the Late Coalition period (AD 1275-1325) onward (e.g., Anschuetz 1998, 2001; Anschuetz et al. 2006). Potential maize production in these areas will be less sensitive to rainfall, but still sensitive to temperature.

Quantitative data for agricultural production in crop fields grown using formative Pueblo agricultural technologies currently are lacking. A recently completed hydrological study at the northwest end of the proposed NRG study area, however, has developed GIS-based protocols for evaluating the relationship between precipitation moisture, physiography, soils, and runoff flows at the scale of whole watersheds (Miller 2007). The study’s contributions include methods for quantifying the minimum precipitation during rainfall and snowfall events needed to produce economically useful runoff in various physiographic settings. This work also provides frameworks for evaluating the percentage of runoff moisture, as well as which lands would be unavailable to Pueblo farmers, based on storm intensity.

Geomorphological investigation within the NRG study area provides important information concerning the hydrological and depositional dynamics of the alluvial fan environments that flank major stretches of the Rio Grande valley where extensive archaeological field complexes dependent upon runoff irrigation are found (Banet 2006). This work, which documents the life history of an alluvial fan agricultural field area, which was intermittently used to grow corn between the late 13th and early 17th century, provides a framework for modeling the spatial structure and the frequency of use of field use.

As mentioned above, new types of economic transactions become archaeologically visible in the NRG after 1300 (Kohler et al. 2004) and begin to supplement the reciprocal exchanges that we have been modeling for the AD 600-1280 period in SW Colorado. Given the manifold economic uncertainties confronting the NRG populations dependent on runoff-water management, these patterns raise the possibility that the people used their enhanced exchange networks to redistribute their numbers among multiple contemporaneous centers to accommodate local residential movement at demographic scales and frequencies unprecedented in either the Mesa Verde or the NRG before the thirteenth-century migrations. In addition, the seeming over-construction of villages—there are at least 75 villages of up to 2000 rooms in the Tewa Basin alone (Anschuetz 2005, 2007) relative to the local area’s agricultural potential (Anschuetz 1998, 2007)—suggests that the NRG’s people significantly modified their community organizations between AD 1250 and 1350 (Anschuetz and Wilshusen 2007).


In overview, VEP II examines a millenium of the late prehistory of the eastern Pueblo Southwest as a unit, building on foundations laid by Adler (1996), Adams and Duff (2004), our own earlier work, and three generations of our predecessors. This involves using common currencies for measuring production and population sizes by creating estimates in commensurable fashions. This research will bring old and disused records into a system in which they can contribute to contemporary research concerns. This effort in turn provides a framework on which others can build, lending archaeological research in this area a more strongly cumulative and increasing-returns (in the sense of Bowles 2004:12ff) character. Our research will examine the settlement patterns in these areas against a simple household-oriented ABM with an optimality ruleset for settlement location, used successfully in our research to date, but also against more complex models involving group processes that have never been employed computationally in archaeology. Our successes (and our failures) in this arena will likewise provide models from which others can learn.

Although this is an ambitious project, it is not nearly as expensive as it might be because we are drawing on a great deal of research already conducted under federal auspices—such as the surveys on Mesa Verde—and on close collaboration with employees of the USGS, NPS, and USFS, largely at no cost to this proposal. 

VEPII Project Area SM.png165.04 KB

The Village Project

The Village Project is designed to help archaeologists understand the factors influencing settlement patterns of small-scale agrarian peoples. Although such societies are becoming increasingly rare, they represent the norm throughout most of the Neolithic period the world over.

This project uses agent-based modeling to investigate where prehistoric people of the American Southwest would have situated their households based on both the natural and social environments in which they lived.

We seek to understand general processes in the environments of southwestern Colorado between A.D. 600 and A.D. 1300. Agent-based models allow us to study a system characterized by high degrees of interaction between the landscape as it was affected by climate change and by the actions of farmers, and among the farmers themselves, as they sought to make a living in this marginal farming area.

Department of Anthropology, PO Box 644910, Washington State University, Pullman WA 99164-4910, 509-335-3441, Contact Us