Habitation Sites and Biotic Communities

The prehistoric peoples of San Diego County drew their sustenance from the region’s natural plant and animal communities. There was considerable spatial variability in the distribution of those resources, and it is reasonable to suppose that settlement strategies were closely attuned to such variability. Less certain are several other matters. Which specific biological settings were valued or avoided? On what spatial scale was settlement adapted to biological setting? Were individual settlements located primarily to maximize access to a particular resource, or to maximize the diversity of the accessible resources?

Several studies in San Diego County have addressed the question of correlations between archaeological site locations and biotic communities:

Allen Jay Prosser (1975) discussed site distributions in Cuyamaca Rancho State Park in relation to several environmental variables, including elevation, which served as a substitute variable for biological communities. Prosser reported a bimodal pattern of site frequencies by elevation. The lower mode corresponded to the main range of coast live oak (Quercus agrifolia), and the upper mode matched the ranges of black oak (Q. kelloggii) and interior live oak (Q. wislizenii).
Ronald V. May (1980, 1987) collected data on the frequencies of different site types within five plant associations in the Table Mountain area. According to his statistics, habitation sites (base camps and campsites) occur more frequently with grassland and leafy shrub associations, while other site types (quarries, flaking stations, workshops, and roasting pits) occur more frequently with agave/yucca associations.
John R. Cook and Scott Fulmer (1980) analyzed site distribution patterns for the McCain Valley Study Area, a belt of BLM land in eastern San Diego County, on the basis of a statistical sample of survey transects. They found a strong positive correlation between site locations and upland desert scrub, and a negative correlation between sites and mountain chaparral.
M. Steven Shackley (1980) studied site locations in Cuyamaca in relation to six biotic associations. He observed that nearly all of the village sites were situated in close proximity to junctions between three different associations.
William R. Graham (1981) analyzed site data from a comprehensive survey of the Laguna Mountain Recreation Area. Among other variables considered were the on-site vegetation community and the distance from meadow edge. Graham found that sites occurred more frequently in grasslands settings and in the vicinity of meadow land, and less frequently in brush lands, than would be predicted by a random location model.
Mary Robbins-Wade (1990) evaluated the location of recorded sites (primarily Archaic-period) on Otay Mesa, in particular their proximity to the ecotones defined by major physiographic features. She reported an association of habitation sites with the ecotones that exist between open areas of grassland or sage scrub and brushier communities of the foothills and canyons. Contrary to her expectation, Robbins-Wade did not find a statistically stronger association with ecotones for residential bases as against field camps.
Lynne E. Christenson (1990) studied a statistically random sample of records for Late Prehistoric sites in the southwestern portions of the county in relation to present and projected prehistoric plant communities. She compared site locations with a stratified sample of non-site locations in the same region. Christenson concluded that sites were found less commonly than non-site locations in modern chaparral communities and more commonly in riparian, grassland, coniferous forest, and oak woodland settings. Somewhat surprisingly, she found little relation between projected prehistoric plant communities and site frequencies.
Koji Tsunoda (2006) used a GIS database of environmental and archaeological variables to analyze relationships between site locations and 10 plant communities in Cuyamaca Rancho State Park. Sites were found to be most numerous in woodland areas, but they occurred most densely in the more limited areas of upland grassland and montane meadows. Sites were fairly numerous in the extensive chaparral areas, but their density was low.
Jerry Schaefer and Don Laylander (2014; Laylander et al. 2015) did a GIS analysis of recorded site locations in the Jacumba/McCain Valley area of southeastern San Diego County. They compared the distributions of sites (analyzed by types and by attributes) with the distributions of five large habitats and 29 more specific vegetation units. Among the findings were that sites were overrepresented in areas of pinyon-juniper woodland and Upper Sonoran woody scrub, and underrepresented in areas of Upper Sonoran mixed chaparral.

Some potential problems may be considered in interpreting the findings of such site location studies:

The archaeological record may be significantly biased by differences in site survival and discovery that relate, directly or indirectly, to the sites’ biotic settings.
The level of generalization in the available mapping of biotic communities may mask patterns of archaeological interest. Shackley (1980) found that the existing mapping of the vegetation communities in Cuyamaca was not sufficiently detailed and accurate for the scale of his settlement study. He prepared revised mapping on the basis of his own observations on the ground.
There may be important discrepancies between modern and prehistoric biotic distributions. Long-term climatic changes may be a factor, at least for periods prior to the Late Prehistoric period. Historic-period changes have been substantial, including the effects of introduced plants, livestock grazing, fire suppression, and water control. John Marvin Dodge (1975:109) studied the effects of historic-period land use and fire history on the vegetation of San Diego County and postulated natural distributions for the major vegetation communities, primarily on the basis of soils units. As compared with modern distributions, Dodge’s model suggested that under natural conditions chaparral was somewhat less widespread, particularly in the western portions of the county, while grassland and woodland were more common.
Prehistoric cultural activity may itself have altered biotic distributions. Associations between habitation areas and biotic communities may in some cases reflect patterns of intentional or unintentional prehistoric alteration of the environment rather than locational selection. Shackley (1980) found that all of the sites in his Cuyamaca study area were located, at least in part, in grasslands, but noted that grassland is an early succession stage after fire and that aboriginal peoples in the region apparently used fire to promote favored resources.
The geographical scale or scales involved in prehistoric locational choices is not certain. Should analysis be based on the on-site biological characteristics of habitation sites (as several of the studies have done), on the characteristics of an area within an easy-access radius of such sites, on a daily foraging catchment, or on some other scale?
When significant associations are found between site locations and biotic communities, they may reflect the correlations of both these variables with some third factor, rather than a direct causal relationship. For instance, if sites are found to be associated with the riparian biotic community, is the reason a high value attached to the biotic community, or does it relate to the availability of water, or perhaps access to convenient travel routes?
Associations between site locations and biotic settings need to be tested against random patterns. For instance, if all habitation sites in a study area are found to lie within 5 kilometers of an ecotone, the significance of that fact can only be evaluated when the proportion of all potential site locations in the study area that share that characteristic is known.

PROSPECTS

Future archaeological investigations may be able to shed light on the extent to which locations for habitation sites were selected on the basis of their proximity to a particular biotic community or to a variety of such communities. Relevant evidence may include site locations, present and projected prehistoric distributions of biotic communities, the distribution of other environmental variables potentially affecting location selection, and factors biasing site survival and discovery. The availability of GIS databases offers an enhanced potential to address these issues more rigorously.