Testing the SRS Predictive Site Model ___________________________________________________________________________________

During spring 2002 I took a brief foray into prehistoric archaeology and conducted an interesting statistical analysis of the predictive site model used on the Savannah River Site by the Savannah River Archaeological Research Program (SRARP).  The model is used to predict the location of sites on the SRS and manage archaeological resources.  Simply put, the model is based on the premise that prehistoric sites with greater content complexity are located in close proximity to major water sources.  Conversely, as distance from water sources increases, prehistoric sites become less complex, contain fewer components, and exhibit less artifact deposition and diversity.  


Distribution of sites on the SRS based on archaeological sensitivity zones.


To statistically test this premise I analyzed the distribution of a sample of sites in relation to water using basic linear regression analyses and chi square tests.  The results, summarized in a brief interim report, clearly demonstrate that proximity to water determines prehistoric site location and complexity.  Proximity to water also influenced historic site location on the SRS until road systems and modern well technology altered the settlement patterns. 


Regression analysis results, number of recorded sites plotted by distance from water, full coverage and predictive model site samples.


Groover, Mark D.  2002  Evaluating the Predictive Site Locational Model: Preliminary Results of Full Coverage Survey on the Savannah River Site, 1998-2002.  Interim report on file, Savannah River Archaeological Research Program, South Carolina Institute of Archaeology and Anthropology, University of South Carolina, Columbia.

Predictive model report in pdf file


Back to Research Page