Geographical Information Systems (GIS) spatial analysis related to the distribution of running water, bedrock fractures related to groundwater, and prehistoric archaeological sites is a useful tool in prehistoric occupation site prediction. This GIS analysis produced a map showing three zones of probability for prehistoric archaeological site occurrence. The map displays a high correlation between a test set of known prehistoric sites and the high and medium probability of occurrence areas. The test results show that 85% of 46 test sites in the 200 km² study area in the Lehigh Valley section of the Great Valley physiographic province, Pennsylvania, fell within high to medium probability zones. A total of 72 prehistoric sites are known within the study area. A sample set of 26 sites is used to create the predictive model. The remaining 46 sites are used to test this model. GIS is used to correlate the sample set of 26 prehistoric sites to mapped surface water and bedrock fractures related to groundwater resources. The spatial relationship between the location of bedrock fractures, surface water, and prehistoric archaeological sites, is used to synthesize a statistical equation (W_b=P * S_d) to predict the probability of prehistoric site occurrence in 100 m² grid cells. These variables are chosen because the availability of water is the most important predicator of site location. The bedrock lithology of the study area includes Upper Cambrian to Upper Ordovician limestones, siltstones, argillites, and conglomerates. The validity of these probability zones proves that GIS spatial analysis is a useful tool in prehistoric archaeological site prediction.