This research endeavors to explain why Ohio’s prehistoric inhabitants placed their earthworks on the Central Ohio landscape at those specific locations. It constructs a probabilistic model based on the locations of known archaeological sites from the Early and Middle Woodland period. Known locations were compared to a set of randomly placed comparison points using digital GIS landscape coverages, and remote sensing imagery to find areas where unknown earthwork sites were more likely to occur. The sample of nine known earthwork sites and nine randomly placed points were located in an arbitrary area of 3825 km2 in Central Ohio.
For each of the 18 points, landscape and descriptive variables were collected to find a set of traits for each site. The variables were assessed statistically using Monte Carlo analyses and Multiple Components Analyses in SAS to see if differences existed between the two populations. While some of the variables did not have statistical significance, visual appraisal of the data did reveal patterns, for example, viewshed and gateway direction. For other variables, the statistical analyses did find that known archaeological sites were more likely to be found in certain environmental areas compared to the random points.
Of the variables that did have statistical significance, a number of them were then selected, intersected and clipped in the GIS coverages to identify Similar Environmental Groupings on the Central Ohio landscape. These Similar Environmental Groupings were areas where certain variables co-occurred on the landscape. The GIS identified 69 such Similar Environmental Groupings, with 18 of those of special interest based on the variables tested. These Similar Environmental Groupings were places where Early or Middle Woodland earthwork sites were more likely to be located based on the statistical analyses.
Visual searching of the 69 Similar Environmental Grouping areas was conducted on various remote sensing imagery with the goal of finding vegetation or soil signatures that would indicate a buried archaeological earthwork structure. While no unknown sites were found visually, ground truthing with various geophysics techniques may find such buried evidence. This research has shown how difficult it is to pinpoint archaeological sites as there are so many variables and pragmatic events to consider. Nevertheless, the GIS model presented here is a dynamic one that can be adapted for further research.