Predictive modeling is an important tool in devising survey and excavation strategies. Predictive models can be either raster or vector based, and range from very simple to extremely complex. Predictive modeling is as valid for pre-contact resources as it is for historic resources. Generally, precontact predictive models rely on such factors as distance to permanent water, slope, aspect, soils, geology, wildlife habitat and seasonal botanical resource availability.
Historic predictive models rely in part on these themes, but often better data is available to aid in predicting historic resource presence in a project area. Some of these datasets were discussed in the Historic Data Assimilation Section, and include historic aerial imagery, Government Land Office (GLO) cadastral survey maps, historic topographic maps, Sanborn Fire Insurance maps, and the like. The best historic resource predictive models also utllize available data to help rule out areas where cultural resources may be present. Such datasets include buried utilities data and transportation data, as well as aerial imagery or a comparison of old topographic maps versus modern maps which allow the user to identify man-made topographic features such as road cuts and fills and other areas of large scale ground disturbance.
The two graphic examples below represent cultural resource predictive modeling at its most basic level. The first image is a georeferenced 1943 aerial photograph which aided in the development of a field inventory survey, and provided insight into the location of historic agricultural sites not readily apparent on the ground surface today. The second image below contains a georeferenced 1916 GLO map with two alternative project transportation corridors overlaid. The georeferenced GLO aided in cultural resource survey and site inventory efforts, as well as in the selection of a project alternative. Note that the Alternative 2 corridor is located in what once was the 1916 river channel. This alternative corridor is actually the existing road prism, which has since been filled in by a series of road-realigning efforts in modern times. As such, it has little potential to contain intact pre-contact or historic site deposits. Assuming accuracy of the historic hand-drawn GLO (and this one has proven to be highly accurate), we also know that the river channel has changed over time, and how much.