Spatial computer models were applied to a route location problem in eastern Wisconsin. Data included soil associations, land cover, slope, and land use stored by 1/9 km cells. Models were constructed to represent land use ideas supported by various sectors of society. Routes automatically selected by each model were plotted on a medium intensity soil map to test their effectiveness in avoiding prime agricultural land.
In an area that was mainly classified in U.S. Department of Agriculture capability classes I and II (82 percent), routes selected by the various models included from 46 percent to 86 percent classes I and II land. Other tests indicated improvements in model performance with increasing length. Subtle differences in models provided significantly different route locations within 17 km. Major differences in models showed significantly different locations within 6 km.
Revised from a land use symposium paper published in Proceedings of the American Society of Photogrammetry, Fall Convention, 1975.
Notes
Revised from a land use symposium paper published in Proceedings of the American Society of Photogrammetry, Fall Convention, 1975.