Developing and evaluating a method to disaggregate legacy SSURGO soil maps

Abstract

For soil mapping, legacy (survey) data has its greatest potential in the disaggregation of complex soil mapping units. Disaggregation in this context is possible by relating different soils within a single map unit to other known characteristics of the landscape such as slope and curvature. For the United States soil map SSURGO, virtually all current soil map units have multiple soil types (known as "series") described within them, indicating there is great potential for such disaggregation. In this study SSURGO map units are disaggregated using descriptions of soil map units and Digital Elevation Models (DEMs) for three locations in the central United States. The model presented disaggregates soil map units using information already within the soil description such as the curvature and slope in which a certain soil series is found. First, slope and curvature are calculated using National Elevation Dataset 10m resolution DEMs. Next, soils are assigned preferentially using fuzzy membership functions to the slope and curvature values for each map unit until all the possible soils are disaggregated. The area percentage of each new soil is calculated, and the model is adjusted until the SSURGO-stipulated percentage of each component is reached. The method worked best for county size areas and the newly created maps increased the number of map polygons by over 100%. The new maps were then compared to soil point data and showed similar accuracy to SSURGO and better accuracy than a similar digital soil mapping approach.