Wednesday, November 4, 2009: 11:30 AM
Convention Center, Room 405, Fourth Floor
The current Soil Survey is provided as polygon coverage, new GIS technologies and software in combination with high resolution spatial data has increased the pressure to improve the soil data delivery format and add raster based products. This would require development of methods to translate qualitative soil-landscape models to numerical ones. The objective of this research is to refine methods for creating spatially distributed raster based predicted soil maps and property maps from the current SSURGO soil polygon maps. Howard County, Indiana was the focus of this study. The county was divided in 7 zones based on SSURGO and suggestions from the Indiana NRCS Soil Scientist staff. Approximately 360 georeferenced soil descriptions were collected throughout the county, representing major soils Blount, Pewamo, Fincastle, Brookston, and Crosby covering 80% of the area. The methodology defined as Terrain Attribute Soil Mapping (TASM), consists of (i) advanced terrain analysis tools to characterize landscapes numerically, (ii) tacit knowledge to formalize analysis of soil-terrain relationships, and (iii) fuzzy membership techniques to assign fuzzy soil classifications to pixels to generate spatially continuum soil property maps. TASM is a unique combination of technologies within the framework of Digital Soil Mapping (DSM) and emphasizes soil-terrain attribute relationships at the landscape scale. The validations results indicate substantial agreement between TASM and SSURGO maps (overall accuracy 0.77%). The kappa coefficient was 0.74% suggesting the substantial agreement was not random. One of the advantages of TASM map is its raster format, especially, the soil membership values at each pixel that can be used to generate soil property maps that would recognize the known fact to the soil scientist that soil properties within and between map units vary on a continuum basis. The methodology described provides powerful tools for the Soil Survey to deliver raster based soil maps and soil properties.