Digital Modeling of Soils in a Production Soil Survey on the Culberson Gypsum Plain, West Texas

Traditional soil survey techniques have been successfully used for years to produce high-quality maps and information for landowners. Unfortunately, it can be difficult to extract the underlying soil-landscape models used to develop these maps, especially where physical limitations such as poor air photography and limited physical access conspired to degrade the consistent application of the soil-landscape model to map creation. The steadily increasing wealth of digital earth surface data, coupled with soil modeling software based on fundamental pedological principles, now allows for field soil scientists in production settings to at least partially overcome the limitations imposed by traditional methods. Modeling software requires that field soil scientists understand the environmental setting for each soil component to be mapped, and explicitly record within a knowledge document their understanding by relating each component to available GIS input data. The knowledge document consists of 1) a list of soil/land components to be mapped, 2) a catalog of environmental data layers necessary to model the distribution of each component, 3) a set of rules with critical values of at least one environmental data layer that distinguishes two components that are adjacently juxtaposed on the landscape. The software is able to implement the rules over land large tracts quickly and consistently. The resulting raster maps can reflect the soil scientist's latest and best understanding of soil landscape relationships. Soil scientists can then focus their time and energy on validating and documenting their soil-landscape models, rather than laboriously translating their mental constructs into maps. This process is illustrated for a production soil survey area in west Texas.