SOIL DRAINAGE CLASS PROBABILITY MAPPING USING A SOIL-LANDSCAPE MODEL

The direct application of quantitative soil-landscape models for soil mapping has been limited by technological constraints. This study combines a statistically based soil-landscape model and geographic information system (GIS) technology to create soil drainage class maps. An existing soil-landscape model that predicts soil drainage class from parent material, terrain, and surface drainage feature proximity variables was used. A digital geographic database of parent material, terrain, and surface drainage feature proximity variables stored in a geographic information system were used as model inputs. Combinations of these landscape variables were defined by overlaying the digital maps and by applying the soil-landscape model to create predictive maps of soil drainage class probability and most-likely soil drainage class. The modeled soil drainage class map agreed with an Order II (1:20000 scale) soil survey for 67% of the study area. A majority of the disagreement was attributed to areas predicted as somewhat poorly to moderately well drained by the model and well drained by the soil survey. This technique consistently assigns soil drainage class based on landscape attributes, documents the data and decision criteria used for drainage class assignment, estimates the uncertainty associated with drainage class assignment, and generates a digital map for GIS applications.