INCORPORATING VARIABILITY INTO SOIL MAP UNIT DELINEATIONS

Published soil surveys comprise a nationwide soils data base available to users. Although adequate for farm planning and agricultural production, the surveys lack information on spatial and temporal variability of soil attributes. Such information is increasingly needed for modeling of water flow and contaminant transport, in geographical information systems (GIS) applications, and for environmental impact assessment. The proposed approach combines spatially interpolated distributions of measured values with soil map unit delineations within a GIS framework. Data analysis provides insights into the potential continuity and variability of attributes within and between the map units. The result is a map that preserves the map unit boundaries but incorporates spatial variability of attribute data into map unit delineations. The approach is illustrated using bulk density and hydraulic conductivity data for surface soil horizons at a farm scale. Semivariograms based on measured values are constructed and cross-validated. Subsequently, GIS overlays of soil survey attributes and kriged overlays of measured data can be prepared. These GIS overlays can be combined pixel by pixel according to a simple rule. Results indicate that spatial distributions of bulk density and hydraulic conductivity are similar to original map unit delineations, but include aspects of the attribute continuity and variability from geostatistical analysis. Such combined representations of bulk density and hydraulic conductivity are superior, because they contain more information than either the soil survey map, or the kriged interpolation of discrete data.