After more than 100 years of effort, modern soil surveys have been completed for all 88 counties in Ohio. Digitizing of spatial and attribute data for Ohio will be completed in 2006. While complete coverage of the State will be available on a county-by-county basis, seamless soil survey maps at Order 2 scale covering multiple counties are not. Major challenges include map unit contrasts across county lines, and disparate intensity, vintage, and quality of survey information between counties. Application of soil resource data by a variety of users, particularly land use planners, will be adversely impacted unless progress in soil survey activities continues.

Future updating activities by the cooperating agencies are being organized on a Major Land Resource Area (MLRA) basis. In the future, individual survey areas will more closely match natural landscape divisions, rather than administrative boundaries. Beyond cross-county edge-matching, major tasks for future survey projects will include rationalizing consistent geographic and taxonomic units, and dealing with significant differences in survey intensity resulting from varying goals and approaches of survey teams of disparate vintage.

Generally, the second century of soil survey in the USA is not yet well served with a modern coherent framework of operational techniques and approaches. Geospatial and pedometric techniques are likely to contribute significantly to future survey activities, but these approaches have not yet been used for routinely updating areas as large as an MLRA. Geospatial and pedometric techniques hold great promise for improving the efficiency of survey, and in producing a more reliable product. However, pedometric techniques have been generally been applied only for small areas, or at small map scales, and institutionalization of these techniques is progressing slowly in the soil survey agencies, where they are often seen as experimental.

A soil-landscape modeling approach is being used to facilitate updating of soil surveys on a MLRA basis in Southeastern Ohio, MLRA 126. A case study involves Belmont and Noble counties, two adjacent areas with similar “soil forming factors”, but contrasting survey vintages and map unit intensity. Soil landscape models have been constructed for each survey area individually and in aggregate. Models were constructed using the SCORPAN approach. Base data layers included the existent soil surveys, specific terrain-based climate attribute surfaces (e.g. solar radiation), historic vegetation, land use and land cover information, terrain attributes from digital elevation models, and surficial geology derived from interpolation of large amounts of stratigraphic data. Models have been constructed using classification trees. Methods for assessing uncertainty in the use of classification trees have been explored.

The models have been applied to identify areas of major discrepancy between the adjacent survey areas, to redefine rational mapping units and mapping intensity, to locate areas with significant mapping uncertainty, and to guide further field investigations. It is suggested that a soil landscape modeling approach can be applied more widely for soil survey updates, and an operational framework is proposed.