The Dusty Trail to Digital Soil Survey in California.
David W. Howell and David W. Smith. USDA Natural Resources Conservation Service, 1125 16th Street, Room 219, Arcata, CA 95521
Soil survey is undergoing dramatic changes which will stand out when we look back at the continuous progress of methods and technology used by soil scientists. One challenge is to implement new technologies into a well-established workflow without dramatically impacting production of stated public goals. In the United States the National Cooperative Soil Survey is implementing new technologies in a variety of ways. Public agencies, universities, and other organizations are using methods similar to those we have adopted in California. In some cases we have adopted methods and tools developed in other parts of the country. Our purpose is to describe the trail that we followed through the new technologies. Our use of these new digital tools is becoming clear, but the trail has not always been obvious. It has been a dusty trail. Beginning in 2002 the USDA Natural Resources Conservation Service in California began planning for the implementation of complete, start-to-finish, digital soil mapping at the field soil survey office level. This would include digital soil survey delineation, digital conversion of existing unpublished paper soil maps, and statistical landscape modeling of selected soil properties. Digital soil survey delineation required new hardware, software, data, work processes, training, and support. In short the workflow of soil survey offices would change dramatically. Digital GIS tools would be substituted for long-standing analog methods. These substitutions needed to be implemented in a phased approach because of budget constraints. The new methods were introduced at a variety of stages in ongoing projects in order to reduce the impact on production. Training was coordinated with national training development and timed to coincide with individual soil scientist and soil survey project needs. Onsite training and support is now augmented with remote network-based support. Digital conversion of unpublished paper soil survey maps provided cost-saving methods to augment the digital data available during ongoing projects as the soil scientists began to create new digital soils data in adjacent areas. This is a transitional, temporary service which will conclude when unpublished paper maps have been converted. This digital conversion consists of scanning aerial photographs with the soil lines on them. The non-rectified photos are orthorectified using image processing software. The soil lines are then extracted and vectorized to be incorporated into an overall digital soil survey. In addition, previously published soil survey data have been converted to digital form. Statistical soil-landscape modeling is being developed to produce estimates of the geographic distribution of selected soil properties. These digital raster models will be used initially to provide pre-mapping estimates to help soil scientists plan their field work. The models are intended to increase sampling efficiency and to increase explicit understanding of soil-landscape relationships. We hope that eventually these model outputs will provide a new product verified by soil survey field work. These new products will be raster-based estimates of soil properties that vary continuously over landscapes, rather than represent soil properties as polygons. We feel that these raster maps will augment polygon data initially, but may eventually replace the polygon form of soils maps. Raster soil property modeling is being developed rapidly around the world. As the input data representing the soil-forming factors improve, so will the model estimates. Quality standards for model properties and for interpreting the results for practical application will be active development topics over the next few years.