Development of the GIS-Based Remote Area Soil Proxy (RASP) Model.
Toby Rodgers1, Crystal Briggs1, Jon Riedel2, Alan Busacca3, and Bruce Frazier3. (1) USDA-NRCS, 2021 E. College Way Ste. 106, Mount Vernon, WA 98273, (2) National Park Service, 7280 Ranger Station Rd., Marblemount, WA 98267, (3) Washington State Univ., PO Box 6420, Pullman, WA 99164-6420
Remote areas such as wilderness and National Parks in the western United States have historically been excluded from soil inventories due to the investment of time and resources required to map such rugged and distant terrain. A Geographic Information System (GIS)-based Remote Area Soil Proxy (RASP) model was developed to map soil distribution in remote areas of Washington State. Our goal was to demonstrate an adaptable GIS-based model for predicting soil-landscape relationships that can create a defensible soil distribution map expeditiously. Our RASP model is designed to create soil maps of remoteareas relying on available GIS data and tacit knowledge gathered in the field by trained soil scientists to create rational soil forming factor digital proxies. The Pasayten and Sawtooth wilderness areas (254,000 ha total), and Thunder Creek watershed (30,000 ha) in North Cascades National Park, Washington served as study areas for development and application of the model. Our model utilizes a 30 meter by 30 meter digital elevation model (DEM) and can be adapted to take advantage of higher resolution DEM's and other digital data if available. We found landscape stability, parent material, and vegetation to be important factors of soil formation and selected appropriate digital data layers to serve as proxies for these factors. A landform map created by Jon Riedel, North Cascades Park geologist, was used as a proxy for topography and parent material, and time as indicated by landform stability. A vegetation layer interpreted from remotely sensed data was used as a proxy for vegetation and climate. Additional topographic attributes, such as slope and wetness index, were calculated from the DEM and incorporated into the model. Threshold data values and combinations, verified through field examinations, were utilized to group recognizable soil-landscape continua into meaningful map units.