Mapping the Willamette National Forest.

Digital soil mapping (DSM) can be used to create accurate soil maps of large areas with rugged and remote terrain.  DSM can be especially useful for creating modern soil maps in areas with some existing soils information.  This project uses random forest (RF) decision trees to create a digital soil map in conjunction with ongoing NRCS soil mapping of the Willamette National Forest (WNF) at the family level of soil taxonomy.  Combinations of elevation, slope, aspect, geology, landform, vegetation, temperature, precipitation, and soil resource inventory data were used as predictive variables in mapping.  Predictive maps were tested using randomly generated points throughout the WNF.  The most accurate map will then be given to the NRCS to assist in their mapping efforts.  The map produced through this study will be used to fill in rugged and remote areas, particularly wilderness regions which are difficult to map using traditional methods.