100718 Prediction of Volcanic Ash Distribution and Depth in Western Montana and the Eastern Idaho Panhandle, USA.

Poster Number 460-637

See more from this Division: SSSA Division: Forest, Range and Wildland Soils
See more from this Session: Digital Soil Mapping of Forest Soil Properties Poster

Wednesday, November 9, 2016
Phoenix Convention Center North, Exhibit Hall CDE

Robert Vaughan, RedCastle Resources Inc., USDA Forest Service-RSAC, Salt Lake City, UT, Vince Archer, USDA Forest Service, Missoula, MT, Jay Skovlin, USDA-NRCS, Missoula, MT, Kevin Megown, USDA Forest Service-RSAC, Salt Lake City, UT and Paul Maus, RedCastle Resources, Inc., USDA Forest Service-RSAC, Salt Lake City, UT
Abstract:
Knowledge of the spatial distribution of volcanic ash-influenced soils is integral to elicit effective forest management and planning in the Inland Pacific Northwest United States. Volcanic ash distribution and ash mantle thickness were modelled across several forests in western Montana and the eastern Idaho panhandle. A Random Forest binomial probabilistic model was used to predict ash presence while regression was used to predict ash mantle depth. Ash mantle presence and depth were garnered from georeferenced soil pedon data (n=2,235) collected by the NRCS and USFS during Soil Surveys and Soil Resource Inventories. Topographic environmental covariates used in the model were developed from 30-m digital elevation models while imagery-based indices (Landsat, 1984-2015) were derived using Google Earth Engine in an effort to reduce model error from forest management and fire scars. Preliminary results suggest that elevation and the first principle component of an 80th percentile median value August Landsat composite image (1984-2015) ranked highest in predictive variable importance. Unexpectedly, slope aspect does not appear to have any predictive value for ash presence. An assessment of map accuracy will be developed.

See more from this Division: SSSA Division: Forest, Range and Wildland Soils
See more from this Session: Digital Soil Mapping of Forest Soil Properties Poster

Previous Abstract | Next Abstract >>