Terrain Attribute Modeling of Volcanic Ash Distributions in Northern Idaho

Abstract

Volcanic ash mantles many landscapes of the Inland Pacific Northwest region of the United States. Because of the close link to forest productivity in the region, understanding processes that have affected the present‐day distribution and characteristics of these ash mantles is important for forest soil management. Presence or absence, thickness, and degree of ash mantle mixing were evaluated at 84 randomly selected stratified sites in the Palouse Range of northern Idaho. A 1‐m digital elevation model (DEM) was generated for the Palouse Range using light detection and ranging (LiDAR) data and resampled to 10‐, 15‐, 20‐, and 30‐m grid resolutions. Terrain attributes derived from these DEMs were used to model volcanic ash mantle presence or absence, thickness, and degree of mixing using classification and regression trees. Model accuracy for ash presence was assessed using 572 data points collected by the NRCS as part of a soil survey update. Overall, elevation was the single variable most related to the presence or absence, degree of mixing, and thickness of a volcanic ash mantle; other terrain attributes had less predictive value in modeling ash mantle characteristics. The 30‐m grid resolution provided the best model of ash presence or absence, with 78% accuracy, indicating good promise for digitally mapping Andisols and related soils across the region. The various grid resolutions had little effect on the outcome and predictive ability of the models, and the overall accuracy of the models varied by only 2%. Moister, higher elevation plant communities provide a protective forest canopy and thick litter layer, which result in a relatively undisturbed ash mantle. At lower elevation where forest canopy is less dense, ash mantles are thinner, highly mixed, or absent.