Page Banner

United States Department of Agriculture

Agricultural Research Service


item Pachepsky, Yakov
item Pierson, Frederick - Fred
item Spaeth, Stephen
item Weltz, Mark

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 6/1/2004
Publication Date: 9/25/2004
Citation: Pachepsky, Y.A., Pierson Jr, F.B., Spaeth, S.C., Weltz, M.A. 2004. Using regression trees to estimate surface water runoff and soil erosion for rangelands. [Meeting Abstract]. European Conference on Machine Learning, September 27, 2004, Bled, Slovinai. Paper #88, CDROM.

Interpretive Summary:

Technical Abstract: Estimates of surface runoff and soil erosion are needed for both rangeland management and rangeland productivity evaluation. Such estimates are made using various models ranging from purely empirical to mechanistic. Our hypothesis was that the direct use of more easily obtainable available data in machine-learning predictive tools may be a viable option for large scale estimates of erosion and runoff. Regression trees were selected because of their ability to discover relationship structures specific for subsets of the whole database, transparency of results, and ability to select the most influential input variables. Data from 442 erosion events from 26 locations in the Western United States were used. Regression trees were developed (a) with input variables that could be obtained from public sources and (b) with input variables that had to be measured on-site. The jackknife cross-validation was used to trim the trees. About 60 % of the variability in runoff could be explained using soil basic properties. Adding information about on-site measured ground cover and surface roughness helped to explain about 70% of the variability. Predictions of the sediment yield were worse as only 40% of variation could be explained. Regression trees were easy to interpret, and grouping had a clear physical meaning in most cases.

Last Modified: 10/19/2017
Footer Content Back to Top of Page