Daniel Nath1, Andrew Manu2, and Amy Norton1. (1) Iowa State Univ, 409 Steinbeck #10, Ames, IA 50014, (2) Iowa State Univ, 100 Osborn Dr, 1126A Agronomy Hall, Ames, IA 50011
Terrain analysis has been a powerful tool for predicting soil properties. This study was conducted to determine optimal grid resolution for predicting multiple soil properties in a low relief landscape. Multiple primary and secondary terrain attributes were created from a high accuracy Digital Elevation Model (DEM). Regression was then used to determine the grid resolution at which specific attributes best predict soil properties. The variance explained by the model ranged from 23% to 70% by changing the grid resolution alone. Prediction of some properties was not affected by grid resolution. The statistical significance of terrain attributes varied at different grid resolutions. Relief is an important contributor to processes that modify the landscape and must be considered when attempting to model those changes.
Back to Symposium--High Intensity Soil Surveys: II
Back to S05 Pedology
Back to The ASA-CSSA-SSSA International Annual Meetings (November 12-16, 2006)