Predicting Soil Health and Function from Remote-Sensed Evapotranspiration and Terrain Attributes in the Glaciated Plains of Eastern North Dakota.

The benchmark Barnes soil series is an extensive upland Hapludoll of the northern Great Plains that is both economically and ecologically vital to the region. Destabilizing effects of tillage coupled with wind and water erosion have degraded Barnes soil quality, but with unknown extent or severity. In order to better document this phenomenon, the landscape requires an efficient method to rapidly evaluate soil health status. Consequently, novel methods are necessary to foster improved soil map resolution and ultimately assess the potential for irreversible state change. Prediction of soil properties at the landscape scale is possible for physiographic regions with similar and consistent soil-forming inputs and can be modeled with digital terrain and climatic derivatives from high resolution remote-sensed data. Therefore, soil property spatial relationships are currently being investigated to revise map unit composition of upland Barnes landscapes affected by accelerated soil erosion. Soil samples from three regionally extensive map units, stratified by evapotranspiration values, were collected to 50 cm and analyzed for chemical, morphologic, and physical properties germane to soil edaphic function. One meter digital elevation models were processed to calculate relative position and wetness terrain attributes for each sample location. Regression models for soil property target variables were optimized and subsequently fitted to a spatial prediction function to generate surfaces for varying soil health properties to 50 cm. These results may prove useful for regional soil map unit disaggregation efforts, precision agriculture applications, property value assessment, and improved land management.