Submitted to: Agronomy Abstracts
Publication Type: Abstract Only
Publication Acceptance Date: 8/1/2008
Publication Date: 10/1/2008
Citation: Abdu, H., Robinson, D., Seyfried, M.S., Jones, S. 2008. Using Geophysical Images of a Watershed Subsurface to Predict Soil Textural Properties. Agronomy Abstracts, 661-3, Joint Annual Meeting of ASA-CSSA-SSSA, and GCAGS, Houston, Texas, October 2008. Interpretive Summary:
Technical Abstract: Subsurface architecture, in particular changes in soil type across the landscape, is an important control on the hydrological and ecological function of a watershed. Traditional methods of mapping soils involving subjective assignment of soil boundaries are inadequate for studies requiring a quantitative assessment of the landscape and its subsurface connectivity and storage capacity. Geophysical methods such as electromagnetic induction (EMI) provide the possibility of obtaining high resolution images across a landscape to identify subtle changes in subsurface architecture. In this work we show how EMI can be used to image the subsurface of a ~38 ha watershed at the Reynolds Creek Experimental Watershed near Boise,Idaho. We also collected soil samples for textural and water content analysis. We present an imaging approach using kriging to interpolate, and Sequential Gaussian Simulation (SGSIM) to estimate the uncertainty in the maps. We then use the EMI maps as exhaustive secondary information in order to predict clay percentage and water content at unsampled locations using different multi-variable kriging methods. Using cross-validation techniques, we compare the accuracy of kriging methods such as ordinary kriging, regression kriging, and ordinary co-kriging in predicting the property of interest. Such spatially-detailed soil property maps can be useful in studying the discrepancy between measured hydrographs and model predictions, where average values used for soil moisture and soil hydraulic parameters can lead to large deviations. Accounting for the spatial variability of hydraulic properties is important in understanding runoff production, and the role of organizational patterns of soil moisture on catchment runoff.