| The world's agricultural soil resources continue to degrade due to rates of soil erosion and organic matter oxidation exceeding those found in native systems. This degradation is generally the result of intensive soil tillage, resulting in decreased amounts of soil residue cover, high runoff rates, and decreased soil organic matter. Degradation pressure is likely to increase in the future due to harvesting of plant residues for bioenergy and increased peak intensity of rainstorms due to climate change, increasing soil detachment and transport.; Developing effective solutions to these problems require the use of models to estimate the extent and severity of these impacts across the landscape. Once impacts are known, appropriate management practices can be selected to counteract these effects. However, implementation of these models will require data on land management practices at the field level, a resolution not currently possible with any routinely collected dataset. This requires the development of new methods to efficiently gather a large amount of quality data.; To this end, methodologies were developed for automatically determining management unit boundaries, corresponding crop and residue cover, and estimating surface roughness. Testing was conducted on agricultural fields in Boone, Hamilton, and Story counties in central Iowa with management unit area estimated to within 10% of the mean and management unit centroids located within 30 m on average and 165 m at a maximum. Crop cover was correctly identified better than 95% of the time. Errors were within USGS standard where applicable. Methodologies for determining residue cover were also determined, with an RMSE of 0.08 before green vegetation emergence and 0.11 after emergence. These errors compare favorably with the +/- 0.10 error in determining residue cover using published measurement methods. Radar estimation of surface roughness was not successfully demonstrated, as no difference was observed in radar backscatter between different surface roughness values. |
