Widely used soil maps tend to be old; for example, the mean age of Iowa’s soil maps are 30 years old and made with 1920’s technology. They are not accurate enough for providing the next generation of better land management. I am interested in improving soil maps using machine learning tools to provide new levels of accuracy and detail.
In my M.S. program, I evaluated soil change using advanced statistical techniques. Now I am incorporating my background with spatial analysis to better analyze and predict soil properties. Based on my experience, I am convinced that incorporating terrain attributes in our calculations provides the information we need to reach our goals of more accurate and detailed soil maps.
Iowa has some of the most productive soils in the globe; over 90 percent of its land is devoted to agriculture. Especially here, soil maps can give us a comprehensive understanding of the soil system that would lead to improved soil nutrient management, crop management, and soil and water conversation.