Automated determination of management units for precision conservation

Abstract

The use of models to target conservation efforts, or “precision conservation,” has become a research interest due to its ability to evaluate different conservation practices in multiple areas and determine where conservation expenditures will provide the best return. To determine if automation of the data collection process is possible, the intersection of USDA common land unit maps, which define the boundaries of agricultural lands, with vegetative cover maps, represented by the USDA cropland data layer, was analyzed for accuracy on approximately 80 central Iowa fields over two years. The resulting dataset successfully determined mean field centroid within 3 pixels, area with an r² of 0.99, crop cover with 95% accuracy, and when compiled over a period of years, crop rotations. The value of the resulting map can also be extended with the use of other technologies to determine field average of properties such as residue cover, tillage, yield, and biomass production. The dataset can also serve as an input for numerous precision conservation models and enable the ability to create maps of other field level processes or properties, such as residue cover, tillage, or fertilization, by defining the boundaries over which management uniformly occurs.