The variability of crop yield in dryland production is primarily affected by the spatial distribution of plant-available water even for seemingly uniform fields. The most productive midwestern soils, which are loess caps over glacial till or outwash, can have a wide range of water-holding capacities in individual fields because of landscape processes and management. An inverse yield model was created as a robust method to quantify the spatial and temporal role of plant-available water on large agricultural fields to improve management options in precision agriculture. Plant-available water maps for a field were estimated from yield maps using inverse water-budget modeling based on measurements of solar radiation, temperature, precipitation, and vapor pressure deficit. The model presented in this paper was applied to 5 yr of corn ( Zea mays L.) yield-monitor data from a field in Waunakee, WI, having three soil mapping units, Plano silt loam (fine-silty, mixed, mesic Typic Argiudoll), St. Charles silt loam (fine-silty, mixed, mesic Typic Hapludalf), and Griswold loam (Fine-loamy, mixed, mesic Typic Argiudoll). The comparison of measured and inverse-modeled plant-available water suggests that the simple inverse yield model produces reasonable results in drier years with uncertainties of about 28 mm of plant-available water. The model helped to quantify the role of plant-available water in determining crop yield. Because of limited input requirements, the model shows promise as a practical tool for using precision farming to improve management decisions, and as a tool to obtain input for landscape-based models.
