Simulating Soil Nitrogen Availability and Loss in Illinois Corn Production Systems.

Maintaining the balance between providing enough nitrogen (N) to crops while avoiding environmental losses has proven difficult despite significant field research and modeling efforts. In the U.S. Midwest, an important question often facing growers is how much soil N is available to support crop growth following different N application times and forms, and different early-season weather. With the long-term goal of developing an online decision-support tool for estimating in-season soil N status to help make informed N management decisions, the aim of this study was to develop a real-time modeling framework for simulating soil N availability. The DSSAT model was calibrated using six Illinois field experiments consisting of different combinations of N fertilizer source and timing. Soil inorganic N concentrations were determined at 0–30 cm and 30–60 cm depths at 10- to 14-day intervals from before planting through pollination. The calibrated model was applied to 5406 points for the 2015 and 2016 corn growing seasons. The current model is reasonably predictive (R2 = 0.76; nRMSE = 15.6%) for estimating soil N status in the corn growing season at three field locations in Illinois. This presentation will discuss possible sources of model error and bias in the robust model validation process against multiple field observations. In addition, the process of integrating DSSAT with geospatial soil and real-time weather datasets along with grower-entered management information in the form of an online App will be discussed.