PROCEDURE TO USE A CROP MODEL TO IDENTIFY WATER–STRESSED AREAS IN SOYBEAN FIELDS USING ON–FARM DATA

Abstract

In the past several years, researchers have developed procedures to use crop models to identify various yield–limiting factors within soybean fields. These procedures have relied upon extensive research–oriented databases, which are not practical to develop for a large number of farms. Limitations in on–farm data may limit the widespread use of crop models to evaluate on–farm variable rate management. However, one practical application may be to use models to identify areas within a field where water stress is the predominant yield–limiting factor. This procedure would allow targeting data collection and scouting efforts to other areas in the field. This study outlines a crop model–based screening procedure to identify areas in a soybean field where water stress is likely the predominant yield–limiting factor. The procedure is based on publicly available data, including a soil–type map and several years of daily weather data, as well as readily available on–farm data such as final yield maps. The procedure is demonstrated on a producer field near Mansfield, Illinois. Data were collected on two fields (North–DN and South–DS) from 1994 to 1999. Each field was divided into 72 grids 0.4–ha in size, and model inputs were developed from the available data. Three parameters (maximum potential rooting depth and two tile drainage parameters) of CROPGRO–Soybean model were calibrated to minimize error between predicted and measured yields in each grid over three years of final yield data. In the DS field, predicted yields in more than 62% of the grids were within .10% of the measured yield in each year (1995, 1997, 1999). In the DN field, predicted yields in more than 60% of the grids were within .10% of the measured yield in two (1994, 1996) of the three years of data that were used in the analysis. Results of simulations enable users to screen grids for problems unrelated to water stress.