Soybean (Glycine max) yield prediction from current and historical weather data using CROPGRO model

Abstract

A technique has been developed for obtaining reliable estimates of soybean [Glycine max (L.) Merr.] yield during the growing season by combining historical and present weather with CROPGRO Soybean model. The procedure is to place the reference year's daily weather into the model up to the time the yield prediction is to be made and sequences of historical data (one sequence per year) after that time until the end of growing season to give yield estimates. The change in predicted yield in response to current and forecast weather as the growing season progresses has been discussed for 2 contrasting years 1991 and 1995 at Jabalpur (Madhya Pradesh) with low and high yield respectively. Yield in 1991 has been determined before the end of the grain filling stage (20-25 days in advance before physiological maturity), whereas in 1995 it is at end of grain filling stage (10-12 days before physiological maturity). Weather forecast is relatively more valuable during reproductive stage in 1991 and 1995. The longer the forecast period, higher its value and earlier in the growing season its relative importance starts.