EVALUATION OF THE GROUNDNUT MODEL PNUTGRO FOR CROP RESPONSE TO WATER AVAILABILITY, SOWING DATES, AND SEASONS

Field experiments were conducted during the period from 1987 to 1992 a t four locations in India to collect data to test and validate the gro undnut (Arachis hypogaea L.) model PNUTGRO for its capability to predi ct phenology, growth, and yield. Groundnut (cv. Robut 33-1) was grown during the rainy and post-rainy seasons at these sites under various m anagement practices such as sowing dates and differential irrigation. Using the data sets from several years, the model was calibrated for g enetic coefficients of cvs. Robut 33-1 and TMV 2 determining their phe nology and growth, as well as for soil physical parameters influencing the soil water balance. The model was validated for cv. Robut 33-1 ag ainst independent data sets obtained from field experiments conducted during the later years. The model predicted the occurrence of flowerin g and podding within +/- 5 days of observed values at locations where growth stages were recorded most frequently. Predictions of growth sta ges beyond podding were less accurate because of difficulties, associa ted with the indeterminate nature of the crop, to record growth stages after pod growth has started in the soil. Changes in vegetative growt h stages, total dry matter accumulation, growth of pods and seeds, and soil moisture were predicted accurately by the model. Predicted pod y ields were significantly correlated (r2 = 0.90) with observed yields. These results indicate that under biotic stress-free situations, the m odel PNUTGRO can be used to predict groundnut yields in different envi ronments as determined by season, sowing date, and moisture regimes.