Nain, A.S.Rawat, Himanshu2018-04-162018-04-162016-07http://krishikosh.egranth.ac.in/handle/1/5810043663Soybean, an important kharif crop of Madhya Pradesh is grown on 5.5 million hectares. The crop is heavily supporting the economic conditions of the farmers as well as the state. Many agrobased industries are using soybean as raw product. However, due to its cultivation in rainfed ecosystem there is large year-to-year variability in productivity and production. In view of large variability, there is greater need to develop a system for timely and accurate estimation/prediction of productivity and production of soybean. Therefore, an attempt has been made in the present study to devise an approach for large (regional scale) area yield estimation. The approach includes i) zonation of study area (districts in the different zones) on the basis of interannular variability in soybean yield arising due to varying weather conditions, ii) calibration and validation of crop simulation model CROPGRO on farmer’s field conditions, iii) use of CROPGRO simulation model on zone for simulating response of soybean crop to ambient environmental conditions, iv) computation of yearto-year deviations in observed yields and simulation yields, v) relating observed yield deviations with simulation yield deviations for prediction of yield deviations, vi) estimation of technological trend yields, vii) incorporation of predicted deviations into trend yields for predicting zone level soybean yields and, viii) aggregation of zonal yield at regional scale using area weightage method. The present study was conducted in the Ujjain district for the calibration and validation of CROPGRO simulation model on farmer’s fields, while 22 districts of Western Madhya Pradesh were selected for the regional yield prediction for the period of 13 years (2001-2013) and yield forecast for two years (2014-16). Cultivar JS 335, which is grown over large area was selected for the model calibration and regional yield prediction. The soil of Ujjain and other districts of the Western MP is clay. The soil is black in colour and is widely known as Black Cotton soil or Regur soil. The calibrated and validated CROPGRO simulation model was used to simulate the response of soybean crop at zone level by applying zone level average conditions. The zonation of 22 districts yielded 3 clusters of districts on the basis of similarity in interannual yield deviations, which were mapped with the help of GIS software and were further divided into four zones (zone 1A, 1B, 2 and 3) based on geographical discontinuity. The zone level soybean yield prediction for a period of 13 years (2001-13) shows quite good agreement between observed soybean yield and predicted yield with RMSE ranging from 11.3 % to 17.3% and R2 value from 0.64 to 0.73. Similarly zone level soybean yields were also forecasted for two years (2014-15) by adopting same approach. The zone level predicted soybean yields were aggregated at region level by applying area weightage method and were compared with observed regional soybean yields. The RMSE value for predicted soybean yield at regional scale was found to be 11%, which is considerable low as compared to CV of observed yield and trend yields. Therefore, it can be concluded that zone based approach together with CROPGRO-simulation model can be used for regional soybean yield prediction and forecasting with quite high accuracy.ennullThesis