Weather based yield crop forecasting: Agro-Physics Approach
Dr. Vinay kumar
Department of physics, Janta Vedic College, Baraut, Baghpat, U.P., India
The present paper deals with use of non-linear regression analysis for developing Wheat yield forecast model for Jhansi (India). Weather indices based regression models were developed using weather indices as independent variables while character under study such as crop yield was used as dependent variable for Wheat Crop i.e. two step non-linear forecast model. Technique of forecasting using non-linear approach and using weather indices will enrich the knowledge in developing customized models on forecasting for different types of crops and for different locations. For that, trend analysis has been done through linear and non-linear approaches. In which for each weather variable two indices have been developed, one as simple total of values of weather parameter in different weeks and the other one as weighted total, weights being correlation coefficients between detrended yield and weather variable in respective weeks. Weather indices based regression models were developed using weather indices as independent variables while detrended yield (residuals) was considered as dependent variable. Time series yield data of 40 years (1970-2010) and weather data for the year 1970-71 to 2009-10 have been utilized. The models have been used to forecast yield in the subsequent three years 2008-09 to 2009-10 (which were not included in model development). The approach provided reliable yield forecast about two months before harvest.
Keywords: Forecasting, Nonlinear regression model, Weather Indices Approach, Weather variables.
