Pre-harvest forecasting of rice (Oryza sativa L.) yield based on weather variables and technological trend

To develop the pre harvest forecasting model of rice yield using weather variables and technological advances, 33 years yield data of Kheda district, Gujarat from 1967-68 to 2001-02 were collected. The weekly averages of weather variables viz., bright sunshine hours (BSS), rainfall (RF), maximum (MaxT) and minimum temperature (MinT) and morning relative humidity (MRH) from 23(rd) to 42(nd) standard meteorological weeks (MSW) of the respective years, were considered in the study. The week-wise, crop stage-wise and generated weather variables (weighted) criteria were used. Time trend was also included as independent variable in all the criteria. To provide pre-harvest forecasts different week-intervals were considered. Step wise regression technique was employed for all forecasting models. Among these the models, the model based on week-wise criteria using original weather variables of 16 weeks provided a reliable pre-harvest forecasting of rice in Kheda district. The forecasting can be done four weeks before expected harvest (i.e. 3(rd) week of September/ at the end of 38 th MSW). The average absolute mean forecast errors of the selected model for the subsequent three years were found to be 10.97 percent. The pre-harvest proposed forecast model accounted 75.4% variation in rice yield.