Vol. 11, Issue 10 (2022)

The urgent need of crop modeling under changing climatic conditions and increasing abiotic as well as biotic new stress an study entitled as “Development of regression model to predict Bt cotton (Gossypium spp.) yield using meteorological variables for Parbhani and Nanded locations” has been conducted at Department of Agricultural Meteorology, Vasantarao Naik Marathwada Krishi Vidyapeeth, Parbhani during 2019-20 to analyze the stepwise regression techniques for meteorological variables with the yield and to develop the regression model to predict the yield with meteorological variable. Weekly weather data (24^th MW to 02^nd MW) of 11 weather variables over a span of 15 years period (2002-03 to 2016-17) along with kharif, season Bt cotton production data for Parbhani and Nanded locations. All the information for significance of stepwise regression and multiple regression model was presented. Statistical analysis carried out by using Microsoft office and SPSS22 copyright Inc., USA Based Application guide.

For development of regression model initially correlation was worked out between yield and weather parameters based on 15 years data, a positive and significant relationship between rainy days (0.544*) and Bt cotton yield, while negative and significant relationship between bright sunshine hour (-0.548*) and Bt cotton yield was found, while relationship of wind speed (-0.717**) was negative and highly significant with Bt cotton yield. At Nanded, rainy days (0.616*) and Bt cotton yield had a favorable and significant association. Also, there was a negative and significant association between minimum temperature (-0.586*) and Bt cotton yield. Stepwise regression models were derived which revealed that for 15 years data both developed models are good fit. The R² value is high, residual error and error percent are low revealed the best fit model to predict Bt cotton yield. As a result, model number 3 (R²=0.864) revealed the best fit model to predict Bt cotton yield, if we have a 15-year average data set for Parbhani. Maximum temperature, evening relative humidity and wind speed was found significant impact on Bt cotton yield at Parbhani. For Nanded, models 2(R²=0.604) and 3(R²=0.762) are suitable for predicting Bt cotton yield. Rainfall, rainy days and minimum temperature was found significant impact on Bt cotton yield at Nanded. Multiple regression models were also best fit to predict Bt cotton yield of 15 years data at Parbhani (R²= 0.961) and Nanded (R²= 0.819).