Loading...
Theses
Browse
1 results
Search Results
ThesisItem Open Access Some statistical and soft computing models for crop yield and weather forecasting(G.B. Pant University of Agriculture and Technology, Pantnagar - 263145 (Uttarakhand), 2021-08) Bhatt, Neha; Shukla, A.K.Crop yield forecasting and weather forecasting are two important aspects for developing economy so that adequate planning exercise is undertaken for sustainable growth and overall development of the country. However ,such forecast studies need to be done on continuing basis and for different agro-climatic zones due to visible effect of changing climatic conditions and weather shifts at different locations and area . Therefore, the present study was undertaken for forecasting the yield of two major crops viz. Rice and Wheat and to forecast the eight weather parameters namely Maximum temperature, Minimum temperature, Humidity A.M, Humidity P.M, Rainfall, Sunshine Hrs, Wind speed and Evapotranspiration the data for the last 20 years (w.e.f 2000- to 2020) of yield and eight weather parameters were obtained from G.B. Pant University of Agriculture and Technology ,Pantnagar, District Udham Singh Nagar, Uttarakhand, India. Statistical and soft computing techniques namely ETS ,TBATS, ARIMA, MLR, FIS ,ANN and ARIMAX were used to develop the forecasting models using R and MATLAB softwares. Correlations between yields of Rice and Wheat with the weather parameters were also calculated. On the basis of the present study following conclusions were drawn: Rice Yield was positively significantly correlated with Minimum Temperature, Humidity P.M., Rainfall, Number of Rainy Days and Sunshine Hours during the cropping season . Wheat yield was positively significantly correlated with Maximum Temperature , Minimum Temperature, Rainfall, Number of Rainy Days and Sunshine Hours during the cropping season . ARIMA model was found to be the best fit monthly weather prediction model for Maximum Temperature, Minimum Temperature, Humidity A.M., Humidity P.M., Sunshine Hours, Wind Velocity and Evapotranspiration as compared to ETS ,TBATS and ARIMA models. ANN model was found to be the best fit daily weather prediction model for Maximum Temperature, Minimum Temperature, Humidity A.M., Humidity P.M., Sunshine Hours and Evapotranspiration as compared to ETS and ARIMA models. For Rice yield prediction on the basis of seasonal average of weather data, various models were developed using ANN, FIS and ARIMAX techniques and it was found that ARIMAX-RY-II( ARIMAX-RICE YIELD – Based on 5 weather parameters) was the best fit model for Rice yield prediction. For Wheat yield prediction on the basis of seasonal average of weather data, various models were developed using ANN, FIS and ARIMAX techniques and it was found that FIS-WY (FIS- WHEAT YIELD) was the best fit model for Wheat yield prediction. For predicting the Rice yield on the basis of monthly weather data during the rice crop season , different MLR models were developed on the basis of different combinations of weather parameters and the best fit MLR-RY-A model with R 2 = 0.62 was selected It was found that November Maximum Temperature, September Minimum Temperature, July Humidity A.M., August Humidity A.M., August Sunshine Hours , November Sunshine Hours, July Maximum Temperature, October Humidity A.M. , November Humidity A.M. , September Rainfall and September Evapotranspiration could be used as best predictors for estimating the rice yield. For predicting the Wheat yield on the basis of monthly weather data during the Wheat crop season , different MLR models were developed on the basis of different combinations of weather parameters and the best fit MLRWY- A model with R 2 = 0.45 was selected . It was found that October Maximum Temperature March Maximum Temperature, March Humidity A.M., November Humidity P.M., April Sunshine Hours and April Wind Speed could be used as best predictors for estimating the wheat yield.
