Modelling wheat yield based on weather parameters at its phenological stages
Loading...
Date
2022-08
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
G.B. Pant University of Agriculture and Technology, Pantnagar, District Udham Singh Nagar, Uttarakhand. PIN - 263145
Abstract
Crop yield prediction is an important aspect for a developing country like India in such a way that it helps decision makers, frame policies and strategies related to distribution, marketing and storage of agricultural products which ultimately lead to the sustainable growth and development of the country. The agricultural sector is severely affected by short term weather fluctuations and long term climate variations. Weather variability during important growth stages of a crop can result in uncontrolled crop yield variations. Wheat (Triticum aestivum) is one of the most widely grown cereal crops and an important staple food next to rice in India. The present study attempted to develop wheat yield prediction models for Udham Singh Nagar district of Uttarakhand state based on weather parameters during different growth stages of wheat. Maximum temperature, Minimum temperature, Relative Humidity A.M, Relative Humidity P.M, Total rainfall, Sunshine hours, Wind velocity and Evapotranspiration were the weather parameters considered for the study. Statistical and soft computing techniques namely Multiple Linear Regression, Artificial Neural Network and Ridge Regression were employed in the study using R software and SPSS software package. Correlations between rabi wheat yield and weather parameters during different growth stages of wheat were also analysed.
The following conclusions were drawn from the study:
•Correlation between rabi wheat yield and maximum temperature during the Dough stage was found to be positive but there was a negative correlation in the case of minimum temperature during the Dough stage.
•Rabi wheat yield was found to be negatively correlated with minimum temperature during the milking stage whereas, rabi wheat yield was found to be positively correlated with morning relative humidity during the tillering stage and evening relative humidity during the Crown Root Initiation stage.
•MLR-W (MLR model developed by using weather parameters at different growth stages of wheat used directly as predictors) model could perform better than the other two models developed using MLR method
•ANN-WI (ANN model developed by using weather indices as predictors) model could perform better than the other two models developed using ANN.
•RR-D (Ridge Regression model developed by using deviations of weather parameters from optimum value during important growth stages of wheat as predictors) model could perform better than the other two models developed using Ridge Regression
•Evaluation based on statistical indices and error percentage during validation revealed that, ANN-WI (ANN model developed using unweighted and weighted weather indices as predictors, R2 = 0.96) out performed MLR-W model and RR-D model.
•Crop yield prediction models based on weather parameters during important growth stages of the crop could provide reliable results.