Mining Effect of Temperature and Rainfall to Develop an Empirical Model for Wheat Yield Prediction

Abstract

Background: Crop yield is affected by several agronomic factors such as soil type and date of sowing, and meteorological factors such as temperature and rainfall. While the agronomic factors are responsible for inter-region variations in yield, the year-wise yield variation in a particular region may be attributed to meteorological factors. Various Data Mining Techniques can be applied to analyse the effect of these factors on crop yield. Objective: To develop a model for prediction of Block-wise average wheat yield in Patiala district of Punjab, India. Method: Sampling is used for the collection of the yield data, and the data concerning temperature and rainfall is obtained from Indian Meteorological Department, Pune. The data is then pre-processed and analysed to study the effect of phase-wise average temperature and total rainfall on the wheat yield. The factors that are found to significantly affect yield are used for building a model for yield prediction. Results: It is found that the average temperature and the total rainfall for the whole wheat growing season are not much of help in explaining the variations in yearly wheat yield. The temperature and rainfall have different effects at different stages of plant growth and the yield is affected accordingly. It is inferred that that the average temperature and the total rainfall during the vegetative phase, and the grain development and ripening phase are the most important parameters for prediction of wheat yield. Conclusion: The stepwise selection mechanism is used to choose the variables whose inclusion explains the maximum variance in yield. The model is evaluated based on different parameters and is found to explain 95.6% of the yearly variations in yield.