WEATHER-DRIVEN EPIDEMIOLOGICAL MODELLING FOR PREDICTION OF BACTERIAL LEAF BLIGHT OF RICE IN THE STATE OF TAMIL NADU

Abstract

Bacterial Leaf Blight (BLB) incidence in rice is a major factor, delimiting rice yield. The pathogen, Xanthomonas oryzae pv oryzae is known to cause huge losses both during the initial stages and later stages of the crop. In order to manage the disease, a timely detection is necessary. BLB in rice is highly influenced by the prevailing weather factors. Many studies have reported, temperature, relative humidity, rainfall and wind speed to affect the disease occurrence and severity. In this study, we have studied the effect of weather factors on the incidence of BLB in Madurai district of Tamil Nadu from 2015-2024. We observed maximum temperature, minimum temperature, morning and evening relative humidity and rainfall as the major weather factors affecting disease develop to formulate a multiple linear regression (MLR) model for forecasting the occurrence of the disease. While carrying out the statistical analysis we identified multicollinearity problems between two weather factors and based on the previous literatures, we decided to remove one of the correlated weather variable to develop the MLR model. The developed model exhibited a high R2 of 0.917 and an adjusted R2 of 0.851. This indicated that the model could explain 91.7% of the variation observed in BLB disease incidence in the region. The developed regression model provides a preliminary framework for weather-based prediction of rice BLB incidence in the study region. Including the major weather factors in the development of a disease prediction model will facilitate early assessment of disease risk and hence, facilitate timely disease management decisions.