Prediction model for gray leaf spot disease of fodder Sorghum

Abstract

Gray leaf spot (caused by Cercospora sorghi Ellis & Everh.) is one of the most destructive foliar disease in fodder sorghum. Prediction model for gray leaf spot disease in fodder sorghum was developed based on 10-year (2010–2019) disease and weather data. It can occur during week-29 (third week of July) to week-43 (last week of October) in a year. Gray leaf spot had very strong, significantly negative correlation (r) with weekly average temperature (r = − 0.77), while moderate correlation was found with weekly average relative humidity (r = − 0.41), rainy days (r = − 0.44), sunshine hours (r = 0.35) and wind speed (r = − 0.36). Rainfall (r = − 0.29) and evaporation (r = − 0.23) had low correlation with gray leaf spot severity. Prediction model was developed using multiple regression coupled with goodness of fit statistics using R2 and Akaike information criteria (AIC). The prediction model included weekly average temperature, relative humidity and rainy days as predictor variables with R2 and AIC value of 0.70 and 819.03 respectively. The model was validated using leave one out cross validation (LOOCV) strategy and difference between observed residual root mean square error (RMSE) and as predicted using LOOCV was taken as indicator of model accuracy. Residual RMSE was 12.05 while that predicted through LOOCV was 12.46, indicating that model will perform fairly well on an independent data set. This prediction model can be used for efficient prediction to manage gray leaf spot disease of fodder sorghum.