A Logistic Regression Model for Predicting Sclerotinia Stem Rot in Egyptian Clover (Trifolium alexandrinum L.)

Abstract

Background: Stem rot caused by Sclerotinia trifoliorum is the most damaging disease of Egyptian clover (popularly called as berseem), which is widely grown as a leguminous, winter season fodder crop in India. Stem rot management currently relies on fungicides which have negative effect on livestock and environmental health. In this study, in order to rationalize the fungicide use for stem rot management, a prediction model which assesses the high risk of stem rot ( greater than 20% incidence) was developed. Methods: The disease and weather data was collected from week-50 (second week of December) to week-14 (first week of April) during 2010-11 to 2019-20. The model was developed using logistic regression modeling approach. Result: The model included increasing weekly average temperature (between 8-25oC) and wind speed (between 1-7 km/hr) as key predictor variables. Goodness of fit statistics such as number of concordant pairs (83.8%), discordant pairs (16.1%), Somers’ D (0.68), Gamma (0.68) and Tau-a (0.33) indicates high accuracy of the model. The model had high area under receiver operating characteristic curve value of 0.84 during development and 0.82 on cross validation indicating that it will perform fairly well on an independent dataset. Thus, these macro climatic-weather variables can be used to predict high risk ( greater than 20% incidence) of stem rot, which ultimately will rationalize use of fungicides for stem rot management. This is the first model to predict Sclerotinia stem rot in Egyptian clover based on weather variables in India and probably around the world.