Prediction model for assessing powdery mildew disease in common Oat (Avena sativa L.)

Abstract

Powdery mildew (caused by Blumeria graminis DC. f. sp. avenae Em. Marchal) is the most important disease of common oat (Avena sativa L.) in cooler and humid regions of the world including India. In spite of this, no prediction model for assessing the high risk (>30% severity) of powdery mildew in common oat is available. In the present study, a logistic regression model which assesses the high risk of powdery mildew in common oat was developed using weather and disease data collected from 15 years (2004–05 to 2018–19) observations in a monitoring experiment conducted at Palampur, India. The model incorporated increasing weekly average temperature (between 11.5 and 21.9 °C) coupled with decreasing relative humidity (between 40 and 60%) and sunshine (between 5.4 and 8.7 h) as key predictors for high (>30% severity) powdery mildew severity. The model was validated for its accuracy using cross validation technique with area under receiver operating characteristic curve value (AUC) of 0.89 during development and 0.87 on cross validation. Since the model depends on weekly values of commonly available macro-climatic weather variables, the fungicide spraying can be done in advance which will help in reducing the losses caused by powdery mildew of common oat. To our knowledge, this is the first model for predicting powdery mildew disease of common oat in India and probably around the world.