Prediction model for bacterial blight (Xanthomonas axonopodis pv. malvacearum) of cotton for central India

Important meteorological variables from weather data were identified, correlated and included in a multiple stepwise linear regression model for forecasting of bacterial blight disease of cotton after 45 days of sowing. Detailed analysis of the data for the four years (1998–2002) collected at the Cotton Research Station, Surat (Gujrat) showed that maximum and minimum temperature and relative humidity (morning) are positively co-related to disease intensity. In present paper three different linear models based on (a) current week weather parameters, (b) previous week weather and (c) previous week disease intensity in combination with weather variables were developed and compared to improve disease forecasting of bacterial blight and validated with field data of subsequent years and recent 2004–05 data. Using model ‘C’ it is now possible to predict the disease intensity one week in advance, providing sufficient time for contingency plan with plant protection inputs to restrict and manage the disease growth.