FORECASTING LATE BLIGHT OF POTATO IN PLAINS OF WEST BENGAL USING JHULSACAST MODEL

Abstract

Late blight is the single most important impediment in potato cultivation throughout the world. The management of this devastating disease depends on regular application of fungicides at short intervals throughout the growing season resulting in health and environment hazards. However, timely forecasting of the disease may result in lesser application of fungicides with effective management of the disease. Out of fourteen forecasting models tested for predicting late blight in plains of West Bengal, none of the model (except Wallin, Bhattacharyya’s II forecast and Dutch rules) could predict late blight. JHULSACAST model was fine-tuned by making modifications in input parameters (relative humidity, temperature duration and amount of rainfall) by utilising weather data for 7 years collected at Bidhan Chandra Krishi Vishavidhyalaya, Kalyani, West Bengal, India. The modified JHULSACAST model could predict late blight within 14 days with accuracy of 100% if 7-day moving relative humidity (≥90%) period ≥105 hours coupled with 7-day moving congenial temperature (7.2 to 26.6^oC) ≥150 hours prevail for 7-consecutive days OR if 5-day moving relative humidity (≥90%) period ≥65 hours, 5-day moving congenial temperature (7.2 to 26.6^oC) ≥105 hours for 5 consecutive days and sum of two consecutive days rainfall (≥2.5 mm) prevail. These models were validated and predicted late blight successfully for two years.