Forecasting hourly risks of pear rust development using real-time weather interpolations in web-based information system

A forecasting model of pear rust disease was implemented in the web-based plant disease forecasting system, www.epilove.com, which is based on the real-time weather observation data collected from 33 automated weather stations in Gyeonggi province, Korea. The model predicts the initiation date of teliospore release of Gymnosporangium haraeanum from trees of Juniperus chinensis using the average of daily mean air temperatures in March. Then the infection risk of pear rust is determined and categorized into four levels as ‘Zero', ‘Low', ‘Intermediate', and ‘High', depending on rainfall hours during the period of 10 days before and 15 days after the date of teliospore release. The performance of disease forecasts was evaluated in a pear orchard in 2007-2009 by applying Hexaconazole 2% when the “High” risk level was first forecast. The efficacy of disease control was 89.9-100% in 2007-2009, indicating that the disease forecasting system would be helpful for pear growers in reducing fungicide applications without decrease in disease control efficacy.