VIEW ARTICLE

Ecology and Epidemiology

Field Testing a Computerized Forecasting System for Rice Blast Disease. Choong- Hoe Kim, Graduate assistant, Department of Plant Pathology and Crop Physiology, Louisiana Agricultural Experiment Station, Agricultural Center, Louisiana State University, Baton Rouge 70803; D. R. MacKenzie, and M. C. Rush. Professors, Department of Plant Pathology and Crop Physiology, Louisiana Agricultural Experiment Station, Agricultural Center, Louisiana State University, Baton Rouge 70803. Phytopathology 78:931-934. Accepted for publication 3 February 1988. Copyright 1988 The American Phytopathological Society. DOI: 10.1094/Phyto-78-931.

A computer program written to predict blast occurrence based on microclimatic events was tested as an on-site microcomputer in upland and flooded field plots in 1984 and 1985. Two battery-operated microcomputer units continuously monitored air temperature, leaf wetness, and relative humidity and interpreted the microclimate information in relation to rice blast development by calculating daily values (0 to 8) of blast units of severity (BUS). Accumulated daily BUS values were highly correlated with blast development on the two rice cultivars Brazos and M-201 grown under upland conditions. Use of cumulative BUS values to predict the logit of disease proportions gave average coefficients of determination (R²) of 71 to 91%, depending on cultivar and year, compared to 61 to 79% when days were used as a predictor of logit disease severity. BUS predicted logit disease severity less accurately with Brazos, which showed field resistance at mid-season, than with M-201. Cumulative BUS values in flooded plots were higher than those in the upland plots. No significant correlation was observed between cumulative BUS values and logit disease under flooded conditions because both cultivars were resistant to blast when flooded.