Modeling the Effects of Cloudy Weather on Regional Epidemics of Soybean Rust.

Abstract

This study simulated daily development of soybean rust (SBR), caused by Phakopsora pachyrhizi, based on cloud cover conditions. Results from a previous study that determined the relationship between shading and apparent infection rates were applied in this study to simulate SBR progress on a regional scale using a semi-empirical logistic model parameterized according to the observed cloud cover conditions. Depending on local weather data availability, cloudy days were assumed to be either (i) the days with less than 2 h of full sun or (ii) the days with complete cloud cover as measured by three daily observations. Estimated disease progress and final estimates of epidemic intensity were verified by 30 reports of seasonal disease progress in 11 regions of Brazil and South Africa from 2002 to 2007. The model predicted final disease severity and the observed final severity fall into a linear relationship with correlation coefficient r = 0.96 and a slope close to 1. Severe SBR epidemics occurred when 19.5 or more cloudy days were recorded during the period from initial disease detection to the date of final disease assessment near the end of a growing season in Brazil and South Africa. Mild epidemics were observed with less than eight cloudy days in a season.