The Development and Validation of A Simple Yield Loss Prediction Model for Soybean Rust.

Disease control measures are economically viable when yield loss is sufficient to justify the cost of the control measures. Models that predict expected yield loss can serve as useful management decision aide tools. The objective of the current study was to develop and validate a simple yield loss prediction model for soybean rust, based on the mechanism by which the disease reduces soybean yield. In a two year study conducted in Brazil, SBR-induced yield loss was found to be due to: (i) accelerated leaf drop, (ii) reduced green leaf area (GLA) and (iii) reduced photosynthetic capacity of GLA. The reduced photosynthetic capacity of GLA has been quantified in controlled and field experiments. The three factors have been integrated into an effective leaf area duration (ELAD) value. A strong correlative relationship between ELAD and yield has been developed using data from Brazil (R = 0.90). Next, independent studies were conducted to validate the accuracy of model predictions over a range of environments, cultivar maturities and row widths, in the U.S.  Trials were planted in Quincy, FL in 2007 and in Tifton, GA in 2008.  In 2007 the trial included determinate and indeterminate MG 5 cultivars planted in 15 and 30 inch rows.  In 2008 MG 7 and MG 8 cultivars were planted in 36 inch rows.  Phenology, leaf area, disease severity and yield were measured.  None of the trials had severe SBR epidemics resulting in rather limited yield losses. The difference in observed versus predicted data shall be discussed.