Submitted to: International IUPAC Symposium on Mycotoxins and Phycotoxins
Publication Type: Abstract Only
Publication Acceptance Date: May 21, 2004
Publication Date: May 21, 2004
Citation: Dowd, P.F. 2004. Reliability of a computer program for predicting mycotoxin levels in Midwestern U.S.A. maize. Proceedings of the International IUPAC Symposium on Mycotoxins and Phycotoxins. Abstract D-31. p. 102. Technical Abstract: Management of preharvest mycotoxin problems is dependent on the ability to detect conditions or factors that contribute to the problem. Intensive data collection over the period from 1992-1999 allowed for the development of a computer program that appears to have use in predicting the presence of causative fungi at silking, and potential levels of mycotoxins several weeks before harvest. Once the program was developed, it was validated on a few to several commercial fields planted with different hybrids from 2000-2003. Weather data, insect data, fungal presence, and mycotoxin levels were recorded, and predicted mycotoxin levels were compared with actual mean values from fields. The program was adjusted as necessary when predictions differed greatly from actual values. In 2000, predicted values of ca. 2.0 ppm of fumonisin were relatively close to actual values. In 2001, predicted values of ca. 0.5 ppm for fumonisin for ears with low insect damage and 2.0 for ears with high insect damage were close to actual values in most cases. In 2002, predicted values of fumonisins were relatively low for most fields, but adjustments yielded values that were relatively close to actual values for both low and high insect damage areas of ca. 1 to 15 ppm. The mycotoxin analysis for 2003 is pending. The program also predicted aflatoxin problems in areas of the Midwestern U.S.A. that did have aflatoxin problems in 2002. Although the program has done a relatively good job for fumonisin, the variation of susceptibility between hybrids in some cases has caused "outlier" problems. As part of the conversion to a Windows version, an additional module was added that allows for adjustment for hybrid outliers based on current data availability. The Windows version also includes extensive visual help sections and storage of input data so that it can be retrieved and utilized in adjoining fields. The program is being implemented as part of a comprehensive mycotoxin management program for Midwestern U.S. maize.