Submitted to: Corn Dry Milling Conference Proceedings
Publication Type: Abstract only
Publication Acceptance Date: 5/30/2003
Publication Date: 5/30/2003
Citation: Dowd, P.F. 2003. Present status of a management-oriented computer program for predicting mycotoxin levels in preharvest Midwest corn [abstract]. Corn Dry Milling Conference Proceedings, p. 13. Interpretive Summary:
Technical Abstract: The feasibility of an integrated management plan for mycotoxins in Midwest corn is presently being evaluated in a joint study involving farmers, industry, farm service, university extension, and ARS. An important component of this plan is a computer program targeted toward farmers forecast potential inoculum levels at silking, and potential mycotoxin levels prior to harvest based on weather, hybrid, insect, and soil data. The initial program was based on data intensively collected from hybrids from three major seed companies planted in central Illinois in the 1990's, and included fumonisin values determined by HPLC analysis. To develop the aflatoxin model for the program, aflatoxin values were taken from the literature for 1983 and 1988 and supplemented with data collected in the 1990's (nearly all negative for aflatoxin). The generalized version of the program predicted fumonisin values that were generally close to actual field values (ranging from ca. 0.1 to 8.0 ppm) during the period in the 1990's from which data was collected on which the program was based. Since then, the program has been validated in studies involving several commercial fields in 2000-2002. In 2000, a general fumonisin predicted value of ca. 1.5 ppm was close to actual values ranging 0.8-1.6 ppm for hybrids previously examined. In 2001, a general fumonisin predicted value of ca. 0.5 ppm was close to actual values ranging from 0.2-0.6 ppm in all fields but one (which was lower) that were relatively free of insect damage. Also in 2001, later planted corn that sustained more insect damage had values of 2.2, 2.2, and 4.2 ppm, which were close to the predicted value of ca. 2.0 ppm. In 2002, actual fumonisin values ranged from 0.7 to 15.5 ppm. Predicted values were overall about 1/2 to 1/4 lower than actual values. The program has now been adjusted so that predicted values for 2002 more closely follow trends found in 2002 (without significantly changing predictions for other years), which again included a wide range of hybrids, planting dates, and insect damage. Because the data set used to develop the aflatoxin model was not as extensive as that used to develop the fumonisin model, the accuracy of aflatoxin values predicted has been less precise than for fumonisin values. However, the program correctly predicted no aflatoxin in the years from early/mid to late 1990's, none in 2000, and low levels of aflatoxin in the study area in 2001 and 2002. Using National Weather Service data available from areas in the Midwest that had high levels of aflatoxin in 2002, the program predicted aflatoxin levels near 100 ppm. Obtaining weather data in a timely manner continues to be a challenge for real time prediction. Conversion of the original DOS version of the program to Windows is nearing completion thanks to efforts of collaborators at Illinois Central College. Some sample screens will be shown.