Evaluating risk of aflatoxin field contamination from climate change using new modules inside DSSAT:

Aflatoxins affect the health of close to 70 percent of the population of the world through contaminated food. Smallholder farmers in developing countries can be especially hard hit, since they consume a high proportion of what they produce without a clear knowledge of the level of contamination their harvest might have. Climate change can cause dramatic shifts in the level of contamination and the frequency of that high levels of aflatoxins are found in harvested foods, particularly maize and groundnuts. In this paper, we introduce new software that is able to estimate potential field concentrations of aflatoxins based on weather, and then apply the software to the question of how projected changes in climate will affect the occurrence of aflatoxins in six countries. The analysis is done at a very fine geographic resolution so that problem areas within countries are also identified. For rainfed groundnuts, baseline period calculations using the module show fairly high frequency of expected contamination levels above 4 ppb for Burkina Faso and Niger (39 and 56 percent), while Nigeria has a more modest estimate of 14 percent. However, factoring in climate change, we find great variation in projections. One of the five climate models used in the analysis projects a much wetter region which serves to drive down aflatoxin concentrations steeply. However, others have lower or even negative projections for changes in rainfall and coupled with temperature increases (large in some climate models), three of the five climate models project rising aflatoxin concentrations. The frequency of projected contamination levels above 4 ppb in rainfed maize are high in the baseline for Niger, at 43 percent, though Niger grows little maize. Burkina Faso, Nigeria, Guatemala, and Honduras all have more modest projections in the baseline (8, 9, 4, 10), while Nepal has just a trace above 0. Aflatoxin concentrations are projected to rise with climate change by all 5 models for Nepal, Guatemala, Honduras, and Nigeria, but only rise for 3 models for Niger and 4 of the 5 models for Burkina Faso. We use regressions with weather variables on projected aflatoxin concentrations levels above 4 ppb to better understand critical levels of rainfall and temperature that could trigger local crises with aflatoxins in on-farm consumption of harvested foods. At the end of the paper, we examine why aflatoxin concentrations in Nepal as reported by the modeling results appear low despite aflatoxins being a significant issue for the country.