Abstract

Batches of grain are accepted or rejected based on average mycotoxin concentrations in a composite grain sample. Spatial analysis of mycotoxins in two grain bulks was performed to determine the spatial distribution of toxins, whether they were co-located and the proportions of grain over legislative limits. The 2D distribution of deoxynivalenol (DON) and ochratoxin A (OTA) in a truck load of wheat grain was analysed, as was the distribution of fumonisins (FB1 and FB2) in a 3D maize grain pile. The data had been previously analysed, but results here show that highly skewed data would need to be transformed to investigate spatial autocorrelation properly. In the truck of wheat grain, DON and OTA showed co-variation and, in contrast to previous studies, OTA showed spatial structure when converted to normal scores. Spatial analysis of the maize pile showed that FB1 and FB2 contamination levels were each highest near the outer face and base of the grain pile. Simulations for both grain bulks showed that, for average toxin concentrations close to legislative limits, the proportion of grain over the legislative limits can vary greatly and could be very small when toxin contamination is highly positively skewed. The implications of the results for management were considered. Post-harvest, strategically placed sensors could be used to monitor environmental conditions within the stored grain in real time and detect the first signs of spoilage allowing swift remediative action so less grain is wasted. Pre-harvest approaches for mycotoxin management are suggested as additional food waste reduction strategies. (C) 2021 IAgrE. Published by Elsevier Ltd. All rights reserved.