3D ecosystem modeling of aeration and fumigation in Australian grain silos to improve efficacy against insects

Abstract

With continued population growth, more food production will be required with lower resource inputs. A significant drain on resources is post-harvest loss due to insects, which results in loss of product, quality and market access, and increased grain spoilage. Aeration and fumigation are key tools to control insect growth in stored grains and grain foods. The implementation of aeration strategies in Australia is made difficult by the warm subtropical climate, meanwhile the success of fumigation is being threatened by the spread of insect resistance to the fumigant phosphine. This dissertation project seeks to improve the understanding of aeration and fumigation by modifying the Maier-Lawrence (M-L) 3D ecosystem model by adding insect growth equations and quantifying fumigant loss from sealed bulk grain silos. The improved model was used to examine aeration under Australian conditions, validate its capability to accurately describe fumigant concentrations during silo fumigation, determine the extent to which fumigations are influenced by operational variables and environmental conditions, and validate its capability to describe fumigant concentrations post-fumigation in order to investigate the time needed to clear a grain storage silo of fumigant in order to assure worker safety. Six aeration strategies were evaluated under Australian conditions. Of these strategies, two were found to be effective in lowering temperatures, i.e., fans were turned on when ambient temperature was less than 20°C, and less than internal grain temperature. The strategy based on temperature differential was the most effective because it cooled the interior of the grain mass with the least amount of energy (using the fewest fan run hours, and reaching 15°C an average of 11 days after than the fastest strategy). Using a 0°C temperature differential was the most effective strategy in terms of reducing insect growth. The expanded (M-L-P) model was then validated based on experimental fumigant concentrations. The model was effective in reproducing average fumigant concentrations and fumigant trends vertically through the grain mass, but was not able to reproduce lateral fumigant variations. Verifications of the model with two different periods of phosphine release (i.e., 24h and 30h) were tested. Based on a 24h phosphine release period the predicted Ct-product differed from the experimental value by 0.9%. A 30h release period predicted a Ct-product that differed by 4.3% from the experimental value but it was more accurate during the times of highest concentration. Increases in leakage reduced fumigant concentrations, but the size of the effect decreased as leakage grew. Increasing temperature and wind speed in the model led to decreased phosphine concentrations, with temperature changes having a more significant impact overall than wind speed changes for the conditions investigated. Decreasing silo surface area to volume ratio dampened the impact of changing weather conditions on phosphine concentrations. A fumigant venting experiment was conducted in a silo at Lake Grace, Australia, to investigate full scale desorption. Two equations estimating fumigant desorption were tested, with an average of 65.5% and -86.3% error. The length of venting periods was simulated to quantify hours needed to mitigate hazardous conditions. For the scenarios investigated 10 to 24 h of venting were needed to reduce residual fumigant concentration below 0.3 ppm depending on simulation assumptions.