|
| Pesticide resistance is a major challenge to increasing the resilience and sustainability of current food production systems. Preserving the susceptibility of pest organisms to chemical products is a key factor to optimize a pesticide-based strategy. However, resistance management strategies (RMSs) must consider unique species biologies, multiple resistance mechanisms, environmental factors, and pest management practices, which can make their implementation complex. Here, we develop a method to help manage this complexity using a grid-based simulation framework for pesticide resistance evolution including population growth and dispersal dynamics. This framework was applied to the fall armyworm, Spodoptera frugiperda, for which resistance evolution is a major concern. We explored the sensitivity of 13 parameters dealing with landscape structure, dispersal rate, chemical treatment protocols, chemical degradation rate, dose-response curves and transition rates (i.e., flux between sub-population driven by the mutation rate). From the sensitivity analysis of simulations, we computed heat maps of the influence of each parameter on a set of variables (total pest population size, fully resistant population size, and resistance frequency). Assuming a large but realistic range for each parameter, Sobol's sensitivity index showed that resistant transition rate (from phenotypically susceptible to resistant sub-populations) and pesticide properties (in particular, degradation rate and dose-response curve) are more important in the outbreak of resistance compared with resistance ratio (i.e., the benefit of being resistant over susceptible in terms of fitness), chemical application intensity and landscape composition. In addition, using Pareto optimality, we assessed the performance of different pesticide application regimes according to total population size, population size of resistant individuals, the total amount of pesticide, and the total area of host plants suitable for S. frugiperda. Across the wide parameter space explored, we revealed the high volatility of outputs suggesting that the performance of treatment protocols depends on the ecological context. Nevertheless, despite this variability, a “windowing” management strategy consisting of a single pesticide group applied per insect generation, provided the optimal control of S. frugiperda and resistance evolution outcomes. This work provides a set of tools to test RMS scenarios for the control of S. frugiperda and to understand how variabilities may arise at different management steps and geographical scales. | |
|