The present study is the first modeling effort at a global scale to predict habitat suitability of fall armyworm (FAW), Spodoptera frugiperda and its key parasitoids, namely Chelonus insularis, Cotesia marginiventris,Eiphosoma laphygmae,Telenomus remus and Trichogramma pretiosum, to be considered for biological control. An adjusted procedure of a machine-learning algorithm, the maximum entropy (Maxent), was applied for the modeling experiments. Model predictions showed particularly high establishment potential of the five hymenopteran parasitoids in areas that are heavily affected by FAW (like the coastal belt of West Africa from Côte d'Ivoire (Ivory Coast) to Nigeria, the Congo basin to Eastern Africa, Eastern, Southern and Southeastern Asia and some portions of Eastern Australia) and those of potential invasion risks (western & southern Europe). These habitats can be priority sites for scaling FAW biocontrol efforts. In the context of global warming and the event of accidental FAW introduction, warmer parts of Europe are at high risk. The effect of winter on the survival and life cycle of the pest in Europe and other temperate regions of the world are discussed in this paper. Overall, the models provide pioneering information to guide decision making for biological-based medium and long-term management of FAW across the globe.
Keywords: climate change; decision support; fall armyworm; machine-learning algorithm; pest management.