Integrated modelling of the life cycle and aeroecology of wind-borne pests in temporally-variable spatially-heterogeneous environment

Integrating population ecology and aeroecology is important for understanding dispersal dynamics of pest insects that depend on aerial habitats. Sugarcane aphid (Melcuiaphis sacchari) is an economic pest of sorghum in Asia, Africa, Australia, and South America and recently invaded North America, affecting over 90% of the continent's sorghum production. Long-distance, wind-aided dispersal together with the ability to overwinter in the southernmost sorghum-producing areas appear responsible for their geographic spread in North America. We developed a spatially-explicit, individual-based, stochastic model that integrates the life cycle and aeroecology of sugarcane aphids to forecast regional infestations of sorghum fields. We parameterized and calibrated the model to represent environmental conditions and sorghum phenologies of the southern to central Great Plains of the USA where sorghum is cultivated. We validated the model by comparing simulated spatial-temporal patterns of aphid infestations to georeferenced field data on infestations collected as part of an extensive field project. We assessed potential usefulness of the model by analysing regional patterns of aphid infestations on sorghum simulated under climatic conditions recorded in the southern to central Great Plains in 2017. On a regional scale, the simulated northward advance of aphid infestations generally followed the seasonal development of sorghum, with the main invasion front moving northward from southern to central Texas, then northwest towards the panhandles of Texas and Oklahoma (northernmost and westernmost regions, respectively, of these states), and then northward again into central and western Kansas. Within this general regional trend, the stochastic nature of simulated infestations resulted in wide variation within each region with regard to time lags between the appearance of sorghum and initial infestation, which led to wide variation in infestation levels. The geographical distribution of simulated aphid infestation probabilities approximately mirrored the proportion of land cover in sorghum production, with a prominent infestation hot spot covering west-central Kansas and the panhandles of Texas and Oklahoma. Models capable of simulating both local population dynamics as well as long-range, multi-generational dispersal of wind-borne pests show promise as a component of adaptive areawide pest management programs.