Revisiting fall armyworm population movement in the United States and Canada.

Introduction: Biophysical approaches validated against haplotype and trap catch patterns have modeled the migratory trajectory of fall armyworms at a semi-continental scale, from their natal origins in Texas or Florida through much of the United States east of the Rocky Mountains. However, unexplained variation in the validation analysis was present, and misalignments between the simulated movement patterns of fall armyworm populations and the haplotype ratios at several locations, especially in the northeastern US and Canada, have been reported. Methods: Using an expanded dataset extending into Canada, we assess the consistency of haplotype patterns that relate overwintered origins of fall armyworm populations to hypothesized dispersal trajectories in North America and compare the geographic distribution of these patterns with previous model projections. Results and discussion: We confirm the general accuracy of previous modeling efforts, except for late in the season where our data suggests a higher proportion of Texas populations invading the northeast, extending into eastern Canada. We delineate geographic limits to the range of both overwintering populations and show that substantial intermixing of the Texas and Florida migrants routinely occurs north of South Carolina. We discuss annual variation to these migratory trajectories and test the hypothesis that the Appalachian Mountains influence geographic patterns of haplotypes. We discuss how these results may limit gene flow between the Texas and Florida natal populations and limit the hereditary consequences of interbreeding between these populations.