Diverse geographic, environmental, and ecological factors affect gene flow and adaptive genomic variation within species. With recent advances in landscape ecological modelling and high‐throughput DNA sequencing, it is now possible to effectively quantify and partition their relative contributions. Here, we use landscape genomics to identify determinants of genomic differentiation in the forest tent caterpillar, Malacosoma disstria, a widespread and irruptive pest of numerous deciduous tree species in North America. We collected larvae from multiple populations across Eastern Canada, where the species experiences a diversity of environmental gradients and feeds on a number of different host tree species, including trembling aspen (Populus tremuloides), sugar maple (Acer saccharum), red oak (Quercus rubra), and white birch (Betula papyrifera). Using a combination of reciprocal causal modelling (RCM) and distance‐based redundancy analyses (dbRDA), we show that differentiation of thousands of genome‐wide single nucleotide polymorphisms (SNPs) among individuals is best explained by a combination of isolation by distance, isolation by environment (spatial variation in summer temperatures and length of the growing season), and differences in host association. Configuration of suitable habitat inferred from ecological niche models was not significantly related to genomic differentiation, suggesting that M. disstria dispersal is agnostic with respect to habitat quality. Although population structure was not discretely related to host association, our modelling framework provides the first molecular evidence of host‐associated differentiation in M. disstria, congruent with previous documentation of reduced growth and survival of larvae moved between natal host species. We conclude that ecologically mediated selection is contributing to variation within M. disstria, and that divergent adaptation related to both environmental conditions and host association should be considered in ongoing research and management of this important forest pest.

The forest tent caterpillar, Malacosoma disstria Hübner (Lepidoptera: Lasiocampidae), is an irruptive forest pest found throughout North America. Widespread species such as M. disstria are exposed to historical and contemporary processes that are not uniform and can generate regionally distinct genomic variation. Previous analyses used a short mitochondrial fragment to infer broad-scale phylogeographic patterns in M. disstria, whereas nuclear markers have been previously applied only in a smaller geographic region. In this study, we quantified M. disstria population variation with genome-wide single nucleotide polymorphisms and cytochrome c oxidase from mitochondrial DNA. Using highly variable genome-wide markers, we resolved clear genomic differences among populations of M. disstria east of the Rocky Mountains that were not detected using mitochondrial variation alone. We also did not detect host-associated divergence in either our genomic or mitochondrial data. Our results highlight the utility of genome-wide markers to resolve intraspecific population structure within a widespread species and support the need for further biogeographic sampling of this forest insect pest.