Research Forester Pacific Northwest Research Station, Alaska, United States
Abstract: Climate warming has dramatically altered the distributions and population dynamics of forest insect defoliators. These insects undergo periodic outbreaks that result from fine-scale density-dependent and coarse-scale density-independent dynamics. Regional outbreaks of hemlock sawfly (Neodiprion tsugae Middleton) and western blackheaded budworm (Acleris gloverana [Walsingham]) have recently defoliated forests across the pacific coastal temperate rainforest at a magnitude not previously reported in Canadian and United States aerial detection surveys. The objective of this study was to identify and evaluate how insect population pressure, climate, and host availability drive (1) the spatial extent of defoliation, (2) the temporal dynamics of defoliation, and (3) how spatial and temporal dynamics interact to drive outbreak activity in each species. We employed an ensemble machine learning approach (random forest and gradient boosting) to classify and predict insect outbreak activity based on aerial insect detection survey data (1980-2022) using climate and forest data at 1 km2 resolution from California through British Columbia to Alaska. Spatial models were fit using 30-year seasonal climate normals; temporal models were fit using the detrended deviation from 30-year normals. Results from this analysis indicate that hemlock sawfly outbreaks predominantly occurred from northern BC through the Alaska panhandle and are restricted to a narrow range of summer and winter temperatures, above average summer precipitation, below average winter precipitation, and a high volume of hemlock. Western blackheaded budworm outbreaks show a broader distribution along the maritime zones of Vancouver and Haida Gwaii Islands in BC and the Alaska panhandle. Activity was restricted to broader population specific temperature tolerances, above average summer precipitation, below average winter precipitation, and a high volume of hemlock. Our analysis of the temporal dynamics of hemlock sawfly outbreaks revealed that periods with dry winters and warm and wet spring and summer conditions drove outbreak events. Similarly, western blackheaded budworm outbreaks tended to coincide with periods with warm and dry springs, cool summers, and wetter winters. Both species exhibited significant yet minor signs of density-dependent spread. Finally, hemlock sawfly irruptive dynamics were better described by the spatial model, while western blackheaded budworm dynamics were better described by the temporal model. Our results demonstrate the overarching importance of climate in driving the biogeography and irruptive dynamics of these insect species and lay the groundwork for future studies investigating insect population dynamics, ecosystem effects, and the potential impacts of climate change.
