A dynamic outbreak distribution model (DODM) for an irruptive folivore: The western spruce budworm

Abstract

Forest disturbances caused by irruptive insect outbreaks have become more frequent due to the non-stationary and warming climate that favours insect development and survival, as well as forest management practices that increase host tree susceptibility. However, predicting and managing the impacts of these highly mobile pests is a challenge due to their uncertain distributions. Dynamic species distribution models (DSDM) acknowledge the non-stationary environment and species’ ecological niches and facilitate niche modeling by matching species observational data relative to environmental data, leading to better outcomes and reducing over-prediction in conservation efforts. The western spruce budworm (WSB; Choristoneura freemani Freeman) is a native forest pest with irruptive dynamics characterized by periodic landscape-scale outbreaks. WSB feeds primarily on Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) foliage and can cause extensive defoliation throughout western North American forests. Using decades-long presence and absence records of WSB outbreaks detected via aerial surveys and environmental variables, we developed dynamic outbreak distribution models (DODMs) to predict the distribution of outbreaks over time. Our results showed that DODMs provided robust temporal trends of WSB distributions while generating predictions that support documented northward shifts in epidemic populations. We found that in addition to the environmental variables known to affect WSB survival, phenological synchrony between bud burst and the emergence of larvae in spring was the most important predictor of outbreak distributions. We also identified high-risk areas for future WSB outbreaks and predicted range expansions and contractions under different climate change scenarios. Our study highlights the importance of using presence-absence data and incorporating temporal trends in environmental variables for better assessment of forest pest potential outbreak distributions. These findings have implications for proactive management strategies to prevent outbreaks and mitigate the impacts of WSB on forest ecosystems.