Species distribution model identifies influence of climatic constraints on severe defoliation at the leading edge of a native insect outbreak

Abstract

Eastern North America is in the midst of a spruce budworm (Choristoneura fumiferana Clem., SBW) epidemic. SBW is one of the most important forest insects with regard to outbreak coverage and impacts to the forest industry in this region. Numerous bioclimatic, vegetation, and spatial–temporal variables can influence the distribution and outbreak patterns of phytophagous insects such as SBW. Published models have yet to agree on consensus drivers of SBW defoliation. Spatially-explicit, fine-scale predictive models are needed, however, to aid in our understanding of current and future SBW impacts. We used a species distribution model - MaxEnt to examine the relationships between the 1972–1981 SBW defoliation and 13 predictor variables for the island of Newfoundland, Canada the leading edge of the current SBW outbreak. The optimal model for predicting the probability of severe SBW defoliation included precipitation in May and June, precipitation in July and August, and maximum temperature in June and July. The probability of severe defoliation rapidly increased when precipitation ranged at 75.0 mm – 91.0 mm in May and June but showed a complex relationship with precipitation in July and August. Probability also increased with maximum temperature in June and July until temperature reached 18.0 °C. Defoliation probability decreased with increasing distance from major roads and river channels, suggesting possible corridor effects. Counter to research in other study areas, host tree species was not an important predictor of defoliation in our analysis. We then used this MaxEnt machine learning algorithm to predict the probability of severe defoliation for the forests of the entire island on a 2 km resolution raster. Areas with high probability (>0.80) were concentrated in the northeast part of the island where black spruce (Picea mariana (Mill.) BSP) dominates the forest landscape. The predicted severe defoliation area (probability 0.4–1.0) for black spruce was 1.4 times the area for balsam fir, which is the dominant canopy forming tree species on the island and preferred host species for SBW. The higher predicted probability of severe defoliation in black spruce stands relative to balsam fir stands may arise due to the unique combination of climatic conditions and geographical distributions of two species on the island. Species distribution models for irruptive insect impacts (i.e., defoliation) are relatively rare but may be an important tool in forecasting future insect damage.