Conservation of Terrestrial Salamanders Through Hemlock Woolly Adelgid Management in Eastern Hemlock Forests within Great Smoky Mountains National Park

Hemlock woolly adelgid (Adelges tsugae; HWA), an invasive aphid-like arthropod, was first documented on the east coast of the United States in the 1950s. HWA is an herbivore which primarily feeds at the needle base of hemlock tree species (Pinaceae: Tsuga). With no evolutionary defenses and few biotic controls, the eastern and Carolina hemlock (Tsuga canadensis and Tsuga carolinensis) serve as the primary diet of HWA in eastern North America. The invasive pest began to spread rapidly throughout the hemlock’s range causing defoliation and death of the trees within 4 – 10 years. With the loss of the foundational species, Tsuga canadensis, several microenvironmental changes were documented. Microenvironmental changes in response to biological invasions and anthropogenic forestry practices can lead to shifts in populations of physiologically sensitive taxa such as salamanders and their prey, terrestrial arthropods.

National Park Service staff at Great Smoky Mountains National Park manage HWA by treating eastern hemlocks with the neonicotinoid pesticides, imidacloprid and dinotefuran. To measure indirect effects of eastern hemlock mortality, and HWA management, this study measured several parameters in hemlock-dominated stands that have been repeatedly treated by the NPS and stands which were untreated and where hemlock woolly adelgid has reduced the hemlock canopy. Our major objectives were to assess microenvironmental and vegetative community differences between managed and un-managed eastern hemlock stands and analyze those differences with respect to arthropod and woodland salamander abundance and/or diversity. A mixed effects ANOVA was used to compare mean soil organic matter (or duff) pH, substrate volumetric water content, vegetative litter depth, temperature, and arthropod diversity and abundance between managed and un-managed stands. A mixed effects linear model using elevation range as a random effect or block was used to model salamander abundance with the aforementioned continuous variables. While the microenvironmental parameters were not significantly different between stand types, order-level richness of arthropods, and woodland salamander abundance did significantly differ (α = 0.05). According to the linear mixed effects model, substrate moisture and forest management were the strongest predictors of salamander abundance (α = 0.05).