Loss of eastern hemlock (Tsuga canadensis) to the hemlock woolly adelgid (Adelges tsugae, HWA) is a striking example of ecological transformation by an invasive insect. Salvage logging, a common human response to forest insect invasion, accelerates this transition. However, it is unclear whether the compound disturbances of logging and HWA result in lasting differences compared to HWA only. Therefore, in 2003 we initiated the Harvard Forest Hemlock Removal Experiment, a controlled, experimental removal of hemlock that contrasts standing death by girdling (as HWA would produce) with hemlock removal by logging (compounding insect disturbance with human disturbance). The experiment also includes unmanipulated hemlock and hardwood forest reference plots. After 15 years, we asked: do the trajectories of reorganization and community change differ between the girdled and logged treatments, or does the girdled treatment just lag the logged treatment but follow the same pattern? We hypothesized that for most metrics of change including ecosystem function, forest structure, and community composition, the rate and magnitude of change would be larger for logging than the insect alone as the forest re-organizes during the first 15 years after disturbance.
Results/Conclusions
Logging and girdling treatments removed about 70% of the aboveground biomass. After 15 years, aboveground live biomass stocks in the girdled and logged treatments increased but were still only 40-50% of their pre-treatment levels. However, additional biomass was stored in standing dead wood (girdled) and in long-lasting wood products (logged). Over 15 years, woody aboveground production in the girdled and logged treatments was less than the reference plots. Community composition shifted after logging or girdling; species turnover (mostly from species gains) was similar in magnitude but occurred 1-2 years later in the girdled treatment compared to the logged treatment. Major species responding to disturbance included the mid-successional tree Betula lenta, and several shrub and herbaceous species, including three woody invasive species. The vertical structure of the girdled plots was more complex than the logged plots, with a larger and more persistent pool of dead wood. In particular, the girdled treatment had more than five times as many large ( >40 cm) snags than the other treatments. In conclusion, although the community composition will likely converge in the girdled and logged treatments, the compound disturbance of insect + logging hastened forest change and reduced the structural complexity of the forest.
