You are here

Harvard Forest >

Harvard Forest Symposium Abstract 2010

  • Title: Overstory Vegetation Response to Experimental Hemlock Removal
  • Primary Author: Audrey Barker Plotkin (Harvard Forest)
  • Additional Authors: Susan Irizarry (Clemson University); Alanna Kassarjian (Simmons College)
  • Abstract:

    Experimentally assessing the system-wide consequences of the loss of foundation species is of pressing concern. The Harvard Forest Hemlock Removal Experiment is a large-scale, long-term experiment aimed at elucidating the population-, community-, and ecosystem-level consequences of the loss of eastern hemlock (Tsuga canadensis), the dominant late-successional tree of northeastern North American forests, due to an exotic pest, the hemlock woolly adelgid (Adelges tsugae).



    Two hemlock removal treatments – (1) girdling all hemlock seedlings, saplings, and mature trees and (2) logging all hemlocks > 20-cm diameter and merchantable hardwoods – are matched with two control treatments. The hemlock controls will eventually be infested by the adelgid and the hardwood controls represent the future of hemlock stands in eastern North America. All overstory trees >5cm dbh were mapped and measured in 2004 and 2009. Logging and girdling manipulations were completed in early 2005.



    By 2009, both the girdled and logged plots had each lost 67-72% of their original basal area (Figure 1). The pace and structure of tree mortality differed between the two treatments, however. In the logged plots, basal area loss was immediate, much biomass was removed from the site, and remaining dead wood was downed. In the girdled plots, death occurred over two years and most dead wood is still standing. In addition, the shift in species dominance was more dramatic in the girdled treatment, as nearly all hemlock died. In contrast, the logging treatment left small-diameter hemlock and removed some stems of other species, so loss was spread among the full suite of species in the site. These differences in overstory change have resulted in a quicker pace of tree regeneration and understory vegetation response in the logged plots versus the girdled plots. It is not yet clear, however, whether the ultimate trajectories of change will diverge between the two treatments.



    Six years into this study, hemlock woolly adelgid was observed throughout the site in 2009. Almost half of the hemlock trees examined in the hemlock control plots were infested, albeit at low levels. Thus, the hemlock ‘controls’ are now adelgid plots, and going forward can be used to distinguish the effects of physical death of hemlock alone (logging or girdling) from additive and/or interactive effects of the adelgid on the forest ecosystem.

  • Research Category: Large Experiments and Permanent Plot Studies

  • Figures:
  • C:dataSimes_hemlockPubs&AbstractsHFSymposiumAbstractssimes-abp-symp10.pdf