Seed Bank in Hemlock Removal Experiment at Harvard Forest 2004-2015

The impending loss of hemlock trees due to hemlock woolly adelgid (Adelges tsugae) infestation prompted the need to identify species of plants able to colonize areas where hemlock has been removed. We investigated species present in the seed bank (in 2004 and 2010), the seedling bank (in 2004 and 2010), and seed rain (from 2005-2015) in the Harvard Forest Hemlock Removal Experiment. In 2004, Kelley Sullivan and Aaron Ellison determined the distribution of seeds buried in the soil in the six hemlock and two hardwood stands, before canopy treatments were applied (data files hf105-01, hf105-02). In 2010, Elizabeth Farnsworth and Aaron Ellison examined the distribution of plant species present five years after the simulated hemlock removal experiment was performed and six years after the baseline, pre-treatment study of the seedbank (data files hf105-03, hf105-04, hf105-05). In the 2010 study, we hypothesized that the composition of the seed and seedling banks and seed rain would diverge among the two treatments and the controls, based on the differential impacts of harvesting versus adelgid attack on the standing vegetation. We further hypothesized that the older (deeper) strata of the seedbanks of the treatment plots would have been exhausted over time, thus yielding poor germination relative to the top strata that continually receive seed rain. Identical methods of characterizing the composition of the seedbank and the aboveground vegetation at seedbank core locations were applied in both 2004 and 2010. Additional comparisons were made with long-term data collected on overall plant species composition in the Simes plots (datasets HF 106, HF126), and with data on seed rain (data file hf105-05). This study provides a unique temporal documentation of changes in the seedbank, seed rain, and vegetation under conditions of a changing overstory.