Formation of a Mid-Atlantic ghost forest

Keryn Gedan and Abigail Philbrick, Biological Sciences, George Washington University, Washington, DC, Sergio Fagherazzi, Boston University, Matthew L. Kirwan, Physical Sciences, Virginia Institute of Marine Science, Gloucester Pt., VA

Background/Question/Methods
Ghost forests are formed when sea level rise and saltwater intrusion kill trees at the coastal interface. While the formation of ghost forests reduces coastal forest habitat, it also promotes the expansion of tidal marshes, which are rapidly losing habitat due to sea level rise. The Mid-Atlantic is a hot spot of sea level rise and provides a window through which to view the future of coastal change. For two years, we monitored 20 x 20 m forest plots that span a salinity stress gradient at the Virginia Coast Reserve Long-Term Ecological Research site to understand ghost forest formation in a chronosequence, or space-for-time substitution, study design. Twenty-four plots spanned three zones of coastal forest: healthy forest, stressed forest, and ghost forest. Within these plots, we monitored tree and shrub growth and survival, the spatial distribution of shrubs and the invasive common reed (Phragmites australis) in the subcanopy, and the community composition of understory vegetation to reconstruct the changes in vegetation structure that occur during ghost forest formation. At the same time, we have monitored edaphic conditions of soil moisture and salinity and light availability to associate changes in the plant community with abiotic drivers.
Results/Conclusions
The species composition and density of trees shifted markedly across the site, representing the die-off of mature trees, lack of forest regeneration, and a loss of deciduous tree species during transition from healthy to stressed to ghost forest. The ghost forest contained many dead trees and primarily live individuals of only two tree species, loblolly pine (Pinus taeda) and red cedar (Juniperus virginiana). In the subcanopy, the invasive common reed Phragmites australis was dominant in the ghost forest, whereas the shrub Morella cerifera was more abundant in the stressed forest. These observations suggest that in the chronosequence, Morella shrubs proliferate early as the forest is stressed and are quickly replaced by Phragmites australis during ghost forest development. Halophytic species appeared in the understory of the ghost forest but were rare in other parts of the forest, representing the switch from a terrestrial to a tidally influenced system. Lastly, in only two years of observations, changes in the canopy, subcanopy, and understory vegetation indicated that the study forest is rapidly converting to marsh. Using several methods, we estimate a rapid pace of change that may convert our 3 ha forested study site to marsh in as little as a quarter-century.