A 14-year study of vegetation on St. George Island conducted with graduate students and volunteers

Background/Question/Methods

Long-term studies are deceptively difficult to conduct and much can be learned from the missteps of others.  Here I describe a study of vegetation on barrier islands, ongoing since 1998, as an example of difficulties associated with the design, maintenance, quality control, data management, and productivity of a long-term study.   The census was designed to quantify how species in different dune habitats vary through time in a dynamic habitat of moving sands and often harsh climate.  Over 14 years, the study has expanded to also address questions about hurricanes, restoration and conservation, climate change, and even facilitative interactions among plant species.

Permanent replicate plots were established in the three characteristic habitats on barrier islands: fore-, inter- and backdunes.  Each plot consisted of a 7 x 7 grid of stakes, with each stake 10 m apart.  Annual censuses have been conducted in a 1 m²quadrat located at each stake.  There are currently 3 replicate plots in each habitat for a total of 441 quadrats.  Percent cover by species is determined for each plot each fall, using a combination of students in a graduate course and volunteers.  Elevation and some basic soil information have also been irregularly determined for each quadrat location.

Results/Conclusions

The ecological results are that annual variation in dune vegetation is species-specific, can be extreme, and is correlated with both spring drought and summer and fall hurricanes.  The data have also been used to predict long-term patterns associated with climate change and identify habitat-specific species to use for dune restoration.  We are expanding the work to now understand the reciprocal interactions between vegetation, succession, and habitat-specific changes in dune geomorphology.

Relevant to this symposium, there are significant difficulties associated with conducting and maintaining long-term studies with students and volunteers.  Some are quality-control issues related to the fieldwork, but others have to do with data maintenance and management issues.  Training is necessary to maintain consistent field methodologies with inexperienced workers. Standards for data entry and checking must be established.  Data must be amenable to later editing when errors in identification occur, with sufficient documentation of such changes.  Our experience should strongly encourage others to initiate long-term studies, but greater forethought will increase both the productivity and efficiency of such studies.