Attributes and Analysis Direction.

David L. Richardson, John H. Porter, and H. H. Shugart.

Permanent vegetation monitoring plots were established on Parramore Island at the Virginia Coast Reserve Long-Term Ecological Research site (VCR/LTER) between October 1992 and April 1993. The one hundred 1/12th hectare plots (16.29 m. radius) were randomly distributed over the island, covering vegetation zones ranging from salt and freshwater marsh to shrub thickets and maritime forest (Figure 1). The purpose of the plots is to: 1.) establish baseline data on the distribution patterns, densities, and growth and mortality rates of the tree and shrub species found on Parramore Island and representative of those of the other islands of the VCR as well; 2.) serve as "ground-reference" controls for analyzing the island's vegetation using remote sensing techniques and for evaluating these techniques; 3.) provide the sorts of data needed to parameterize forest succession models (FORET and ZELIG) for predicting species distributions and dynamics on the islands of the VCR; and 4.) quantify rates and patterns of shifts from one ecotone to another over time so that, in conjunction with other studies, the processes driving these shifts can be better understood.

The initial location of each plot was chosen at random and then staked out in the field. The center and edge of each plot were drawn over 2 m. resolution images derived from a 1:20,000 aerial photo in which individual and clustered tree canopies and internal marsh patterns were visible. Using these plot maps along with a compass and rangefinder, the plot centers were located and marked by driving a length of rebar into the ground and affixing a numbered aluminum tag and plastic flagging to it. Any noticable deviation from the preselected location of the plot's center was noted (e.g., shifted due to impenetrable Smilax patches), and readjusted once the fieldwork was complete (Figure 2).

Within each plot, every tree and shrub that had a diameter at breast height (dbh) of 2.5 cm. or greater was tagged, its species identified, and its distance and azimuth from the plot center, height, and dbh (for tree species) or basal diameter (bsd, for shrub species) measured (Figure 3). In instances where a tree either forked or grew horizontally below dbh level, each stem or vertical growing branch was tagged and measured as an individual tree but was noted to be a part of the same tree as the other stems. This situation only regularly occurred with red cedar, Juniperus virginiana. For shrubs, the stem with the maximum bsd was the one measured and tagged. Additionally, the total number of stems at dbh level in incremental 3 cm. size classes was tallied.

Also within each plot, estimates of ground cover and measurements of light levels were made at between three to six randomly located subsites 2 m. in diameter. The ground cover was determined as a percent of the cover falling into the categories of grasses, rushes and sedges, vines and briers, miscellaneous herbaceous plants, woody species, and bare ground (including open water). In each category the maximum height of the vegetation was measured. A ceptometer was used to measure light levels present at 20 cm. intervals from the ground up to 120 cm., plus at 160 and 200 cm., as well as an "open sun" reading taken in direct sunlight for comparison between subsites.

We are currently using data from the permanent plots to assay the accuracy of recent land-cover classifications of Parramore using remote sensing techniques and to quantify the descriptions of what each cover class comprises. While preliminary agreement between the permanent plots and TM image-based maps is good, permanent plots located within patchy ecotone transitions, such as on small marsh hammocks and pimples, are very often "lost" within the larger 30 m. x 30 m. TM footprints.

Current efforts are also underway to analize the community structure and species associations using various SAS and GIS packages. Baseline data showed 11 tree and 3 shrub species to be present on the island (not counting Iva frutescens which never attained minimum dbh), but only 5 tree and 2 shrub species were commonly found throughout the survey. They were: Juniperus virginiana, Pinus taeda, Prunus serotina, Persea palustris, Ilex opaca, Myrica cerifera, and Baccharis halimifolia. For tree species, basal area estimates range from zero in the marshes and along the beach to over 50 m2/ha along the ridge of Little Beach, and stem density estimates range from zero to over 3,000 stems/ha (Figure 4). For shrub species, basal area estimates range from zero in the marshes and along the beach to over 100 m2/ha within the shrubline transition zones between forest and marsh, and patch density estimates range from zero to over 1,000 patches/ha. Preliminary principal components analysis of the dataset, looking just at the internal structuring of the different woody species, indicates that while species associations may and do seem to exist (Figure 5), these associations are not inherently exclusive of other species; rather, the distribution and range of any particular assemblage is likely controlled by environmental variables acting independantly on each species of the assemblage. We hope that by quantifying and including such environmental variables in the ongoing analysis, we will be able to come up with a predictive model for species and association distributions for Parramore and the other islands of the VCR/LTER.

Special thanks to Bruce P. Hayden and Patrick N. Halpin for invaluable assistance with the recent and ongoing analyses of the collected data.