Site suitability was calculated by examing factors that favor eelgrass establishment success. These factors were analyzed using available geospatial data. Five data layers were used in identifying suitable sites based, in part, on work completed by Dr. Fred Short and his colleagues (Short and others 2002):
Bathymetry
The bathymetry data were created by the National Oceanic and Atmospheric Association (NOAA) Coastal Services Center from depth soundings distributed through the NOAA Geophysical Data System (GEODAS). Values were extracted from the GEODAS database using the default HYD93 download and tabulated with header information. These data are believed to be complete within the area of interest. In cases where multiple soundings occurred within 50 meters, older soundings were discarded. The remaining data available from 1934 to 1996 surveys were included. Depth values were created by building a Triangulated Irregular Network (TIN) and converted to a grid using the methods described in the metadata provided with the downloadable bathymetry data for Rhode Island (www.edc.uri.edu/restoration/html/spatial/download.htm).
Temperature
Researchers at the Brown University Department of Geological Sciences Planetary Geosciences Group provided a consolidated raster data layer identifying areas that reached or exceeded 25 degrees Celcius based on Thematic Mapper satellite imagery of temperature. The data were compiled from 26 seasonally diverse dates of information from the Landsat 4, 5, and 7 sensors, covering cloudless scenes from 1984 through 2002. Areas remaining below 25 degrees Celcius received a value of 1 (suitable). Areas where temperatures reached or exceeded 25 degrees Celcius received a value of 0 (unsuitable).
Light
Light data were interpolated from the following sources (Granger
and others 2000):
The light extinction coefficient
(k) at each station and date was calculated and then a grid for each date was
interpolated using the Inverse Distance Weighting (IDW) algorithm. The maximum
extent of each interpolation was dictated by the station locations. Areas where
the solution for depth allowed 20 and 50 percent
light on each particular date. Areas that received less than 20 percentlight
were given a value of 0, areas that received at least 20 percent light were
given a value of 1, and areas that received at least 50 percent light were given
a value of 2. Next, the grids representing
individual dates were multiplied together using the map calculator function
in ArcView. This yielded a grid depicting the number of surveys at which each
area received a certain percentage of light. A high value indicated that the
area received 50 percent light for a majority of the survey dates while a value
of 0 indicated that the site received less than 20 percent light during at least
one survey. A final grid for each light
data set was produced where a value of 0 was given to all areas that received
less than 20 percent light, a value of 1 was given to areas that received at
least 20 percent light or 50 percent light for less than ½ the surveys,
and a value of 2 was given to all areas that received 50 percent light for more
than ½ the surveys. Current Eelgrass
Location In order to avoid
disturbing existing eelgrass beds, data on current eelgrass distribution were used.
Since the boundaries of existing eelgrass
beds are not well defined within the study area, each bed was given a 100-meter
buffer and removed from the analysis. Current eelgrass locations
were obtained from a variety of data sources, many of which are available from
the Web. This project used eelgrass location data from the following sources:
Each of these points, lines,
and polygons were buffered by 100 meters. Areas within the buffer were considered
unsuitable and received a value of 0. Areas beyond the buffer received a rating
of 1 (suitable). Historic Eelgrass
Location Considering where eelgrass
was known to exist historically allows one to understand where characteristics
of the water and sediment were suitable for eelgrass to exist. Anecdotal information
suggests that Narragansett Bay was once filled with eelgrass where depth and
other conditions were appropriate. Historic eelgrass locations have been captured
by a project funded by the Rhode Island Aqua Fund (RIAF; a program funded through
a 1988 bond issue that protects the resources of and reduces pollution in Narragansett
Bay).
Data for the coastal ponds
were developed from work by URI documented by Dr. Thorne-Miller and colleagues
in 1983. The authors reported and mapped eelgrass distributions in Ninigret,
Green Hill, Potter, Point Judith, and Trustom Ponds from aerial photography
taken during the summers of 1978, 1979, and 1980. Historical data were digitized
for only Ninigret and Green Hill Ponds for this calculation, because the other
ponds were not included in other input data (e.g., current eelgrass). Since
these data will be used in the calculation to identify historical eelgrass presence/absence,
habitats were digitized from the published hard copy maps by visually estimating
boundary locations. This method yielded reasonable data for identifying suitable
and highly suitable areas for eelgrass restoration. All areas known to support
eelgrass historically were buffered by 100 meters, and areas known to support
eelgrass historically were given a value or 2. All other areas received a value
of 1. Assembly These layers were multiplied
together to obtain a suitability ranking. Areas within the grid with the highest
value (Gridcode = 8) are considered the "Most Suitable" for establishing eelgrass transplants.
Areas with a value of 4 are considered "Very Suitable" and areas with a value of 2 are
considered "Suitable" for transplaning eelgrass. Areas that score a value of 1 are considered
"Somewhat Suitable" and areas scoring 0 are "Unsuitable" for eelgrass transplants.
The final scored grid was converted
to a shapefile for use in ArcView software without the need for any other software extensions.
Granger, S., B. Buckley, M. Traber, M. Brush, M. Richardson, and S. Nixon.
2000. An assessment of eutrophication in Greenwich Bay. RI SeaGrant. Kopp, B., A. Doherty, and S. Nixon. 1997. A guide to the site-selection for
eelgrass restoration projects in Narragansett Bay, RI. Final Report to the RI
Aqua Fund Council. 22pp + App. Oviatt, C., A. Keller and L. Reed. 2002. Annual Primary Production in Narragansett
Bay with no bay-wide Winter-Spring phytoplankton bloom. Estuarine, Coastal and
Shelf Science 54: 1013-1026. Short, F.T., R.C. Davis, B.S. Kopp, C.A. Short and D.M. Burdick. 2002. Site
selection model for optimal restoration of eelgrass, Zostera marina L. Marine
Ecology Progress Series 227:263-267. Thorne-Miller, B., M. M. Harlin, G. B. Thursby, M. M. Brady-Campbell, and B.
A. Dworetzky. 1983. Variations in the distribution and biomass of submerged
macrophytes in five coastal lagoons in Rhode Island, U.S.A. Botanica Marina
26:231-242.
For further details on the data used in the suitability calculation, please see the metadata
records for those data sets. These are available with the Seagrass Site Selection Tool on CD-ROM, which
can be requested from www.edc.uri.edu/restoration/html/spatial/habmodel.htm.
References