METHODS
Monitoring methods

We use several methods to monitor each reef site. In photostation surveys, the primary goal is to identify the species of scleractinian corals and hydrocorals present and quantify their abundance. Chain transect surveys are used to determine the relative abundances of different substrate types as well as coral species. Video surveys of transects and photostations are also recorded to preserve information in digital format. Larger scale haphazard photographic surveys are used to determine whether the photostations are representative of the surrounding reef. Finally, a 1000 m2 swim survey on each reef provides a virtually complete list of coral species present on the reef. Most of the data presented here are taken from photostation surveys. A major feature of our photostation method is that the boundaries of the actual areas sampled are permanently marked by stainless steel stakes cemented into holes drilled into limestone substrate so that the stations can be relocated.

Chain transects were established by stretching a 25-m line between two eyepins permanently cemented into the seafloor. Chains are attached to the 25-m line at the beginning of each meter, allowed to drop to the seafloor, and made to conform to the three-dimensional substrate directly below the 25-m line. The number of chain links overlying each different type of substrate are counted and recorded. This process is replicated three times for each transect. The effect is that of "slicing" through the reef and tracing the resulting outline. In this survey, all scleractinian corals and hydrocorals are identified to species. Other categories of organisms and substrate types were recorded, but are not considered in this paper. Three 25-m chain transects have been surveyed annually since 1989 in each of the Biscayne N.P. sites; one 25-m chain transect was surveyed in each of the N.M.S. sites beginning in 1991.

Photostations are set up on a four-stake system, with four stainless steel stakes cemented into the substrate defining the x- and y-coordinates of a 2.00 x 2.25 m (4.5 m2) PVC frame which serves as a positioning grid. Rings to support the grid from underneath are slipped over the stakes and screwed in place at notches filed into the stakes so that the z-coordinates of the grid are defined as well. This grid provides the boundaries and supports for a smaller PVC camera frame (0.50 x 0.75 m, or 0.375 m2) with a 35 mm Nikonos V camera with a 28 mm underwater lens mounted 1.0 m above the plane of the frame. The camera frame outlines the actual two-dimensional area sampled with each photograph. Twelve photographs, each covering 0.375 m2, are required to survey each 4.5 m2 area. In each of the photostations at Carysfort and Looe Reefs, the 4.5 m2 grid is laid down twice for a total area of 9.0 m2 (24 photographs). In the photostations at Biscayne N.P., the grid is laid down three times for a total area of 13.5 m2 (36 photographs).

Before 1991 the placement of the 4.5 m2 grid in Carysfort and Looe Reefs was guided by two stakes that defined the corners along one side of the large frame during placement. In 1991 those photostations were improved by the implantation of stainless steel stakes that more precisely and repeatably define the z-axis, thereby bringing all photostations up to the same standard of the four-stake system for future surveys. All data presented here from Carysfort and Looe Reefs were collected from photographs using the two-stake system. The sites in Biscayne N.P. have been based since their establishment upon the four-stake system. The photostations on Looe Reef were surveyed in 1984, 1985, 1986, and 1991; Carysfort Reef photostations were photographed in 1984, 1985, 1986, 1988, and 1991; and Biscayne N.P. photostations were surveyed in 1989, 1990, and 1991.

The 0.50 x 0.75 m camera frame provides a reliable scale within each photograph that is accurate at the level of the frame. In order to create an objective measure of scale at the level of the substrate, in 1991 "scale pins" were added to all photostations: stainless steel rods measuring 1/4" in diameter were driven into 3/16" holes drilled into areas of bare, stable substrate and the distances between the pins measured to the nearest 0.5 cm with tree calipers. A set of four pins was established within one frame of each set of 12 photographs.

In order to test whether our photostations are truly representative of the surrounding reef, we have conducted haphazard surveys of the areas in which our photostations are located. Using either a Hasselblad camera mounted on a 4 m2 frame, or a Nikonos camera mounted on a 1 m2 frame, photographs representing 60 to 100 m2 were taken of the reef surrounding each photostation. The corals in these photographs were analyzed and quantified in the same manner as photographs from the photostations. This survey was performed once on each reef site for comparison with the coral composition of the corresponding photostation for that year, and is not repeated on a yearly basis.

Data analysis

The data collected by the chain transect method consists of an enumeration of coral species and substrate types, their sequence, and their length underneath the chain. Since there is some variation in the placement of the chain under the transect line (due to wave surge, etc.), every transect is counted three times within a one to five day period so that confidence intervals can be calculated for transect results. For each transect we calculate species richness, Shannon-Weiner species diversity, percent cover of coral species, and relative abundance of substrate types. Since we obtain three replicate counts for each transect each year, we are able to statistically compare these values between years. Since the transect length is 25 m, but the distance travelled by the chain over the length of the transect is greater than 25 m, this method also yields information about the three-dimensional topography of the reef, a parameter not directly described by our other sampling methods taken alone.

The photostation surveys are performed using Kodak Tri-X black and white negative film. Each survey is repeated with Kodak Ektachrome color slide film to produce a backup set of color photographs. Each photograph includes the outline of the camera frame, which carries identifying information on the specific reef site, frame number, and date of the photograph (see examples in Figure 4). Coral colony and frame outlines are traced from prints onto mylar sheets, coral species are identified and assigned a unique colony number that can be followed through time, and the colonies are digitized.

Colony areas in the photographs represent projected surface area: that is, three-dimensional objects projected onto two dimensions in a non-orthographic fashion (objects closer to the camera appear to be larger than objects of the same size but further away from the camera). While orthographic surface area, the result of stereo photo images, would more accurately represent absolute percent cover (Done, 1981), both methods underestimate the biological surface area of a living coral. This is increasingly the case for colonies with greater three-dimensional relief, especially for a branching coral such as Acropora palmata where living tissue on the undersides of branches would not be visible at all, regardless of the photographic system employed. While projected surface area is a reasonable estimate of percent cover, it is true that estimates of change in coral cover by this method will be conservative.

The camera frame defines an area of 0.50 x 0.75 m at the plane of the frame, but because of the phenomenon of parallax, a photograph actually includes an area greater than 0.50 x 0.75 m below the plane of the frame down to the sea floor. Adjacent photographs therefore contain overlap. In order to ensure that individual colonies are counted and measured only once, the photographs are corrected for parallax. Areas of overlap in adjacent frames are removed by vertical or horizontal cut lines at the left and bottom edges of each camera frame where they overlap with the right and top edges of adjacent frames.

Cumulative species richness and percent cover are calculated on a frame by frame basis. Percent cover for individual species, cumulative species diversity (based either on the number of colonies of each species or the area covered by each species), and year to year comparisons for percent cover are analyzed on a colony by colony basis.

To homepage for coral reef ecology.
To personal page for Ouida Meier.

This work was performed while Ouida was a graduate student of James Porter's at the University of Georgia.