Methods

Data Collection

Img 1: Diver using a tape measure during an underwater survey on a coral reef.

The dataset used in this study is the 2016 Florida Keys Reef fish Visual Census dataset from the National Coral Reef Monitoring Program. The reef fish visual census dataset is a continually growing dataset whose purpose is to provide information on reef fish species composition, abundance, density, and size. Surveys were conducted throughout the US Caribbean and the Gulf of Mexico, beginning in 1978 and continuing to the present.

The Reef fish Visual Census data collection method was modified from Bohnsack and Bannerot (1986). In this method, the diver remains at a fixed site and surveys fishes within an imaginary vertical cylinder. The imaginary cylindar is centered on the diver and has a diameter of 15 m, which is marked out using a surface float tied to the bottom and a tape measure. Two teams, each with two divers, are deployed to survey at each observation site. Their data is averaged to reduce variability at the site level. The divers' collect the latitude and longitude of the site using a GPS, fish information (species, count, time-seen, and length), and environmental data (water depth, habitat type, underwater visibility).

Dives were terminated if visibility was less than 5 m, if the depth was greater than 99 m, or if there were strong bottom currents and divers were unable to stay fixed in place. In the Florida Keys region, these surveys took place at a total of 797 sites (Fig 1).

Fig 1: Observation sites along the Florida Keys Reef Tract.

Data Preparation

Manipulation of the Reef Visual Census Data

The Reef fish Visual Census dataset was collected in a "long" format; however, for the analyses that I wished to run on this dataset, it needed to be in a "wide" format. Therefore I performed a long-to-wide transformation using the dcast() function in R. I also removed variables that were unnecessary for this study such as the length of each fish.

Addition of the Environmental Data

Four raster datasets were used to collect environmental data on the study region, in addition to the observational data collected through the reef visual census. In order to add these values as variables in the existing dataset, the existing dataset had to be clipped to only contain sample sites that are within the borders of the rasters. The dataset was clipped using the Clip tool in ArcGIS Pro, and the remaining 545 observation sites were coloured by subregion (Fig 2). Values from each raster were extracted at the location of each of these 545 observation points in R, using the sf and raster packages.

Calculation of Species Richness and Simpson's Diversity Index

Species richness and Simpson's diversity index were calculated for each site using the species occurrence data (Eq 1 & 2). Species richness is the number of unique species at each site. Simpson's diversity index is the probability that if you chose two individuals at random they would be different species. Simpson's index was chosen over Shannon's index because Shannon's index emphasizes species richness, which I already had a separate metric for, while Simpson's index emphasizes species evenness.

Fig 2: Observation sites along the Florida Keys Reef Tract, clipped to within the boarder of the rasters and coloured by subregion.

Environmental Data

Below are the rasters used to provide data on water depth, mean summer levels of dissolved oxygen, mean summer water temperatures, and mean summer water salinity (Fig 3). The cell size of these rasters is roughly 5m by 5m, giving a relatively good resolution.

Fig 3: Rasters of water depth, mean summer dissolved oxygen, mean summer water temperature, and mean summer salinity.

Reef habitat types were categorized based on patchiness and vertical relief (Fig 4), and given the following codes:

ISOL_HR / _MR / _LR :

SPGR_HR / _MR / _LR :

CONT_HR / _MR / _LR :

RUBB_LR :

- Isolated patchy reef with high / moderate / low vertical relief

- Spur-groove reef with high / moderate / low vertical relief

- Contiguous reef with high / moderate / low vertical relief

- Rubble reef with low vertical relief

Habitats-categorized-by-vertical-relief-and-degree-of-patchiness-following-Franklin-et.png

Fig 4: Reef habitat types, categorized by patchiness and vertical relief (Ault, 2017).