Overfishing can be a major pressure on coral reef systems, reducing levels of biodiversity and typically resulting in shifts in fish size, abundance, and species composition, altering the ecological balance on the reef. Overfishing occurs as a result of a combination of an overabundance of fishers and overcapitalization of the fishing fleet relative to the available fish stock.

Analysis Method

Threats to coral reefs from overfishing were evaluated on the basis of population density within 30km of a reef location, adjusted by the area of shelf (up to 30m depth) within 30 km of the reef location. The analysis was calibrated using observation of coral reef fish abundance from surveys. The management effectiveness of marine protected areas was included as a factor mitigating threat.

DATA SOURCES USED IN OUR ANALYSIS OF THE OVERFISHING THREAT:

1. Population density—U.S. DOE, “LandScan,” 2001.
2. Shelf area—Developed at WRI on the basis of data from the Danish Hydrological Institute (DHI), ”MIKE C-MAP” depth points and data on coastline location—NASA, “SeaWiFS” and NIMA, “VMAP,” 1997.
3. Coral reef fish abundance—Reef Environmental Education Foundation (REEF) website http://www.reef.org (accessed 10 February 2003.)
4. Target fish geographic ranges - FAO. 2002. Living Marine Resources of the Western Central Atlantic, (The) . FAO Species Identification Guides for fishery purposes. Report GIS data (unpublished).
5. Ecological Units—K.J. Sullivan Sealey and G. Bustamante Setting geographic priorities for marine conservation in Latin America and the Caribbean (Arlington, Virginia: The Nature Conservancy, 1999) project GIS data (unpublished)
6. Target species - List of 16 species recommended by Phil Kramer (AGRAA and The Nature Conservancy) and discussed by Reefs at Risk workshop participants. (See Table 7.)

TABLE 7: Reef fish target species

Developing a coastal population adjusted for shelf area:

Our indicator of overfishing pressure is based on a ratio of human coastal population within 30 km of a coral reef adjusted by the coastal shelf area within 30 km of the reef.

Steps:

1. We calculated the shelf area (number of 1 km resolution shelf cells) within 30 kms of each cell;
2. In order to weight resource availability (i.e. factor into the analysis the fact that the same population would exert greater pressure on a smaller-sized available fishing area) we adjusted the population density according to the shelf area (<30 m) within a distance of 30 km. A population cell where there was a high surrounding shelf would be reduced, but a cell where there was no or little shelf would remain the same, with values inbetween altered proportionally. Each cell was reclassified to the order of 100% (for lowest shelf area) down to 50% (for highest shelf area);
3. Human population was identified (clipped) within 10 km of the coastline.
4. The coastal population value was then multiplied by the reclassified shelf layer (a percentage) to give new values for coastal population weighted by shelf area.

Population pressure per unit of shelf area was then calculated as the population (adjusted by shelf) within 30 km of a grid cell, which is our proxy indicator for overfishing threat to coral reefs. Coral reefs were then overlaid and classified using country-level expert opinion/literature as a guide. This threat estimate (THR_OVF_RAW) was adjusted for areas that have active coastal management. Specifically, in Marine Protected Areas (MPAs) rated as having full or partial management effectiveness, the overfishing threat level was reduced by one grade (i.e., from high to medium or from medium to low). This reduction results in an estimate of overfishing threat adjusted for management (THR_OVF _ADJ). Coral reef locations are then overlaid and classified by this adjusted threat estimate.

Calibration

For each ecological unit, the average number of species observed from the REEF survey sites was calculated for that unit. This average was then divided by the number of species expected to be found, to give a value (when multiplied by 100) of the average percent of target species observed. This data set was then used to calibrate the population density modeling.

Validation

We subsampled the REEF surveys done by experts and for each survey site calculated the average number of our target species observed. We then looked for a reverse correlation between the estimated threat level and the average number of target species observed (i.e. the higher the threat level, the lower the number of target species that would be observed). We found that there was a strong correlation, showing that the higher the threat level, the fewer (on average) target species would be seen.