Widespread diversity deficits of coral reef sharks and rays

A global survey of coral reefs reveals that overfishing is driving resident shark species toward extinction, causing diversity deficits in reef elasmobranch (shark and ray) assemblages. Our species-level analysis revealed global declines of 60 to 73% for five common resident reef shark species and that individual shark species were not detected at 34 to 47% of surveyed reefs. As reefs become more shark-depleted, rays begin to dominate assemblages. Shark-dominated assemblages persist in wealthy nations with strong governance and in highly protected areas, whereas poverty, weak governance, and a lack of shark management are associated with depauperate assemblages mainly composed of rays. Without action to address these diversity deficits, loss of ecological function and ecosystem services will increasingly affect human communities.R code for the analysis in this paper is provided. All data files used in the analysis are also provided. It will be necessary to specify the location of the data files in the R code to enable it to run. README.txt file contains additional information.Funding provided by: Paul G. Allen Family FoundationCrossref Funder Registry ID: http://dx.doi.org/10.13039/100000952Award Number:General BRUVS methods FinPrint sampling of coral reefs used Baited Underwater Video Stations (BRUVS). BRUVS consisted of a weighted metal frame holding a compact high-resolution video camera (typically a GoPro) and a 1 m arm holding a bait bag containing 1 kg of crushed oily fish in front of the camera. Where possible, at least 50 BRUVS were deployed at a reef (defined as a single isolated reef, or a patch of a large reef ~10 km2). Four to eight BRUVS were deployed simultaneously at a reef, with all BRUVS set at least 500 m from each other. BRUVS were deployed between 0-40 m depth and for a minimum of 60 minutes. This deployment time was sufficient to adequately sample the core coral reef shark and ray species. At each BRUVS drop, the date, depth, time of deployment and retrieval, GPS coordinates, sea conditions, and weather conditions were collected. All videos were read by trained annotators using either Event Measure (https://www.seagis.com.au/event.html) or FinPrint Annotator (https://github.com/GlobalFinPrint/Finprint-Annotator). BRUVS that landed with a severely obstructed view were not annotated and were removed from further analyses. All sharks and rays observed were identified to the lowest possible taxon (mostly to species), and the time they entered the video frame and the maximum number of each species in a single frame (MaxN) in each video were recorded. MaxN is a conservative measure of relative abundance that ensures individuals are not double-counted. Species identifications were verified by a senior annotator. Information on deployments and the results of the video annotations as part of the FinPrint project were entered into a central database. In addition to data included in the FinPrint database, we gathered data from regions in which the FinPrint project had been unable to collect data, or where data were not included in the FinPrint database. This data set includes 35 reefs, including those from the eastern tropical Pacific (Costa Rica 3 reefs, Mexico 1 reef, Clipperton Island 1 reef), Fiji (3 reefs), Indonesia (3 reefs), Red Sea (Saudi Arabia, 5 reefs, Sudan 1 reef), United Arab Emirates (1 reef), Maldives (5 reefs), Madagascar (2 reefs), Australia (2 reefs), Brazil (2 reefs), USA (1 reef), Puerto Rico (2 reefs) and Colombia (Seaflower Biosphere Reserve, 3 reefs). These reefs were used only in the assemblage analysis, and not in the estimation of population depletion. Habitat determination A screenshot from all deployments was taken once the BRUVS settled on the benthos and analysed using BenthoBox software (www.benthobox.com). Relief and habitat were determined for each deployment. A 20-square grid (five across, four up) was played over the screenshot from each deployment. All squares containing benthos were categorised into one of six relief scores ranging from 0 (flat) to 5 (complex). The average relief score for all squares containing benthos was then calculated for each deployment. Habitat was similarly assessed. The most abundant habitat type within each square containing habitat was identified based on a pre-determined list: hard coral, soft coral, bleached coral, unconsolidated (sand/rubble), consolidated (rock), seagrass, turf algae, macroalgae, and other (cnidarians, sponges, etc.). The percent cover of each habitat type was then calculated by the number of squares with that as the most abundant habitat type over the total number of squares containing benthos. For example, in a video where 13 squares contained benthos and five had hard coral as the most abundant habitat type, hard coral was considered to account for 38.5% of benthic cover for the deployment