Résumé
For the first time ever at such a spatial scale, the EPICURE project is using the complementary exosystemic, functional and fisheries approaches to analyze and compare the partially exploited fish‐populating structure of the Geyser, Zélée, and Iris banks. This project focuses on three main points, which are 1) a habitat map of the three banks and an assessment of the live coral coverage as a proxy of the Geyser bank’s health, 2) a description of fish assemblages and a functional analysis relative to habitat type and time and 3) an estimate of the abundance index in structure size of the main exploited stocks. Overall, 42 habitats over 268 km² were mapped on the Geyser bank with 534 field verifications. On the Zélée bank, 24 habitats were mapped over 183 km² with 204 field verifications. Finally, 36 habitats were characterized on the 235 km² of the Iris bank with 98 field verifications. For each bank, two maps were produced, 1) one characterizing low typology levels (N4) describing main geomorphological structures and 2) a second map of a higher typology level (N5) including the type and the coverage proportion of benthic populations. Based on the constructed maps, 230 fish count stations with STAVIRO video system (SVR) and 35 stations with diving count approach (UVC) were located on the Geyser bank to estimate fish assemblages according to habitat type. 52 STAVIRO stations were also recorded on the Iris bank. Benthic community and coral coverage are very different depending on their location to the north, east, south or center of the Geyser bank, a fact that was never revealed by previous studies. In fact, the high coral biodiversity in the south (live coral coverage is greater than 70% among which more than 17% are of plate Acopora, Acropora hyacinthus) and in the center lagoon, are important elements that should be considered for the management of this bank. It is also observed that reef patches isolated in the center lagoon, long ago identified but never explored, were mapped on the three banks and fish were quantified on the Geyser and Iris banks. We counted more than 41,500 reef patches on the three banks, all situated around 30‐40 meters deep with peaks at 15‐30 meters on the Geyser bank and at 80 meters deep on the Iris bank with peaks sometimes reaching 30 meters. Thanks to the fine mapping of this project, such structures were able to be precisely located for fish counting with video and diving approaches. Results are surprising because such habitats have among the highest species richness and functional diversity and some of the largest biomass measured on the bank. Finally, the scientific campaigns of the EPICURE project allowed the description and measurement of size‐structured abundance indexes for the main fished species on the Geyser and Iris banks for the first time. The complementarity of the used methods, UVC and SVR, allowed us to obtain species abundance indexes (density and biomass), permitting us to obtain a health estimate of fisheries sensitivity of the fish stocks on both of these banks. The main results highlight overexploitation of the biggest four families (Lutjanidae, Serranidae, Lethrinidae, Carangidae), and smaller commercial species stocks that are better structured in size and globally more abundant. It was not possible to evaluate Zélée fish stocks. Apart from the Zélée bank, all of the objectives of the project were reached, and even more concerning the mapping of habitats. Future scientific surveys will therefore benefit from this solid base at the scale of these banks for further assessments.