Tropical marine ecosystems hold major biodiversity hotspots1 and provide a significant share of global fish catch2. Meanwhile, they are increasingly threatened by anthropic pressure including overfishing, global change, invasive species introduction, habitats destruction and pollution3. In particular, on-going global ocean warming is expected to severely affect species distribution, abundance and extinction rates but also trophic interactions and entire food webs balance4,5. These threats are critical especially for human populations that rely heavily on marine resources and depend on small-scale fisheries (SSF) or tourism for their livelihoods such as tropical developing states or small tropical islands6–8.
Tropical coastal environments form mosaic of interconnected mega-habitats extending from the shoreline to the open ocean. This complex structure greatly influences the dynamics of fish assemblages9. In recent years, mesophotic reef ecosystems (MREs) have gained attention10–14, not least because their depth may offer protection from anthropic stressors14,15. MREs occur in tropical and subtropical regions and are characterized by the presence of light-dependent corals and associated fauna at depths below 30-40 m extending to 150 m in areas with high water clarity12,13,16. MREs are known hot spots of tropical fish diversity and host fish communities ecologically distinct from shallow water reefs17. The mesophotic zone usually encompasses the shelf-break, a transition area from shelf to ocean characterised by a rapid change in the topography with a steep slope. The stiffness of the slope is associated with turbulent mixing enhancing primary productivity and therefore attracting prey and predators18–20. It concentrates diverse fishing resources over a relatively narrow area, sustaining important multispecific reef fisheries21–24. However, so far few study actually quantified the relative importance of mesophotic reefs, in particular because consistent observations extending from the shoreline to the shelf break are lacking25.
Oceanic islands and shallow seamounts also act as topographic anomalies that trigger complex physical processes increasing primary production and concentrating higher trophic levels. This phenomenon, known as the Island Mass Effect (IME26) is originated by the turbulence created by the island bathymetry, which uplift nutrient-rich water into the photic zone, enhancing primary production27. Oceanic islands and shallow seamounts are important environments for maintaining local biodiversity and non-resident migrating top predatory species28. IME aggregative effect on top predators supports commercial, artisanal and recreational fisheries29,30, which play an important role in the local socio-economic life of insular populations31. So far, most studies on the IME focused on physical-biogeochemical processes32,33. They showed that primary productivity is most enhanced on the leeward side of islands30,34. However, since fewer studies focused on higher trophic levels, the response of fish is generally depicted as symmetrical around islands27. No studies, for instance, determined if fish follow the pattern of primary productivity and concentrate downstream of islands.
Yet, this kind of knowledge is essential to assist decision making in conservation policies to protect biodiversity and the sustainability of fishing and diverse marine uses. Protective management is generally achieved through the creation of Marine Protected Areas (MPAs) delineating permitted and non-permitted zones according to pre-defined management objectives35. However, in some cases, the consequences of establishing MPAs are not adequately thought out, and a poorly planned MPA can be detrimental for local populations that rely on marine resources36. Indeed the decision support tools used to design MPAs rely on available data. To coherently manage the use of maritime space and achieve ecological, economic and social objectives, Marine Spatial Planning (MSP) is increasingly used as a strategic alternative aiming at integrating MPAs in a broader context37. MSP is a complex process requiring the use of optimization solvers that ultimately requires large quantities of spatially explicit cross-disciplinary knowledge and data (ecological, legal, social, economic)38. One of the main challenges to improve knowledge of tropical ecosystems and their resources and implement MSP is thus the data collection39.
Comprehensive monitoring are required to provide ground information for sustainable management40,41. Fish assemblage data are often used to help understanding how human activities influence marine ecosystems42,43 or as a measure of ecosystem health44 and as a basis for managerial decisions45. A variety of methods is used to assess tropical fish populations, including fishing gears or visual observations, each presenting its own pros and cons46. Fishery-dependent methods provide long time-series, wide spatiotemporal coverage but are biased by, among other, gear selectivity47. Scientific fish catches are more reliant but have a limited spatiotemporal coverage47. Tropical reef fish communities are also classically described via direct in situ observations through diver-based underwater visual census (UVC)48. Scuba diving is constrained by a set of limitations including underwater time and maximal diving depth and visibility49–51. As a result, most UVC-based studies are restricted to near shore shallow waters and provide punctual small-scale information whereas species richness and patterns of distribution is heavily influenced by the range of the sampling area52. To overcome part of these limitations and bias, underwater video techniques are increasingly used, whether stationary or towed, remotely operated or autonomous, baited or not53. The use of video increases sampling range and is more time efficient than diver-based observation54 but each technique has different limitations and combining underwater video and other sampling methods is therefore recommended55.
Active acoustics, in particular multi-frequency, is a powerful tool allowing simultaneous and continuous observation of the distribution of a variety of marine communities and abiotic characteristics56–58. Acoustics range of observation reaches several hundreds of meters below the surface, which allows prospecting the pelagic domain59. However, the ability to discriminate acoustically among taxa remains coarse and works best in relatively low-diversity temperate systems with a few well-defined and acoustically distinct groups56. Acoustic species discrimination remains challenging in highly diverse tropical ecosystems. Moreover, acoustic needs to be coupled with other observational methods to perform species identification, classically extractive one such as trawls and nets60. However, trawling is not always possible in topographically complex environments or in MPAs61–63. To lift out this lock, the combination of acoustic methods with non-extractive optical methods has emerged. These methods were mostly applied in temperate water64–69 whereas to date, only few studies focused on tropical waters70–72.
In this context, we combine multifrequency acoustic and video observation to provide a comprehensive vision of fish distribution around a tropical oceanic marine ecosystem. The study area, Fernando de Noronha Archipelago (FNA) located ~350 km off the coast of Brazil (Fig. 1), is a typical low productivity and high biodiversity system34,73,74, representative of tropical ecosystems. Like many other tropical small islands, the local population of FNA relies on artisanal fisheries for protein income75 and the economic activity is mainly based on tourism. Tourism generates demographic pressure and all its externalities, amplifying the demand for fish but also enhancing marine related activities such as recreational fishing or diving76,77. Beside, FNA is protected by a series of legal instruments regulating the uses of the marine environment and marine resources. Indeed, FNA is bathing in an Environmental Protection Area (EPA) where sustainable use of marine resources and tourism is allowed. The EPA includes a smaller no-take MPA, the “National Marine Park of Fernando de Noronha (PARNAMAR)”, covering about 70% of the main island and the coastline from the shore to the 50 m isobaths78.
On the base of three surveys combining multi-frequency active acoustics and underwater videos, we propose to address the questions identified above. Among other, we show that the combined approach allows for a comprehensive description of the distribution of fish biomass and fish assemblages. We provide the first biomass estimation of the black triggerfish Melichthys niger, a key tropical player. Comparing the effects of euphotic and mesophotic reefs we show that more than the depth, the most important feature is the topography with the shelf break as the most important hotspot. We complete the portrait of the IME showing that the effect is asymmetrical. While the primary productivity is higher downstream, fish concentrate upstream. Finally, we show the comprehensive information gained by our approach is directly usable to implement scientific-grounded MSP.