The Abrolhos shelf extends ~ 200 km offshore and is SWA’s most biodiverse region, encompassing a large mid-to-outer shelf hard bottom domain with reefs and rhodolith beds (~ 20,900 km2)5,6. Fine-sediment dissipative beaches and a large estuary with mangroves dominate the coastline, and terrigenous-mixed sediments predominate in the inner shelf27. This large and complex seascape (Fig. 1) comprises a representative experimental setting for understanding the distribution and abundance of reef fishes in different habitats, as well as for exploring the drivers and spatial scaling of beta diversity in reef fish assemblages. The high richness of reef fishes off coral reefs that we found in Abrolhos was unexpected, and sheds new light toward the integration of phenomena that occur at different scales and across distinct habitats and groups of organisms11,20. From a practical standpoint, our results are relevant to improve marine management in complex tropical seascapes with rhodolith beds23 and other large marginal habitats.
The high richness of reef fishes in rhodolith beds, where fish biomass was smaller than on reefs (Supplementary Fig. S1 online; Fig. 4), seems to be primarily related to the much larger area of rhodolith beds, as well as to the broader depth and cross-shelf range of this hard-bottom habitat, contrasting with reefs. Rather than being a regional idiosyncrasy, the relatively larger area and cross-shelf range of non-reef habitat used by reef fishes seems to be recurrent in tropical shelves across all ocean basins8,9,23. However, due to logistical constrains and to the apparent smaller relevance of marginal habitats to fish and other reef-associated organisms, these habitats are still much less sampled than the iconic shallow water reefs20, with the exception of mangroves and seagrass beds3,8,9.
Compositional variability in biological communities is strongly dependent on spatial scale. Accordingly, beta diversity is expected to be high at biogeographic and local scales, while turnover tends to be lower at regional scales28,29. Reef fish assemblages tend to vary sharply at small spatial scales (< 1 km) due to variation in habitat structure, exposure to wave energy and currents, and stochasticity in population dynamics16, but may be markedly similar over broader biogeographic areas21,30, indicating non-random species associations and convergent ecosystem functioning under similar biotic and abiotic forcing. In Abrolhos, distance among sites was a weak predictor of beta diversity at the seascape level, and turnover between reef sites was also relatively low, akin to the turnover between reefs and inshore rhodolith beds with high amounts of sand and low macroalgae cover (Fig. 1). On the other hand, the relatively high compositional variation at the seascape level responded positively and non-linearly to habitat structure and to the variation in the amount of light reaching the bottom (Supplementary Fig. S2 online). Light at bottom integrates depth and turbidity in a single and ecologically meaningful variable, and captures the major ecological features across coral reef depth gradients14,31. The effects of light at bottom over reef fishes seems to be largely indirect and associated with trophic (e.g. grazing and predation on small invertebrates) and non-trophic (e.g. shelter from predators) connections with the benthos31.
Reef fishes have a relatively high dispersion potential, both as larvae and as swimming adults, and their distribution at the seascape level is largely constrained by large scale environmental filtering16,21. The high turnover recorded at the seascape level, with highly dissimilar reef fish assemblage structures, indicates that reefs and rhodolith beds have distinct functional properties. Turf dominance on reefs (~ 50% of turf cover) may be associated with the high abundance of roving herbivores in this megahabitat, and to the lower turnover between reef sites. Turf algae is the main trophic connection with herbivorous fish, which typically avoid macroalgae32, and was positively correlated with light at bottom, which may be associated with the higher reef fish turnovers observed at the seascape level.
Most biological assemblages include a smaller number of abundant species and a long tail of rare species, and coral reef fishes are no exception. Abundance distributions tend to be more skewed toward commonness than the Gaussian and less so than the lognormal distribution, which would be the sum and the product of random variables, respectively33. While there were more species with maximum abundances recorded in pinnacles and fringing reefs, nearly one-third of the 49 species recorded in both benthic megahabitats had maximum abundances in rhodolith beds (Fig. 2). This trend is remarkable, as it evidences that reefs may be the marginal (“suboptimal”) habitat for several reef fishes in the SWA. The maximum abundances recorded in rhodolith beds are not related to small juveniles, but the role of rhodolith beds as a critical habitat for juvenile reef fish is unclear and deserves further investigation. We observed small surgeonfish, parrotfish and grunts in rhodolith sites with dense algal canopies (see Fig. 5), which may function as structural refugia against predators3,4,13. Akin to mangroves, juvenile reef fish may not depend on rhodolith beds, but the presence of large expanses of hard bottom with dense algal canopies may enhance diversity and biomass in reefs through the exchange of propagules, individuals and nutrients. In addition, rhodolith beds are a better connectivity matrix than soft sediments for adult reef fish migration toward spawning grounds near the shelf edge19, as recorded in Abrolhos for the red (Epinephelus morio) and black (Mycteroperca bonaci) groupers34.
The trophic structure of reef fish assemblages was also largely driven by benthic megahabitat and cross-shelf gradients. The low abundance of planktivorous fishes in the Abrolhos reefs has been previously remarked35, and is herein confirmed as a regional trend, once this trophic guild was also minor in rhodolith beds. For instance, large schools of forked-tailed plankton-feeding damselfishes (Chromis spp.), wrasses (Clepticus spp.) and groupers (Paranthias spp.), which are common along the entire West Atlantic16, are lacking from Abrolhos, where only sparse individuals occur in offshore sites. Several microcarnivores (e.g. Prognathodes brasiliensis, Serranus chionaraia) and other deeper-water dwellers (e.g. Rhomboplites aurorubens, Chromis enchrysura) recorded only in offshore rhodolith beds add to the habitat and cross-shelf variation, albeit these species may associate to reef structures across their distribution range21. The dissimilar trophic structure of fish assemblages in rhodolith beds was dominated by an impoverished assemblage of benthic-feeding carnivores (e.g. Calamus sp., Cantherhines macrocerus, Malacanthus plumieri) and carnivorous fish that feed both in the bottom and in the water column (e.g. Balistes vetula and Carangoides crysos, respectively) (Fig. 3). Together with the lower biomass of herbivores (Fig. 4), this trend indicates that most of the energy that flows through fishes in rhodolith beds comes either from small prey captured among the calcareous nodules, or from secondary production in the water column. Relatively longer pelagic food chains in more offshore sites may partially explain the lower reef fish biomass in this megahabitat, added to a clear constrain to fish herbivory within rhodolith beds’ extensive canopies of unpalatable macroalgae such as Sargassum, Lobophora, Padina, Stypopodium and Dictyota (see Fig. 5). Herbivores’ biomass may be underestimated by BRUVs, but the trophic comparisons between habitats (reef vs rhodoliths) should remain valid. However, caution must be taken when comparing baited-video surveys with underwater visual censuses.
Fish biomass in the vast offshore rhodolith beds was consistently higher than that of inshore reefs and rhodolith beds, and seconded that of the outer arc reefs within the ANMP. Considering the several maximum abundances recorded in rhodolith beds (Fig. 2) and the large area of this megahabitat, it is likely that population sizes of several commercially important species (e.g. groupers and snappers) are larger in rhodolith beds than in reefs9,20. Fish biomass was five-fold higher in the reefs protected by the ANMP, indicating an overwhelming effect of management regime. The exclusive and frequent occurrence of the endangered Caribbean reef sharks (Carcharhinus perezi) within the ANMP adds to the positive effects of the enforced no-take status of this area. Rhodolith beds became the most important reef fishing grounds in the Abrolhos shelf since the 1980s, following the overfishing of inshore reefs12,31. However, besides being rarely targeted by enforcement operations, which should regularly restrain illegal fishing gear (e.g. drift nets for lobsters, ‘hookahs’) and seasons’ violations, rhodolith beds and shelf-edge reefs are the least represented habitats in the existing MPA network6,12.
It is unlikely that further sampling will change our overall conclusions, nor will have any impact in the recommendations for managers and policy-makers. An additional caveat about the generality of the results presented herein is that rhodoliths occur under a wide spectrum of environmental conditions, from shallow high-energy to deep low-energy settings, from the tropics to polar latitudes23, and in ocean basins with very different pools of fish species. Therefore, some rhodolith beds may not be suitable as reef fish habitats. Despite the low SWA-endemic fish biomass (~ 2%) in rhodolith beds, endemics comprised 15% of the fish assemblages in this megahabitat. This finding provides additional support for the idea that rhodolith beds are extremely relevant to the conservation of the biodiversity of the small, unique and highly threatened Brazilian reefs. Rhodolith beds comprise a major but belittled reef fish habitat within the SWA, and should be urgently prioritized in marine spatial planning, licensing and fisheries management. While carbonates’ mining and oil and gas exploitation in reefs are banned in Brazil, these activities are steadily growing in rhodolith beds, which are often categorized as rubble with low importance for biodiversity conservation36. Quantitative assessments covering the broad spectrum of reef fish habitats are needed for robust inferences about rarity/conservation status and the habitat range of these vertebrates, which play a major role in healthy ecosystem functioning.
Source: Ecology - nature.com
