Preserving marine ecosystems and the services they provide has called for conservation strategies focused on protecting biodiversity from the effects of human exploitation of living and non-living resources (Briscoe et al., 2016; Magris et al., 2020). To that end, these strategies have often relied on the effectiveness of area-based measures (e.g., marine protected areas - MPAs) to protect biodiversity ‘hotspots’, defined as areas where diversity is particularly rich, including rare and endemic species (Costello et al., 2022). While consistent with the strategies adopted in terrestrial ecosystem conservation, and likely useful for coastal ecosystems, protecting areas considered biodiversity hotspots may prove insufficient when it comes to the open ocean (Briscoe et al., 2016). In essence, that is because ecosystem services (including food provisioning) are not only maintained by biodiversity alone but by a combination of biophysical processes that result in productivity enhancement and energy transfer through the trophic web, that sustain rich and abundant communities (Mann and Lazier, 2005; Scales et al., 2014). As noted by Briscoe et al. (2016), ‘a strategy that focuses on protecting processes, patterns, and features that promote enhanced biological productivity in addition to biodiversity, will have a greater probability of including important conservation features’.

Deep-ocean habitats are facing growing human pressures originated from climate change and different forms of use of valuable living and non-living resources (Ramírez-Llodra et al., 2010; Levin and LeBris, 2015; Danovaro et al., 2017). Global warming has affected the deep-sea, which is gradually becoming warmer, more acidic, less oxygenated and even more food-limited (Thurber et al., 2014; Sweetman et al., 2017; Cheung et al., 2022). Simultaneously, human activities such as fisheries and mining seriously threaten structurally-complex biogenic habitats, including cold-water coral reefs, which host locally abundant and diverse communities on deep continental margins, seamounts and ridges, with little chances of recovery (Boschen et al., 2016; Clark et al., 2019; Carreiro-Silva et al., 2022). In that sense, deep-sea conservation has become of major concern around the globe (Barbier et al., 2014), raising important debates about cost-effective conservation strategies (e.g., Dunn et al., 2018).

The definition of priority conservation features and the delimitation of areas where these features could be more effectively protected, are critical steps for the development of deep-sea conservation strategies (Combes et al., 2021a). Deep-ocean basins are mostly oligotrophic, except for localized regions where combinations of topography, circulation, geological or biological processes create local phenomena that promote a relatively stable food supply and sustain abundant/rich communities (e.g., hydrothermal vents, food-falls, seamounts and others) (Genin, 2004; Ramirez-Llodra et al., 2010). These are ‘productivity hotspots’ in the deep-ocean that encompass trophic-wide aggregations of primary producers, secondary and tertiary consumers, and top predators, some of them targets of commercial fisheries (Levin and Dayton, 2009; Briscoe et al., 2016). Conservation strategies that prioritize regions identified as productivity hotspots, could favor the protection of unique biodiversity features while preserving current and future uses of the deep-sea.

The spatial delimitation of productivity hotspots in the deep-ocean, as defined above, requires the characterization of localized biophysical processes and associated biological communities, usually achieved by integrated field studies (e.g., Kenchington et al., 2014). In poorly explored ocean basins, however, these studies tend to be insufficient to support conservation initiatives, justifying the use of alternative spatial analyses that can approximate the localization of potential productivity hotspots. Often, these analyses explore spatial patterns of regionally available surrogate environmental variables (e.g., Anderson et al., 2011; Misiuk and Brown, 2024), in particular those that describe seafloor geomorphic features and can be derived from bathymetry maps (Wilson et al., 2007; Harris et al., 2014; Walbridge et al., 2018). These terrain variables have been commonly used to describe deep-sea habitats and assess relevant drivers of deep-sea benthic species distribution (Yoklavich et al., 2000; Brown et al., 2011; Anderson et al., 2011; Misiuk and Brown, 2024). They can also delineate a variety of deep-sea geological features that originate from tectonic, sedimentary, oceanographic and biological processes (e.g., cold-water coral reef growth) interacting at multiple spatial and temporal scales (Blondel, 2001; Camerlenghi, 2018). Some geoforms are characterized by abrupt topographies (e.g. seamounts, canyons, shelf breaks), which modify the flow of deep-water masses and interpose the flux of particles suspended on oceanic pelagic layers, promoting mechanisms that spatially aggregate plankton, micronekton and their predators (e.g., Fernandez-Arkaya et al., 2017). These mechanisms operate mostly at intermediate depths (~200 – 400 m) by inducing (a) the upwelling/uplift of nutrient-rich deep waters up to the photic layer, strongly and persistently enough to sustain phytoplankton and zooplankton growth, and (b) the flow-driven processes of biomass accumulation in relatively small areas (Genin, 2004). High order consumers, including meso- and bathypelagic fishes, squids and deep-sea shrimps and crabs may couple their behavior and life history patterns with these biophysical processes, benefiting from the increased prey supply and forming stable patches of elevated density, which collectively typify deep productivity hotspots in space (Healey et al., 1990). If these organisms are commercially valuable (e.g., orange roughy, toothfishes, alfonsinos, aristeid shrimps, geryonid crabs and others) such patches tend to be exploited by fishing fleets, and sustain deep-water fisheries (Koslow et al., 2000; Clark, 2009; Company et al., 2012; Salois et al., 2023).

This study will explore the premise that historical records of fishing operations in association with deep abrupt topographies may operate as effective surrogates for oceanographic processes that define the spatial distribution of productivity hotspots. Also, because these areas include conservation features and food provisioning (i.e., exploitable biomass of fishing resources), their spatial delimitation is intended to support area-based management and conservation strategies in the deep-sea. The study will focus on the South American continental margin off Southeastern and Southern Brazil (SW Atlantic), here referred to as the Brazilian Meridional Margin (BMM, sensu Alberoni et al., 2019). The slope region is characterized by a complex topography and a variety of geological features that include sedimentary slopes, submarine canyons and pockmarks (Harris et al., 2014). In association with these features, distinctive benthic habitats and communities have been described, particularly cold-water coral banks, reefs and cold-seeps (see reviews in DeLeo et al., 2020; Kitahara et al., 2020; Shimabukuro et al., 2020). Bottom fisheries surged in this region in 2000, targeting concentrations of predatory demersal fish (e.g., the monkfish Lophius gastrophysus), crustaceans (Chaceon spp. and aristeid shrimps) and squids (Illex argentinus) (Perez et al., 2009a). For ten years, this activity affected over 10% of the available and previously untouched slope area of the BMM, being responsible for important biomass reductions of the main targeted stocks, and the increased mortality of noncommercial fauna including cold-water corals (Perez et al., 2013, 2020; Kitahara et al., 2020). After that period, fishing pressure on the slope grounds reduced to relatively low levels (Costa et al., 2024), but the region continued to be increasingly exposed to other anthropogenic pressures (Halpern et al., 2015), most notably, offshore exploration of oil and gas (Perez et al., 2020) and global warming [i.e., the region is under the influence of a major marine hotspot of the Atlantic Ocean (Franco et al., 2020)].

Based on these cumulative human pressures and environmental sensitivity criteria (e.g., biodiversity representation, population connectivity and threatened species), priority areas for implementing new MPAs have been proposed within the BMM slope region (Magris et al., 2020; Costa, 2024). However, no concrete initiatives in that direction exist to date, but only management measures and regulations placed on slope fishing and oil and gas exploration activities (Perez et al., 2020). The delimitation of productivity hotspots could inform area-based management and conservation strategies, such as the implementation of MPA networks along the slope region (Costa, 2024), the limitation of bottom fishing areas (Perez et al., 2009a) and the exclusion of ecologically significant regions from the areas leased for deep-sea oil and gas exploration (Bernardino and Sumida, 2017). These strategies, and others, are expected to emerge in the context of the recently established Brazilian Marine Spatial Planning process (Gerhardinger et al., 2019), in which areas defined as deep-sea productivity hotspots could be particularly useful to develop conservation plans in the oceanic sectors of Brazilian EEZ. Around the globe, criteria defining deep-sea MPAs location, both within and outside EEZs, have gradually incorporated ecosystem productivity as a conservation feature (e.g., Briscoe et al., 2016; Visali et al., 2020) and some high-seas MPAs proposed under the BBNJ framework are productivity hotspots, e.g., the Costa Rica Dome, the South Tasman Sea, the Emperor Seamount Chain, the Mascarene Plateau, and the Gulf of Guinea (PEW, 2024). Despite the availability of data on habitats and biodiversity in these areas, their selection and spatial delimitation have relied on mapping environmental variables that best represented desired conservation features. Because habitat and biodiversity descriptions in the BMM slope region are scarce and scattered (Sumida et al., 2020) using the same approach, as proposed in this study, seems necessary and timely.

During one decade (2000 – 2009) fully monitored demersal fishing operations captured fish, crustaceans and cephalopods on the largely unexplored deep seafloor of the Brazilian Margin, covering a wide range of latitudes, depths and geoforms (Perez et al., 2009a; Sant’Ana and Perez, 2016). Because these were commercial operations seeking profit, they identified aggregations of targeted species, the spatial distribution of which were shown to be affected by geoforms extending along- and across the shelf break and slope. Fishes, deep-sea shrimps, crabs and squid exhibited variable responses to these features. Notwithstanding, there were general patterns emerging particularly in models that explained higher fractions (>50%) of the total variance, namely those fit to the abundance of all catches combined (65.6%), to the abundance of fishes (77.3%), and to the probability of positive catches of deep-sea shrimps (67.3%) ( Table 3 ). These models highlighted the importance of shelf break embayments of Santos Basin and, within them, the sectors incurvated and excavated by erosive action of the Brazil Current ( Figure 2 ). The upper ‘moderate slopes’ (300 – 500 m depths) were particularly favorable for fishes and squid concentrations, whereas the lower ‘moderate slopes’ (600 – 800 m depths) increased the probability of deep-sea shrimp positive catches. Because these patterns resulted from the analyzes of catches of different fishing fleets, their interpretation require an initial consideration as to what extent they may reflect the effect of fishing gears and the spatial dynamics of fishing operations rather than the species abundance patterns.

Firstly, it is important to note that the international fleets covered a previously unexplored region of the BMM, to identify profitable grounds of mostly virginal resources (see review in Perez et al., 2009a). The spatial patterns of their operations throughout this period, therefore, is the outcome of a process of finding high concentrations of the targeted species, which, in essence, optimized the delimitation of potential productivity hotspots. Vessels operating different gear had almost exclusive targets (i.e., the taxonomic groups considered) and their operations exhibited spatial patterns that tended to maximize the catch of these targets (Sant’Ana and Perez, 2016). As a consequence, it was possible to observe that catches of taxomonic groups tended to occur in most geoforms explored by the fleets (binomial model, Figure 8 ), but their abundance was higher only in specific ones (lognormal model, Figure 9 ), possibly due to natural conditions that favored their concentration. Demersal fishes could be an exception, since they were targeted by both upper slope trawlers and gillnet vessels. The latter, however, was responsible for nearly 70% of the catches and the effect of the spatial patterns of their operations predominated in the analysis. Therefore, it is plausible to conclude that the spatial patterns of abundance of the different taxonomic groups were not significantly affected by variable efficiencies of different fishing gear or spatial operation patterns of different fleets, but, on the other hand, these elements possibly affected results obtained in the analysis of all catches combined. Nevertheless, when these results were analyzed in the context of those obtained in the analyzes of the taxonomic groups separately, the spatial patterns of the areas regarded as productivity hotspots remained robust.

Demersal fishing was widely recorded in most cross-shelf geoforms (classes). However, fishing hauls and catches concentrated preferentially over few slope geoforms distributed along the entire BMM following topography breaks, i.e., the shelf break (300 – 500 m depths) and terraces break (600 – 800 m) ( Figure 3 ). Higher abundances of the taxonomic groups were observed over these preferential geoforms, but only in certain portions of their spatial distribution ( Figure 7 ). Therefore, it is uncertain whether the species aggregating factor was directly linked to the preferential geoforms or these were components of more complex biophysical processes operating on the slope of the BMM. Previous studies using the same catch records revealed that suitable habitats for deep-sea shrimps and monkfish were mostly driven by variables related to water masses impinging on the seafloor, surface productivity and deep-sea circulation (e.g. mixed layer depth, seafloor salinity and temperature, current speed and direction) (Sant’Ana, 2023). In the deep-sea, these variables interact with geomorphology, and particularly with abrupt topography, contributing to localized biophysical phenomena that may promote aggregation of predatory megafauna by enhancing prey supply in an otherwise generally impoverished deep seafloor (Genin, 2004; Maier et al., 2023). In that sense, the effect of geoforms on the occurrence and abundance of demersal megafauna in the BMM may be associated with biophysical processes resulting from their interaction with along-slope and across-slope circulation processes.

Deep-sea abrupt topographies (e.g. shelf break-slopes, seamounts, canyons) are known to aggregate zooplankton and fish through biophysical coupling mechanisms involving enhanced population growth and/or accumulation processes (reviewed by Genin, 2004). Growth of predatory fish populations has been attributed to (a) localized upwelling events that promote enhanced biological productivity propagating through the food web or enhancement of organic matter sink of dead organisms and fecal pellets during productivity peaks, or (b) other mechanisms that increase the supply of prey through zooplankton and fish responses to deep-water flow dynamics, referred to as ‘trophic subsidy’ (Gooday and Turley, 1990; Genin, 2004). Aggregations often reported on shelf break and slope regions have been associated with vertical water flow (upwelling or uplifting) that can either enrich waters in the photic zone long enough to enhance primary productivity and enable energy to be transferred through local food web, or provoking active responses of zooplankton that tends to become spatially aggregated and promptly available for predators.

In the Brazilian Meridional Margin, the offshore geostrophic flow of the BC transports generally warm and oligotrophic water masses. Yet topographically-induced mesoscale (100s of km) and sub-mesoscale (10s of km) processes of the BC result in cross-shelf transport and upwelling/uplift of nutrient-enriched SACWs that drive the formation of patches of oceanic biological productivity over mostly oligotrophic shelf break and slope regions (Pereira et al., 2024). Along these regions the uplift of SACWs has been described (Mesquita et al., 1983; Brandini, 1990; Campos et al., 2000) and attributed to the effect of surface wind stress (Ekman transport), pressure gradients derived from the interaction between the BC flow and the shelf break topography, and eddies and meanders of the BC (Campos et al., 2000; Palma and Matano, 2009; Combes et al., 2021a, 2023; Pereira et al., 2024). All these processes tend to describe an enhanced onshelf transport of bottom waters across the slope and shelf break benthic boundary layer in the Santos Basin region, between 25° – 21°S, overlying sectors where the abundance of catches was highest (embayments A, B, C, Figure 10 ). In this region, onshelf transport of bottom waters seems to be favored by an important change in coastline orientation that modifies the along-shore pressure gradient and, through geostrophy, promotes onshore bottom flow and uplift (Palma and Matano, 2009). In addition, the shelf break sinusoidal morphology produces a series of abrupt changes in orientation that may also locally enhance cross-shelf transport (Combes et al., 2021a). Cyclonic eddies originate from the strengthening of the BC flow, and propagate downstream of regions where the coastline changes orientation (~23°S), particularly within shelf break embayments (‘shelf break bights’, Combes et al., 2023). These eddies promote cross-shelf water exchange and vertical mixing, entrapping for several weeks cold, fresh, nutrient-rich shelf waters over the shelf break embayments (Campos et al., 2000; Combes et al., 2023; Amorim et al., 2024). The spatial connection between these shelf break vertical water flow processes and concentrations of predatory megafauna ( Figure 10 ) suggests the incidence of a biophysical aggregating mechanism (see below) and support the spatial delimitation of mesoscale productivity hotspots at the Santos Basin shelf break embayments. At a sub-mesoscale, within the embayments, it is plausible that the Brazil Current flow, past the protrusion sector, detaches from the shelf break, accelerating the uplift flow over the incurvate and steeper sector of the embayment ( Figures 7 , 10 ) where demersal fishes seem to concentrate (Combes et al., 2023). This uplift flow may be decelerated over the sediment terraces (sediment deposition area) that extends downstream within the embayments, a process described by models that compare deep-water upwelling in narrow vs. wide continental shelves (Jiang et al., 2011). Interestingly, in Figure 4G , the ‘flow direction’ terrain variable demonstrates that water flow over terraces in fact differs from the continental slope general flow.

The uplift of nutrient-rich SACW up to the photic zone (at surface or subsurface depths) can enhance phytoplankton growth leading to the formation of patches of herbivorous zooplankton that support growth of predatory fish populations (Genin, 2004). In the BMM, maxima of chlorophyll-a concentrations have been reported year-round in association with BC eddies over shelf break embayments at surface (visible in satellite imaging, Combes et al., 2023; Amorim et al., 2024) and subsurface layers (Brandini, 1990; Pereira et al., 2024). The propagation of energy down to predatory fishes dwelling on or near the slope seafloor, however, would require pelagic-benthic coupling mechanisms, such as a ‘biological active transport’ (Trueman et al., 2014). This transport is provided by diel vertical migrators including zooplankton, fishes and squids, that during daytime may reach slope depths above 500 m becoming available as prey to both benthic and benthopelagic feeders (Trueman et al., 2014). The most abundant species caught on the shelf break and slope within the embayments are benthic (monkfish) and benthopelagic feeders (argentine hake, codling and the argentine squid). All of them include pelagic (euphasid shrimps, engraulids, myctophids, rough scad Trachurus lathami) and demersal prey (Merluccius hubbsi, Cynoscion guatucupa, Synagrops bellus, Illex argentinus and others) in their diets, which indicate they have access to vertically migrating prey or migrate themselves following them ( Figure 11 ) (Haimovici et al., 1993; Santos and Haimovici, 1997; Nascimento, 2006; Valentim et al., 2008; Muto and Soares, 2011; Costa et al., 2019). The productivity hotspots of Santos Basin embayments could, therefore, be sustained by biophysical shelf break uplift processes coupled with biological active transport of pelagic and demersal fish vertical migrators, some of them targeted by bottom fisheries (Perez et al., 2009a).

The scarlet shrimp (Aristaeopsis edwardsiana) is the dominant deep-sea shrimp species in Santos Basin shelf break embayments (Dallagnolo et al., 2009). The species is an active swimmer that preys on both infauna invertebrates and pelagic fishes and cephalopods (Rezende et al., 2014). It forms year-round reproductive aggregations concentrated in a narrow depth zone (600 – 700 m depth), under the influence of the AAIWs that extends underneath the SACWs (> 500 m depth, Dallagnolo et al., 2009; Pezzuto and Dias, 2009; Sant’Ana, 2023). This zone is also characterized by concentrations of cold-water corals at Campos and Santos Basins, that thrive on zooplankton and particulate organic matter on 500 – 800 m depths (Kitahara et al., 2020; Bernardino et al., 2023). In association with cold-water corals, there is a diverse fish, crustacean and cephalopod fauna, all susceptible to deep-sea shrimp trawling (Perez et al., 2013). It is uncertain whether these deep fauna aggregations benefit from the biophysical shelf break uplift processes, or may be under the influence of other topography-related trophic subsidy mechanisms that may affect the transition zone between SACW and AAIW. Below intermediate slope depths (>500 m), trophic subsidy is often related to enhanced horizontal flux due to currents impinging on sloping topographies, tides and internal waves that can accelerate the near bottom flow of food particles and zooplankton, and laterally advect vertical migrating micronekton. Cold-water corals grow faster under the enhanced prey flow regime (Maier et al., 2023), aggregating demersal fauna including deep-sea shrimp prey. Moreover, such prey flow can directly benefit demersal predators that hunt pelagic prey advected by currents in-between periods of rest in less dynamic regions near the seafloor (the ‘feed-rest’ strategy Genin, 2004). The sepiolid Heteroteuthis dispar, and the squid Pyroteuthis margargaritifera are water column vertical migrators that advect during the day over abrupt topographies, such as seamounts (Clarke, 2007). These pelagic cephalopods have been reported in the diet of the scarlet shrimp, suggesting they can prey on pelagic micronekton advected by currents onto the slope benthic boundary layer, while also feeding on small invertebrates that live in the sediment (Rezende et al., 2014).

Squid and deep-sea crabs showed a less clear association with the productivity hotspots. The argentine squid (Illex argentinus) has a complex population structure on the BMM, formed by a resident spawning group as well as large north Patagonian spawning migrators that reach the region during the austral winter-spring months (Perez et al., 2009b). The latter sustained elevated catches between 28° - 30°S and 400 – 700 m depths, as well as the abundance peaks observed in the upper slope of the shelf break embayments D and E ( Figure 8D ). Winter-spring reproductive migrators of the southern population travel within the cold SACWs (Haimovici et al., 2008) and are possibly transient components of southern shelf break embayments productive areas. A similar link between life history patterns and the Gulf Stream shelf break exchange process was demonstrated for the co-generic species I. illecebrosus in the NW Atlantic (Salois et al., 2023). Deep-sea crab catches in embayments A-D were exclusively sustained by the royal crab (Chaceon ramosae), whereas in embayment F only the red crab (Chaceon notialis) was caught (Perez et al., 2009a). Both species exhibit complex reproductive bathymetric and latitudinal migrations, when mature and immature males and females tend to segregate in space and time (Pezzuto and Sant’Ana, 2009). Modelling the effect of geoforms on these population strata distribution could inform better if and how deep-sea crab species relate do shelf break embayments productive areas.

The analysis of historical demersal fishing operations delimited potential mesoscale and sub-mesoscale productivity hotspots in the BMM. In these hotspots, topography-induced biophysical processes may be the source of biological productivity sustaining patches of predatory demersal fishes, crustaceans and squid of commercial interest. Integrated geological, physical and biological field studies are required, however, to confirm and properly describe general mechanisms. Recent efforts have been conducted to explore habitats in the slope areas of Santos Basin, mostly driven by the need to characterize cold seeps and cold-water coral zones in association with pockmarks and other geological features (e.g. Bendia and Carrerette, 2022; Perez et al., 2023). The present study contributes to these efforts adding spatial patterns of predatory megafauna that can be integrated with the understanding of habitat structure and biophysical processes of the BMM. More importantly, these patterns strengthen the concept of productivity hotspot in the conservation debate of marine areas off the Brazilian margin (Briscoe et al., 2016).

The occurrence of vulnerable marine ecosystems (e.g., cold-water coral banks) along the slope regions of Santos, Campos and Espírito Santo basins have long justified calls for conservation action, including de implementation of marine protected area networks that could protect these ecosystems from the expanding oil and gas exploration activities (Bernardino and Sumida, 2017; Bernardino et al., 2023). The region has also been prioritized for conservation in models that combined biodiversity representation, population connectivity and threatened species, with human cumulative impacts, including fishing (Magris et al., 2020; Costa, 2024). This study demonstrates that exploiting deep-sea fishing resources and protecting biodiversity in the region may be related to the existence of mesoscale and sub-mesoscale biophysical processes that fuel productive deep ecosystems in the generally oligotrophic western boundary limb of the South Atlantic subtropical gyre (Pereira et al., 2024). We advocate that these unique productivity hotspots couple pelagic and benthic processes and should be taken into consideration in systematic conservation planning actions (Levin et al., 2017). Strategies to elaborate these actions, however, often penalize areas where there could be economic losses, excluding them from priorities areas for conservation (Combes et al., 2021b). Because biological productivity and biodiversity uniqueness are often indissociable in the deep-sea (e.g., Rosa et al., 2008), area-based initiatives towards conservation of oceanic regions and the high seas may need to conciliate both properties (Briscoe et al., 2016; Karim and Cheung, 2024).