Tropicalization of demersal megafauna in the western South Atlantic since 2013

Catches throughout the study period reached maximum levels in 2006–2012, decreasing sharply thereafter reaching low levels in 2019 (Supplementary Fig. 1). The whitemouth croaker (Micropogonias furnieri) and the argentine croaker (Umbrina canosai) were the dominant species in the catches. Jointly, they represented, on average, over 50% of the total landed biomass in the period. This biomass included other 78 species: 62 teleosts, 3 elasmobranchs, 8 crustaceans and 5 molluscs. Overall, catch composition maintained a 1.5:1 ratio of species with warm- and cold-water affinities from the beginning of the time series until 2012. After that, warm-water species abundance increased in the catches changing the resulting ratio to 4.1:1 in 2019 (Fig. 2).

Mean temperature of the catches

Annual MTC oscillated around 21 °C (SD = 0.63 °C) between 2000 and 2019. Until 2013, the MTC time-series exhibited peaks (2005, 2010) and troughs (2008, 2013), but no particular trend was evidenced. After 2013, MTC increased continuously reaching maximum values in 2019 (Fig. 3). The segmented regression model defined one significant discontinuity in 2012 (95% CI: 2010–2015), which delimited an early period (2000–2012) when MTC oscillated with no significant trend (p-value = 0.789), from a late period (2013–2019) when MTC increased sharply at a 0.41 °C yr⁻¹ (p-value < 0.001) (Table 1). Similar catch warming trends have been described in Large Marine Ecosystems around the globe¹⁷ and in more limited regions including the Aegean and Ionian Seas^18,19, the Yellow and East China Seas²⁰ and the Bay of Biscay (Cantabrian Sea—NE Atlantic)⁴⁶. Considering the entire time-series, the MTC increase rate in the BMM was equal to 0.57 °C. decade⁻¹, exceeding estimates for the world ocean (0.19 °C. decade⁻¹) and for non-tropical regions (0.23 °C decade⁻¹)¹⁷, as well as for the regions above, except the Aegean Sea (1.01 °C decade⁻¹) and the Ionian Sea (1.17 °C decade⁻¹)¹⁹. During the 2013–2019 period, the decadal MTC increasing rate (4.11 °C decade⁻¹) largely exceeded any regional estimate reported, a pattern consistent with the expected ecosystem changes in a region of intense ocean temperature increase (Fig. 1)^23,24.

SBT values remained relatively stable from 2001 to 2010 but increased continuously from 2013 to 2019 (Fig. 3), as confirmed by the significant positive linear trend (0.077 °C yr⁻¹, p-value = 0.012) (Table 1). MTC variation was significantly explained by SBT with 0 and 1-year time-lag (p-value = 0.001 and p-value = 0.023, respectively) (Table 2).

These results were also consistent with global patterns¹⁷ but particularly relevant were the trends described in the Argentinean–Uruguayan Common Fishing Zone (AUCFZ, Fig. 1), where an MTC warming trend was described from 1985 to 2017²¹. This trend was explained by sea surface temperature (SST) variation which also increased steadily since mid-1990. Authors suggested that an important ‘oceanographic change has occurred in the region, which modulated the MTC index’. Because the MTC series analysed in the BMM is shorter (2000–2019) than the one analysed in the AUCFZ (1973–2017), and considered temperature at the sea bottom rather than at the surface (SST), a direct comparison between regions is not fully informative. However, the INALT20 model-derived SBT time-series, available since 1958⁴⁷ (Fig. 3), showed that positive anomalies became frequent from 1994 onwards, as observed in the AUCFZ SST time-series, indicating that such oceanographic change was noticed in both adjacent regions of the SWAO. Notwithstanding, in the BMM a noticeable steady increase of both MTC and SBT seems to take place from 2012 onwards, suggesting a more recent regional shift. If only this time period is analysed in the AUCFZ time series (Fig. 4 in ref. ²¹) positive anomalies would predominate approximately from 2010 onwards and years of maximum MTC anomalies would occur after 2014, coinciding with those identified in the BMM time series. This suggests that the signal of the second MTC shift is also present in the AUCFZ. Several studies have demonstrated that an SWAO general warming process is associated with the poleward expansion of the Brazil Current and the Brazil–Malvinas Confluence^26,48. Artana et al.²⁷ showed that this feature migrated southward between 1997 and 2006, oscillating widely thereafter (with southernmost positions in 1998, 2004, 2011, 2015 and 2017) and that Brazil’s Current transport volumes tended to increase from 1998 to 2016. These trends are consistent with (a) periods of intensified SBT positive anomalies in the BMM time-series, and (b) the positive effect of SBT and BCt on MTC variation. BCt affected MTC with a 3-year (p-value = 0.031) and 4-year (p-value = 0.029) time-lags (Table 2). The reason for such a delayed response is uncertain, but possible explanations could be related to the complex physical interactions between the Brazil Current and the BMM shelf waters.

Gianelli et al.²¹ argued that whereas many species in this dynamic transition region may be adapted to environmental oscillations, such a sustained oceanographic change would gradually provoke ‘unprecedented changes in the composition and structure of ecological assemblages’. The caveat here, however, is that fishing data may affect the MTC analyses in different ways, e.g. through market-oriented behaviour of fishing fleets, which tend to establish temporal and spatial strategies in pursuit of profitable concentrations of their main fishing targets⁴⁶. In the BMM, demersal fishing fleets explore a great variety of resources available in geographical space and different seasons in order to attain economic stability⁴². By doing so, they tend to integrate, in their catch composition, a wide spectrum of megafauna communities. However, trawl and gillnet vessels have developed a variety of métiers, i.e. a combination of target species, fishing area, gear, and time of the year^49,50, whose operational patterns in the BMM could partially modulate MTC variation. For instance, if fishing operations of a particular métier aiming at an abundant cold-water species predominated in relation to operations of other métiers during a year, an environment-independent MTC drop would be observed in such a year. We assessed these fishery-dependent effects in two ways. Firstly, an annual index of métier diversity (Dm, based on Simpson species diversity index) was computed and used to express the effect of dominance (low Dm values) vs. evenness (high Dm values) of métiers in the catches. Annual Dm did not exhibit any particular trend along the analysed time-series (p-value > 0.08, Table 1) and affected negatively MTC with a 4-year time-lag (p-value = 0.002, Table 2). This approach was first proposed by Gianelli et al.²¹, who found similar results in the AUCFZ, i.e. the effect of fishing métiers on MTC time-series was either non-significant or in an opposed direction to that exerted by the ocean temperature. Secondly, we submitted the MTC time-series to a species sensitivity analysis, showing that the accentuated positive trend of MTC, observed between 2013 and 2019, remained unchanged no matter which species we removed from the analysis (Table 3). Between 2000 and 2012, when no particular trend was evidenced in the MTC time series (Table 1), the estimated slope of the linear model increased, becoming significantly positive, when the codling Urophycis mystacea was excluded from the time-series (Table 3). This is an abundant slope species (mean thermal preference = 16 °C) whose catches remained above average between 2007 and 2013 (Supplementary Fig. 3), mostly through the activity of a double-rig trawl métier which included slope species in the period (DR_1, Supplementary Table 3).

In the AUCFZ, the exclusion of the most abundant argentine hake (Merluccius hubsi; cold-water affinity) and the whitemouth croaker (Micropogonias furnieri; warm-water affinity) from the catch time-series changed considerably MTC variation²¹. Important catch reductions of the former have been attributed to overfishing, which has also a potential for modulating MTC time-series. This can be the case of several cold- and warm-water species that largely contributed to BMM demersal catches during the studied period, whose exploitation regimes were categorized as unsustainable⁵¹. In this region, for instance, important abundance declines of the cold-water argentine croaker (Umbrina canosai, from 2010 onwards) and the monkfish (Lophius gastrophysus, from 2000 onwards) in the BMM (Cardoso et al., unpublished results) could modulate MTC leading towards a catch warming scenario. However, at least in the latter species, biomass levels have remained extremely low after 2010, a period when the fishing effort was maintained below critical levels, suggesting that factors other than fishing pressure (i.e. ocean warming) could be driving temporal patterns of the species abundance. In any case, as pointed out in other MTC study regions^19,20, overfishing effects on MTC seems hardly dissociable from, for instance, poleward retractions of cold-water species and, in fact, may have a synergistic effect.

Catch composition analysis

Changes in species abundances in the catches of the demersal fisheries in the BMM evidence strong contrasts between the early (2000–02) and late (2017–19) periods of the time-series. These periods were aggregated into two largely dissimilar year-groups by the multiple regression tree-principal coordinate analysis (Fig. 4), which discriminated an initial scenario (Group I), when annual catches were characterized by 15 main species, most of them with cold-water affinity (species on quadrants 1 and 2 in Fig. 4b), from a late scenario (Group IV) defined by scores attributed by eight species mostly with warm-water affinity (species in quadrant 3, Fig. 4b). Such a contrast was also corroborated by (a) the elevated contributions of these year-groups to the estimated total beta diversity, statistically significant in 2000 (14.2%), 2001 (10.8%), 2002 (9.2%) and 2019 (10.1%) (Supplementary Fig. 4), and (b) the Temporal Beta Diversity indices (TBI) comparing years within Groups I and IV, which resulted in significant losses in species abundance (Fig. 5). It is important to note that mean biomass gains and losses were significant between these groups (second largest among the periods compared), and dominated by warm- and cold-water species, respectively (Fig. 6, Supplementary Table 2). The argentine croaker (Umbrina canosai) concentrated 15.4% of cold-water species biomass losses in the catches, along with the argentine hake (Merluccius hubbsiI), the argentine stiletto shrimp (Artemesia longinaris), the monkfish (Lophyus gastrophysus) and others (Supplementary Table 2, Supplementary Fig. 3). The whitemouth croaker (Micropogonias furnieri) concentrated 25.6% of warm-water species biomass gains in the catches, followed by the grey triggerfish (Balistes capriscus) and the spotted pink shrimp (Penaeus brasiliensis) (Supplementary Table 2, Supplementary Fig. 2). Jointly, these biomass gains and losses contributed to a warming of the catches between the two extreme periods, and supported the process of tropicalization, as revealed by the MTC analysis.

Catch composition analysis also suggested that the important changes in the demersal assemblages, as proposed by Gianelli et al.²¹, may have taken place in the BMM between 2003 and 2012. Unlike in the MTC analysis, however, a more precise shift period was not evident. The progression of years in the 2-D ordination plot (Fig. 4) suggested a temporal modification in catch composition from the initial scenario (Group I), when cold-water species were abundant in the catches (species on quadrants 1 and 2 in Fig. 4b), to two intermediate scenarios (Groups II and III) when these species were gradually less abundant and substituted by other cold-water species (on quadrant 4 in Fig. 4b). TBIs calculated between years within Groups I and II produced three significant comparisons, two indicating losses (2000/2003, 2000/2006) and one indicating gains (2000/2007) (Fig. 5). There were important mean biomass gains between these two periods (Fig. 6), concentrated in the warm-water whitemouth croaker (M. furnieri, 39.0%) and striped weakfish (Cynoscion guatucupa, 10.3%), and the cold-water argentine croaker (14.9%) and codling (13.1%) (Supplementary Table 2). Biomass changes were limited between years within Groups II and III. Cold-water species, chiefly the codling (U. mystacea) and the argentine croaker (U. canosai), dominated both gains and losses of biomass, respectively (Fig. 6, Supplementary Table 2), but TBI comparisons indicated that these were not significant changes (Fig. 5). The last transition in the catch composition (Groups III and IV) was marked by great biomass losses mostly of cold-water species (Fig. 6), including the codling (U. mystacea, 16.3%), Argentine croaker (M. furnieri, 15.3%), the Argentine stiletto shrimp (A. longinaris, 9.7%), the Argentine hake (M. hubbsi, 6.9%), the monkfish (L. gastrophysus, 2.4%) and others (Supplementary Table 2). TBI comparisons also indicated species losses, but they were significant only in relation to the year 2019 (Fig. 5).

Interpreting such changes in demersal catch composition, in light of the warming trend in the SWAO, required prior consideration of the physical processes associated with known patterns of the spatial distribution of fish and shellfish populations⁴⁶. In the BMM, demersal fauna diversity tends to change from typically subtropical in the South Brazil Bight (23°S–28°S) to a mixed subtropical/warm-temperate towards the ‘central shelf’ subregion (south of 28°S, Fig. 1), which comprises the continental shelf area off southern Brazil, Uruguay and northern Argentina³³. In this subregion, off southern Brazil, Martins and Haimovici³⁹ described four teleost fish demersal assemblages formed by species with similar temperature affinities, whose latitudinal and bathymetric distribution are associated with seasonal interactions of coastal, subantarctic and subtropical shelf water masses. A ‘cold shelf assemblage’ was shown to expand over mid-shelf bottoms during the austral winter, as driven by the increased influence of subantarctic shelf waters and the northward displacement of Subtropical Shelf Front. This assemblage contained some abundant cold-water species present in the demersal catches, including the argentine croaker (U. canosai) and the argentine hake (M. hubbsi), which have accounted for important biomass losses (>20%) in the BMM. In addition, a ‘coastal’ and a ‘warm shelf’ assemblages were shown to expand southwards over the shelf during the austral summer. These assemblages contained fish species with warm-water affinity, including the whitemouth croaker (M. furnieri), which alone accounted for 25% of biomass gains in demersal catches. In an ocean warming scenario, induced by the southward displacement of the Brazil Current and its influence over the shelf, a southward retraction of the ‘cold water shelf’ assemblage and expansion of the ‘coastal’ and ‘warm water shelf’ assemblages would be expected, justifying the observed MTC trends and temporal patterns of species abundance in the BMM demersal catches.

Deviations from this general pattern, however, were also characterized partially because targeted species may display different levels of adaptation and respond differently to a warming environment¹³. For instance, the whitemouth croaker (M. furnieri) contributed significantly to warm-water species biomass gains in the period, but also to biomass losses. The species exhibits a complex stock structure that includes three spatially delimited stocks: one occupying the South Brazil Bight (‘Southeastern Brazil Stock’) and two extending over the central shelf off southern Brazil (‘Southern Brazilian Stock’) and at the AUCFZ (‘common Argentinean—Uruguayan stock’)⁵². Despite its warm-water affinity, the species exhibits a wide thermal tolerance, making it plausible that these stocks display some level of adaptation to local conditions and respond differently to the ocean warming process they have been exposed to in the SWAO. In addition, the La Plata River and the Patos/Mirim Lagoon systems are important nursery grounds for this species that may also respond to other climate-change-related effects such as freshwater discharge variability in these systems²¹. In that sense, the general tropicalization scenario characterized in the BMM demersal catches may be affected in different ways by multiple specific population processes operating at smaller spatial scales. An additional source of error may derive from thermal preferences being attributed to demersal species from sea surface temperature distribution¹⁷. How much this affected MTC temporal trends in the BMM is uncertain but it was likely dampened by (a) the large proportion of catches originating from species in shallow shelf areas where temperatures are more homogeneous throughout the water column and (b) a positive correlation between SST and SBT values derived from the INALT20 model (r = 0.84) suggesting that, regionally, MTC calculated with surface temperatures should still be related to thermal changes on the seafloor. Finally, annual MTC may also be modulated by spatial factors (e.g. latitude, depth) and trophic levels^22,46. Addressing these components is critical for a more comprehensive analysis of temperature-related community changes in the BMM.

Notwithstanding such limitations, historical catch data has proven to be an effective proxy for global climate effects on marine ecosystems regionally, with the advantage of further signalling future changes in the economic performance of current fishing regimes. How will the demersal fishing industry adapt to changes in the availability of traditional and non-traditional targets in the BMM? Which métiers will no longer be viable and which ones may emerge to explore expanding stocks of subtropical species? What adaptive measures can be incorporated into fishing management regimes (both national and transnational) to attain ecological and economic objectives in the coming decades? These are critical questions that could influence industry adaptive strategies and guide management measures over the next decades in the BMM, but whose answers will require extended analyses, with a database expanded to include spatial components and fisheries economic descriptors.

As a preliminary approach, however, we may infer that most métiers have included cold-water species among their most frequent targets (Supplementary Table 3) and will potentially face the need to adapt their fishing strategies. Particularly relevant seems the future performance of double-rig trawlers (métier DR_1) that have largely dominated the demersal fishing activity in the study period (Supplementary Table 1). Their operations, however, have long been driven by a generalist fishing behaviour and will likely be adaptable to new opportunities, as seen in the past⁴⁹. Gillnet fishing (métier GN_1) was the second most active demersal fishing activity heavily focused on the whitemouth croaker (Supplementary Table 3). Whereas this may appear as a winner species under an ocean warming scenario, it is uncertain how long could it sustain fishing pressures increased by the effect of the progressively scarce cold-water target-species (e.g. métier GN_2 and others)⁵³. Trawling for the warm-water sea-bob shrimp (Xyphopenaeus kroyeri – métier DR_2) can be an opportunity for trawlers in the future, especially those targeting cold-water shrimps (A. longinaris and Pleoticus muelleri—DR_3) (Supplementary Table 3). Such a perspective, however, will likely be limited by the increased competition with artisanal fisheries that have long exploited this coastal species in the BMM⁵⁴. Finally, the likely decline of the argentine croaker in the catches may threaten the future viability of trawl (métiers PT_2, ST_2, ST_3) and gillnet (métier GN_2) operations that currently concentrate in the central shelf region. These catches have been mostly sustained by a population that migrates between the northern Patagonian shelf and southern Brazil⁵⁵ suggesting that, in the future, these métiers could only be viable south of Brazilian waters. These are some possible scenarios that indicate how demersal fisheries can be reshuffled at the BMM, calling for adaptive management that will require rigorous assessments of sustainable catches of subtropical species, and include actions that redistribute effort in areas that take into account species with changing distribution patterns, including transboundary stock management⁵⁶.

Source: Ecology - nature.com