4.1. Northern spatial shift of small pelagics
The northward shift in S. aurita distribution observed over the two decades of sea surveys was concomitant with the important catch increase recorded in Morocco at the end of the studied period in 2015 (FAO, 2016). As for many marine species worldwide (Perry et al., 2005; Cheung et al., 2013), we hypothesize that the exceptional SST warming likely forced this shift observed south of Cape Blanc, the strongest ever recorded in the tropics during the 1982–2021 period (Fig. S1). These facts could account for the exceptional presence of S. aurita up to Casablanca during the Moroccan-led small pelagic fish assessment in 2015. The DFN scientific catch data indicate a northward shift of 180 km for S. aurita along the studied period. The synergistic effects of climate variability and change on species fitness (Lima et al., 2022) mediate small pelagic habitat suitability. It was well-established that short-lived species, like small pelagic fish, are highly sensitive to short-term environmental variations, and these fluctuations play a crucial role in shaping their distribution in the immediate term (Fréon et al., 2005). However, the significance of considering the impact of long-term environmental changes on small pelagic fish habitats must be recognized. Over time, gradual environmental modifications can influence these species' overall habitat suitability.
Similar to S. aurita, five additional small pelagic fish species displayed significant northward trends in their distribution, while two others, including S. maderensis, exhibited no significant trend. As deduced from its acoustic-based biomass, S. maderensis displayed no significant shift in its northern limit and overall distribution. Given that all surveys were consistently conducted during the fourth quarter of the year, it is unlikely that the observed shifts are attributable to irregular survey timing or a regional increase in stock biomass (FAO 2018; Fig. S2).
The assessment of latitudinal shifts in the fishery has previously employed calculating a monthly catch center of gravity (Zeeberg et al., 2008). However, it is important to note that the use of catch rates per vessel as an index of fish abundance could be disputed. This is particularly true in Mauritania waters, where the fishery is primarily industrial. The catch rates depend on fish abundance and are influenced by catchability (Brehmer and Gerlotto, 2001) and socio-economic factors (Ba et al., 2017). Some stochasticity must be expected in the observations from the sea surveys; this means that the species' northern limit was not necessarily observed. The records from DFN sea surveys were based on trawls carried out to identify all pelagic fish schools. The watch officer started scientific fishing operations only when significant acoustic densities of pelagic fish were observed. Consequently, low densities or fish presence inshore in the survey area have probably been overlooked (Brehmer et al., 2006; David et al., 2024).
Any shift in the distribution of sardinella and other small pelagic stocks could lead to significant economic and social instabilities. In Mauritania and southern Morocco, fisheries are responsible for a significant part of the gross domestic product, particularly as major investments have been made in establishing fishmeal factories (Corten et al., 2017; Deme et al., 2023) to process this species. Recently, the pressure on these stocks has increased because of the emergence of fish meal factories and the arrival of semi-industrial vessels, i.e., purse seiners and trawlers (Braham et al., 2024). In Senegal during the BSS period, no specific change in the focus on targeted species has been observed because both sardinella remain the most important species in biomass, even though local authorities have supported a pronounced alternative policy in favor of fish farming over the past years. At the sub-regional level, due to the various threats to the fish resources, a growing concern in all national fisheries centers, is to set up common monitoring of transboundary exploited fish before the beginning of a major fisheries conflict (Spijkers et al., 2018).
4.2. Spatial considerations
The Cape Blanc boundary at around 21°N roughly corresponds to the position of the Cape Verde Frontal Zone (Zenk et al., 1991) that occurs in November off Mauritania and separates the North and South Atlantic Central Waters in the eastern North Atlantic Subtropical Gyre (Martínez-Marrero et al., 2008). Until recently, this location has been regarded as a “faunistic limit” for many benthic (Intes and Le Loeuff, 1984), planktonic (Weikert, 1982), and ichthyoplanktonic (Hamann et al., 1981) populations. It was also considered the northernmost distributional area for several small pelagic fishes in northwest Africa (Binet, 1988). DFN sea surveys now indicate several key species were moving beyond this boundary.
The observed warming patterns appear to significantly influence the spatial restructuring of small pelagic fish populations, including S. aurita. Notably, the northern distances observed in the spatial shift of the pelagic fish (between 40 and ~ 200 km) align closely with the magnitude of the isotherms displacements (40–145 km) since 1995 (Fig. 5; Tables S3). SST is likely a primary driver behind the movements of S. aurita, as suggested by previous studies (Sabatés et al., 2006; Montero-Serra et al., 2015; Tsikliras, 2008). It is important to note that the observed changes are not solely attributed to SST; various physical and ecological factors, including regularly increasing winds and decreased primary productivity south of Cape Blanc, contribute to this transformative shift.
The trend in S. aurita biomass showed that the portion of the stock north of Cape Blanc rose from 38 to 65% from 1995 to 2015 (Fig. 7), with a significant difference (p < 0.001) between the first and the last part of the time series. Notably, the northern part (Fig. 4b, area 1), characterized by a strong summer coastal upwelling (Benazzouz et al., 2014), was the region with one of the greatest increases in SWS intensity across the entire Canary Current ecosystem over the last 34 years, including the BSS period. To examine the hypothesis of a northern population displacement (Fig. 2a, Fig. 7) triggered by strong and/or rapid environmental changes, particularly the quick rise in SST observed in 1994-95 and the intensified upwelling in 1998 (Fig. 4b and 4d), we employed two distinct bootstrap procedures. Across all analyses, we observed consistently significant to highly significant changes in SWS and upwelling index in all areas. In contrast, no change was recorded for SCC except in area Senegal (area 5) and SST exhibited significant change only in LEM period and not BSS one (Table S4).
While the entire CCLME was notably impacted by long-term warming (Fig. S1), the heightened upwelling in the northern part of the region (North of Cape Blanc) serves to mitigate coastal heating and maintain productivity. Consequently, this area offers more conducive habitat conditions and a relatively stable thermal environment for small pelagics. In stark contrast, the southern part of the system (area 5, Senegal) experiences a modest increase in SST coupled with a decline in phytoplanktonic biomass (Fig. 4). This observation aligns with the findings of Gomez-Letona et al. (2017) over a shorter period (2002–2016). It implies that this region is likely becoming a less productive habitat for some species.
The acoustic biomasses of S. aurita and S. maderensis during the BSS period (Fig. S3) showed that S. aurita biomass was decreasing, with the lowest value observed for the last year (2015). During this 21-year period, the average SWS north of Cape Blanc was higher than 7 m s− 1, well over the optimum environmental window of 5 m s− 1 defined for the recruitment success of small pelagic fishes (Cury and Roy, 1989; Diankha et al., 2018). Therefore, environmental conditions are likely not favorable for all species according to their reproductive success (Balde et al., 2019).
4.3. Ecological responses
The absence of a significant shift in the distribution of S. maderensis might be explained by their high physiological adaptability to environmental disturbances (Ba et al., 2016) and the absence of rivers or estuaries in the West Sahara (areas 2 and 3) providing a natural habitat for its life cycle. In contrast, S. aurita, while opportunistically taking advantage of primary enrichment (Roy et al., 1989; Diankha et al., 2018), is more sensitive to warming, making it less adaptable to strong and continuous changes in ocean temperature as reported south of Cape Blanc. This difference in habitat shift can be explained by the environmental sensitivity of S. aurita, which exhibited less phenotypic plasticity (Baldé et al., 2019) than S. maderensis (Ba et al., 2018) and, therefore, more prone to spatial adaptation.
Brachydeuterus auritus, Chloroscombrus chrysurus, Selene dorsalis, and Sphyraena guachancho are known to be present in brackish waters during their juvenile stage (Sloterdijk et al., 2017). For this reason, their northern border of distribution remains mainly south of the Cape Verde Frontal Zone and notably further south than those of the other species considered in this study (Fig. 3) that thrive in the richest part of the coastal upwelling, beyond the transition zone between southern and northern CCLME. Especially Selene dorsalis did not move significantly. This species is considered semi-demersal, at least the least pelagic among the eight selected in this study, with a diet based on zoobenthos as nekton (Diouf, 1996).
The spatiotemporal shift in low thermal tolerance species, such as S. aurita, off Mauritania appears primarily controlled by thermal constraints rather than water-enrichment gradients (Thiaw et al., 2017). Alternatively, the shift in these species might be related to expanding and raising of oxygen minimum zone (Stramma et al., 2008; Machu et al., 2019), occurring on the shelf to the south of the system. Such changes in water temperature and dissolved oxygen levels might influence trophic relationships, leading to major changes in the composition of planktonic prey species. Sardinella aurita is an opportunist feeder (Fréon, 1988), primarily feeding on phytoplankton (mainly diatoms) in offshore waters (Nieland, 1982), but also copepods, which are the two most abundant phytoplankton and zooplankton groups in the study area. Any alteration in the plankton community can disturb the whole food web (Beaugrand and Reid, 2003), influencing the small pelagic fish community, their abundance, and distribution (Brosset et al., 2016; Brodeur et al., 2018). However, the observed shift in S. aurita beyond 26°N was not related to the primary production level, as there was no significant increase in SSC to the north of Cape Blanc (Table S4). The complex interplay between trends in SWS and planktonic productivity (Fig. 6b, c) underscores indirect biogeochemical processes and variations in planktonic compositions that likely vary seasonally, consistent with Bakun’s theory of summer warming (Bakun et al., 2015). Identifying the determinism of the phenological shift is challenging, as SST increase can modulate population characteristics, predator-prey interactions, intra- and inter-specific interactions, and induce potential physiological changes in marine organisms (Little et al., 2020).
Similar to S. aurita, five other commercial species showed significant northward trends in their distribution, both in the DFN data and in the Moroccan acoustic surveys, which detected exceptionally high densities of small pelagic fish off Casablanca (33.5°N) in 2015 (Mamza et al., 2015). Among them, C. chrysurus, B. auritus, and S. guachancho have been regularly observed north of 20°N since 2003. Spatial shifts in species assemblage within West Africa are likely to impact ecosystem composition, productivity, fishing endeavors, and national economies. With additional warming, it was hypothesized that species currently constrained by their limited northward distribution might expand their range further northward due to a scarcity of estuarine habitats. Consequently, an abrupt northward displacement in the distribution of both S. maderensis and S. dorsalis could be anticipated, along with the possibility of a stock collapse, given their reliance on brackish water during various life cycle stages.
Finally, for the whole sub-region, small pelagic fish, because of their abundance, partly regulate the trophic dynamics of the upwelling ecosystem, in a wasp-waist trophic control (Cury et al., 2000). As a result, any change in their abundance might disturb the whole food web, triggering unknown changes in biomass production and species composition, which, in turn, might impact the entire fisheries sector.