Estuaries are threatened by intense and continuously increasing human activities. Here we estimated the sensitivity of fish assemblages in a set of estuaries distributed worldwide (based on species vulnerability and resilience), and the exposure to cumulative stressors and coverage by protected areas in and around those estuaries (from marine, estuarine and freshwater ecosystems, due to their connectivity). Vulnerability and resilience of estuarine fish assemblages were not evenly distributed globally and were driven by environmental features. Exposure to pressures and extent of protection were also not evenly distributed worldwide. Assemblages with more vulnerable and less resilient species were associated with estuaries in higher latitudes (in particular Europe), and with higher connectivity with the marine ecosystem, moreover such estuaries were generally under high intensity of pressures but with no concomitant increase in protection. Current conservation schemes pay little attention to species traits, despite their role in maintaining ecosystem functioning and stability. Results emphasize that conservation is weakly related with the global distribution of sensitive fish species in sampled estuaries, and this shortcoming is aggravated by their association with highly pressured locations, which appeals for changes in the global conservation strategy (namely towards estuaries in temperate regions and highly connected with marine ecosystems).
Estuaries are highly productive and valuable ecosystems1, albeit not especially diverse. But their functioning and services are threatened by continuously increasing human activities2 while the coherence between estuarine biota sensitivity, threats and conservation is poorly known.
Anthropogenic activities have caused loss of estuaries’ areas and connectivity with adjacent ecosystems, habitat loss (e.g. wetlands) and degradation (e.g. water quality), depletion of important species and accelerated species invasions3. Global fish catches in estuaries (and sea) have been increasingly dominated by less vulnerable species while more vulnerable fish became over-exploited or depleted due to “fishing down food webs” processes4. But disentangling human- from naturally-induced changes in estuarine biodiversity is complex since estuaries are naturally dynamic and stressed (i.e. Estuarine Quality Paradox)5. Estuaries are intrinsically linked with marine and freshwater ecosystems, and their fish assemblages include resident species, frequent migrants or occasional stragglers from adjacent ecosystems, as well as migratory diadromous species6, 7. Thus estuarine communities are potentially impacted by human activities in and around estuaries which affect species that colonize estuaries and their environmental conditions8, 9.
The susceptibility of a system or community to perturbation increases with an increase in its exposure and intrinsic vulnerability and with a decrease in its protection or adaptive capacity10, 11. Over a third of the world’s oceans show medium-high to very-high level of human activities and pressures, with coastal ecosystems (0–200 m) showing high levels of both land- and ocean-based anthropogenic activities and pressures12. Human-driven impacts on marine fishes include for instance lower biodiversity (i.e. taxonomic, functional and phylogenetic)13 and biomass14, 15, as well as changes in composition like less large-body and low resilience marine fishes16, 17. Similarly worrisome are freshwater ecosystems, with 65% of global river discharge and habitats under moderate-to-high levels of activities and pressures, explaining the global freshwater biodiversity crisis18. For instance, freshwater ecoregions with lower percentages of free-flowing distances show lower percentages of endemic freshwater and diadromous fishes19.
Despite the commitment of nations worldwide to protect ecosystems (Convention on Biological Diversity and Aichi Biodiversity Targets), and an increase in protected land and sea in the last half-century (doubling each decade)20, there are serious aquatic conservation shortfalls. Many protected areas exist in name only (“paper parks”) and are not (or are insufficiently) managed. Although estuaries represent most of coastal ecosystems, only a low number of estuaries and percentage of their areas is protected, and the coverage of sensitive assemblages is mostly unknown and insufficient, at least in some regions21, 22. Meanwhile, conservation of marine fishes (which are dominant in estuaries) is deficient, for instance marine protected areas currently provide low coverage for most species and their evolutionary history, including those with high taxonomic and functional sensitivity10, 23, 24.
Worldwide, estuarine fish assemblages show strong spatial patterns and environment relationships. Biogeographical region and environmental features of estuaries (e.g. temperature, connectivity and area) regulate patterns of species richness25, 26, composition and functional traits27, 28. Functional traits determine the way species use resources and their tolerance to environmental conditions. For instance, body size is a key trait as it directly relates with other traits such as mobility, trophic interactions, age at first reproduction and rate of population growth29. Moreover body size in aquatic ecosystems is globally unevenly distributed27, 28, 30. Therefore, we may also expect global patterns in such covarying traits and in the resulting species sensitivity and response to changes, which can be measured for instance with: ‘species vulnerability’ (species intrinsic extinction vulnerability to fishing)31, and ‘species resilience’ (species productivity or resilience to fishing)32 - both based on ecology and life history traits (Table 1).
Here, we analyse for the first time the susceptibility of biodiversity in estuaries to stressors from multiple ecosystems, at a worldwide extent. Our aim is to identify global conservation pitfalls in estuaries by assessing whether more vulnerable and less resilient fish assemblages are associated with particular environmental features, high levels of human pressures and/or low levels of protection, potentially making them more susceptible to disturbances. To this end, using publicly available data and spatial analysis, for a set of estuaries distributed worldwide (Fig. 1) we characterized the fish assemblages (Supplementary methods and Table S1), the vulnerability and resilience traits of their species (Table 1) as well as ecosystem features (Supplementary methods and Table S2). We also estimated intensity of human activities and pressures, as well as protection in and around each estuary (i.e. for marine, estuarine and freshwater ecosystems due to their inherent connectivity; but separately per ecosystem). Finally, we used correlations, linear- and linear mixed models to: identify links between environmental conditions, human pressures and protection worldwide; and to assess if more vulnerable and less resilient assemblages occur in estuaries with particular conditions and if they are exposed to low levels of protection and high levels of exposure to human pressure worldwide.
Results and Discussion
Briefly, this study shows that the vulnerability and resilience traits of fish from the sampled estuarine assemblages worldwide are particularly associated with geography and certain environmental gradients (Fig. 2; Table 2). Additionally, in short, exposure to human driven pressures and extent of protection in sampled estuaries are inversely related and are also heterogeneously distributed across the globe (Figs 1 and 3; Table 3), with many estuaries with intense human pressure supporting more vulnerable and less resilient fish species, despite no strong relation with extent of protection (Fig. 2; Table 4). In all, results provide insight into global conservation needs of estuaries based on the vulnerability and resilience of their fish assemblages.
Results show that human pressure directly in sampled estuaries (Hestuary) and in the adjacent marine and freshwater ecosystems (Hmarine, Hfreshwater) are moderately correlated (pairwise Pearson correlations 0.38–0.52), and strongly correlated with mean intensity from these three ecosystems (pairwise Pearson correlations 0.73–0.84 with Hmean, and 0.54–0.95 with Hweighted-mean) (Fig. 1; Table 3). It is acknowledged that human pressures are intense in coastal zones, especially near heavily populated zones33,34,35,36, and estuaries are directly exposed due to location and with their surrounding ecosystems impacted by cumulative aggregated activities12, 37. The data used represent a set of human activities and human-induced pressures, therefore they only indirectly inform on human-induced impacts, due to possibly different mitigation measures and local context. The current analysis could benefit from a higher spatial resolution of data on human pressures38, but such data are not available for all regions across this global extent. Still, our results indicate that the fish assemblage in a given estuary tends to receive a similar degree of exposure to human pressures from these three ecosystems. This represents an added challenge to the already complex conservation and spatial planning of estuaries9, namely in scenarios with high pressures in estuary and surrounding ecosystems.
Furthermore, our results also indicate that existing conservation efforts in sampled estuaries are partially related with those in the surrounding ecosystems: i.e. the extent of coverage by protected areas (PA) from the three ecosystems is moderately correlated (pairwise Pearson correlations 0.11–0.60 for PAall and 0.32–0.57 for selected PAI–IV, lower between marine and freshwater ecosystems) (Table 3). The observed relation could be generated by a consistency of conservation policies: the level of conservation policies/investments in a given country/region/continent seems roughly similar across ecosystems, which can be due to socio-economic and political context. Still, potential protection should not be confounded with realized protection. Benefits from marine protection are acknowledged to depend on protection level and effectiveness which is influenced by factors including enforcement, stakeholders engagement, presence of no-take zones, surrounding human pressure, size, isolation and age39, 40. Furthermore, connectivity and interactions across the marine-estuarine-freshwater gradient should be central in conservation planning, since estuaries are known to assemble fish species from these ecosystems (marine species often dominant6, 7) with connectivity essential for maintaining species life cycles and ecosystem functioning9, 41, 42. But despite increasing management plans that include continuous protected areas across more than one ecosystem, conservation across land-river-marine realms and across freshwater-estuarine-marine realms are insufficiently applied36, 42.
In the set of sampled estuaries, there is a decoupling of the coverage by protected areas and intensity of human pressure: within each ecosystem, there is −0.14 to −0.49 pairwise Pearson correlation between protection and human pressure (Table 3). But this global negative relationship between protection and human pressure is not present in all continents, it arises amid considerable variability (Supplementary Fig. S1) and may be due to the current set of analysed estuaries - since different continents show disparate intensity of human pressures and extent of coverage by protected areas (Fig. 3), distribution and number of samples (due to uneven distribution of adequate fish assemblage data for this analysis; Supplementary Table S2). This advises caution in interpretation of the obtained global patterns. Still, the sampled estuaries covered all continents and the full spectrum of pressure intensity (from very low to very high as described in the work by Halpern and colleagues for marine ecosystems12) reinforcing the present results. Indeed, it is recognized that many protected areas are placed in zones with intense human pressures33, 35, 36, but large marine and land reserves are also often strategically placed in zones where conflicts with multiple human activities are minimized in advance (such as marine offshore zones, or higher and unproductive lands), which might decrease effectiveness of protected areas43, 44. Nevertheless, it has been previously shown that coastal marine reserves (usually more exposed to human threats) are as effective in protecting biodiversity as those placed offshore or in less-developed locations43, therefore protecting locations highly exposed to human pressures should be pursued in a global conservation strategy.
Intensity of human pressures in and around sampled estuaries shows a clear geographical pattern (higher in Europe and Asia, intermediate in Africa, North America and South America, and lower in Oceania), regardless of the variable considered (Hmarine, Hestuary, Hfreshwater, Hmean or Hweighted-mean) (Fig. 3). This pattern reflects the mean human population density of continents (http://data.worldbank.org). Meanwhile, percentage of coverage by protected areas is higher in sampled estuaries of Oceania and lower in Africa and Asia. The difference between estuaries in different continents observed here reflects the known vastly divergent protection regimes for marine ecosystems (0–200 nautical miles) implemented in those continents (with Oceania notably standing out), and are less akin to terrestrial protection (which is broader in Central and South America)20. Here, the difference in protection of sampled estuaries between continents is especially evident for protected areas of I–IV IUCN management categories (selected PAI–IV), which are known to represent approximately 25–55% of protected areas in all continents, but dominate in Oceania (around 85%) and are scarce in Africa (around 10%)20. It is widely acknowledged that despite efforts to reach conservation targets (10% of sea and 17% of land by 2020 - Convention on Biological Diversity), efforts are geographically imbalanced globally, and coverage by marine no-take areas is reduced (0.08%) and many or even most protected areas are inadequately managed, lacking an integrative network design45.
Overall, there are contrasting scenarios of protection and human pressures in sampled estuaries across the globe, notably (Fig. 3): a) high human pressure and medium protection of estuarine fish assemblages in Europe (high coverage by PAall, but low coverage by selected PAI–IV); b) high pressure and low protection of estuarine fish assemblages in Asia; c) low pressure and high protection of estuarine fish assemblages in Oceania. The low coverage of sampled estuaries by protected areas with stricter measures (PAI–IV) observed in most continents, especially in estuaries with intense human pressure, highlights a likely conservation shortfall regarding many estuarine fish assemblages (as seen in fish assemblages of other aquatic ecosystems10, 24), and argue in favour of urgently revising management and conservation plans.
In our dataset, estuaries in higher latitudes (and with lower temperature) tend to have higher intensity of human pressure, but the latitude cline is weakly and ambiguously related with the coverage by protected areas (Table 3). The observed latitudinal increase in human pressure in and around estuaries resembles the reported latitudinal increase in GDP per capita, but contrasts with the acknowledged latitudinal decrease in population density46. Nevertheless, this latitudinal trend should be viewed with caution - in our study, the location of sampled estuaries is imposed by fish assemblage data, and scarce data on fish assemblages in some regions results in a smaller representation of such regions (e.g. tropical and subtropical Asia where human pressure is often high, and some very high latitude regions where human pressure is low12); whereas human pressures in rapidly developing regions may be underestimated. Additionally, results show that sampled estuaries with higher connectivity with the marine ecosystem (tidal regime, estuary type, mouth width) and larger area (of the estuary and drainage basin) tend to have higher human pressure (especially Hestuary), likely because they attract larger human populations, but they also have higher protection (when considering PAall), possibly due to higher conservation obligations.
Fish species ‘vulnerability’ and ‘resilience’ traits are inversely correlated in the surveyed estuaries (Table S3), since both are based on life history and ecological characteristics (with four shared parameters)31, 32. Moreover, body size is used to parameterize vulnerability, and several parameters in vulnerability and resilience are acknowledged to covary with size29, resulting in size, here, being positively correlated with vulnerability and negatively with resilience (Table S3). For instance, geographical range size of marine fishes has been shown to increase with adult (e.g. size, schooling behaviour) and larval traits (e.g. pelagic larval duration) that together affect dispersal and post-dispersal persistence of new populations30, 47. Similarly, larger species have been reported to have higher fecundity, older maximum- and first maturity-age [also lower von Bertalanffy growth coefficient (K)] and slower intrinsic population growth rate29, 48.
Sampled estuaries worldwide are on average dominated by species with low vulnerability, and high to medium resilience (Table 5) (and concurrently by species with small to medium maximum size27). Accordingly, it has been shown that abundance decreases with the increase in body mass for trophic webs generally29 and that opportunistic and periodic life-history strategies dominate in European estuaries, where equilibrium strategy is rarer (large generation time and age-specific survivorship, small fecundity, chiefly marine stragglers)28.
Moreover, vulnerability and resilience traits differ among sampled estuaries globally, and their distributions relate with environmental conditions (as shown with linear- and linear mixed models) and with human pressures (as shown with Pearson correlation). Explicitly, opposite relationships are evident in fishes with different degrees of vulnerability (namely low versus other higher categories); as well as in fishes with different degrees of resilience (namely high versus other lower categories) (Fig. 2; Table 2). The proportion of fishes with low vulnerability (and high resilience) decreases in estuaries in Europe, and markedly in estuaries that are from higher latitudes (lower temperatures) and that have higher connectivity with the marine ecosystem (wide tidal amplitude, and in permanently open estuaries) (Fig. 2; Table 2). Meanwhile, the inverse spatial pattern and trait-environment relationship occurs for fishes with higher vulnerability (and lower resilience) (Fig. 2; Table 2). This global pattern of vulnerability and resilience mirrors the global pattern previously observed for body size in these estuaries27 due to trait covariation. Several mechanisms have been proposed for the latitudinal and temperature cline in marine fishes body size49, including energetic and biotic advantage of smaller fish at higher temperature versus larger fishes at lower temperatures30. Simultaneously, it is known that estuaries with less connectivity with the marine ecosystem hinder colonization by marine fishes50 (which tend to be larger than freshwater fishes in this database27, with both marine and freshwater fishes in European estuaries previously reported as larger than residents28). The observed link (of fish vulnerability and resilience with environmental conditions in estuaries worldwide) is further supported by previous regional evidence that life-history strategies relate with environmental conditions in estuaries and river basins, these strategies being consistent with climate regime and historical events (chiefly stability of suitable conditions)28, 51. Overall, present results should be seen as a first attempt to identify the main current conservation concerns for estuaries at a global extent, although the observed trait patterns and trait-environment relationships might be influenced by some data limitations (i.e. spatial differences in availability of assemblage studies, and of survey sampling method, effort and coverage of estuarine habitats). Still, the observed patterns seem broadly supported by their compliance with above-mentioned previous studies.
The currently known global pattern of fish body size in estuaries is based on inter-species variability27, but analysing intra-species variability (i.e. size-frequencies per estuary) would expand knowledge of trait-environment relationships, since estuaries are typically nurseries. Additionally, vulnerability and resilience traits measured here are species-specific30, 31 but developing size-specific vulnerability traits would allow considering influence of life-stage and size on response to disturbances.
Our approach revealed that intense anthropogenic pressures in and around sampled estuaries overlap many estuarine fish assemblages with higher sensitivity traits, and this occurs in: estuaries in particular regions (high latitude, especially Europe) and estuaries with certain environmental features (high connectivity with the marine ecosystem - open and with wide tidal amplitude), as modelled (Fig. 2; Table 2). Sampled estuaries with greater human pressure tend to have species with higher vulnerability and lower resilience (Fig. 2; Table 4). This overlap raises some global conservation concern, especially since in and around the sampled estuaries the percentage of coverage by protected areas slightly tends to decrease with the increase of human pressure (Table 3), but is poorly related with assemblage sensitivity (Fig. 2; Table 4). Also concerning is the small percentage of coverage provided by protected areas with IUCN management categories I-IV (PAI-IV, that restrict human activities and more likely benefit biodiversity) in sampled estuaries of most regions - except Oceania. A mismatch between protected areas and desirable conservation, aiming at preserving global fish biodiversity, has been also reported for other aquatic ecosystems. For instance, there is poor protection of distribution range of most marine species24, and of impacted marine zones with high endemism52, biodiversity35 or high taxonomic and functional sensitivity (although not species rich)10.
Taxonomic biodiversity, especially hotspots52, 53 is prominent in conservation since it is acknowledged that maintaining high species richness expectedly improves community resilience to environmental stress, and conserving endemism presumably safeguards genetic variability, with both likely preventing biodiversity erosion52, 54. In contrast, little attention is given in conservation to traits and functional diversity, despite their role in maintaining ecosystem functioning and stability55. Although estuaries are not typically highly taxonomically diverse, they support high productivity and ecosystem services1, 2, 9, and therefore have high conservation value. Globally, estuarine fish species richness is known to increase towards the equator (which is a general ecological rule56) and in open systems57, advocating the conservation value of systems with those characteristics. Present results reinforce the value of estuaries with high connectivity with the marine ecosystem and in addition support the conservation value also of temperate estuaries, as species in those systems have higher vulnerability and lower resilience. Moreover, sampled estuaries in higher latitudes are more exposed to human pressure but not especially covered by protection. Results highlight that, in many regions, efforts are needed to apply effective conservation measures within existing protected areas since little coverage is provided by protected areas with IUCN management categories I-IV.
Global conservation strategies should cover a network of locations/habitats (spatially nested within biogeographical regions) that protects several aspects of biodiversity - e.g. range rarity, low species resilience and high endemicity, taxonomic/functional diversity and sensitivity, as well as species vulnerability. Such strategies should consider that biogeographical region and ecosystem features regulate estuaries’ species richness25, composition58, functional traits27, as well as vulnerability and resilience (present study). Still, further research is needed for prioritizing particular sites, especially to account for effects of habitat complexity at local scales59 and on links between taxonomic and functional diversity.
Materials and Methods
We compiled a comprehensive database for estuaries distributed worldwide (Fig. 1) on (a) fish assemblage composition in estuaries (Supplementary methods and Table S1), (b) ecosystem features of the sampled estuaries (Supplementary methods and Table S2) and (c) traits of the sampled fishes (Table 1). The database included 2434 taxa for 378 estuaries worldwide. Since estuaries are transition ecosystems we characterized human pressures as well as protection in and around each estuary (i.e. for marine, estuarine and freshwater ecosystems, but characterized separately per ecosystem).
Fish assemblage data
We compiled a database of studies of fish assemblages in individual estuaries, aiming at a wide characterization of each estuary’s fish community (i.e. we excluded studies on dominant/selected taxa) and habitats (e.g. subtidal, tidal flats, creeks) but in some cases a complete characterization of habitats was not possible. To minimize sampling effects of different gear types we considered only active gears (e.g. trawl-, seine-, cast- nets, or trap-like gears such as enclosure nets/traps) and considered only surveys where total sampled area could be estimated so that it could be used to minimize sampling effort bias in subsequent analysis. Moreover, some estuaries are represented by more than one study in the database, and when possible, an estuary’s fish assemblage reported in a given study was treated separately by type of survey. Therefore, in this database, a sample consists of the fish assemblage sampled in a given estuary and survey (530 samples in 378 estuaries).
For each taxa we characterized ‘intrinsic extinction vulnerability to fishing31, coded using five categories (Table 1): low (<30%), low to moderate (30–40%), moderate to high (40–60%), high to very high (60–70%) and very high (>70%). This aggregate trait was parameterized with maximum body length, age at first maturity, von Bertalanffy growth parameter k, natural mortality rate, maximum age, geographic range, annual fecundity and strength of aggregation behaviour31. We also characterized ‘species productivity or resilience to fishing’32, i.e. minimum population doubling time, coded using four categories (Table 1): high (<1.4 yr.), medium (1.4–4.4 yr.), low (4.5–14 yr.) and very low (>14 yr.). This aggregate trait was determined through intrinsic rate of increase, von Bertalanffy k, fecundity, age at maturity and maximum age32. Finally, we also characterized the maximum body size of each species (small: <15 cm; medium: 15–50 cm; large: 50–100 cm; very large >100 cm). Traits were recorded using information available in FishBase (www.fishbase.org) and additional literature. Trait values were not available for <10% of the taxa (i.e. genus or families), which accounts for a mean of 5% and 7% per sample, for vulnerability and resilience, respectively. However, this percentage is consistent across continents, except it is higher in Africa, and lower in North America than Europe, due to the percentage of taxa resolved at species level (ANOVA and Tukey HSD, P < 0.05).
To evaluate the preponderance of the different trait categories in estuaries, we determined the “relative taxa richness” of each trait category per sample: i.e. the proportion of the taxa richness of a given trait category (e.g. high resilience) relative to the total observed taxa richness (i.e. richness = number of taxa). We used proportions to standardize among assemblages with different number of taxa resulting from different sampling effort. Moreover, we used taxa richness rather than abundance, since abundance data are available for less estuaries and we previously showed27 they both describe these assemblages in the same way.
Biogeographical and environmental data
We determined a set of biogeographical and environmental variables for each estuary in the database (Supplementary methods and Table S2). Biogeographical location was characterized using continent and marine biogeographic realm60. Energy and productivity were described with latitude and temperature at the mouth of the estuary, and primary productivity of the adjoining marine and terrestrial ecosystems. Ecosystem size was described using area of the estuary and of the adjoining freshwater ecosystem (drainage basin), and continental shelf width was used as a proxy for the area of the adjoining marine coastal ecosystem. Hydrological connectivity of the estuary with the marine ecosystem was depicted with estuary type (open or temporarily-open), estuary mouth width and tidal range (macro-, meso- or microtidal). Finally, habitat suitability of the estuary was described in terms of salinity type (regular, regular-to-hyperhaline or hyperhaline).
Human activities and pressures data
We characterized the potential level of exposure to human activity and pressure of the fish assemblage in and around each estuary (from marine, estuarine and freshwater ecosystems). We used data on drivers with acknowledged effect on ecosystem degradation, from reliable data sources and available at suitable coverage and resolution. Exposure to pressure in the marine ecosystem was measured with the index of cumulative human impact developed by Halpern and colleagues12. This index is based on 17 anthropogenic drivers of ecological change representing four main aspects: general, climate change, fishing and pollution. For each estuary in our database, we determined exposure to human pressure in the marine ecosystem (Hmarine) as the mean index in the coastal marine ecosystem (i.e. shallower than 200 m depth - continental shelf) within an influence radius defined by the size of that ecosystem (with 20, 40, 125, 440, 600 and 980 km radius respectively applied to the 25th, 50th, 75th, 90th, 95th and 100 percentiles of continental shelf width in our database) (Supplementary Fig. S2).
To estimate exposure to human pressure in the estuarine ecosystem we used human population density around the estuaries. Human population density reflects a range of human driven impacts generated by multiple activities in and around estuaries (e.g. urban, industrial, rural, harbour, water use, resource exploitation) - for example, in previous studies Pearson correlation between human pressure and overall pressure was R2 = 0.6361 and R2 = 0.5162. For each site in our database, we quantified human pressure in the estuarine ecosystem (Hestuary) as the mean population density (data for year 2000; http://sedac.ciesin.columbia.edu/data/set/gpw-v3-population-density 63) within an influence radius defined by estuary area (with 1, 5, 10, 30, 40, 265 km radius respectively applied to the 25th, 50th, 75th, 90th, 95th and 100 percentiles of estuary area in our database) (Supplementary Fig. S2).
We evaluated exposure to human pressures in the freshwater ecosystem based on the cumulative incident threat index to river biodiversity developed by Vörösmarty and colleagues18. This index comprised 23 geospatial drivers under four themes: catchment disturbance, pollution, water resource development and biotic factors. For each estuary in the database, human pressure in the freshwater ecosystem (Hfreshwater) was quantified as the mean index within an influence radius covering the drainage basin area (with 15, 35, 85, 270, 315 and 1345 km radius respectively applied to the 25th, 50th, 75th, 90th, 95th and 100 percentiles of drainage basin area in our database) (Supplementary Fig. S2).
To assess the potential protection of the fish assemblage in each estuary we determined the location of protected areas worldwide, by combining spatial data from the World Database on Protected Areas64 and MPAtlas65. Following an approach used in previous studies10, 24, two alternative selections of the database were done to address differences in terms of protection level: (a) PAall - considering all protected areas; and (b) PAI-IV - considering only protected areas classified with IUCN management categories I-IV (respectively strict nature reserve or wilderness area, national park, national monument or feature, habitat/species management area) which are protected areas that restrict human activities (e.g. fishing). For each estuary in the database, we quantified the extent of coverage by protected areas within an influence radius, in three ways, namely: PAmarine - i.e. protected areas in the marine ecosystem shallower than 200 m (continental shelf); PAestuary - i.e. protected areas in and around the estuary; PAfreshwater - i.e. protected areas in and around the estuary but excluding the marine ecosystem. We used the radii used previously in the estimation of pressures. The extent of coverage by protected areas was calculated in area (km2) and in percentage (% of area that is protected within the influence radius). The lack of geospatial vector data for all estuaries and corresponding rivers/drainage basins precluded a more refined estimation of PAestuary and PAfreshwater. All pressure and protection data were compiled in ArcGIS for desktop version 10.4 (http://desktop.arcgis.com) using a Cylindrical Equal Area projection.
Environmental, pressure and protection variables were logx + 1 transformed to reduce skewness and the effect of extreme observations66, 67. In addition, each pressure variable was normalized (scaled) so that 0 represents the lowest pressure and 1 the highest (from each value we took the minimum and divided by the range). Based on human pressures in the three ecosystems (Hmarine, Hestuary, Hfreshwater) we also calculated: the mean of the three ecosystems (Hmean), and the weighted mean of the three ecosystems (Hweighted-mean) where, for each estuary, the weight of each ecosystem is given by the percentage of fish from that ecosystem in the estuarine assemblage (i.e. % marine fish for Hmarine, % resident fish for Hestuary, % freshwater fish for Hfreshwater and % diadromous fish for Hmean).
We first examined the pairwise Pearson correlations between: all environmental variables, pressure variables and protection variables (R package stats); between all traits (relative taxa richness of trait categories); as well as between all traits and pressure/protection variables. To avoid effects of multicollinearity, several environmental variables were excluded from subsequent analyses, namely: latitude (with temperature), estuary mouth width and drainage basin area (with estuary area) (Table 3). We then used linear models (LM) to disentangle the relationship of fish traits (response variables) with all biogeographical and environmental variables (predictors). Additionally, since some estuaries have more than one sample in our database, we used linear mixed models (LMM) which were formulated similarly to the linear models but also included estuary as a random predictor. In both LM and LMM, sampling effort (i.e. total sampled area) was always included as a predictor to account for differences in effort between samples in our database. To attain robust estimates of the importance and parameter of each predictor, we implemented a multi model approach using: hierarchical partition of variation (R package relaimpo; only for LM) and multimodel inference (R package MuMIn; for both LM and LMM) which evaluate predictor importance respectively based on R2 and Akaike information criteria. Each trait category (e.g. proportion of taxa with low vulnerability) was modelled as a separate response variable, and for each trait category, we fitted two alternative models: with and without the biogeographical variables. As a note, categorical environmental variables were considered as continuous in correlation analysis and as ordered factors in linear- and linear mixed models (tidal regime: microtidal - 1, mesotidal - 2, macrotidal - 3; estuary type: temporarily open - 1, open - 2 salinity type: regular - 1, regular to hyperhaline - 2, hyperhaline - 3). A significance level of 0.05 was considered in all statistical analyses.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Costanza, R. et al. The value of the world’s ecosystem services and natural capital. Nature 387, 253–260 (1997).
Barbier, E. B. et al. The value of estuarine and coastal ecosystem services. Ecol Monogr 81, 169–193 (2011).
Lotze, H. K. et al. Depletion, Degradation, and Recovery Potential of Estuaries and Coastal Seas. Science 312, 1806–1809 (2006).
Cheung, W. W. L., Watson, R., Morato, T., Pitcher, T. J. & Pauly, D. Intrinsic vulnerability in the global fish catch. Mar Ecol Prog Ser 333, 1–12 (2007).
Elliott, M. & Quintino, V. The Estuarine Quality Paradox, Environmental Homeostasis and the difficulty of detecting anthropogenic stress in naturally stressed areas. Mar Pollut Bull 54, 640–655 (2007).
Elliott, M. et al. The guild approach to categorizing estuarine fish assemblages: a global review. Fish Fish 8, 241–268 (2007).
Potter, I. C., Tweedley, J. R., Elliott, M. & Whitfield, A. K. The ways in which fish use estuaries: a refinement and expansion of the guild approach. Fish Fish 16, 230–239 (2015).
Robins, P. E. et al. Impact of climate change on UK estuaries: A review of past trends and potential projections. Estuar Coast Shelf S 169, 119–135 (2016).
Elliott, M. & Whitfield, A. K. Challenging paradigms in estuarine ecology and management. Estuar Coast Shelf S 94, 306–314 (2011).
Parravicini, V. et al. Global mismatch between species richness and vulnerability of reef fish assemblages. Ecol Lett 17, 1101–1110 (2014).
Cinner, J. E. et al. Evaluating social and ecological vulnerability of coral reef fisheries to climate change. PLoS One 8, e74321 (2013).
Halpern, B. S. et al. A Global Map of Human Impact on Marine Ecosystems. Science 319, 948–952 (2008).
D’Agata, S. et al. Human-mediated loss of phylogenetic and functional diversity in coral reef fishes. Curr Biol 24, 555–560 (2014).
Maire, E. et al. How accessible are coral reefs to people? A global assessment based on travel time. Ecol Lett 19, 351–360 (2016).
Mora, C. et al. Global human footprint on the linkage between biodiversity and ecosystem functioning in reef fishes. PLoS Biol 9, e1000606 (2011).
Henriques, S. et al. Structural and functional trends indicate fishing pressure on marine fish assemblages. J Appl Ecol 51, 623–631 (2014).
Mellin, C. et al. Humans and seasonal climate variability threaten large-bodied coral reef fish with small ranges. Nat Commun 7, 10491 (2016).
Vörösmarty, C. J. et al. Global threats to human water security and river biodiversity. Nature 467, 555–561 (2010).
Liermann, C. R., Nilsson, C., Robertson, J. & Ng, R. Y. Implications of Dam Obstruction for Global Freshwater Fish Diversity. BioScience 62, 539–548 (2012).
Deguignet, M. et al. United Nations List of Protected Areas (UNEP-WCMC, 2014).
Prates, A., Gonçalves, M. & Rosa, M. Panorama da Conservação dos Ecossistemas Costeiros e Marinhos no Brasil (Ministério do Meio Ambiente, Brasília, Brazil, 2012).
Turpie, J. K. et al. Assessment of the conservation priority status of South African estuaries for use in management and water allocation. Water SA 28, 191–206 (2002).
Mouillot, D. et al. Global marine protected areas do not secure the evolutionary history of tropical corals and fishes. Nat Commun 7, 10359 (2016).
Klein, C. J. et al. Shortfalls in the global protected area network at representing marine biodiversity. Sci Rep 5, 17539 (2015).
Vasconcelos, R. P. et al. Global patterns and predictors of fish species richness in estuaries. J Anim Ecol 84, 1331–1341 (2015).
Attrill, M. J., Stafford, R. & Rowden, A. A. Latitudinal diversity patterns in estuarine tidal flats: indications of a global cline. Ecography 24, 318–324 (2001).
Henriques, S. et al. Biogeographical region and environmental conditions drive functional traits of estuarine fish assemblages worldwide. Fish Fish (2017).
Teichert, N. et al. Living under stressful conditions: Fish life history strategies across environmental gradients in estuaries. Estuar Coast Shelf S 188, 18–26 (2017).
Woodward, G. et al. Body size in ecological networks. Trends Ecol Evol 20, 402–409 (2005).
Kulbicki, M., Parravicini, V. & Mouillot, D. Patterns and processes in reef fish body size in Ecology of fishes on coral reefs (ed. C. Mora) 104–115 (Cambridge University Press, 2015).
Cheung, W. W. L., Pitcher, T. J. & Pauly, D. A fuzzy logic expert system to estimate intrinsic extinction vulnerabilities of marine fishes to fishing. Biol Conserv 124, 97–111 (2005).
Musick, J. A. Criteria to Define Extinction Risk in Marine Fishes. Fisheries 24, 6–14 (1999).
Ban, N. C., Alidina, H. M. & Ardron, J. A. Cumulative impact mapping: Advances, relevance and limitations to marine management and conservation, using Canada’s Pacific waters as a case study. Mar Policy 34, 876–886 (2010).
Stelzenmüller, V., Ellis, J. R. & Rogers, S. I. Towards a spatially explicit risk assessment for marine management: Assessing the vulnerability of fish to aggregate extraction. Biol Conserv 143, 230–238 (2010).
Coll, M. et al. The Mediterranean Sea under siege: spatial overlap between marine biodiversity, cumulative threats and marine reserves. Glob Ecol Biogeogr 21, 465–480 (2012).
Batista, M. I., Henriques, S., Pais, M. P. & Cabral, H. N. Assessment of cumulative human pressures on a coastal area: Integrating information for MPA planning and management. Ocean Coast Manage 102, 248–257 (2014).
Halpern, B. S. et al. Global priority areas for incorporating land-sea connections in marine conservation. Conserv Lett 2, 189–196 (2009).
Claudet, J. & Fraschetti, S. Human-driven impacts on marine habitats: A regional meta-analysis in the Mediterranean Sea. Biol Conserv 143, 2195–2206 (2010).
Edgar, G. J. et al. Global conservation outcomes depend on marine protected areas with five key features. Nature 506, 216–220 (2014).
Claudet, J. et al. Marine reserves: size and age do matter. Ecol Lett 11, 481–489 (2008).
Vasconcelos, R. P., Reis-Santos, P., Costa, M. J. & Cabral, H. N. Connectivity between estuaries and marine environment: Integrating metrics to assess estuarine nursery function. Ecol Indic 11, 1123–1133 (2011).
Beger, M. et al. Conservation planning for connectivity across marine, freshwater, and terrestrial realms. Biol Conserv 143, 565–575 (2010).
Huijbers, C. M. et al. Conservation Benefits of Marine Reserves are Undiminished Near Coastal Rivers and Cities. Conserv Lett 8, 312–319 (2014).
Minns, C. K. et al. Direct and indirect estimates of the productive capacity of fish habitat under Canada’s Policy for the Management of Fish Habitat: where have we been, where are we now, and where are we going? Can J Fish Aquat Sci 68, 2204–2227 (2011).
Wood, L. J., Fish, L., Laughren, J. & Pauly, D. Assessing progress towards global marine protection targets: shortfalls in information and action. Oryx 42, 340–351 (2008).
Kummu, M. & Varis, O. The world by latitudes: A global analysis of human population, development level and environment across the north–south axis over the past half century. Appl Geogr 31, 495–507 (2011).
Luiz, O. J. et al. Adult and larval traits as determinants of geographic range size among tropical reef fishes. Proc Natl Acad Sci USA 110, 16498–16502 (2013).
Sunday, J. M. et al. Species traits and climate velocity explain geographic range shifts in an ocean-warming hotspot. Ecol Lett 18, 944–953 (2015).
Fisher, J. A. D., Frank, K. T. & Leggett, W. C. Global variation in marine fish body size and its role in biodiversity–ecosystem functioning. Mar Ecol Prog Ser 405, 1–13 (2010).
James, N. C., Cowley, P. D., Whitfield, A. K. & Lamberth, S. J. Fish communities in temporarily open/closed estuaries from the warm- and cool-temperate regions of South Africa: A review. Rev Fish Biol Fisher 17, 565–580 (2007).
Mims, M. C., Olden, J. D., Shattuck, Z. R. & Poff, N. L. Life history trait diversity of native freshwater fishes in North America. Ecol Freshw Fish 19, 390–400 (2010).
Selig, E. R. et al. Global priorities for marine biodiversity conservation. PLoS One 9, e82898 (2014).
Guilhaumon, F., Gimenez, O., Gaston, K. J. & Mouillot, D. Taxonomic and regional uncertainty in species-area relationships and the identification of richness hotspots. Proc Natl Acad Sci USA 105, 15458–15463 (2008).
Myers, N., Mittermeier, R. A., Mittermeier, C. G., da Fonseca, G. A. & Kent, J. Biodiversity hotspots for conservation priorities. Nature 403, 853–858 (2000).
Mouillot, D., Graham, N. A., Villeger, S., Mason, N. W. & Bellwood, D. R. A functional approach reveals community responses to disturbances. Trends Ecol Evol 28, 167–177 (2013).
Gaston, K. J. Latitudinal gradient in species richness. Curr Biol 17, R574 (2007).
Vasconcelos, R. P. et al. Global patterns and predictors of fish species richness in estuaries. J Anim Ecol 84, 1331–1341 (2015).
Henriques, S. et al. Processes underpinning fish species composition patterns in estuarine ecosystems worldwide. J. Biogeogr. 44, 627–639 (2017).
Hillebrand, H. & Blenckner, T. Regional and local impact on species diversity - from pattern to processes. Oecologia 132, 479–491 (2002).
Spalding, M. D. et al. Marine Ecoregions of the World: A Bioregionalization of Coastal and Shelf Areas. BioScience 57, 573–583 (2007).
Vasconcelos, R. P. et al. Assessing anthropogenic pressures on estuarine fish nurseries along the Portuguese coast: a multi-metric index and conceptual approach. Sci Total Environ 374, 199–215 (2007).
Cabral, H. N. et al. Ecological quality assessment of transitional waters based on fish assemblages in Portuguese estuaries: The Estuarine Fish Assessment Index (EFAI). Ecol Indic 19, 144–153 (2012).
Center for International Earth Science Information Network - CIESIN - Columbia University, United Nations Food and Agriculture Programme - FAO & Centro Internacional de Agricultura Tropical - CIAT. Gridded Population of the World, Version 3 (GPWv3): Population Count Grid, http://dx.doi.org/10.7927/H4639MPP, NASA Socioeconomic Data and Applications Center (SEDAC), Date of access: 01/01/2016, http://sedac.ciesin.columbia.edu/data/set/gpw-v3-population-count (2005).
International Union for Conservation of Nature (IUCN) & United Nations Environment Programme - World Conservation Monitoring Centre (UNEP-WCMC). The World Database on Protected Areas (WDPA), Date of access: 01/01/2016, www.protectedplanet.net (2016).
Marine Conservation Institute. MPAtlas, Date of access: 01/01/2016, www.mpatlas.org (2016).
Clarke, K. R. & Warwick, R. M. Change in marine communities: an approach to statistical analysis and interpretation (Plymouth Marine Laboratory, UK, 2001).
Zuur, A. F., Ieno, E. N. & Smith, G. M. Analysing Ecological Data (Springer, 2007).
The authors would like to thank several authors for providing supplementary material to their publications - particularly Trevor Harrison for supplementary material to Harrison (2003). We would also like to thank FishBase (http://www.fishbase.org/) for providing their database on fish species. Research was financed with national funds through Fundação para a Ciência e a Tecnologia (FCT) via project PTDC/MAR/117119/2010, MARE was funded with project UID/MAR/04292/2013, R.P.V. with Investigador FCT Programme 2013 (IF/00058/2013) and S.H. with a Post Doc grant (SFRH/BPD/94320/2013), all from FCT.
Electronic supplementary material
About this article
Scientific Reports (2018)