Conceived and designed the experiments: JJCM PM KI BK LBC. Performed the experiments: JJCM GP KI BK GP TT LBC CH AH AS AB JG DIG GB YS. Analyzed the data: JJCM GP KI BK LBC CH AH AS AB JC EK EGC ALK. Contributed reagents/materials/analysis tools: JJCM PM GP KI BK JG DIG. Wrote the paper: JJCM GP KI BK TT. Provided data: JC EK RR-R AM.
The authors have declared that no competing interests exist.
Assemblages associated with intertidal rocky shores were examined for large scale distribution patterns with specific emphasis on identifying latitudinal trends of species richness and taxonomic distinctiveness. Seventy-two sites distributed around the globe were evaluated following the standardized sampling protocol of the Census of Marine Life NaGISA project (
The study of biological diversity or biodiversity has gained strong scientific interest in recent decades (13,029 and 31,691 references, respectively, in Web of Science in the last decade), due to the consequences that diversity loss might have on humanity
To detect changes in natural communities, and unequivocally relate them to anthropogenic impacts or climate disruptions, proper biological baseline data are of utmost importance. Very few long-term/large-scale data sets are currently available (but see
Intertidal rocky shores assemblages are appropriate to study changes driven by global-scale anthropogenic impacts and climate change effects due to their ecological characteristics and accessibility
Current paradigms of latitudinal diversity gradients postulate an increasing number of species increases from the pole to the equator
This study was carried out as part of the research conducted by the Laboratorio de Ecología Experimental, approved by and under the guidelines of the Departamento de Estudios Ambientales, Universidad Simón Bolívar, Caracas, Venezuela.
Surveys were done at 72 rocky intertidal sites (
1 = Gulf of Alaska, 2 = Gulf of California, 3 = Northeast U.S. Continental Shelf, 4 = Scotian Shelf, 5 = Caribbean Sea, 6 = Patagonian Shelf, 7 = South Brazilian Shelf, 8 = Celtic-Biscay Shelf, 9 = Mediterranean Sea, 10 = Benguela Current, 11 = Aghulas Current, 12 = South China Sea, 13 = Kuroshio Current.
LMEs | Abb. | Sites | Replicates per site | Ocean | Countries | Bottom type |
Gulf of Alaska | GoA | 11 | 5 | Pacific | USA, Canada | Bedrock, Sandstone and Boulders |
Agulhas Current | AgC | 7 | 10 | Indian | South Africa | Boulders and Sandstone |
Celtic-Biscay Shelf | CBS | 2 | 5 | Atlantic | England | Bedrock |
Northeast U.S Continental Shelf | NCS | 2 | 5 | Atlantic | USA, Canada | Cobbles and Bedrock |
Caribbean Sea | CbS | 29 | 10 | Atlantic | Colombia, Venezuela, Trinidad & Tobago | Bedrock |
Benguela Current | BgC | 7 | 10 | Atlantic | South Africa | Boulders, Sandstone and Rocky reef |
Mediterranean Sea | MdS | 3 | 5 | Mediterranean Sea | Italy | Bedrock and Sandstone |
Scotian Shelf | StS | 1 | 5 | Atlantic | Canada | Cobbles |
South China Sea | SCS | 1 | 3 | Pacific | Vietnam | Bedrock |
Patagonian Shelf | PaS | 5 | 10 | Atlantic | Argentina | Bedrock |
Kuroshio Current | KuC | 1 | 5 | Pacific | Japan | Bedrock |
South Brazil Shelf | SBS | 1 | 5 | Atlantic | Brazil | Bedrock |
Gulf of California | GoC | 2 | 5 | Pacific | Mexico | Loose boulders |
Abb = Abbreviation code for LMEs, SST = Sea Surface Temperature.
Data were collected following the standardized NaGISA protocol
Fourteen environmental variables were examined to determine the most important drivers for describing trends in species numbers and composition of assemblages associated with intertidal rocky shores. Variables were estimated for sampling site within each LME using different sources and were grouped into variables related to either “natural” or “anthropogenic” influences (
Variable | Short | Description | Reference |
Sea-surface temperature | SST | Average of monthly values of the MODIS Aqua mission from July 2002 to December 2009 | |
Chlorophyll- |
CHA | Average of monthly values of the MODIS Aqua mission from July 2002 to December 2009 | |
Chlorophyll- |
CHAa | Numbers of events that surpassed 2 standard deviations of the average chlorophyll-a for a given year | |
Rainfall | RAI | Average of monthly accumulated rainfall from January 1979 through September 2009 obtained using the TOVAS web-based application | |
Rainfall anomalies | RAIa | Numbers of events that surpassed 2 standard deviations of the average rainfall for a given year | |
Photoperiod | PHO | Common astronomical formulae were used to compute the difference between the sunrise and sunset time | |
Inorganic pollution | INP | Urban runoff estimated from land-use categories, US Geologic Survey ( |
|
Organic pollution | ORP | FAO national pesticides statistics (1992–2001), ( |
|
Nutrient contamination | NUTC | FAO national fertilizers statistics (1993–2002), ( |
|
Acidification | AC | Aragonite saturation state 1870–2000/2009, 1 degree lat/long resolution | |
Invasive species incidence | INV | Cargo traffic 1999–2003 | |
Population pressure | HUM | Estimated as the sum of total population adjacent to the ocean within a 25 km radius. LandScan 30 arc-second population data of 2005 were used. | |
Shipping activity | SH | Commercial ship traffic 2004–2005 | |
Ocean-based pollution | OBP | Modelled as a combination of commercial shipping traffic data and port data |
Because of the coarse spatial resolution of the environmental data and the land-mask imposed to the models from which data were derived from, most variables were not predicted precisely for the shore sampling sites. When a site was no farther than 50 km from the model, a spline interpolation was used to the raster data to compute its value at the coordinate of the sampling site. Furthermore, LMEs were used as the scale of comparison in this study because of the potential inaccuracy of satellite-derived data from optical sea-surface properties (e.g., chlorophyll-a, primary productivity) on small spatial scales
Since different sampling efforts were used in different LMEs, the number of taxa at each LME was interpolated using saturation curves (i.e., UGE method
Biological data from each site were transformed to presence-absence data and a similarity matrix was constructed based on the taxonomic dissimilarity coefficient
Environmental variables were normalized to a common scale. Geographic coordinates were included in this matrix and considered in further analyses in order to detect possible effects of distances among sampling sites. Redundant variables were identified using multiple correlation analysis (i.e., draftsman plots) after square-root transformation of skewed variables and excluded from the analysis. It is important to note that whenever latitude or longitude were used, these variables conserved their sign. To select the combination of variables that best matched the biological distribution patterns, a similarity matrix of environmental variables based on Euclidean distances was linked to the taxonomic dissimilarities patterns (
A total of 801 taxa were identified from 1499 sample quadrats. The number of observed and standardized taxon richness varied among LMEs (
LMEs | n | S | UGE (n = 20) | Dominant group | Grazers | Other important species |
Gulf of Alaska | 110 | 106 | 45 | Brown and red algae (Phaeophyceae) | Littorinidae, limpets and chitons (Lottiidae and Littorina) | |
Agulhas Current | 70 | 110 | 86 | Red algae |
Littorinidae and limpets |
|
Celtic-Biscay Shelf | 20 | 45 | 45 | Brown and red algae |
Littorinidae, limpets and snails |
|
Northeast U.S Continental Shelf | 20 | 47 | 47 | Brown algae |
Littorinidae and limpets |
|
Caribbean Sea | 154 | 261 | 120 | Brown, red and green algae (encrusting coralline algae) | Littorinidae, sea urchins, limpets, snails and chitons |
|
Benguela Current | 70 | 97 | 75 | Brown and red algae |
Littorinidae, limpets, snails and chitons |
|
Mediterranean Sea | 40 | 65 | 57 | Brown and red algae (Corrallinaceae) | Littorinidae, Sea urchins, Limpets and Snails |
|
Scotian Shelf* | 10 | 7 | n/d | Brown algae |
||
South Chine Sea* | 7 | n/d | Barnacle | Limpets (Patellogastropoda) | Saccostrea (bivalve) | |
Patagonian Shelf | 59 | 35 | 30 | Mussels |
Limpets |
|
Kuroshio Current* | 5 | 4 | n/d | Sponges |
Limpets and chitons |
Patellogastropoda |
South Brazil Shelf* | 5 | 34 | n/d | Barnacles |
||
Gulf of California | 20 | 8 | 8 | Cyanophyceae | Snails |
Includes total number of quadrats (n), total number of observed taxa (S), estimators of number of taxa for a standard sampling size of 20 quadrats based on saturation curves (UGE method) and most common species or taxa per LME. Asterisks denote LMEs with fewer than 20 quadrats (n<20), for which no UGE was calculated (n/d)
No latitudinal patterns were found using UGE-standardized richness estimates as indicated by a low Pearson's correlation index (ρ = −0.12;
A constrained ordination (CAP) of sampling sites using LMEs as predictor factor effectively showed that sites based on assemblage information were strongly grouped by LME (
Canonical analysis of principal coordinates (CAP) plots generated from taxonomic dissimilarity coefficients (
The latitudinal trend in taxonomic composition was largely due to the presence of prominent taxonomic groups in the LMEs as indicated by SIMPER analysis (
GoA | AgC | CBS | NCS | CbS | BgC | MdS | StS | SCS | PaS | KuC | GoC | |
Patellidae Corallinaceae Siphonariidae | Littorinidae Fucaceae Trochidae | Fucaceae Littorinidae Ulvaceae | Rhodomelaceae Corallinaceae Dictyotaceae | Patellidae Corallinaceae Mytilidae | Rhodomelaceae Patellidae Corallinaceae | Fucaceae | Chthalamidae Ostreidae Patellogastropoda | Rhodomelaceae Mytilidae Ulvaceae | Halichondriidae Patellogastropoda Polychaeta | Chthalamidae Neritidae Turridae | ||
Phaeophyta Chlorophyta Rhodophyta | Littorinidae Fucaceae Trochidae | Fucaceae Littorinidae Cladophoraceae | Dictyotaceae Echinometridae Zoanthidae | Patellidae Sabellariidae Cryptonemiaceae | Dictyotaceae Cystoseiraceae Chlorophyta | Fucaceae | Ostreidae Patellogastropoda Chthalamidae | Chlorophyta Hildenbrandiaceae Balanidae | Patellogastropoda Lottiidae Polychaeta | Neritidae | ||
Phaeophyta Chlorophyta Rhodophyta | Patellidae Corallinaceae Siphonariidae | Ulvaceae Mytilidae Lottiidae | Corallinaceae Dictyotaceae Echinometridae | Patellidae Mytilidae Buccinidae | Dictyotaceae Cystoseiraceae Mytilidae | Fucaceae | Ostreidae Chthalamidae Patellogastropoda | Mytilidae Chlorophyta Ulvaceae | Halichondriidae | Neritidae | ||
Phaeophyta Chlorophyta Rhodophyta | Patellidae Corallinaceae Siphonariidae | Trochidae Patellidae Littorinidae | Corallinaceae Dictyotaceae Echinometridae | Patellidae Corallinaceae Trochidae | Patellidae Corallinaceae Dictyotaceae | Fucaceae | Chthalamidae Ostreidae Patellogastropoda | Chlorophyta Balanidae Siphonariidae | Halichondriidae Patellogastropoda Polychaeta | Chthalamidae Neritidae | ||
Phaeophyta Chlorophyta Rhodophyta | Patellidae Corallinaceae Siphonariidae | Fucaceae Littorinidae Trochidae | Fucaceae Littorinidae Ulvaceae | Patellidae Mytilidae Trochidae | Patellidae Cystoseiraceae Chthalamidae | Fucaceae | Ostreidae Chthalamidae Patellogastropoda | Mytilidae Ulvaceae Chlorophyta | Halichondriidae Patellogastropoda Lottiidae | Chthalamidae Neritidae Turridae | ||
Phaeophyta Chlorophyta Rhodophyta | Corallinaceae Siphonariidae Rhodomelaceae | Littorinidae Fucaceae Rhodomelaceae | Fucaceae Littorinidae Ulvaceae | Rhodomelaceae Dictyotaceae Echinometridae | Rhodomelaceae Dictyotaceae Cystoseiraceae | Fucaceae | Ostreidae Chthalamidae Patellogastropoda | Rhodomelaceae Chlorophyta Balanidae | Halichondriidae Patellogastropoda Lottiidae | Chthalamidae Neritidae Turridae | ||
Phaeophyta Rhodophyta Chlorophyta | Patellidae Corallinaceae Siphonariidae | Fucaceae Littorinidae Trochidae | Fucaceae Littorinidae Ulvaceae | Corallinaceae Echinometridae Ralfsiaceae | Patellidae Buccinidae Actiniidae | Fucaceae | Ostreidae Patellogastropoda Chthalamidae | Ulvaceae Hildenbrandiaceae Siphonariidae | Halichondriidae Patellogastropoda Lottiidae | Neritidae | ||
Phaeophyta Chlorophyta Rhodophyta | Corallinaceae Patellidae Siphonariidae | Littorinidae Trochidae Rhodomelaceae | Littorinidae Ulvaceae Rhodomelaceae | Corallinaceae Rhodomelaceae Dictyotaceae | Patellidae Corallinaceae Mytilidae | Corallinaceae Patellidae Rhodomelaceae | Chthalamidae Ostreidae Patellogastropoda | Mytilidae Rhodomelaceae Ulvaceae | Halichondriidae Patellogastropoda Lottiidae | Chthalamidae Neritidae | ||
Phaeophyta Chlorophyta Rhodophyta | Corallinaceae Patellidae Siphonariidae | Fucaceae Trochidae Littorinidae | Fucaceae Ulvaceae Littorinidae | Corallinaceae Rhodomelaceae Dictyotaceae | Patellidae Corallinaceae Mytilidae | Corallinaceae Patellidae Rhodomelaceae | Fucaceae | Mytilidae Rhodomelaceae Ulvaceae | Halichondriidae | Neritidae | ||
Phaeophyta Rhodophyta Sessilia | Patellidae Corallinaceae Trochidae | Littorinidae Fucaceae Trochidae | Fucaceae Littorinidae Lottiidae | Corallinaceae Dictyotaceae Echinometridae | Patellidae Trochidae Buccinidae | Patellidae Dictyotaceae Cystoseiraceae | Fucaceae | Ostreidae Patellogastropoda | Halichondriidae | Neritidae | ||
Phaeophyta Chlorophyta Rhodophyta | Corallinaceae Patellidae Siphonariidae | Littorinidae Fucaceae Trochidae | Fucaceae Littorinidae Ulvaceae | Corallinaceae Rhodomelaceae Dictyotaceae | Patellidae Mytilidae Corallinaceae | Corallinaceae Patellidae Rhodomelaceae | Fucaceae | Chthalamidae Ostreidae Patellogastropoda | Mytilidae Rhodomelaceae Ulvaceae | Chthalamidae Neritidae Turridae | ||
Phaeophyta Chlorophyta Rhodophyta | Corallinaceae Patellidae Siphonariidae | Littorinidae Fucaceae Trochidae | Fucaceae Littorinidae Ulvaceae | Corallinaceae Rhodomelaceae Dictyotaceae | Patellidae Mytilidae Corallinaceae | Corallinaceae Patellidae Rhodomelaceae | Fucaceae Corallinaceae | Ostreidae Patellogastropoda Chthalamidae | Mytilidae Rhodomelaceae Ulvaceae | Halichondriidae Patellogastropoda Lottiidae |
Comparisons are columns vs. rows, meaning families or taxa are more abundant in an LME by column compared to the LME of the intersecting row.
A constrained ordination (CAP) of sites based on the environmental variables showed clear differences among sites located in different LMEs (
Canonical analysis of principal coordinates (CAP) generated from Euclidian distances of the environmental matrix using LMEs as predictor factors. Green triangle = Gulf of Alaska, Yellow square = Agulhas Current, Red square = Mediterranean Sea, Blue triangle = Celtic-Biscay Shelf, Green diamond = Gulf of California, Blue diamond = Northeast US Continental Shelf, Inverted blue triangle = Caribbean Sea, Blue circle = Benguela Current, Green square = South China Sea, Green circle = Kuroshio Current, Blue square = Patagonian Shelf, Empty blue circle = Scotian Shelf.
Correlation of the matrices of environmental variables with the biological, by means of a BIOENV routine, indicated that the variables that best explained patterns of spatial distribution of LMEs, based on their biological information (ρ = 61.1%) were: photoperiod, rain fall anomalies, SST, chlorophyll-a anomalies and the index of inorganic pollution (
Number of variables considered | Correlation | Selections |
5 | 0.611 | PHO, RAla, SST, CHAa, INP |
5 | 0.598 | PHO, RAla, SST, CHAa, CHA |
5 | 0.578 | PHO, RAla, SST, CHA, INP |
4 | 0.567 | RAIa, SST, CHA, INP |
The overall intertidal rocky shore assemblage descriptions provided here correspond well with documented species lists for some LME's (e.g. Caribbean Sea, Gulf of Alaska, and the South African Agulhas and Benguela Current), despite the often small sampling effort in our study. For the Caribbean, for example, similar assemblage description, based on dominant species, was obtained from more detailed studies with more effort
The proposed cline in species diversity from low to high latitudes for most terrestrial and some marine groups
Alternatively, the lack of latitudinal patterns found in this study might be due to low sample size in some LMEs. Small sample sizes are likely to omit rare species in a given assemblage, which would result in underestimations of species richness for those particular LMEs.
In this study, despite the fact that no latitudinal gradient was found in terms of the univariate estimator of taxon richness, a clear latitudinal pattern was found for the multivariate aspect of taxonomic composition of intertidal rocky shores assemblages (
Differences in taxonomic composition among LMEs demonstrated that spatial distribution patterns of these assemblages were not homogeneous over large spatial scales. Consequently, null models, predicting uniform assemblage patterns over large spatial scales, could be discarded
Through correlation analyses (i.e. BIOENV), six environmental variables were identified as potential drivers of spatial distribution patterns of intertidal rocky assemblages. Of those, five are considered “natural” variables, and only one (inorganic pollution index) was directly related to anthropogenic influences. There was no evidence to unequivocally separate environmental models and neutral models to explain taxonomic composition, because assemblages were highly correlated with latitude (
Correlative analysis does not establish cause and effect. However, the identification of correlated drivers can give us some insight into which variables may be most influential. Actual cause-consequences relationships between environmental (anthropogenic or natural) drivers and rocky shore assemblages at global scales are further complicated due to the inherent complexity of spatial and temporal variation in which these assemblages naturally fluctuate
(0.03 MB DOC)
We greatly appreciate the invaluable help of all volunteers who assisted the different sampling teams around the world. We also wish to thank the Census of Marine Life and its administrative staff who made the implementation of a global network possible. We are grateful to Dr. Brian Helmuth and an anonymous reviewer who greatly contributed, through their comments, to improve the quality of this manuscript.