Species richness and abundance in north and northeast Atlantic fish communities
Patterns in marine fish species richness, distribution and community composition in the north and northeastern Atlantic have been identified and related to density, depth, maximum body size and temperature using data from 30 bottom trawl surveys containing almost 46 thousand trawl hauls obtained within an area from the Gulf of St Lawrence in Canada to Guinea in Africa. The objectives were to generate baseline from which future changes in fish diversity can be evaluated, to identify the major environmental drivers explaining the current distribution of fish species, and to test some of the general hypothesis about the latitudinal gradient in marine fish biodiversity with special focus on temperature and density. Previous work to identify such patterns has either been based on survey data from fewer areas or has used observed or inferred presence/absence of species from general species inventories with little or no standardization of sampling effort across areas.
The major changes in fish community composition in the northeastern Atlantic occur between the Faroe Islands and Scotland, and between Morocco and Mauritania. Small species are more dissimilar across space than large species, suggesting a general increase in distribution area with species size.
We calculated the taxonomic dissimilarity of the trawl survey catches from the presence/absence of demersal and pelagic fish species in the surveys while accounting for their taxonomic relatedness. Analysis of the rank correlation between sets of environmental variables and the dissimilarity matrix of the trawl surveys revealed that the highest correlation (0.926, p<0.001) was due to average sea surface temperature (SST). Temperature was the most important parameter separating the boreal (SST<9) from the temperate and tropical surveys (SST>10). Separate ordinations for small (<55cm) and large (>55cm) species revealed considerable more taxonomic separation across surveys for small species than for large. Analysis of taxonomic relatedness can be used to define faunal boundaries, where environmental or other features create natural limits for species dispersal. Such boundaries are important to consider when predicting and interpre ting changes in species distribution and abundance generated e.g. by global warming.
The number of fish species of different size increases with sampling effort and temperature. It declines with depth, is significantly affected by mesh size and temperature range, and shows a parabolic relationship to species size.
A Generalized Additive Model (GAM) explains 80% of the deviance in richness as an effect of log species size, temperature, temperature range, area swept by the gear, depth and mesh size. The high proportion of deviance explained suggests that the model can be used to extrapolate species richness within the limits of temperature, depth, mesh size, and area swept covered by the observations. The model may thus be used be used to improve predictions of the effect of climate change and other drivers on fish species diversity.
A comparison of different hypotheses about the origin of the pattern in fish species richness in the north Atlantic suggests that a neutral assumption best explains the pattern observed.
Many hypotheses have been proposed to explain the decline in species richness from equator towards the poles that can be observed for both marine and terrestrial biota. We have tested a few of the mechanistic theories of species richness using Generalized Linear Models to analyse the fish survey data. We find that a neutral assumption, where species of similar size are assumed to be competitively equivalent, provides a better explanation of the latitudinal gradient in marine species richness than models where speciation is explained by either sea surface temperature or by latitude and depth. However, additional work needs to be conducted to model species dispersal and to compare the observed species abundance distributions with those predicted by the neutral theory. When thoroughly tested, models of species richness can be used to infer richness in poorly sampled environments, and help to distinguish between the effects of various anthropogenic drivers of biodiversity.
Fish density scales positively with species richness as would be expected from a curvilinear relationship between richness and ecosystem functioning.
Understanding the consequences of species extinctions due to anthropogenic forcing is important for managing marine ecosystems. Manipulative small scale experiments have previously found that mixtures of species tend to increase ecosystem functioning relative to the level of functioning experienced when the average species is held in monoculture, and that losses of species therefore on average will decrease the functioning of a given ecosystem. Identifying the relationship between richness and function in marine ecosystems is difficult due to sampling artifacts and random inter-annual change in recruitment success, but the relationship between fish density and species richness found in the our analysis of the survey data suggests that a response similar to the one found in the small scale experiments may exist for fish species richness in the northeast Atlantic.
Relevance for Policy:
- Common Fisheries Policy
- Convention on Biological Diversity
- EU Biodiversity Strategy
- Marine Strategy Framework Directive