Limits...
Functional redundancy patterns reveal non-random assembly rules in a species-rich marine assemblage.

Guillemot N, Kulbicki M, Chabanet P, Vigliola L - PLoS ONE (2011)

Bottom Line: First, we found that the relationship between functional and species diversity displayed a non-asymptotic power-shaped curve, implying that rare functions and species mainly occur in highly diverse assemblages.Last, we found little effect of habitat on the shape of the functional-species diversity relationship and on the redundancy of functions, although habitat is known to largely determine assemblage characteristics such as species composition, biomass, and abundance.Our study shows that low functional redundancy is characteristic of this highly diverse fish assemblage, and, therefore, that even species-rich ecosystems such as coral reefs may be vulnerable to the removal of a few keystone species.

View Article: PubMed Central - PubMed

Affiliation: UR-CoRéUs, Institut de Recherche pour le Développement, Noumea, New Caledonia. nicolas.guillemot@gmail.com

ABSTRACT
The relationship between species and the functional diversity of assemblages is fundamental in ecology because it contains key information on functional redundancy, and functionally redundant ecosystems are thought to be more resilient, resistant and stable. However, this relationship is poorly understood and undocumented for species-rich coastal marine ecosystems. Here, we used underwater visual censuses to examine the patterns of functional redundancy for one of the most diverse vertebrate assemblages, the coral reef fishes of New Caledonia, South Pacific. First, we found that the relationship between functional and species diversity displayed a non-asymptotic power-shaped curve, implying that rare functions and species mainly occur in highly diverse assemblages. Second, we showed that the distribution of species amongst possible functions was significantly different from a random distribution up to a threshold of ∼90 species/transect. Redundancy patterns for each function further revealed that some functions displayed fast rates of increase in redundancy at low species diversity, whereas others were only becoming redundant past a certain threshold. This suggested non-random assembly rules and the existence of some primordial functions that would need to be fulfilled in priority so that coral reef fish assemblages can gain a basic ecological structure. Last, we found little effect of habitat on the shape of the functional-species diversity relationship and on the redundancy of functions, although habitat is known to largely determine assemblage characteristics such as species composition, biomass, and abundance. Our study shows that low functional redundancy is characteristic of this highly diverse fish assemblage, and, therefore, that even species-rich ecosystems such as coral reefs may be vulnerable to the removal of a few keystone species.

Show MeSH
Comparing observed and random distributions of functions among species.A: Functional diversity (number of functions/transect) as a function of species diversity (number of species/transect) for the three functional classification schemes (DS, DSH, DSHG). Lines indicate 95% confidence intervals estimated from random distribution (Monte Carlo analysis with 999 permutations). B: Cumulative frequency distributions of observations for which the number of observed functions was significantly lower than the number of functions obtained from a random distribution (i.e. lower than the 95% confidence interval). DS: diet×size; DSH: diet×size×home range; DSHG: diet×size×home range×gregariousness.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3198825&req=5

pone-0026735-g004: Comparing observed and random distributions of functions among species.A: Functional diversity (number of functions/transect) as a function of species diversity (number of species/transect) for the three functional classification schemes (DS, DSH, DSHG). Lines indicate 95% confidence intervals estimated from random distribution (Monte Carlo analysis with 999 permutations). B: Cumulative frequency distributions of observations for which the number of observed functions was significantly lower than the number of functions obtained from a random distribution (i.e. lower than the 95% confidence interval). DS: diet×size; DSH: diet×size×home range; DSHG: diet×size×home range×gregariousness.

Mentions: Increasing the complexity of the functional classification scheme rendered the relationship between functional and species diversity more linear and closer to the 1∶1 line (Figure 3B). If we had chosen the possibly most complex classification where each species would represent a unique function, then observed and randomly generated functional-species diversity relationships would all have followed the same 1∶1 line, and thus would not differ. To some extent, this was observed for the DSHG classification where most observations (78%) fell within the 95% confidence interval estimated from 999 randomly generated relationships (Figure 4). However, none of the 22% of the DSHG observations that fell outside the 95% CI were greater than the upper limit of the CI (Figure 4A); this was true for all classification schemes. Moreover, the number of observations that were smaller than the 95% CI lower limit increased for simpler classifications, for which the general functional-species diversity relationship was therefore further from the 1∶1 line (Figure 4). Whereas 22% of the DSHG observations fell below the 95% CI, the percentage increased to 40% for DSH and 60% for DS (Figure 4B). Thus, there were significantly fewer functions in the observed assemblage than was expected by chance, implying that increasing species diversity primarily increased the redundancy of some essential functions rather than the functional diversity per se. Observations that fell below the 95% CI were not randomly distributed. Cumulative frequency distributions were S-shaped (Figure 4B), implying that past a certain level of species diversity, most observations fell within the CI (Figure 4A). Interestingly, this threshold was similar for all functional classification schemes and could be visually estimated at approximately 90 species (Figure 4).


Functional redundancy patterns reveal non-random assembly rules in a species-rich marine assemblage.

Guillemot N, Kulbicki M, Chabanet P, Vigliola L - PLoS ONE (2011)

Comparing observed and random distributions of functions among species.A: Functional diversity (number of functions/transect) as a function of species diversity (number of species/transect) for the three functional classification schemes (DS, DSH, DSHG). Lines indicate 95% confidence intervals estimated from random distribution (Monte Carlo analysis with 999 permutations). B: Cumulative frequency distributions of observations for which the number of observed functions was significantly lower than the number of functions obtained from a random distribution (i.e. lower than the 95% confidence interval). DS: diet×size; DSH: diet×size×home range; DSHG: diet×size×home range×gregariousness.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3198825&req=5

pone-0026735-g004: Comparing observed and random distributions of functions among species.A: Functional diversity (number of functions/transect) as a function of species diversity (number of species/transect) for the three functional classification schemes (DS, DSH, DSHG). Lines indicate 95% confidence intervals estimated from random distribution (Monte Carlo analysis with 999 permutations). B: Cumulative frequency distributions of observations for which the number of observed functions was significantly lower than the number of functions obtained from a random distribution (i.e. lower than the 95% confidence interval). DS: diet×size; DSH: diet×size×home range; DSHG: diet×size×home range×gregariousness.
Mentions: Increasing the complexity of the functional classification scheme rendered the relationship between functional and species diversity more linear and closer to the 1∶1 line (Figure 3B). If we had chosen the possibly most complex classification where each species would represent a unique function, then observed and randomly generated functional-species diversity relationships would all have followed the same 1∶1 line, and thus would not differ. To some extent, this was observed for the DSHG classification where most observations (78%) fell within the 95% confidence interval estimated from 999 randomly generated relationships (Figure 4). However, none of the 22% of the DSHG observations that fell outside the 95% CI were greater than the upper limit of the CI (Figure 4A); this was true for all classification schemes. Moreover, the number of observations that were smaller than the 95% CI lower limit increased for simpler classifications, for which the general functional-species diversity relationship was therefore further from the 1∶1 line (Figure 4). Whereas 22% of the DSHG observations fell below the 95% CI, the percentage increased to 40% for DSH and 60% for DS (Figure 4B). Thus, there were significantly fewer functions in the observed assemblage than was expected by chance, implying that increasing species diversity primarily increased the redundancy of some essential functions rather than the functional diversity per se. Observations that fell below the 95% CI were not randomly distributed. Cumulative frequency distributions were S-shaped (Figure 4B), implying that past a certain level of species diversity, most observations fell within the CI (Figure 4A). Interestingly, this threshold was similar for all functional classification schemes and could be visually estimated at approximately 90 species (Figure 4).

Bottom Line: First, we found that the relationship between functional and species diversity displayed a non-asymptotic power-shaped curve, implying that rare functions and species mainly occur in highly diverse assemblages.Last, we found little effect of habitat on the shape of the functional-species diversity relationship and on the redundancy of functions, although habitat is known to largely determine assemblage characteristics such as species composition, biomass, and abundance.Our study shows that low functional redundancy is characteristic of this highly diverse fish assemblage, and, therefore, that even species-rich ecosystems such as coral reefs may be vulnerable to the removal of a few keystone species.

View Article: PubMed Central - PubMed

Affiliation: UR-CoRéUs, Institut de Recherche pour le Développement, Noumea, New Caledonia. nicolas.guillemot@gmail.com

ABSTRACT
The relationship between species and the functional diversity of assemblages is fundamental in ecology because it contains key information on functional redundancy, and functionally redundant ecosystems are thought to be more resilient, resistant and stable. However, this relationship is poorly understood and undocumented for species-rich coastal marine ecosystems. Here, we used underwater visual censuses to examine the patterns of functional redundancy for one of the most diverse vertebrate assemblages, the coral reef fishes of New Caledonia, South Pacific. First, we found that the relationship between functional and species diversity displayed a non-asymptotic power-shaped curve, implying that rare functions and species mainly occur in highly diverse assemblages. Second, we showed that the distribution of species amongst possible functions was significantly different from a random distribution up to a threshold of ∼90 species/transect. Redundancy patterns for each function further revealed that some functions displayed fast rates of increase in redundancy at low species diversity, whereas others were only becoming redundant past a certain threshold. This suggested non-random assembly rules and the existence of some primordial functions that would need to be fulfilled in priority so that coral reef fish assemblages can gain a basic ecological structure. Last, we found little effect of habitat on the shape of the functional-species diversity relationship and on the redundancy of functions, although habitat is known to largely determine assemblage characteristics such as species composition, biomass, and abundance. Our study shows that low functional redundancy is characteristic of this highly diverse fish assemblage, and, therefore, that even species-rich ecosystems such as coral reefs may be vulnerable to the removal of a few keystone species.

Show MeSH