Relationships between species feeding traits and environmental conditions in fish communities: a three-matrix approach
|Author(s)||Brind'Amour Anik1, 4, Boisclair Daniel1, Dray Stephane2, 3, Legendre Pierre1|
|Affiliation(s)||1 : Univ Montreal, Dept Biol Sci, Montreal, PQ H3C 3J7, Canada.
2 : Univ Lyon 1, CNRS, UMR 5558, Lab Biometrie & Biol Evolut, F-69622 Villeurbanne, France.
3 : Univ Lyon, F-69000 Lyon, France.
|Source||Ecological Applications (1051-0761) (Ecological Soc Amer), 2011-03 , Vol. 21 , N. 2 , P. 363-377|
|WOS© Times Cited||36|
|Keyword(s)||ecological niche, feeding traits, fish communities, fourth-corner method, functional group, habitat, lake, Laurentian Shield lakes, Quebec, Canada, littoral zone, spatial scale, three-matrix approach|
|Abstract||Understanding the relationships between species biological traits and the environment is crucial to predicting the effect of habitat perturbations on fish communities. It is also an essential step in the assessment of the functional diversity. Using two complementary three-matrix approaches (fourth-corner and RLQ analyses), we tested the hypothesis that feeding-oriented traits determine the spatial distributions of littoral fish species by assessing the relationship between fish spatial distributions, fish species traits, and habitat characteristics in two Laurentian Shield lakes. Significant associations between the feeding-oriented traits and the environmental characteristics suggested that fish communities in small lakes (displaying low species richness) can be spatially structured. Three groups of traits, mainly categorized by the species spatial and temporal feeding activity, were identified. The water column may be divided in two sections, each of them corresponding to a group of traits related to the vertical distribution of the prey coupled with the position of the mouth. Lake areas of low structural complexity were inhabited by functional assemblages dominated by surface feeders while structurally more complex areas were occupied by mid-water and benthic feeders. A third group referring to the time of feeding activity was observed. Our work could serve as a guideline study to evaluate species traits x environment associations at multiple spatial scales. Our results indicate that three-matrix statistical approaches are powerful tools that can be used to study such relationships. These recent statistical approaches open up new research directions such as the study of spatially based biological functions in lakes. They also provide new analytical tools for determining, for example, the potential size of freshwater protected areas.|