Background: In recent years, some ecologists have advocated the use of functional groups instead of direct species in linking site composition to the environment. They could potentially reveal connections between distant sites and aid in the formation of widely-applicable environmental policies. Several studies have compared the efficiency of using functional groups, in which species are grouped based on functional traits, like feeding method or size, to using species directly. However, few have looked at the effect of varying the complexity of functional groups when compared to species data. This study compares functional group classes of varying complexity, with complexity defined as the number of traits considered, to species data. The hypothesis that more complex functional group classes, compared to less complex classes, tend to approach the results obtained when using taxonomy, is tested.
Methods: In testing this hypothesis, this study uses site composition data from aquatic floor (benthic) ecosystems in the Canadian Arctic. Four functional traits were considered important to describe these species: Bioturbation (sediment disturbance), body size, feeding habit and mobility. These traits were used to segregate species into functional groups of varying complexity, with complexity level determined by the number of traits (out of four) being used. Four environmental characteristics were considered for each site: Chlorophyll a, phaeopigments, depth and salinity. In order to test how similar functional group data is to species data, we sought to determine whether the same environmental variables were important in explaining site composition. This was determined by BIO-ENV analyses and Spearman Rank correlations. Mantel permutation tests then determined whether the correlations were significant.
Results: While all levels of complexity, from one to four functional traits, showed some significant correlations (Spearman Rank ≥0.5, p≤ 0.05) between site composition and environmental variables, there was no general trend suggesting functional group complexity correlates with greater similarity to taxonomic data. For presence/absence data, all functional results, regardless of complexity, pinpointed only phaeopigments as important, while presence/absence species data also included chlorophyll a and depth. All results with strong and significant correlations (r≥0.5 p≤ 0.05), regardless of data type or complexity, maintained a measure of food supply (Chlorophyll a or phaeopigments), demonstrating its importance in determining ecosystem composition at these sites.
Limitations: Potential improvements include measuring traits directly from the organisms, considering more environmental variables and increasing the number of functional traits considered. Which traits are considered also vary with each study.
Conclusions: The hypothesis was not validated by the results. When pinpointing the most complex functional group class (the most important variable), rather than a less complex class, it was not guaranteed that the chosen variables would be the same as species data. Some classes of less complexity showed greater similarity to full species data. Some outcomes, like the presence/absence results, also imply certain species redundancies in the ecosystem, particularly regarding depth. These results have implications for the concept of functional redundancies in ecosystems, an important point in developing widely applicable environmental policies.
