Plant assemblages do not respond homogenously to local variation in environmental conditions: functional responses differ with species identity and abundance

Questions We investigated some commonly held assumptions of community assembly theory needed to provide accurate predictions of changes in plant species assemblages across environmental gradients or following environmental change. Do (1) dominant and subordinate species respond in the same way to changes in environmental variables; (2) plant species assemblages show higher interspecific than intraspecific trait responses; and (3) co‐existing dominant species differ in their responses to the same environmental variables? Location Islands in L akes U ddjaure and H ornavan, northern S weden. Methods We explored the responses of forest understorey vegetation assemblages to variation in environmental resources across a chronosequence of 30 lake islands that differ in fire history, above‐ground and below‐ground resource availability and species diversity. For one plot on each island, we measured specific leaf area, leaf dry matter content and foliar N and P of all dominant and subordinate understorey plant species to assess species‐specific and weighted and non‐weighted community‐level trait responses to variation across islands in all major local environmental drivers. Results Consistent with our expectations, we found that species responses to environmental conditions were not homogenous within assemblages, and that responses of dominant and subordinate species differed. Further, intraspecific variation was often an important component of local‐scale plant community‐level responses. Responses were often relatively consistent across species, but dominant species sometimes showed contrasting responses of the same trait to the same environmental factor. Finally, environmental factors that influenced community average trait values also affected functional diversity. Conclusions This study has shown that several common assumptions that underpin community assembly theory do not necessarily hold, and this can cause inaccuracies in predicting plant functional composition responses to changes in environmental variables. Because these assumptions are central to current models that predict vegetation responses to environmental change, it is crucial to further test in which particular environmental context and to what extent these assumptions are critical for model accuracy.