A central goal of ecology is determining how biodiversity influences ecosystem function. There are two non-exclusive hypotheses for how this may occur: 1) the mass ratio (MR) effect, which holds that species affect ecosystem functions in direct proportion to their abundance; and 2) niche complementarity (NC), which postulates that increased species diversity will enhance ecosystem functions through complementary resource use. We tested the relative importance of these two hypotheses on net primary production (NPP) over eight years (sampled between 2008 and 2016) in 81 1m2 long term monitoring plots in alpine tundra at the Niwot Ridge LTER, Colorado, USA. Using structural equation models (SEM), we evaluated how environmental conditions (topography, snow depth), and the community-weighted means (MR) and functional diversity (NC) of five putatively important plant functional traits (specific leaf area, height, leaf area, leaf dry matter content, and chlorophyll content) simultaneously influence NPP. We predicted that overall, MR effects would have a larger influence on NPP than NC effects, but that this relationship would likely vary over time.
Results/Conclusions
Consistent with our predictions, we found that MR effects were the dominant mechanism by which biodiversity influenced NPP in this system. While the patterns varied among different traits, in general, community-weighted means were positively correlated with NPP while functional diversity was either negatively correlated or not correlated with NPP. Of the five traits examined, models with leaf area best predicted NPP, explaining up to 61% of the variation in NPP for a given year. The positive relationship between leaf area community-weighted means and NPP was significant in all eight study years. In contrast, leaf area functional diversity had little influence on NPP and was significantly negatively related to NPP in only two years. While snowpack is considered an important driver of productivity in alpine systems, its influence on NPP was consistently smaller than that of leaf area community-weighted means and was only significant in five of the study years. Our study supports the use of functional traits and SEMs for investigating the biodiversity-ecosystem function relationship and adds to a growing body of literature that suggests that MR effects play a larger role in determining ecosystem functions than NC effects.