Much evidence about effects of plant diversity on ecosystem functioning involves studies in relatively small plots at relatively short time scales (one to several years). However, many ecosystem services are driven by multiple processes across relatively broad scales of space and time. We use effects of plant diversity on ecosystem carbon (C) storage to examine the need to expand the spatial and temporal scales of study, and in several other dimensions as well: metrics of diversity, assessing multiple processes, and integrating diversity effects with other ecosystem controls. We expected that ecosystem C storage would be a function of state factors including climate, topography, time, and organisms. We therefore explored the roles of climate, topography, stand age, and the traits and diversity of the dominant organisms (i.e., trees) on carbon storage in the temperate and boreal forests of Québec. We tested effects of abiotic state factors alone and in combination with indices of diversity and/or community-weighted mean (CWM) plant functional traits on ecosystem C pools, including live tree C, standing dead C, soil organic horizon C, and total ecosystem C. We also tested the effectiveness of different diversity metrics – species richness, phylogenetic diversity, and functional diversity - for predicting these pools.
Abiotic state factors alone explained much of the landscape-scale variance in C storage, with different C pools responding to different variables, as expected. Similarly, CWM plant functional traits were the strongest biotic predictors for all C pools, but the traits relevant for live tree C differed from those relevant for organic horizon C. While diversity metrics also explained some additional variance, they were never the best biotic predictors for any forest carbon pool; in several cases, diversity had negative effects on litter layer C. Of the data sets and diversity metrics we tested, species richness was never the best performing. Our results have several implications. First, at landscape scales, BEF studies need to integrate mechanisms from plot-scale research with other known drivers of ecosystem processes. Second, where ecosystem services result from multiple different processes, no simple relationship may exist with any one biological metric – whether functional traits or diversity. Finally, moving beyond the expediency of species richness will help better understand mechanistic effects of diversity on ecosystem processes and services.