PS 50-60
A state factor approach reveals little effect of tree species richness on forest carbon storage
Tests of the effects of plant diversity on ecosystem processes have typically used synthetic communities, and only rarely have manipulated other state factors that also control ecosystem processes. Although ecosystem ecology has a long history of testing effects of plant traits on ecosystem processes, there is much recent debate regarding the relative strength of effects of plant traits versus plant diversity. Furthermore, translating effects of plant diversity on ecosystem processes into effects on ecosystem services often requires understanding multiple ecosystem controls that may operate differently on different processes. For these reasons, our knowledge of how natural variation in plant traits and diversity affect ecosystem services remains poor. We investigated the role of ecosystem state factors (climate, organism traits, topography, and time since disturbance) on forest carbon storage using data from long-term monitoring plots in temperate and boreal forests of Québec. We used a series of structural equation models of increasing complexity to test the roles of abiotic state factors alone, abiotic state factors plus community-weighted mean (CWM) plant functional traits, and abiotic state factors plus CWM traits plus plant diversity, on four separate ecosystem carbon pools (live plant, standing dead, coarse woody debris, and soil organic horizon).
Results/Conclusions
Models using only abiotic state factors explained 31 and 80% of variance in detrital and living (tree) ecosystem carbon pools, respectively, and 75% of variance in total ecosystem C stocks. We tested for additional variance explained by CWM of a variety of different plant functional traits and found that the best predictors were waterlogging tolerance for total and detrital C and maximum height for living tree biomass, each of which explained an additional 1% of the variance. In comparison, species richness explained less than 1% or no additional variation in any of the carbon pools. These results have several implications. First, and not surprisingly, at the landscape scale, ecosystem carbon storage responds to differing abiotic controls on production and decomposition – as soils became poorly drained, detrital carbon increased exponentially; live tree carbon responded primarily to temperature and precipitation. Second, CWM plant functional traits added minor amounts of predictability to estimates of ecosystem carbon storage, but the traits relevant for pools responding to live biomass were different from those relevant for soil C storage. Finally, in contrast to prior results for productivity, we found little reason to incorporate metrics of plant diversity into predictions of ecosystem carbon storage.