Species‐rich boreal forests grew more and suffered less mortality than species‐poor forests under the environmental change of the past half‐century

Abstract

Climate and other global environmental changes are major threats to ecosystem functioning and biodiversity. However, the importance of plant diversity in mitigating the responses of functioning of natural ecosystems to long‐term environmental change remains unclear. Using inventory data of boreal forests of western Canada from 1958 to 2011, we found that aboveground biomass growth increased over time in species‐rich forests but decreased in species‐poor forests, and importantly, aboveground biomass loss from tree mortality was smaller in species‐rich than species‐poor forests. A further analysis indicated that growth of species‐rich (but not species‐poor) forests was statistically positively associated with rising CO2, and that mortality in species‐poor forests increased more as climate moisture availability decreased than it did in species‐rich forests. In contrast, growth decreased and mortality increased as the climate warmed regardless of species diversity. Our results suggest that promoting high tree diversity may help reduce the climate and environmental change vulnerability of boreal forests.

Spatial variation in climate modifies effects of functional diversity on biomass dynamics in natural forests across Canada

“…The average ring counts of the tree samples for the species with the oldest ages were used to provide a conservative estimate for the time since fire (Chen et al, ; Gutsell & Johnson, ). We used the middle stand age (SA) of a census period (i.e., the period between two successive censuses) to account for effects of forest development processes (representing changes in stem density and composition; Hisano et al, ) on growth, mortality and ΔAGB.…”

“…The two‐way interactions of SA, CMI ave , MAT ave , FD and CWM PC1 were included based on the model selection by the Akaike information criterion (AIC). Based on AIC, the SA, FD and CWM PC1 were transformed by natural logarithmic or squared root functions (Hisano et al, ; Ratcliffe et al, ). All the explanatory variables were centred and scaled (mean = 0, SD = 1) before analysis to allow coefficient comparison.…”

“…Although the maximum variance inflation factors (VIFs) among the explanatory variables were 3.02 for growth, 2.85 for mortality and 3.05 for ΔAGB, there was a positive correlation between CMI ave and MAT ave ( r = .44). To ensure that multicollinearity did not influence coefficient estimates, we compared the outcomes from Equation 1 (accounting for both CMI ave and MAT ave simultaneously) with those from alternative models accounting for either CMI ave or MAT ave individually (individual models) (Supporting Information Figure S5), similar to the previous study (Hisano et al, ). Moreover, we also refitted Equation 1 by ridge regression (penalized parameter estimates; Dormann et al, ) using the glmnet package (Friedman, Hastie, Simon, Qian, & Tibshirani, ).…”

“…Compared with species‐poor forests, species‐rich forests might have less mortality owing to a lower intensity of competition, with niche partitioning among species. However, the limited studies have reported that mortality either increases (Poorter et al, ) or decreases (Hisano, Chen, Searle, & Reich, ) with increasing tree diversity. This discrepancy could be attributed to potentially different responses of mortality to diversity associated with the abiotic environment, as has been suggested for growth (Mori, ; Paquette & Messier, ; Ratcliffe et al, ).…”

“…By explicitly factoring out the effects of abiotic drivers and functional identity (community‐weighted mean of resource acquisition traits) from those of functional diversity, we tested the following hypotheses: (a) the effects of functional diversity on growth would be greater in drier/colder sites than in humid/warmer sites; and (b) mortality would decrease with functional diversity owing to reduced competition, and the magnitude of the decrease would be greater in drier/colder sites than in humid/warmer sites. We modelled changes in biomass with functional diversity, which represents niche complementarity (Laliberte & Legendre, ), while simultaneously accounting for the influence of plot size (Chisholm et al, ), functional identity (representing the selection effect; Ratcliffe et al, ) and stand age (Hisano et al, ).…”