Biodiversity mitigates drought effects in the decomposer system across biomes

Abstract

Multiple facets of global change affect the earth system interactively, with complex consequences for ecosystem functioning and stability. Simultaneous climate and biodiversity change are of particular concern, because biodiversity may contribute to ecosystem resistance and resilience and may mitigate climate change impacts. Yet, the extent and generality of how climate and biodiversity change interact remain insufficiently understood, especially for the decomposition of organic matter, a major determinant of the biosphere-atmosphere carbon feedbacks. With an inter-biome field experiment using large rainfall exclusion facilities, we tested how drought, a common prediction of climate change models for many parts of the world, and biodiversity in the decomposer system drive decomposition in forest ecosystems interactively. Decomposing leaf litter lost less carbon (C) and especially nitrogen (N) in five different forest biomes following partial rainfall exclusion compared to conditions without rainfall exclusion. An increasing complexity of the decomposer community alleviated drought effects, with full compensation when large-bodied invertebrates were present. Leaf litter mixing increased diversity effects, with increasing litter species richness, which contributed to counteracting drought effects on C and N loss, although to a much smaller degree than decomposer community complexity. Our results show at a relevant spatial scale covering distinct climate zones that both, the diversity of decomposer communities and plant litter in forest floors have a strong potential to mitigate drought effects on C and N dynamics during decomposition. Preserving biodiversity at multiple trophic levels contributes to ecosystem resistance and appears critical to maintain ecosystem processes under ongoing climate change.

Keywords: biodiversity; climate change; ecosystem functioning; forest carbon cycling; litter decomposition.

Fig. 1.

Litter C (A) and N (B) loss and drought effect size for C loss (C) and N loss (D) with increasing complexity of decomposer communities. Decomposer community complexity was manipulated by using litterbags of 0.05 mm mesh size [allowing the presence of microorganisms and microfauna (e.g., nematodes and protists) only], litterbags of 1 mm mesh size (allowing additionally the access by mesofauna (e.g., springtails and mites)), and litterbags of 5 mm mesh size [allowing access by the full complexity of decomposers communities including macrofauna (e.g., millipedes and isopods)]. Mean values of all litter treatments across the five locations are shown for control plots (white) and for partial rainfall exclusion plots (orange) in panels (A and B) [n = 15 litter combinations × 18 replicated plots per treatment (control vs. drought) for the five locations = 270, see Table 1 for statistics]. The numbers on top of each pair of columns show the relative difference between the mean values in the control and the drought treatment for each level of decomposer complexity. Dashed lines in panels (A and B) indicate average values of litter C or N loss in the control treatment with the simplest decomposer community. The drought effect size (lnR) of C loss (panel C) and N loss (Hedges d, panel D) are shown for each level of decomposer complexity, with different capital letters indicating significant decomposer complexity differences at P < 0.05 (no significant differences for drought effect size on N loss). Dashed lines in panels (C and D) indicate the zero line (negative effects sizes below and positive effect sizes above the zero line). Linear mixed models, where location was considered as random effect were used for statistical testing and comparisons were done by the least square significant (LSD) method.

Fig. 2.

Net diversity effects in two-, three-, and four-species litter mixtures on C loss (A) and N loss (B) and drought response ratios on C loss (C) and N loss (D). Mean values of all litter mixture treatments (single litter species treatments were used to calculate net diversity effects) across the five locations are shown for control plots (white) and for partial rainfall exclusion plots (orange) [n = 6, 4, and 1 litter combinations × 3 mesh sizes × 18 replicated plots per treatment (control vs. drought) for the five locations = 324, 216, and 54 for two, three, and four species mixtures, respectively., see Table 2 for statistics]. Asterisks indicate significant differences between control and drought treatments for each species richness level separately (*, **, and *** denote significant effects at P < 0.05, P < 0.01, and P < 0.001, and ns = not statistically significant). Panels (C) and (D) show the response ratios of the net diversity effects under drought in three- and four-species litter mixtures compared to the two-species mixtures under control conditions for C loss (C) and N loss (D).

Fig. 3.

Net diversity effects with increasing decomposer community complexity [from left (0.05 mm mesh) to right (5 mm mesh)] on C loss (A) and N loss (B). Mean values of all litter mixture treatments (single litter species treatments were used to calculate net diversity effects) across the five locations are shown for control plots (white) and for partial rainfall exclusion plots (orange) [n = 11 litter mixtures × 18 replicated plots per treatment (control vs. drought) for the five locations = 198, see Table 2 for statistics]. The numbers on top of each pair of columns indicate the relative difference between the mean values in the control and the drought treatment followed by asterisks for statistical significance (* and ** denote significant effects at P < 0.05 and P < 0.01, and ns = not statistically significant) for each level of decomposer complexity.