Discussion

Lacking a general understanding of dominant factors driving forest little decomposition leads to high uncertainties in the estimation of carbon dynamics at both regional and global scales and its feedback to climate changes (Davidson and Janssens 2006, Krishna and Mohan 2017). From the perspective of experimental macroecology, this study took advantage of highly diversified environmental gradients along elevations of multiple mountain ecosystems, and analysed the effects and relative importance of 10 abiotic and biotic factors in decomposition processes both within and among mountains by using standardized litter materials. The results demonstrate a high variation of decomposition rates (0.002 – 0.05) and stabilization (0.006 – 0.35) across study sites. Within mountains, soil microclimate played a dominant role in decomposition processes mainly in the western high-altitude mountains with colder environments, a similar magnitude of importance for non-climatic factors emerged in the western and the northernmost mountain, indicating the spatial variation of driving factors in decomposition processes. Across mountains, although tree richness impacted decomposition rates, and moisture, soil pH, and slope affected stabilization significantly, soil temperature possessed the strongest association with both decomposition rates (0.48) and stabilization (-0.46), supporting the importance of microclimate and environmental range or gradient in controlling decomposition processes (Bradford et al. 2014, Gallois et al. 2023). Our study provides a general understanding of microclimate and non-climatic factors’ effects on decomposition across mountains, shedding light on the underlying mechanisms for forest carbon dynamics under global change.

Decomposition rates and stabilization along elevations

Among mountains, the velocity of litter mass loss is generally higher in the east than in the west due to the premium hydrothermal conditions. The highest decomposition rate reached 0.24, with no significant difference between the southeastern mountains. This result is comparable to a study in Kilimanjaro Mountain which used the same standardized litter. Their maximum mean decomposition rate is nearly the same as our observations. The difference is that the stabilization in Kilimanjaro is generally higher possibly caused by the obvious transition between warm-dry and warm-wet weather during the decomposition period (Becker and Kuzyakov 2018). For both studies, high elevations are projected to promote carbon stocks due to slow decomposition rates and/or higher capacity in stabilization possibly contributed by large macroaggregates (Feyissa et al. 2023).

The dominant role of microclimate in decomposition

Of the 10 mountains, the significant changes in decomposition rates and stabilization with elevation are nearly in line with mountains that experienced significant correlations between decomposition (stabilization) and microclimatic factors. Although non-linear and linear correlations emerged for the same factor among mountains, there is a potential convergence of significant effects of microclimate in mountains located in the west. The western mountains generally possess higher elevations and cooler environments because of the expanded monsoon that was caused by the fast uplift of the Tibetan Plateau at ~ 41 million years ago (Wu et al. 2022). Colder environments are often associated with a higher thermal sensitivity of microbes or soil animals, resulting in increased decomposition rates with temperatures (Koven et al. 2017). In contrast, the expected positive effects from the evident temperature gradient on decomposition rates in warmer tropical environments can be neutralized by thermal-adapted microbes (Koven et al. 2017).
Across mountains, we found a significant change in the slope of the decomposition rate as a function of soil mean temperature across all mountains (the best breakpoint at 19.79 ℃, N = 998, P< 0.001). Below the breakpoint, decomposition rates increased with soil temperatures significantly, and the mean soil temperatures of these western mountains fell into this range. Beyond the breakpoint, the increase in soil temperature no longer accelerates decomposition rates. Of the 10 mountains, this ecotone mountain is the only one with an east-west orientation and acts as a barrier to the southward flow of the cold air (Wang et al. 2016), which also experienced a mean soil temperature lower than the breakpoint. Taken together, these observations stress the key role of microclimate in decomposition processes and imply that there might be a threshold of temperature effects on decomposition rates.

Inconsistent effects of non-climatic drivers on decomposition

Non-climatic environmental drivers (e.g., tree species richness, soil biochemistry, and topography) emerge as significant factors constraining litter decay across mountains. For instance, litter decomposition rates increased with tree species diversity, echoing the facilitation of high plant species diversity on decomposition rates via complementary effects caused by multiple functional traits (Handa et al. 2014). Given temperature and species diversity co-vary at the global scale, i.e., tropical forests often possess higher species diversity than boreal forests, it is essential to consider their interaction to increase the predicting accuracy of carbon stocks in forest ecosystems (Spohn et al. 2023). In complex terrain, topography emerges as a significant driver regulating decomposition through the effect on stabilization of litter-derived organic carbon (Fig. 6). Across the globe, complex terrain accounts for more than 50% of the land surface (Rotach et al. 2014), and the explanatory power of slope on the variance of soil CO2 flux exceeds 50% in such complexed ecosystems (Reyes et al. 2017). Our results thus advocate that the prediction of ecosystem functions at large spatial scales should take the assumption of microhabitat (e.g., topography and canopy cover) into account, which has been currently underrepresented in multiple earth system models (Bonan et al. 2013, Phillips et al. 2019, Ren et al. 2024).

Forest litter decomposition towards experimental macroecology

With massively accumulated evidence of different factors controlling decomposition (Cornwell et al. 2008, Becker and Kuzyakov 2018, Forrester et al. 2023), the incomparability between findings impeded our general understanding of drivers in this critical process because of differentiations in approaches, designs, and methodologies (Spake et al. 2022, Catford et al. 2022). This extent of incomparability can be partially addressed by statistical methods in meta-analyses, while a straightforward and efficient solution is conducting a standardized experimental design at multiple locations at once (“coordinated distributed experiments”) (Vellend 2016). For instance, a meta-analysis focusing on surface litterbags at a global scale (a wide vegetation type) uncovers that 70.2% variation in decomposition rates has been attributed to litter quality but not climate (Zhang et al. 2008). A macroecological experiment spanning from Mediterranean forests to boreal forests in Europe verifies the dominant control of macroclimate in decomposition (Joly et al. 2023). These contrasting results expand our view of decomposition but also indicate that there are intrinsic differences in terms of system characteristics that cause divergent results, such as surface vs. buried decomposition, and forest vs. non-forested ecosystem. Our coordinated distributed experiment controls this difference one step forward to the microenvironment in mountainous forests where the microclimate has been intensively buffered by dense canopies and tree diversity is generally high. As tea bags have been used as standardized litter in many ecosystems (e.g., grassland, alpine and boreal forests) (Althuizen et al. 2018, Chen et al. 2018; von Oppen et al. 2024), future macroecological experiments may consider the transferability of these standardized results to realistic mass loss of local litters, contributing profoundly to current and future dynamics of terrestrial carbon and nutrient cycling.
In conclusion, our macroecological experiment across 10 mountains provides empirical evidence that there is a large spatial variation of soil microclimate and non-climatic factors controlling decomposition processes, with a dominant role of microclimate in the western mountains with colder environments. Across all mountains, drivers differ between decomposition rates and stabilization, with a positive effect of temperature-related variables and tree diversity on decomposition rates and mainly negative effects of microclimate, pH, and slope on stabilization. This coordinated distributed experiment generates a general understanding of abiotic and biotic influences in decomposition, laying underlying mechanisms for carbon dynamics in complex terrain.