Figure Legends

Fig. 1 The geographic location of the 10 mountains and experimental setup. a: the distribution of the studied mountains spanning subtropical and tropical regions in China. From the West to East: GMT: Gaoligong Mountain; YMT: Jade Dragon Snow Mountain; EMS: E’mei Mountain; JFS: Jinfo Mountain; DMS: Daming Mountain; BWL: Bawangling; GS: Guan Mountain; DBS: Dabie Mountain; DYS: Daiyun mountain; TMS: Tianmu mountain. Different elevations are shown in colours. b: An example of decomposing design. c: standard tea bags before burying. d: retrieved tea bags after 90-day decomposition.
Fig. 2 Elevational patterns of decomposition rate (a) and stabilization (b) in each mountain. Colours represent different mountains. Dashed and solid least squares regression lines represent non-significant and significant relationships between elevation and decomposition rates or stabilization, respectively. The elevation of each mountain has been scaled between 0 to 1.
Fig. 3 The relative importance of environmental drivers for explaining the elevational variations of decomposition rate (k ) in each mountain. The importance value of each driver is extracted from AIC weight based on model averaging. Colours indicate the four groups (navy blue-Microclimate; gray-plant; yellow-Soil physiochemistry; red-microhabitat) of environmental drivers. Temp: mean temperature of the growth season; TempV: temperature variation of the growing season; Mois: mean moisture of the growth season; MoisV: moisture variation of the growth season, Tree: tree species richness; Soil P: soil total phosphorus, Cover: Canopy cover; Thick: ground litter thickness.
Fig. 4 The relative importance of environmental drivers for explaining the elevational variations of stabilization (S ) in each mountain. Other explanations can be found in Figure 3.
Fig. 5 Scatter plots showing the effect of four microclimatic drivers on decomposition rates (a, b, c, d) and stabilization (e, f, g, h) in each mountain. The black regression line in each panel represents the relationship across all mountains. The shaded area shows the confidence interval at the 95% level. Solid and dashed lines indicate significant and non-significant effects, respectively.
Fig. 6 Scatter plots showing the effect of three selected non-climatic drivers on decomposition rates (a, b, c) and stabilization (d, e, f) in each mountain. Other explanations can be found in Figure 5.