Drivers of spatial variability of soil respiration along altitudinal gradient in Northwest Caucasus Mountains

Mountains occupy almost a quarter of the land area and store significant pools of soil organic matter (SOM), which is a potential source of atmospheric CO2 under warming climate. However, carbon fluxes in mountain areas with high environmental heterogeneity remain poorly understood, in particular regarding the spatial variability of soil respiration (RS). The study was conducted on the northeastern slope of the Northwest Caucasus Mountains (1260-2480 m a.s.l.; Russia) that crossed five vegetation belts (i.e., mixed, fir and deciduous forests, subalpine and alpine meadows). RS was measured simultaneously (at 10 a.m. on 11 August 2018) across five vegetation belts (at 12 randomly distributed points per belt; totally n = 60) using the closed static chamber technique. As potential drivers of RS spatial variability, soil physico-chemical (temperature, moisture, total and dissolved C and N contents, C:N ratio, pH), soil microbial (microbial biomass C content, basal respiration, enzymatic activities: ?-glucosidase, chitinase and leucine aminopeptidase) and vegetation properties (grasses projective cover, its species richness, Shannon-Wiener diversity index, abundance of graminoids and forbs) were assessed. The RS rate ranged from 1.3-12.7 µmol CO2 m-1 s-1, with average values of 3.7 and 7.3 µmol CO2 m-1 s-1 for forests and grasslands respectively. Stepwise regression and subsequent path analysis showed that key driver of RS spatial variability in forests was temperature-sensitive soil chitinase activity (explained variance 50%), while in grasslands it was graminoid abundance (explained variance 27%). The forest soils are mostly limited in N, therefore RS variability depends largely on SOM-derived CO2 sources, i.e. activity of the N-acquiring enzyme. In the grasslands, extensive network of fine roots and the associated considerable contribution of root-derived respiration to Rs, makes the flux more sensitive to vegetation composition and associated phenology and C allocation patterns. Thus, soil N availability and differences in plant cover play a crucial role in regulation of RS spatial patterns in mountains ecosystems.