3.2. Influencing factors of the spatial variation in soil microbial respiration
Variables measured in this study had different roles in shaping Rm’s spatial variations at the continental scale and latitude < 32.5°N and > 32.5°N regional scales, and soil physicochemical properties had greater unique explanation for this variation than others (Fig. 4). Variation partitioning analysis results demonstrated that across all soil samples all variables measured totally explained 56.1% of the Rm’s spatial variation at the continental scale, in which climate, soil physicochemical and microbial properties explained separately 11.5%, 49.5%, and 42.4% of the Rm’s spatial variations (Fig. 4a). Among all factors, soil physicochemical and microbial properties () uniquely explained 13.61% and 6.13% of Rm’s spatial variation, but climate variables less uniquely explained across the continent. In the latitude < 32.5°N region, all variables explained 77.4% of the spatial variation in Rm, and soil physicochemical and microbial properties explained uniquely 13.3%, and 4.6% of this variation, separately (Fig. 4b), but in the latitude > 32.5°N region, all variables totally explained only 34.6% of the Rm’s spatial variation, and soil physicochemical and microbial properties uniquely explained 13.0%, and 1.3% of this spatial variation, separately (Fig. 4c). These results suggest that the unique role of soil physicochemical properties in regulating the spatial variation in Rm has less change in different regions but the unique role of microbial properties varied.
The piecewise SEM explained 76% (Fig. 5a), 92% (Fig. 5b) and 62% (Fig. 5c) of the Rm’s variations at the continent, latitude < 32.5°N and > 32.5°N regions, respectively, with all models to be significant even when accounting for the random effects of sampling site. Climate was lest important in directly controlling Rm’s variations than soil physicochemical and microbial properties in all case, but it could indirectly affect Rm’s variations by influencing soil physicochemical and microbial properties (Fig. 5), and this indirect effect mainly caused by MAT was stronger in the latitude < 32.5°N region (Fig. 5b). LOC and fungal biomass were most important among soil physicochemical and microbial property, respectively, in all case (Fig. 5). The SEM results showed that importance of climate, soil physicochemical property and microbial property visa different ways in shaping the Rm’s variations varied at different scales.
The boosted regression model analysis was further conducted to quantify the importance of the key influencing factors in shaping the Rm’s spatial variation, and the results demonstrated that the key influencing factors varied at different scales (Fig. 6). At the continental scale, LOC and fungi were more important in shaping the Rm’s variation, respectively contributing 20.5%, and 18.9% relative importance of the spatial variation regulating Rm, (Fig. 6a), which was also shown by the results of Spearman correlation analysis (Fig. S1). While fungi and LOC were more important in the latitude < 32.5°N and > 32.5°N regions, contributing 17.3% (Fig. 6b) and 21.0% (Fig. 6c) relative importance of the spatial variation regulating Rm, respectively.