Primary factors controlling variations in soil microbial respiration at different scales
As our second hypothesis, in this study we found that factors controlling Rm’s spatial variation and underlying mechanisms were different at the different scales, and revealed that the relative importance of climate, soil physicochemical and microbial properties in driving Rm’s spatial variation varied with scales according to variance partitioning results (Fig. 4), although they together affected the Rm’s spatial patterns through their various impacts on microbial activities (Colman & Schimel, 2013). We found that climate affected indirectly Rm’s variation by changing different soil physicochemical and microbial properties at different scales (Fig. 5), although some previous studies demonstrated Rm was related to MAT and MAP at different scales (e.g., Li et al., 2020). At the continental scale, climate affected indirectly Rm by altering mainly LOC, fungal and gram-positive bacterial biomass (Fig. 5a), which was supported by the observation of Wang et al. (2018), while in the latitude < 32.5°N region, the alteration of fungal and gram-positive bacterial biomass by climate played a critical role in Rm’s variation (Fig. 5b), and the changes in LOC and SOC caused by climate controlled Rm’s variation in the latitude > 32.5°N region (Fig. 5c). In comparison with climate and soil microbial variables, the largest total (49.5%) and unique (13.6%) explanation of soil physicochemical property for the spatial variation in Rm (Fig. 5a) suggested that soil physicochemical property had the most importance in regulating the Rm’s spatial variation across China forest ecosystems. As some previous studies reported (Dai et al. 2017; Herbst et al. 2011; Meyer et al. 2017), some physicochemical variables such as the quality and quantity of the substrate as the direct factors of soil microbial metabolisms exert a crucial role in determining the spatial patterns of microbial activity. We further revealed that LOC was the most significant explanatory variable among physicochemical factors in regulating the Rm’s spatial variation based on boosted regression analysis (Fig. 6a), which was supported by the positive relationship between LOC content and Rm (Fig. S1b). This is because high LOC content is usually associated with high microbial availability and C energy, and preferentially utilized by soil microorganisms.
Soil microbial property had a unique explanation (6.13%) for the spatial variation in Rm at the continental scale (Fig. 4a). Our results provide a first direct experimental evidence that soil microbial characteristics play a unique role in regulating Rm’s spatial variation at the continental scale and suggest the relationships between microbial community composition and Rm potentially linked to microbial life strategies and functional capabilities (Takur et al., 2018; Trivedi et al., 2016). Various microbial groups with different C use efficiency (Austin et al., 2004; Waring et al., 2013) and preferences to soil organic matter (Lehmann and Kleber, 2015; Ramirez et al., 2012) maybe lead to this unique explanation of microbial variables to the Rm’s spatial variation, which was in accord with recent studies suggesting the significant role of soil microbial biomass and community in controlling ecosystem multiple functions (Bradford et al. 2017; Delgado-Baquerizo et al., 2016b; Liu et al., 2018b). We quantified the importance of major microbial groups measured by PLFAs in predicting Rm based on the boosted regression analysis and revealed fungi were the most important drivers of the spatial variation in Rm (Fig. 6a), supporting the growing literature that demonstrates the significance of fungi in driving soil functions (de Boer et al., 2005; Dacal et al., 2019; Delgado-Baquerizo et al., 2016b; Wagg et al., 2014). Together with other results, we highlight the importance of including microbial community composition in Earth system model to improve our ability to predict C feedbacks in terrestrial ecosystems.
More importantly, we found an interesting result that microbial variables had larger unique explanations for Rm’s spatial variation in the latitude < 32.5°N region than in the latitude > 32.5°N region (Fig. 4b,c) and for the first time revealed the differences in the primary factors regulating the Rm’s variations in the latitude < 32.5°N and > 32.5°N regions. Microbial property, especially fungal biomass as the primary predictors in regulating Rm’s variations was probably related to higher microbial activity because of high MAT (Fig. 5b; Fig. 6b). Although some previous studies demonstrated that in cropland or grassland ecosystems soil microbial biomass was a strong regulator of soil respiration or litter decomposition at regional or continental scales (Colman & Schimel, 2013; Bradford et al., 2017), they did not identify its relative importance and also not observe that fungal biomass had different importance in regulating Rm’s spatial variation in different regions. Therefore, our findings advanced our understanding of Rm’s spatial variation and its controlling mechanisms at different scales.
Although we have some important findings that to some extent enhance our understanding of Rm’s variation and its mechanisms, here we noted that the unique explanation of soil microbial property may be underestimated because some key enzyme activities associating with SOC process were not measured. Some experiments have emphasized the important role of enzyme activities in controlling Rm through mediating the rate-limiting step of SOC depolymerization (Ali et al., 2018; Dungait et al., 2012; Kandeler et al., 2006). Therefore, we considered that the unique explanation of soil microbial property to Rm’s spatial variation would be higher than our current estimation if some key enzyme activities are measured. Another important aspect to mention here is that we controlled soil samples under 60% water holding capacity, which made Rm independent of soil moisture (Ali et al., 2018; Wang et al., 2018). That is to say, the influence of soil moisture on Rm was not considered in our study, although soil moisture has strong effect on Rm (Chang et al., 2014; Stoyan et al., 2000). Therefore, we should be careful when directly using these results in predicting Rm in field and its dynamics under global change environment.
In conclusion, we highlighted the importance of incorporating Rm’s spatial variation in C-climate models for better predicting responses of forest soil C dynamics to future environmental change using a novel incubation method and for the first time revealed the hump-shaped relationship between latitude and Rm with a latitudinal threshold of 32.5°N at the continental scale. The factors controlling Rm’s spatial variation and underlying mechanisms varied in different regions, although labile organic C was the most important variable in regulating the Rm variation at the continental scale. Importantly, soil microbial variables, particularly fungal biomass, played an important and unique role in regulating Rm’s spatial variation, but their significances were higher in shaping Rm’s variation across the continental and latitude < 32.5°N region than at latitude > 32.5°N region. Overall, our findings suggest labile organic C and fungal have critical roles in controlling Rm in China’s forest ecosystems, and suggest that including Rm spatial variation in Earth system models can potentially improve our capacity to predict changes in soil organic C balance under changing environment.