Introduction
Rising soil microbial respiration (Rm) is an important process that
controls CO2 flux from terrestrial soil to the
atmosphere and soil organic carbon (SOC) dynamics (Bond-Lamberty &
Thomson, 2010; Chen et al., 2015), and is expected to accelerate under
ongoing climate warming (Bond-Lamberty et al., 2018). Given that
globally Rm is the second largest C efflux in terrestrial ecosystems
(IPCC, 2017), a small change in Rm rate will have the potential to
greatly affect atmospheric CO2 concentration and then
modify feedbacks between SOC dynamics and global change (Bond-Lamberty
et al., 2018; Rodeghiero and Cescatti, 2005). Therefore, accurately
predicting Rm rate is urgently needed for projecting atmospheric
CO2 concentration and reducing the uncertainty in the
prediction of SOC dynamics by carbon–climate models (Schmidt, 2011).
Despite the importance of Rm in global C cycling, our understanding of
the spatial variation in Rm and its shaping mechanism at the continental
scale is still quite limited.
Soil Rm is primarily controlled by abiotic and biotic factors, including
climate (Karhu et al., 2014), soil physicochemical properties
(Delgado-Baquerizo et al., 2016a; Wang et al., 2017) and microbial
characteristics (Dacal et al., 2019; Takur et al., 2018). These factors
will inevitably affect the spatial variation of Rm at various scales
(e.g., Ali et al., 2018; Meyer et al., 2017; Ngao et al., 2012).
However, what the importance of these factors in shaping Rm’s spatial
variation remains unclear, although these previous studies significantly
advanced our understanding of Rm variation and its influencing factors.
Furthermore, traditional incubation methods for measuring Rm in previous
studies have been questioned because the incubation temperature differs
from the ambient soil temperature, with a result that reported Rm is
less representative for soil microbial activity (Xu et al., 2017). More
importantly, in most of previous studies soils from various sites with
different mean annual temperature (MAT) were incubated at the same
temperature, resulting in inaccurate estimation of Rm because Rm
generally increases with increasing temperature (Hamdi et al., 2013; Liu
et al., 2018a). In addition, most of these previous studies focusing on
the spatial variation in Rm only at the small (e.g., stand, ecosystem or
regional) scales resulted in difficultly scaling up to the continent,
even global scale. Therefore, it is urgent to provide a clear picture of
spatial variation in Rm at the continental scale. In this study we will
use a novel incubation experiment based on MAT of soil origin sites near
to the ambient soil temperature to estimate Rm at the continent scale.
Growing studies have emphasized the significance of microbial community
composition, activity and diversity (Dacal et al., 2019;
Delgado-Baquerizo et al., 2016b; Takur et al., 2018) in driving Rm. A
recent study noted that at the regional scale microbial biomass C was a
significant variable in explaining basal soil respiration (Ali et al.,
2018) because considerable proportion of Rm is controlled by the enzyme
activities through mediating the rate-limiting step of SOC
depolymerization (e.g., Kandeler et al., 2006). Given that bacteria have
lower C use efficiency and prefer more labile organic matter than fungi
(Lehmann & Kleber, 2015; Waring et al., 2013) and higher soil-C
assimilation of gram-positive bacteria relative to gram-negative
bacteria (Creamer et al., 2015), we assume that differences in microbial
community composition (e.g., fungi:bacteria and gram-positive and
-negative bacteria ratios) may have unique contribution in shaping the
spatial variation in Rm. However, this assumption has not been verified,
particularly at the continental scale. Furthermore, whether their
relative importance in driving Rm’s spatial variation varies among
different scales remains largely explored.
Here, we conducted a novel incubation experiment using mean annual
temperature (MAT) of each soil origin site for 238 soils along a 4,200
km north–south transect of forests in China to explore the spatial
variation in Rm and its underlying mechanisms at different scales. The
MAT of each soil origin site was first used as the incubation
temperature of the corresponding soils to assess Rm’s spatial variation.
In order to identify the key influencing factors and quantify the
relative importance of climate, soil physicochemical and microbial
properties in shaping Rm’s spatial variation, variance partitioning and
boosted regression analyses have been performed. The objectives of this
study were to explore the Rm’s spatial variations in forest ecosystems
along a latitudinal gradient at the continental and regional scales, and
reveal their potential driving mechanisms. In this study, we
hypothesized that Rm would linearly change with increasing latitude and
climate (i.e., MAT and mean annual precipitation) would have important
roles in shaping the Rm spatial variation. Given the metabolic activity
and biomass of soil microorganisms generally increase toward warm, moist
tropical regions (Crowther et al., 2019; Tedersoo et al., 2019), as well
as that microbial property links closely with soil properties in more
productive ecosystems (i.e., tropical forest) (Delgado-Baquerizo et al.,
2019). We further hypothesized that there would be different mechanisms
in explaining Rm’s spatial variations in different regions.