Soil organic carbon (SOC) changes under future climate warming are difficult to quantify in situ. Here we apply an innovative approach combining space-for-time substitution with meta-analysis to SOC measurements in 113,013 soil profiles across the globe to estimate the effect of future climate warming on steady-state SOC stocks. We find that SOC stock will reduce by 6.0 ± 1.6% (mean±95% confidence interval), 4.8 ± 2.3% and 1.3 ± 4.0% at 0–0.3, 0.3–1 and 1–2 m soil depths, respectively, under 1 °C air warming, with additional 4.2%, 2.2% and 1.4% losses per every additional 1 °C warming, respectively. The largest proportional SOC losses occur in boreal forests. Existing SOC level is the predominant determinant of the spatial variability of SOC changes with higher percentage losses in SOC-rich soils. Our work demonstrates that warming induces more proportional SOC losses in topsoil than in subsoil, particularly from high-latitudinal SOC-rich systems.

Soil carbon (C) is comprised of a continuum of organic compounds with distinct ages (i.e., the time a C atom has experienced in soil since the C atom entered soil). The contribution of different age groups to soil C efflux is critical for understanding soil C stability and persistence, but is poorly understood due to the complexity of soil C pool age structure and potential distinct turnover behaviors of age groups. Here, we build upon the quantification of soil C transit times to infer the age of C atoms in soil C efflux (aefflux) from seven sequential soil layer depths down to 2 m at a global scale, and compare this age with radiocarbon‐inferred ages of C retained in corresponding soil layers (asoil). In the whole 0–2 m soil profile, the mean aefflux is 194211021 (mean with 5%–95% quantiles) year and is just about one‐eighth of asoil (14767172547 year), demonstrating that younger C dominates soil C efflux. With increasing soil depth, both aefflux and asoil are increased, but their disparities are markedly narrowed. That is, the proportional contribution of relatively younger soil C to efflux is decreased in deeper layers, demonstrating that C inputs (new and young) stay longer in deeper layers. Across the globe, we find large spatial variability of the contribution of soil C age groups to C efflux. Especially, in deep soil layers of cold regions (e.g., boreal forests and tundra), aefflux may be older than asoil, suggesting that older C dominates C efflux only under a limited range of conditions. These results imply that most C inputs may not contribute to long‐term soil C storage, particularly in upper layers that hold the majority of new C inputs.

Soil biogeochemical processes may present depth‐dependent responses to climate change, due to vertical environmental gradients (e.g., thermal and moisture regimes, and the quantity and quality of soil organic matter) along soil profile. However, it is a grand challenge to distinguish such depth dependence under field conditions. Here we present an innovative, cost‐effective and simple approach of field incubation of intact soil cores to explore such depth dependence. The approach adopts field incubation of two sets of intact soil cores: one incubated right‐side up (i.e., non‐inverted), and another upside down (i.e., inverted). This inversion keeps soil intact but changes the depth of the soil layer of same depth origin. Combining reciprocal translocation experiments to generate natural climate shift, we applied this incubation approach along a 2200 m elevational mountainous transect in southeast Tibetan Plateau. We measured soil respiration (Rs) from non‐inverted and inverted cores of 1 m deep, respectively, which were exchanged among and incubated at different elevations. The results indicated that Rs responds significantly (p < .05) to translocation‐induced climate shifts, but this response is depth‐independent. As the incubation proceeds, Rs from both non‐inverted and inverted cores become more sensitive to climate shifts, indicating higher vulnerability of persistent soil organic matter (SOM) to climate change than labile components, if labile substrates are assumed to be depleted with the proceeding of incubation. These results show in situ evidence that whole‐profile SOM mineralization is sensitive to climate change regardless of the depth location. Together with measurements of vertical physiochemical conditions, the inversion experiment can serve as an experimental platform to elucidate the depth dependence of the response of soil biogeochemical processes to climate change.