Microbial physiology plays a critical role in the biogeochemical cycles of the Earth system. However, most traditional soil carbon models are lacking in terms of the representation of key microbial processes that control the soil carbon response to global climate change. In this study, the improved process‐based model TRIPLEX‐GHG was developed by coupling it with the new MEND (Microbial‐ENzyme‐mediated Decomposition) model to estimate total global soil organic carbon (SOC) and global soil microbial carbon. The new model (TRIPLEX‐MICROBE) shows considerable improvement over the previous version (TRIPLEX‐GHG) in simulating SOC. We estimated the global soil carbon stock to be approximately 1195 Pg C, with 348 Pg C located in the high northern latitudes, which is in good agreement with the well‐regarded Harmonized World Soil Database (HWSD) and the Northern Circumpolar Soil Carbon Database (NCSCD). We also estimated the global soil microbial carbon to be 21 Pg C, similar to the 23 Pg C estimated by Xu et al. (2014). We found that the microbial carbon quantity in the latitudinal direction showed reversions at approximately 30°N, near the equator and at 25°S. A sensitivity analysis suggested that the tundra ecosystem exhibited the highest sensitivity to a 1°C increase or decrease in temperature in terms of dissolved organic carbon (DOC), microbial biomass carbon (MBC), and mineral‐associated organic carbon (MOC). However, our work represents the first step toward a new generation of ecosystem process models capable of integrating key microbial processes into soil carbon cycles.

Soil microorganisms are critical biological indicators for evaluating soil health and play a vital role in carbon (C)-climate feedback. In recent years, the accuracy of models in terms of predicting soil C pools has been improved by considering the involvement of microbes in the decomposition process in ecosystem models, but the parameter values of these models have been assumed by researchers without combining observed data with the models and without calibrating the microbial decomposition models. Here, we conducted an observational experiment from April 2021 to July 2022 in the Ziwuling Mountains, Loess Plateau, China, to explore the main influencing factors of soil respiration (R_S) and determine which parameters can be incorporated into microbial decomposition models. The results showed that the R_S rate is significantly correlated with soil temperature (T_S) and moisture (M_S), indicating that T_S increases soil C loss. We attributed the non-significant correlation between R_S and soil microbial biomass carbon (MBC) to variations in microbial use efficiency, which mitigated ecosystem C loss by reducing the ability of microorganisms to decompose organic resources at high temperatures. The structural equation modeling (SEM) results demonstrated that T_S, microbial biomass, and enzyme activity are crucial factors affecting soil microbial activity. Our study revealed the relations between T_S, microbial biomass, enzyme activity, and R_S, which had important scientific implications for constructing microbial decomposition models that predict soil microbial activity under climate change in the future. To better understand the relationship between soil dynamics and C emissions, it will be necessary to incorporate climate data as well as R_S and microbial parameters into microbial decomposition models, which will be important for soil conservation and reducing soil C loss in the Loess Plateau.

Oxygen (O₂) limitation contributes to persistence of large carbon (C) stocks in saturated soils. However, many soils experience spatiotemporal O₂ fluctuations impacted by climate and land‐use change, and O₂‐mediated climate feedbacks from soil greenhouse gas emissions remain poorly constrained. Current theory and models posit that anoxia uniformly suppresses carbon (C) decomposition. Here we show that periodic anoxia may sustain or even stimulate decomposition over weeks to months in two disparate soils by increasing turnover and/or size of fast‐cycling C pools relative to static oxic conditions, and by sustaining decomposition of reduced organic molecules. Cumulative C losses did not decrease consistently as cumulative O₂ exposure decreased. After >1 year, soils anoxic for 75% of the time had similar C losses as the oxic control but nearly threefold greater climate impact on a CO₂‐equivalent basis (20‐year timescale) due to high methane (CH₄) emission. A mechanistic model incorporating current theory closely reproduced oxic control results but systematically underestimated C losses under O₂ fluctuations. Using a model‐experiment integration (ModEx) approach, we found that models were improved by varying microbial maintenance respiration and the fraction of CH₄ production in total C mineralization as a function of O₂ availability. Consistent with thermodynamic expectations, the calibrated models predicted lower microbial C‐use efficiency with increasing anoxic duration in one soil; in the other soil, dynamic organo‐mineral interactions implied by our empirical data but not represented in the model may have obscured this relationship. In both soils, the updated model was better able to capture transient spikes in C mineralization that occurred following anoxic–oxic transitions, where decomposition from the fluctuating‐O₂ treatments greatly exceeded the control. Overall, our data‐model comparison indicates that incorporating emergent biogeochemical properties of soil O₂ variability will be critical for effectively modeling C‐climate feedbacks in humid ecosystems.

The development of the new process‐based TRIPLEX‐GHG model derives from the Integrated Biosphere Simulator (IBIS), which couples nitrification and denitrification processes to quantify nitrous oxide (N₂O) emissions from natural forests and grasslands. Sensitivity analysis indicates that the nitrification rate coefficient (COE_NR) is the most sensitive parameter to simulate N₂O emissions. Accordingly, we calibrated this parameter using data from 29 global forest sites (across different latitudes) and grassland sites. The average nitrification rate coefficient gradually increases in the order of tropical forest to grassland to temperate forest to boreal forest, and giving means of 0.009, 0.03, 0.04, and 0.09, respectively. This study validated the mean value for each ecosystem at 52 sites globally. Calibration results both indicate the good performance of the model and its suitability in capturing seasonal variation and magnitude of N₂O flux; however, it is limited in modeling N₂O uptake and increments during periods of snowmelt. Additionally, validation results indicate that simulated and observed annual or seasonal N₂O fluxes are highly correlated (R² = 0.75; P < 0.01). Consequently, our results suggest that the model is suitable in simulating N₂O emissions from different forest and grassland land types under varying environmental conditions on a global scale.

Accurately predicting carbon‐climate feedbacks relies on understanding the environmental factors regulating soil organic carbon (SOC) storage and dynamics. Here, we employed a microbial ecological model (MEND), driven by downscaled output data from six Earth system models under two Shared Socio‐economic Pathways (SSP1‐2.6 and SSP5‐8.5) scenarios, to simulate long‐term soil biogeochemical processes. We aim to analyze the responses of soil microbial and carbon‐nitrogen (C‐N) processes to changes in environmental factors, including litter input (L), soil moisture (W) and temperature (T), and soil pH, in a broadleaf forest (BF) and a pine forest (PF). For the entire soil layer in both forests, we found that, compared to the baseline period of 2009–2020, the mean SOC during 2081–2100 increased by 40.9%–90.6% under the L or T change scenarios, versus 5.2%–31.0% under the W change scenario. However, soil moisture emerged as a key regulator of SOC, MBC and inorganic N dynamics in the topsoil of BF and PF. For example, W change led to SOC gain of 5.5%–37.2%, compared to the SOC loss of 15.5%–18.0% under L or T scenario. Additionally, a further reduction in soil pH by 0.2 units in the BF, representing the acid rain effect, significantly resulted in an additional SOC gain by 14.2%–21.3%, compared to the LTW (simultaneous changes in the three factors) scenario. These results indicate that the results derived solely from topsoil may not be extrapolated to the entire soil profile. Overall, this study significantly advances our comprehension of how different environmental factors impact the dynamics of SOC and the implications they have for climate change.