Land use and cover change (LUCC) is a major driver of global carbon cycles, accounting for a substantial portion of historical atmospheric CO₂ increases. The impact of LUCC on soil organic carbon (SOC) is particularly significant, given SOC's vital role in soil health, food production, and climate change mitigation. While existing research has explored the effects of LUCC on carbon dynamics, uncertainties remain, particularly regarding the specific impact on regional carbon budgets and the interplay between direct and indirect climate influences. This study focuses on China, a country that has undergone substantial LUCC in recent decades due to population growth and agricultural intensification, coupled with large-scale reforestation initiatives. The research aims to provide a comprehensive assessment of LUCC's impact on China's land carbon cycle by integrating meta-analysis of existing studies with a process-based ecosystem model driven by high-quality land use data. This integrated approach seeks to refine existing estimates of LUCC's effect on China's carbon balance and to elucidate the key drivers of SOC dynamics under changing land use.

The literature extensively documents the impact of LUCC on global carbon cycles, estimating LUCC-induced carbon emissions to be a significant fraction (25%) of the historical atmospheric CO₂ increase. However, methodologies for assessing these impacts vary, ranging from simplified Tier 1 IPCC methods to more complex modeling approaches using remote sensing and process-based ecosystem models like Dynamic Global Vegetation Models (DGVMs). These models, while valuable, often suffer from uncertainties stemming from input parameters, model errors, and differences in forecasting. Meta-analyses have emerged as a useful tool for synthesizing findings from multiple studies, statistically assessing the impact of LUCC on SOC across various regions and land-use types. However, there's a lack of comprehensive, long-term LUCC databases and oversimplified models for forest carbon dynamics, especially in regions like China, which have seen dramatic land-use transformations. This study addresses these limitations by combining a rigorous meta-analysis with advanced modeling, incorporating comprehensive datasets like the Land-Use Harmonization (LUH2) dataset to drive a process-based land ecosystem model.

This research employed a multi-faceted approach integrating meta-analysis and process-based modeling to estimate the impact of LUCC on China's land carbon dynamics from 1979 to 2014. The meta-analysis involved a three-level hierarchical model, synthesizing data from 132 publications (1248 observations) related to LUCC's effects on SOC in China. This systematic review followed PRISMA guidelines, meticulously extracting effect sizes, accounting for data quality, and addressing potential publication bias using techniques such as funnel plots and statistical tests. The meta-analysis incorporated various environmental variables (Mean Annual Temperature, Mean Annual Precipitation, elevation, bulk density, soil pH, and duration of land-use change) to investigate their influence on SOC dynamics under different land-use conversions (e.g., forest to cropland, grassland to cropland, cropland to forest). Statistical analyses included random forest modeling, simple linear regression, correlation analysis, and structural equation modeling (SEM) to assess the relative importance of different factors and their interactions. For the modeling component, the Community Earth System Model (CESM) version 2, incorporating the Community Land Model (CLM5), was used to simulate carbon cycle dynamics under scenarios with and without LUCC. The model was driven by the LUH2 dataset for land use information and the China Meteorological Forcing Dataset (CMFD) for climate data. Model validation was performed against flux tower data from nine ChinaFlux sites, comparing simulated Gross Primary Productivity (GPP) with observed values. Finally, the Mann-Kendall method was used to analyze temporal trends in carbon cycle variables.

The meta-analysis revealed a substantial 39.2% loss in SOC change in China due to LUCC, with grassland conversion to cropland showing significantly higher losses (56%) than forest-to-cropland conversion (37%). However, the conversion of cropland to forest, grassland, and shrubland resulted in SOC increases of 11%, 13%, and 22%, respectively, highlighting the positive impact of afforestation. Analysis of environmental factors indicated that soil bulk density was the most crucial driver of SOC changes following LUCC, followed by the duration of land-use change, with mean annual precipitation (MAP) playing a lesser role. Indirect climate effects on SOC were more significant than direct effects. The CESM model simulations showed that LUCC significantly increased China's terrestrial carbon sink, nearly doubling the GPP trend compared to a scenario without LUCC. Net ecosystem productivity (NEP) exhibited a positive trend, reaching 0.02 ± 0.12 Pg C yr⁻¹ due to LUCC, significantly higher than the negative NEP (-0.06 ± 0.16 Pg C yr⁻¹) under a LUCC-free scenario. Spatially, the largest carbon sinks were concentrated in southwestern China, while some regions, including central China and the Yangtze River Delta, experienced GPP decline. LUCC led to a 1.6 Pg C loss in topsoil (0-20 cm) soil organic carbon, with variations across different regions. Analysis of carbon fluxes related to litter and coarse woody debris (CWD) indicated higher fluxes in southeastern China and a significant increase in wood product pools due to land cover change.

This study's findings highlight the crucial role of LUCC in shaping China's land carbon cycle. The observed 39.2% loss in SOC due to LUCC is substantial, emphasizing the need for sustainable land management practices. The significant positive impact of afforestation on SOC underscores the effectiveness of reforestation and afforestation programs in mitigating carbon loss. The model simulations, validated by empirical data from the meta-analysis, robustly demonstrate that LUCC has acted as a major driver of increased carbon sequestration in China. Regional variations in carbon dynamics emphasize the spatial heterogeneity of LUCC impacts, highlighting the need for regionally tailored conservation and management strategies. The findings are consistent with previous studies documenting the impact of LUCC on SOC globally but offer a refined estimate specifically for China. Differences in methodologies and study focus likely explain discrepancies between this study's findings and global averages. The significant influence of soil bulk density on SOC dynamics further emphasizes the importance of considering soil properties in land management strategies.

This study demonstrates the profound influence of LUCC on China's land carbon cycle, showing that while LUCC has caused significant SOC loss, it has also accelerated the country's terrestrial carbon sink. Afforestation efforts have played a crucial role in mitigating carbon loss. Future research should focus on improving model accuracy by refining input parameters, enhancing the representation of physical processes, and improving the understanding of above- and below-ground carbon partitioning. The integrated approach of this study—combining meta-analysis with advanced ecosystem modeling—provides a valuable framework for future research on land carbon dynamics in other regions.

Several limitations could affect the interpretation of the results. Uncertainties exist in historical land-use data within the CESM and LUH2 datasets, potentially impacting model accuracy. The model's representation of internal carbon cycling processes might be oversimplified, leading to some uncertainty in the above/below-ground partitioning of the carbon sink. The reliance on published meta-analyses introduces potential biases inherent in the primary studies included. While efforts were made to address these limitations, future studies should aim to refine model representations of physical processes and improve the accuracy of input data.