The effects of land use change on soil organic carbon stock in China: A meta-analysis with the empirical modeling approach

Abstract

China has experienced a pronounced frequency of land use changes (LUC) over recent decades, resulting in notable alterations to soil organic carbon (SOC) stock. Given that the soil carbon (C) pool constitutes the largest C reservoir in the Earth's system, unraveling the spatiotemporal evolution mechanism of SOC stock influenced by LUC becomes imperative. This understanding holds significant importance for addressing global climate change and ensuring the sustainable management of soil resources. Despite hundreds of field studies and no less than a dozen literature reviews conducted, a substantial divergence persists regarding the direction and magnitude of changes in SOC stock associated with LUC. Drawing upon a synthesis of 194 publications encompassing 658 sites across China, this study endeavors to investigate the impact of LUC on the dynamics of SOC stock in topsoil (0–20 cm). The results of the meta-analysis revealed that, on average, LUC led to an increase of SOC stock by 3.50 t ha−1. Notably, significant increments in SOC stock were observed following conversions from cropland to forest (17.74 t ha−1), cropland to grassland (7.64 t ha−1), cropland to plantation (9.27 t ha−1), grassland to forest (10.24 t ha−1), and plantation to forest (13.30 t ha−1). Conversely, a substantial decline in SOC stock was noted after transitions from forest to cropland (−33.57 t ha−1), forest to grassland (−48.61 t ha−1), forest to plantation (−14.76 t ha−1), and grassland to cropland (−11.13 t ha−1). Furthermore, we investigated the temporal dynamics and revealed the principal driving factors influencing SOC stock changes under LUC in China. Sand content and initial SOC stock exhibited a negative correlation with SOC stock changes, while silt content, mean annual precipitation and mean annual temperature showed positive correlations. Additionally, the multiple linear regression models demonstrated robust predictive capabilities for SOC stock dynamics under LUC with the number of sites >30 (R2 of 0.31–0.82). The findings of this study offer valuable insights, serving as a pertinent reference for informed decision making in the enhancement of land use management to augment SOC stock in China.