Assessing and predicting soil carbon density in China using CMIP5 earth system models

Abstract

Soil carbon (SC) is a key component of the carbon cycle and plays an important role in climate change; however, quantitatively assessing SC dynamics at the regional scale remains challenging. Earth system model (ESM) that considers multiple environmental factors and spatial heterogeneity has become a powerful tool to explore carbon cycle-climate feedbacks, although the performance of the ESM is diverse and highly uncertain. Thus, identifying reliable ESMs is a prerequisite for better understanding the response of SC dynamics to human activity and climate change. The 16 ESMs that participated in the fifth phase of the Coupled Model Intercomparison Project (CMIP5) were employed to evaluate the skill performance of SC density simulation by comparison with reference data from the International Geosphere-Biosphere Programme Data and Information System (IGBP-DIS). Although ESMs generally reflect spatial patterns with lower SC in northwest China and higher SC in southeast China, 11 of 16 ESMs underestimated the SC in China, and 5 of 16 ESMs overestimated the SC density as most ESMs had large discrepancies in capturing the SC density in the northern high latitudes of China and the Qinghai-Tibet Plateau. According to a series of model performance statistics, SC simulated by Institute Pierre Simon Laplace (IPSL) Coupled Model had a close spatial pattern with IGBP-DIS and showed higher skills for SC predictions in China relative to other CMIP5 ESMs. The multimodel ensemble average obtained by IPSL family ESMs showed that SC density exhibited increasing trends under both the RCP4.5 scenario and RCP8.5 scenario. The SC density increased slowly under RCP8.5 compared with that under RCP4.5 and even displayed a decreasing trend in the late 21st century. The findings of this study can provide a reference for identifying the shortcomings of SC predictions in China and guide SC parameterization improvement in ESMs.

Keywords: Carbon cycle; Model uncertainty; Multimodel ensemble; Projection; Spatial distribution.