Using explainable AI to diagnose the representation of environmental drivers in process-based soil organic carbon models

Abstract

The accuracy of soil organic carbon (SOC) models and their ability to capture the relationship between SOC and environmental variables are critical for reducing uncertainties in future projection of soil carbon balance. In this study, we evaluate the performance of two state-of-the-art process-based SOC models, the vertically resolved MIcrobial-MIneral Carbon Stabilisation (MIMICS) and the Microbial Explicit Soil Carbon (MES-C) model, against a machine learning (ML) approach on predicting global SOC content. By applying multiple interpretable ML methods, we find that the poor performance of the two process-based models is associated both with the missing of key variables, and the underrepresentation of the role of existing variables such as net primary production (NPP). Soil cation exchange capacity (CEC) is identified as an important predictor missing from process-based models, and soil texture is given more importance in models than indicated by ML results. Although the overall relationships between SOC and individual predictors are reasonably captured, the varying sensitivity across entire predictor range is not replicated by process-based models, most notably for NPP. Observations exhibit a nonlinear relationship between NPP and SOC while models show a monotonic positive trend. Additionally, MES-C largely diminishes interacting effects of variable pairs, whereas MIMICS produces mismatches relating to the interactions between NPP and both soil temperature and moisture. Process-based models also fail to reproduce the interactions among soil moisture, soil texture, and soil pH, hindering our understanding on SOC stabilisation and destabilisation processes. Our study highlights the importance in improving the representation of environmental variables in process-based models to achieve a more accurate projection of SOC under future climate conditions.