Improving soil organic carbon parameterization of land surface model for cold regions in the Northeastern Tibetan Plateau, China

• The soil organic parameterizations of the CLM4.0 and DLM1.0 are insufficient for modeling soil moisture in cold regions with high soil organic carbon.
• A new soil organic carbon parameterization in land surface models is developed.
• The impacts of SOC on soil moisture vary significantly with seasons, which is largest for JJA followed by SON, MAM and DJF.

Most current land surface models (LSMs) show poor performance in soil water and temperature simulations for cold regions because of the inappropriate model structures and ill-parameterizations. This study aims to explore how to parameterize soil organic carbon (SOC) in term of soil moisture simulations using the Community Land Model version 4.0 (CLM4.0) model and the Dynamic Land Model version 1.0 (DLM1.0) model. Firstly, we discuss the sensitivities of modeled soil moisture to SOC based on observations. The SOC parameterizations in both CLM4.0 and DLM1.0 were proved poor against the measured SOC through sensitivity analysis for the Heihe River watershed area (cold but with high SOC). A SOC parameterization was developed and tested based on multiple year observations. Our findings are twofold: (i) the soil organic parameterizations in CLM4.0 and DLM1.0 are inappropriate for simulating soil moisture in cold regions with high SOC, and the modified SOC parameterization significantly improves the soil moisture simulations; (ii) the impacts of SOC on soil moisture vary significantly with seasons, which is largest for JJA (June, July and August) followed by SON (September, October and November), MAM (March, April and May) and DJF (December, January and February). This study highlights that the impacts of SOC should be fully considered for LSMs soil moisture simulations, especially for cold areas with high SOC.