Impact of modified parameterizations in CLM5.0 on soil hydrothermal dynamics in permafrost regions of the Qinghai–Tibet Plateau

Abstract

Accurate understanding and modeling of soil hydrothermal dynamics in permafrost regions is essential for reliably assessing future permafrost changes and their impacts. However, the inadequate representation of soil water‒heat transport processes in current land surface models (LSMs) introduces large uncertainty in simulating permafrost dynamics, particularly on the Qinghai–Tibet Plateau (QTP). In this study, we modified the parameterizations of soil thermal conductivity, unfrozen water and soil evaporation resistance in version 5.0 of the Community Land Model (CLM5.0) and assessed their effects on soil hydrothermal dynamics in permafrost regions on the QTP using in-situ measurements at the depths of 10–40 cm. The results showed that soil temperature was more sensitive to the modified soil thermal conductivity and unfrozen water schemes, with average RMSE reduced by approximately 0.60 °C compared to the default CLM5.0. Soil moisture was mainly affected by the unfrozen water scheme during freezing and by the optimized soil evaporation resistance scheme during thawing, with maximum accuracy improvements of 8% and 25%, respectively. All three schemes significantly improved soil thermal conductivity simulations, reducing RMSE by over 80%. Overall, our modifications remarkably reduced simulation errors compared to the default schemes, improving the average accuracy of soil temperature, soil moisture and soil thermal conductivity by approximately 16%, 21% and 81% respectively. Additionally, this study emphasized the importance of accurately representing permafrost-related processes in LSMs, as they significantly affected simulation results. Specifically, soil thermodynamics is strongly sensitive to subtle changes in soil moisture transport processes, such as the hysteresis effect of unfrozen water content, and parameterizations of snowpack and vegetation. Therefore, future work should focus on enhancing the accurate representations of these processes and optimized parameters in LSMs to improve the simulation accuracy in permafrost regions on the QTP. This study enhanced the understanding of soil hydrothermal processes in LSMs and provided valuable insights for the future model development for permafrost regions under the context of climate change.