UCLA

Abstract. Frozen soil processes are of great importance in controlling surface water and energy balances during the cold season and in cold regions. Over recent decades, considerable frozen soil degradation and surface soil warming have been reported over the Tibetan Plateau and North China, but most land surface models have difficulty in capturing the freeze-thaw cycle and few validations focus on the effects of frozen soil processes on soil thermal characteristics in these regions. This paper addresses these issues by introducing a physically more realistic and computationally more stable and efficient frozen soil module (FSM) into a land surface model—the third-generation Simplified Simple Biosphere model (SSiB3-FSM). To overcome the difficulties in achieving stable numerical solutions for frozen soil, a new semi-implicit scheme and a physics-based freezing-thawing scheme were applied to solve the governing equations. The performance of this model, as well as the effects of frozen soil process on the soil temperature profile and soil thermal characteristics, were investigated over the Tibetan Plateau and North China using observation and models. Results show that the SSiB3 model with the FSM produces more realistic soil temperature profile and its seasonal variation than that without FSM during the freezing and thawing periods. The freezing process in soil delays the winter cooling, while the thawing process delays the summer warming. The time lag and amplitude damping of temperature become more pronounced with increasing depth. These processes are well simulated in SSiB3-FSM. The freeze-thaw processes could increase the simulated phase lag days and land memory at different soil depths, as well as the soil memory change with the soil thickness. Furthermore, compared with observations, SSiB3-FSM produces a realistic change of maximum frozen soil depth at decadal scales. This study shows the soil thermal characteristics at seasonal to decadal scales over frozen ground can be greatly improved in SSiB3-FSM and SSiB3-FSM can be used as an effective model for TP and NC simulation during cold reasons.