Simulating lake ice phenology using a coupled atmosphere-lake model at Lake Nam Co, a typical deep alpine lake on the Tibetan Plateau

Abstract. Simulating the ice phenology of deep alpine lakes is important and challenging in coupled atmosphere-lake models. In this study, the Weather Research and Forecasting (WRF) model, coupled with two lake models, the fresh-water lake model (WRF-FLake) and the default lake model (WRF-Lake), was applied to lake Nam Co, a typical deep alpine lake located in the centre of the Tibetan Plateau, to simulate its lake ice phenology. Due to the large errors in simulating lake ice phenology, related key parameters and parameterizations were improved in the coupled model based on observations and physics-based schemes. By improving the momentum, hydraulic and thermal roughness length parameterizations, both the WRF-FLake and the WRF-Lake models reasonably simulated the lake freeze-up date. By improving the key parameters associated with shortwave radiation transfer process when lake ice exists, both models generally simulated the lake break-up date well. Compared with WRF-Lake without improvements, the coupled model with both revised lake models significantly improved the simulation of lake ice phenology. However, there were still considerable errors in simulating the spatial patterns of freeze-up and break-up dates, implying that significant challenges in simulating the lake ice phenology still exist in representing some important model physics, including lake physics such as grid-scale water circulation, and atmospheric processes such as snowfall and surface snow dynamics. Therefore, this work can provide valuable new implications for advancing lake ice phenology simulations in coupled models and the improved model also has practical application prospects in weather and climate forecasts.