Sensitivity of Radiation Fog to Minimum Eddy Diffusivity Threshold over North China: An Observational and Modeling study

Radiation fog seriously threatens traffic safety, human health, and the development of low-altitude economy in North China. However, there are still great uncertainties in the numerical forecasting of radiation fog, and turbulence parameterization is one of the key uncertainty sources. In this study, the spatiotemporal variations of turbulent diffusion were analyzed using the 5-level tower turbulence observations in Tianjin during four radiation fog episodes from 2016 to 2019. Based on the observation analysis, an improvement to the minimum eddy diffusivity of heat (K_hmin) in the Yonsei University (YSU) scheme was implemented in the Weather Research and Forecasting (WRF) model. K_hmin is the minimum threshold of the heat eddy diffusion coefficient (K_h) defined in the model (0.01 m² s^-1 in YSU scheme) to avoid the physically unrealistic situations where the model calculates zero turbulent atmosphere under strong stability conditions. However, observations in this study revealed that the 10th percentile of K_h at night was mostly lower than 0.01 m² s^-1, indicating that the model may overestimate the nighttime turbulent diffusion during the fog episodes. Sensitivity experiments demonstrated that changes in K_hmin significantly altered the simulation of nocturnal boundary layer (NBL) structure, while exerting a rather negligible impact during the daytime. Reducing K_hmin resulted in a lower surface temperature and a stronger inversion, thereby facilitating the formation of radiation fog. When K_hmin was reduced from 0.01 to 0.0001 m² s^-1, the nighttime Threat Score (TS) and Probability of Detection (POD) for fog forecasting increased by 0.053 and 0.029, respectively. Conversely, increasing K_hmin led to a weaker inversion and the dissipation of fog. This study highlights the importance of K_hmin in the planetary boundary layer (PBL) scheme for simulating fog, and also provides a novel perspective for improving fog forecasting. Specifically, a smaller K_hmin value may be more appropriate for simulating radiation fog.