High-Resolution Simulation of Dense Fog in Tianjin City: Effect of Horizontal Resolutions From Mesoscale to Large Eddy Scale

Abstract

Fog is one of the most severe weather phenomena affecting the safety of land, sea, and air transportation, with significant impacts on national economies. This study investigates a dense fog event with visibility less than 50 m that occurred in Tianjin city from December 19 to 20, 2016. Using the large-eddy simulation (LES) capability of the Weather Research and Forecasting (WRF) model, a quadruple one-way nesting approach was applied to downscale horizontal resolution from 5 and 1 km at the mesoscale to 200 and 40 m at the large-eddy scale, providing high-resolution simulations of this fog event. Combined with Himawari-8 satellite retrievals, surface meteorological data, 255 m meteorological tower data, and fog droplet spectral data, the study compares the simulation performance of different horizontal grid resolutions during this dense fog event and discusses the interaction between cloud microphysics parameterization and fine-scale turbulence simulations. The results reveal: With increasing resolution, the model becomes more sensitive to fluctuations of meteorological variables within the fog layer, resulting in a more pronounced diurnal variation and a tendency for the fog to dissipate more readily. At a horizontal resolution of 40 m, the WRF-LES simulation yields the smallest errors in boundary layer temperature and humidity profiles, owing to its ability to capture the fine-scale structure of buoyancy oscillations at the fog top. In addition, the onset and dissipation of fog simulation is closely linked to the coordination between fine-scale turbulence simulation and microphysical parameterizations, which are closely related to the autoconversion process. Among the tested schemes, the Purdue Lin scheme at 1 km resolution shows the best agreement with observed results. There are mutual interactions between cloud microphysics and turbulence parameterizations, which together determine the accuracy of fog simulations.