Impact of Horizontal Grid Resolution from Ten-Kilometric to Hectometric Scales on Radiation Fog Forecasting over North China Plain

Abstract The impact of varying horizontal grid resolutions (HGRs), ranging from ten-kilometric to hectometric scales (15 km–500 m), is evaluated statistically for radiation fog numerical forecasting spanning the years 2016–19 and studied in detail for a specific case over the North China Plain (NCP). Results indicate that, although simulations using finer HGR could better represent the spatial distribution of meteorological elements and the influence of small-scale underlying surfaces on local meteorological fields, the fog forecasting skills do not consistently improve. Among the four HGRs examined with our model setup, the optimal HGR for fog forecasting is found to be 5 km, followed by 2.5 km. The impact of model HGR on fog simulation mainly originates from the differences in the response speed of wind advections with different grid spacings. The model generally shows a systematic overestimation of near-surface winds over flat terrain. When the overall wind speed within the study area is low (i.e., weak synoptic forcing), the discrepancies in the simulated wind speed among different HGRs are small. Nonetheless, when the large-scale synoptic wind enters the study area, obvious discrepancies could be seen in simulated wind speed among different HGRs, due to the different grid scales’ response speed to wind advections. With a larger grid scale, the influence of synoptic wind enters the study area more quickly and earlier, resulting in a more pronounced overestimation in wind speed. Since the presence of fog is highly sensitive to dynamic conditions (i.e., wind), this would finally lead to the differences in fog simulation under different HGRs. Significance Statement Numerical forecasting of fog remains a challenge, and how would the model horizontal grid resolution (HGR) influence the precision of fog forecasts remains a critical question and lacks definitive conclusions. This study investigates the impact of different HGRs on the forecasting of fog. Results reveal that enhanced HGR generally improves the model’s ability to capture the spatial distribution pattern of meteorological elements, whereas the skill scores of fog forecasting do not consistently improve. The influence of model HGR on simulating the spatial distribution of fog is more pronounced during the formation and dissipation stage, with simulations at finer HGR could better capture the spatial distribution of meteorological elements. The impact of different HGRs on fog simulation mainly arises from discrepancies in the response speed of wind advection.