Simulation of fog with the ECMWF prognostic cloud scheme

Abstract

AbstractA prognostic cloud scheme has been operational at the European Centre for Medium‐Range Weather Forecasts (ECMWF) since 1995. In this paper the performance of the cloud scheme in simulating fog is assessed. A case‐study is performed with the one‐column version of the ECMWF model in order to analyse how the model reproduces the main mechanisms of fog generation and dissipation. A climatology of the model's fog is produced using data from the ECMWF Re‐Analysis and compared with climatological data. High resolution (T213L31) operational forecasts with the prognostic cloud scheme are compared with synoptic reports of visibility for Europe. In this context the relation between fog and temperature, on the one hand, and fog and wind speed, on the other, is explored in some detail.The comparison between the simulated and the observed visibility, in the one‐column model case‐study, shows that the evolution of fog is properly simulated. the analysis of the behaviour of the different parametrized physical processes suggests that the subtle balance between the various processes, fundamental for a realistic fog simulation, is achieved. This study also shows that in radiation fog a cooling process like radiation, that only changes the temperature without changing the moisture, is rather essential in order to reach saturation. the comparison between the model's fog climatology and the climatological data shows that the model is able to reproduce most of the major fog areas, particularly over the ocean.The analysis of the results from high‐resolution fog forecasts for Europe and the comparison with synoptic observations of visibility shows that the model simulates realistically some of the fog over Europe, particularly at 00 and 06 UTC. At 12 UTC the fog in the model is usually underestimated. In general, the number of situations where the visibility is between 2 km and 11 km is underestimated by the model. It is shown that the model is always slightly closer to saturation than the observations and this might be helping the model in producing fog. the model reproduces the observed behaviour of fog as a function of temperature and wind speed rather realistically.