The London Model: forecasting fog at 333m resolution

Fog formation is particularly challenging for numerical weather prediction models to accurately capture, due to its dependence on unresolved processes, both in the atmosphere and land-surface. To assist with forecasting at key London airports, the Met Office has started routinely running a 333m horizontal resolution model covering the Greater London area. We demonstrate how better representation of surface characteristics, e.g. the orography and surface type, generally lead to improvements in fog forecasting, and how differences in cloud representation lead to improvements in near-surface temperature forecasting. We will also discuss how an 18 month field campaign conducted by the Met Office is leading to advances in the understanding and modelling of fog processes. Early results have shown how accurate representation of pre-fog surface cooling is crucially important for determining fog onset time, and how aerosol-fog interactions affect the development of fog throughout the night, which can be important in determining dissipation time. Finally we will show how some simple improvements to the aerosol-fog parametrization can improve model performance.