Abstract
The prediction of fog remains a challenge due to its time and space variability and to the complex interaction between the numerous physical processes influencing its life cycle. During the first stage of this thesis, the local meteorological and synoptic characteristics of fog occurrence over the Grand Casablanca region (Morocco) are investigated. To achieve this, hourly surface meteorological observations, at two synoptic stations of this coastal region, are used. An objective fog-type classification has been developed in this work and used to discriminate the fog events into the well known types. This fog climatology points out that the fog is often localized and that it is predominantly of advection-radiation type, followed by fog resulting from cloud base lowering and radiation fog. Besides, two different fog types can occur when fog simultaneously concerns the two stations. The advective processes associated with sea breeze circulation during the daylight, followed by the radiative processes often leads to fog formation over this coastal region. Numerical simulations are performed later using the meso-scale non-hydrostatic model Meso-NH. These simulations confirm that the physical processes, governing the life cycle of fog, can be different according to the physiographic features of the area. Moreover, the numerical prediction of coastal fog over heterogeneous area is very sensitive to sea surface temperature, land local topography and land cover. It also depends on the model’s ability to reproduce the sea breeze circulation during the daylight followed by the radiative processes early in the night. The systematic numerical simulations of the fog events that occurred during the winter 2013-2014 indicate the Meso-NH’s ability to well capture the fog occurrence with a relatively high false alarm rate, particularly over the coastal station.