The ESCOMPTE campaign was an intensive period (June and July 2001) of measurements in the Region of Marseille city, with the main objective to observe the conditions leading to photochemical pollution episodes. In the associated UBL/CLU-Escompte experiment, an effort was made on the Marseille urban zone, aimed at documenting the four-dimensional structure of the urban boundary layer in connection with the urban canopy thermodynamics during a summer period of low wind and breeze conditions, from June 5 to July 15, 2001*. Three meteorological stations were installed in order to characterise the urban boundary layer close to the ground. In one of them, the GLM station at the south, a mini-sodar was put on the rooftop of a building (15 meter height). At the same time, a telescopic mast (35 meter height) was installed on the ground close to the mini-sodar building with a sonic anemometer at the top. With the mini-sodar, it was possible to measure up to 150 meters with a 5m vertical resolution the three componants of the wind speed, averaged over 15 mn, and also the root mean square value of the vertical wind speed fluctuations (rmsW). Two main difficulties are encountered in using sodar system. First, the noise emission is a trouble for people living in the vicinity, mainly during night. During this period, we reduced the emission power down to 30% of the maximum, which was sufficient for the sodar noise to be accepted. The second point is the quality of the echo signal. We have not found any problems, this is due to the emission frequency (4500 Hz) which is above the urban background noise frequency range (from 10 to 1500 Hz).

Comparisons between the first level of detection of the mini-sodar (20 m above roof level) and the sonic anemometer show good agreement for the mean wind speed U and the mean direction DU, and also for the rms W values. With the sonic anemometer value of u*, we computed the U(z)/u* vertical profile. It will be shown that this one is not always described by the well known logarithmic function, mainly when the ground influence is low (corresponding to low u* values). In afternoon, the logarithmic profile is more often observed. However, in some cases, we have found two zones with different roughness parameters, one extanding up to about 50 meter height which corresponds to the roughness layer (Rotach, 1999), and the second one above which is the remaining of the surface layer. Values of the displacement height d and the roughness length zo deduced from observed profiles are in agreement with results of other works (Grimmond, 1999). The rms W values increase with height up to about 50 meters when the ground influence is strong, corresponding to high values of u* and to well defined logarithmic profiles. In the other cases, we observed a large scatter in the rms W data and no special trend with height.

* abstract proposed to the AMS Fourth Symposium on the Urban Environment, May 2002, Virgina: 'The UBL/CLU-Escompte experiment : description and first results, P. Mestayer and P. Durand