Evolution of convective boundary layer in deep valley for air quality modeling

Alpine ecosystems are sensitive to air pollution due to emission sources and local meteorology. In this framework, the `POllution des Vallées Alpines' (POVA) program started in May 2000. The main objective of the POVA program is to characterize the pollution sources and study the relationship between atmosphere dynamics and pollution events. Field campaigns have particularly investigated mountain-valley wind systems in the Chamonix and the Maurienne valleys (France). Investigations of smaller scale flow fields were carried out, aiming at detailing valley wind systems.

Data were collected from 25 June through 10 July 2003. This study will focus on a clear day in both valleys.

Mixing depth structure and its evolution are diagnosed from radar profiler data for July 8, 2003 in the Chamonix valley and June 29, 2003 in the Maurienne valley, France. The behaviour of C_n ² peaks coupled with the vertical air velocity variance s_w² were used to locate the height of the mixed layer. Use of the turbulent kinetic energy dissipation rate e to provide mixing-height is discussed. Tethersonde vertical profiles were carried out to investigate the lower layers of the atmosphere in the range of approximatively 400-500 m a.g.l. The tethersonde-derived mixing heights are especially useful to study the reversal of the down-valley wind system during the morning transition (see Figure 1).

Figure 1: Pattern wind structure evolution (horizontal velocity component and direction) from 0625 GMT (U1) to 1107 GMT (D8). Tethered-balloon data, Maurienne valley, June 29, 2003

Analyses and numerical simulations were performed for these events using the Advanced Regional Prediction System (ARPS) developed at the University of Oklahoma with horizontal grid size from 300 to 1000 meters. Results from ARPS were input in a meso-scale chemical-transport model (TAPOM).

Local thermally driven wind circulation within both valleys - up-valley flow during day and down-valley flow during night - was well simulated by the model. Available measurements were used to evaluate the model results. Both ground surface and upper-levels calculated winds showed good agreement with the observations. Meteorological variables and ozone were compared with ground station data (see Figure 2).

Figure 2: Virtual potential temperature (K), Bois du Bouchet monitoring station, Chamonix valley, July 8, 2003