Numerical simulation of the 7 to 9 September 2006 AMMA mesoscale convective system: Evaluation of the dynamics and cloud microphysics using synthetic observations
Résumé
This paper presents a numerical simulation of a Mesoscale Convective System (MCS) observed during the AMMA (African Monsoon Multidisciplinary Analysis) experiment with the BRAMS model (Brazilian Regional Atmospheric Modelling System). The aim is to document the life cycle of the MCS and to identify key cloud microphysical processes and their signatures by making use of synthetic observations calculated from the simulated fields. These observations: ARM (Atmospheric Radiation Measurement) 95 GHz equivalent radar reflectivity factor and Doppler velocity and infrared brightness temperatures in three SEVIRI (Spinning Enhanced Visible and InfraRed Imager) channels centred at 8.7, 10.6 and 12 µm are simulated using respectively Mie scattering theory and FASDOM (Fast Discrete Ordinate Method), a fast radiative transfer code. Synthetic observations and model variables are compared to various measurements from several platforms (W-band and Massachusetts Institute of Technology (MIT) ground-based Doppler radars, soundings, aircraft measurements, and Meteosat Second Generation) to evaluate the model at different scales and to identify the signatures of microphysical properties with a focus on the anvil part of the MCS. A method using both the ARM and the MIT radar data is used to identify the different regimes within the MCS. A relatively good agreement with direct comparisons is found, as well as discrepancies in the microphysical scheme parametrization that clearly need improvements (using in situ measurements). Microphysical signatures are also studied using joint radar reflectivity/Doppler-height histograms. Their analysis shows that the model tends to overplay the role of the riming processes, even in the anvil part of the MCS. Comparisons of the Particle Size Distributions (simulated and measured in situ) show the model's ability to reproduce complex PSDs (e.g. a multimodal behaviour)