Articles | Volume 17, issue 1
Research article
03 Jan 2017
Research article |  | 03 Jan 2017

WRF-Chem model simulations of a dust outbreak over the central Mediterranean and comparison with multi-sensor desert dust observations

Umberto Rizza, Francesca Barnaba, Mario Marcello Miglietta, Cristina Mangia, Luca Di Liberto, Davide Dionisi, Francesca Costabile, Fabio Grasso, and Gian Paolo Gobbi

Abstract. In this study, the Weather Research and Forecasting model with online coupled chemistry (WRF-Chem) is applied to simulate an intense Saharan dust outbreak event that took place over the Mediterranean in May 2014. Comparison of a simulation using a physics-based desert dust emission scheme with a numerical experiment using a simplified (minimal) emission scheme is included to highlight the advantages of the former. The model was found to reproduce well the synoptic meteorological conditions driving the dust outbreak: an omega-like pressure configuration associated with a cyclogenesis in the Atlantic coasts of Spain. The model performances in reproducing the atmospheric desert dust load were evaluated using a multi-platform observational dataset of aerosol and desert dust properties, including optical properties from satellite and ground-based sun photometers and lidars, plus in situ particulate matter mass concentration (PM) data. This comparison allowed us to investigate the model ability in reproducing both the horizontal and the vertical displacement of the dust plume, as well as its evolution in time.

The comparison with satellite (MODIS-Terra) and sun photometers (AERONET) showed that the model is able to reproduce well the horizontal field of the aerosol optical depth (AOD) and its evolution in time (temporal correlation coefficient with AERONET of 0.85). On the vertical scale, the comparison with lidar data at a single site (Rome, Italy) confirms that the desert dust advection occurs in several, superimposed "pulses" as simulated by the model. Cross-analysis of the modeled AOD and desert dust emission fluxes further allowed for the source regions of the observed plumes to be inferred. The vertical displacement of the modeled dust plume was in rather good agreement with the lidar soundings, with correlation coefficients among aerosol extinction profiles up to 1 and mean discrepancy of about 50 %.

The model–measurement comparison for PM10 and PM2.5 showed a good temporal matching, although it revealed a marked overestimation of PM10 and PM2.5 (of the order of 70 % during the dust peak). For PM10, it was also possible to investigate the accordance between the model- and the measurement-based dust PM10, this confirming the model PM10 overestimation to be related to over-predicted dust mass up to a factor of 140 %. In all the model–measurement comparisons performed, the enhanced capabilities of the physics-based emission scheme with respect to its simplified, minimal version were evident and are documented.

Final-revised paper