Comparisons of aerosol optical depth provided by seviri satellite observations and CAMx air quality modelling
Keywords: Aerosol Optical Depth (AOD), Satellite data, Chemical Transport Model, dust outbreak.
Abstract. Satellite data provide high spatial coverage and characterization of atmospheric components for vertical column. Additionally, the use of air pollution modelling in combination with satellite data opens the challenging perspective to analyse the contribution of different pollution sources and transport processes. The main objective of this work is to study the AOD over Portugal using satellite observations in combination with air pollution modelling. For this purpose, satellite data provided by Spinning Enhanced Visible and Infra-Red Imager (SEVIRI) on-board the geostationary Meteosat-9 satellite on AOD at 550 nm and modelling results from the Chemical Transport Model (CAMx - Comprehensive Air quality Model) were analysed. The study period was May 2011 and the aim was to analyse the spatial variations of AOD over Portugal. In this study, a multi-temporal technique to retrieve AOD over land from SEVIRI was used. The proposed method takes advantage of SEVIRI's high temporal resolution of 15 minutes and high spatial resolution.
CAMx provides the size distribution of each aerosol constituent among a number of fixed size sections. For post processing, CAMx output species per size bin have been grouped into total particulate sulphate (PSO4), total primary and secondary organic aerosols (POA + SOA), total primary elemental carbon (PEC) and primary inert material per size bin (CRST_1 to CRST_4) to be used in AOD quantification. The AOD was calculated by integration of aerosol extinction coefficient (Qext) on the vertical column. The results were analysed in terms of temporal and spatial variations. The analysis points out that the implemented methodology provides a good spatial agreement between modelling results and satellite observation for dust outbreak studied (10th -17th of May 2011). A correlation coefficient of r=0.79 was found between the two datasets. This work provides relevant background to start the integration of these two different types of the data in order to improve air pollution assessment.