Caracterização de profundidade óptica de aerossóis atmosféricos no nordeste brasileiro: um estudo através do sistema de modelagem EURAD-IM com assimilação de dados

Abstract

Since numerical weather prediction (NWP) studies began, many efforts have been made in order to obtain improved results with higher accuracy and quality. When compared to meteorological forecast, chemical weather predictions have the same importance as with each day the concern on how atmospheric pollutants on climate change grows. To obtain better prediction, different modeling techniques are adopted. One of them is data assimilation (DA), which matches observational and modeled data in order to obtain more accurate results. Usually this type of technique is use when there is lack of data, which makes traditional modeling less trustworthy. The main goal of the work is to characterize the AOD though a variational data assimilation method in three dimensions (3DVAR) of the EURAD-IM inverse model for the Northeast Brazil (NEB) and remote sensing data. EURAD (European Air Pollution Dispersion) and its inverse form EURAD-IM developed at the University of Cologne, in Germany, is a Eulerian model that operates from local to hemispheric scale and calculates transport, diffusion and chemical transformation of about 70 chemical species. In NEB the short amount and quality of emission sources makes modeling a challenge. For this reason, Aerosol Optical Depth (AOD) derived from the MODerate Resolution Imaging Spectroradiometer (MODIS) on board of AQUA and TERRA were considered. These data were used to investigate and characterize aerosols particles in the region for a ten-year (from 2005 to 2014) period and to validate the model results. In this it was observed that AOD in NEB has a feature of low particle load, with the medium value for the period of 0.2, except during two seasons, named as typical periods. Those periods occur from January to March and from August to October, in all the years. It was also observed that NEB is influenced by different types of particles, sea salt (the main type), biomass burning, urban pollution, dust and dust plus pollution. Also, observed that dust and biomass burning particles from the African continent are advected towards NEB during the typical periods. The simulations with EURAD-IM were performed in the year of 2014 and consider as boundary conditions, data from The Copernicus Atmosphere Monitoring Service (CAMS) and as meteorological driver, simulations from The Weather Research and Forecasting (WRF) model. It was observed that EURAD-IM and MODIS data correlates well, except during the months when AOD showed elevated values. On those EURAD overestimates AOD in NEB. Study case with consideration on DA, showed that 3D-VAR works well for the region. It also showed to be suitable for the NEB providing a good correlation with the observational data and better results when compared with pure model simulations.