资源环境科技发展态势分析平台

The identification and characterization of High Atmospheric Turbidity (HAT) episodes is a key objective of global aerosol monitoring. This study presents a comparison of three different methodologies that were used to identify HAT episodes in the north-central Iberian Peninsula. The first methodology (named C&S inventory) is based on columnar aerosol optical depth (AOD from the Aerosol Robotic Network, AERONET) and surface particulate matter concentrations (PMx from the European Monitoring and Evaluation Programme, EMEP) as well as ancillary information. Another methodology (named SPR) is based on PM surface concentrations levels and ancillary information. Both methods are carefully reviewed by human observers. A third method, based only on fine and coarse mode values of AOD was also analysed. This method (the SDA or Spectral Deconvolution Algorithm) is found to be a good operational candidate for automating the identification of HAT episodes. The three methods allow for the identification of mineral desert dust (coarse type 'D') aerosols and aerosols of fine type, 'A' (i.e. biomass burning or polluted aerosols): their mixture, categorized as 'MD' and 'MA' classes (depending of the prevailing 'D' or 'A' type) is only identified in the C&S and SDA inventories. The three inventories show about 60% coincidence across a 2005-2014 reference period. When the C&S and SDA inventories are compared, the agreement is very high if columnar aerosol data is available: > 90% for desert aerosol type and > 70% for fine aerosol type. The comparative study of these three aerosol inventories was motivated by the need to automate existing methodologies.