Atmospheric circulation evolution related to desert-dust episodes over the Mediterranean

Cases of atmospheric circulation evolution favouring the occurrence of desert aerosol episodes (DAEs) over the broader Mediterranean region were investigated using an objective and dynamic algorithm, with daily satellite data for the period 2000-2013. After identifying strong and extreme DAEs, at a 1 degrees x1 degrees geographical-cell level, 255 dust aerosol episode days (DAEDs) and 148 cases of consecutive DAEDs, namely desert aerosol episode cases (DAECs), are defined. For each DAEC, the evolution of the lower tropospheric circulation 1 and 2 days before, during the initiation and after the cessation of the DAEC, is considered. S-mode factor analysis and k-means cluster analysis are applied on mean sea-level pressure and 700 hPa geopotential height fields obtained from the National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) Reanalysis Project, classifying the 148 cases of atmospheric-circulation evolution into six homogeneous and discrete clusters. The mean intra-annual variation of the DAECs reveals a primary maximum in May (18.9%), and their mean annual number is equal to 11.4 DAECs. On a seasonal basis, the highest percentage of the DAECs is found in spring (51.4%). Maximum duration of the DAECs is 7 days, with 58.8% lasting 1 day. Annually, the mean monthly number of the DAEs varies from 35.8 (September) to 58.0 (April). The western parts of the Mediterranean are affected by the DAEs when cyclonic conditions prevail in the western Mediterranean and northwestern Africa. In contrast, the central and eastern parts of the study region are affected by dust storms when a low-pressure system in the central Mediterranean or central Europe and an anticyclone in the eastern Mediterranean prevail. As to the mean regional intensity (aerosol optical depth at 550 nm) the strong DAEs vary from 0.67 to 0.77, while the extremes vary from 1.14 to 2.06. Generally, strong DAEs are more frequent than extremes (in five out of six clusters).