Multisensor analysis of convection in Mediterranean cyclones

.A recent climatological study (Trigo et al., 1999) shown that the cyclonic activity is a key feature of the meteorology of the Mediterranean basin during the whole year. In particular, during warm months, short-living cyclones (averaged lifetime about 30 hours) rise from well known birthplaces (Gulf of Genoa, Gibraltar area, Atlas Mountains are the most effective) and affect the weather in the Mediterranean area. Following a different approach Porcú et al. (1997) shown that very often such cyclones force convective initialisation and development. Moreover those episodes are responsible for the most severe rainfall events sometime related with flood/flash flood occurrence in coastal and continental areas. The combination of long- lasting, moderate precipitation with heavy showers is the more common mechanism originating such floods. The cloud systems related to the cyclonic structure often develop over the sea: the scarcity of conventional observation available in the Mediterranean basin makes satellites peculiar points of view for a detailed study of these events. A complete analysis will include the use of ECMWF fields to assess the larger scale setting: in particular we'll consider the presence and intensity of Potential Vorticity Anomalies (PVAs) related to the cyclonic depression. PVAs are suspected to be precursor of heavy rainfall episodes (Massacand et al., 1999) and are powerful tools in classify cyclonic systems (Georgev, 1999). A further approach will be developed will include combined SSM/I-Meteosat water vapour retrieval over cloud-free areas, in order to evaluate the potential of convective initiation. In this work we present preliminary results of a multi-sensor, multi-frequency analysis of convective patterns in cyclonic structure. We modified a cloud classification algorithm originally developed for visible-infrared (VIS-IR) data (Porcù and Levizzani, 1992) to include also microwave radiances, to increase the informative content of the classification. In particular the SSM/I brightness temperature at 85 GHz is used together with the equivalent black body temperature as measured by the Meteosat infrared channel. The results show potential in understanding of the convective patterns, especially if embedded in cyclonic cloud bands, as it is common at mid-latitude.