A climatology of mesoscale convective systems over Europe using satellite infrared imagery. I: Methodology

Abstract An automated method aimed at producing a significant European mesoscale convective system (MCS) climatology is presented. It uses Meteosat infrared window channel images and is composed of two main tools: an automated cloud‐shield tracking method and a robust method of discriminating between convective and nonconvective cloud shields. The automated cloud‐tracking method defines cloud systems as connected sets of pixels, named ‘cells’, after temperature and area thresholding and it is based on the overlapping between cells in successive images. It handles splits and merges of cells and takes cell movement into account. It has three parameters: the temperature and area thresholds and a minimum overlapping threshold. It is concluded that it performs a correct tracking at any temperature threshold between −30 ° C and −55 ° C and for an area threshold greater than 1000 km 2 , so that it allows the tracking of MCSs during most of their life cycle. The automated discrimination between convective and non‐convective cloud shields uses a discrimination parameter based on brightness‐temperature gradients on the edges of cells, because strong values of this gradient are observed at the beginning of the life cycle of MCSs when cold anvils develop. A seasonal study, and the sensitivity of the method to the temperature threshold, are presented. The method shows significant quality during the entire warm season (from April to September): it correctly discriminates 80% of MCSs and more than 90% of the most electrically active ones, while showing a low false‐alarm rate around 8%; therefore the method seems to be useful for climatological purposes. Copyright © 2002 Royal Meteorological Society