Méditerranée NM-1988 43

Frontal depressions over the Mediterranean sea and Central Southern Europe

A.A. FLOCAS*

Introduction

In two previous papers (FLOCAS, 1984 a, b) the distribution of fronts over central and southern parts of Europe and the Mediterranean Sea were discussed using daily data for the nine-year period 197 1 to 1 979. This data base of over 1 1 ,000 occurrences offrants included a considerable number (72%) of frontal depressions. This paper examines these cases in order to determine the temporal frequency of frontal depression centres and their mean tracks.

The earliest work of a similar nature was published by VAN BEBBER (1882) who recognized five depression tracks across central and northern Europe. He used data for the years 1876 to 1880 to produce a map which is still used in modern textbooks (e.g. BARRY and PERRY, 1973). He updated his work (1891) using data for the period 1875-1890 to produce monthly charts of mean depression tracks. Unfortunately his area of study was north of 40°N and so generally excluded the Mediterranean apart from the Ligurian sea and the north Adriatic. The coverage was extended into the eastern Mediterranean and the Balkans by WEICKMANN (1922). The Meteorological Office (1 962) analysed depression tracks based on United States sources for the periods 1926-39 and 1945-52. They devised a special classification for the Mediterranean based on five source regions from which fifteen tracks were derived. These source regions are much larger than the 5° latitude/longitude squares used here (see FLOCAS, 1984a) and their method of obtaining frequencies is not explained. It is assumed from the text that depressions were allocated to the main source regions but how they were subsequently tracked is unclear. Four of the Meteorological Office tracks coincide with VAN BEBBER's (his Va, Vb, Vc and Vd) and their area extends further east in the Mediterranean compared with the present study area. Since their analysis is based on all depressions, not only frontal depressions and since their source regions are rather broad and ill-defined it is to be expected that there will be differences between their data and those used in this paper.