ArticlePDF Available

Abstract and Figures

In this study, the climatology of tornadoes, waterspouts and funnel clouds over Greece is presented for the period 1709–2012. The climatology consists of two datasets. An historical dataset (1709–1999) is based on newspaper archives, historical archives, published tornado literature, administrative records and reports of Hellenic National Meteorological Service (HNMS). A recent dataset (2000–2012) is based on newspaper articles, eyewitness reports to the media, HNMS's reports and an open-ended online tornado report database which has been developed and maintained by the Laboratory of Climatology & Atmospheric Environment of the University of Athens. Altogether, 612 Greek tornadic events compose the climatology: 171 tornadoes, 374 waterspouts and 67 funnel clouds. Tornadic events during the past 13 years (2000–2012) have occurred all over the Greek territory and there is frequent tornadic occurrence over north Crete and Corfu Island. Tornadoes are more frequent to occur over NW Peloponnesus followed by south parts of Corfu Island. However, waterspouts are more frequent over north Crete followed by Corfu Island. Tornadic monthly variability depicts a maximum during October, followed by September and November. October is the month with the highest tornado frequency, followed by November and July. The highest waterspout frequency month is September followed by October and December. Tornadoes most commonly develop during the warm time of the day, as more than 75% of all cases occur during 08:00–15:00 hours UTC with a maximum at 12:00 hours UTC. Waterspout frequency of occurrence has two maxima during the day, the first early in the morning (07:00–09:00 hours UTC) and the second after the noon time period (14:00–15:00 hours UTC). The dominant (27.7% of total cases) intensity of tornadoes in Greece is T4 based on the T-scale during the 300-year period (1709–2012); there have been at least 114 injured and 29 deaths.
Content may be subject to copyright.
INTERNATIONAL JOURNAL OF CLIMATOLOGY
Int. J. Climatol. (2013)
Published online in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/joc.3857
A climatology of tornadic activity over Greece based
on historical records
I. T. Matsangouras,aP. T. Nastos,a*H. B. Bluesteinband M. V. Sioutasc
aLaboratory of Climatology and Atmospheric Environment, Faculty of Geology and Geoenvironment, University of Athens, Greece
bSchool of Meteorology, University of Oklahoma, Norman, OK, USA
cELGA – Meteorological Applications Centre, Airport Macedonia, Thessaloniki, Greece
ABSTRACT: In this study, the climatology of tornadoes, waterspouts and funnel clouds over Greece is presented
for the period 17092012. The climatology consists of two datasets. An historical dataset (17091999) is based on
newspaper archives, historical archives, published tornado literature, administrative records and reports of Hellenic National
Meteorological Service (HNMS). A recent dataset (20002012) is based on newspaper articles, eyewitness reports to the
media, HNMS’s reports and an open-ended online tornado report database which has been developed and maintained by
the Laboratory of Climatology & Atmospheric Environment of the University of Athens. Altogether, 612 Greek tornadic
events compose the climatology: 171 tornadoes, 374 waterspouts and 67 funnel clouds. Tornadic events during the past
13 years (2000 2012) have occurred all over the Greek territory and there is frequent tornadic occurrence over north
Crete and Corfu Island. Tornadoes are more frequent to occur over NW Peloponnesus followed by south parts of Corfu
Island. However, waterspouts are more frequent over north Crete followed by Corfu Island. Tornadic monthly variability
depicts a maximum during October, followed by September and November. October is the month with the highest tornado
frequency, followed by November and July. The highest waterspout frequency month is September followed by October
and December. Tornadoes most commonly develop during the warm time of the day, as more than 75% of all cases
occur during 08:0015:00 hours UTC with a maximum at 12:00 hours UTC. Waterspout frequency of occurrence has two
maxima during the day, the first early in the morning (07:0009:00 hours UTC) and the second after the noon time period
(14:0015:00 hours UTC). The dominant (27.7% of total cases) intensity of tornadoes in Greece is T4 based on the T-scale
during the 300-year period (17092012); there have been at least 114 injured and 29 deaths.
KEY WORDS tornadoes; waterspouts; funnel clouds; NCEPNCAR reanalysis; Greece
Received 7 April 2013; Revised 27 September 2013; Accepted 1 October 2013
1. Introduction
Tornadoes (TR), waterspouts (WS) and funnel clouds
(FC) could be characterized as one of the most violent
of all small-scale natural phenomena. They are associ-
ated with extremely high winds, inside and around the
tornado’s funnel, causing extended damage and in many
cases loss of life.
Extreme weather events have always fascinated the
mankind, motivating studies and attempts to understand
the mechanisms that produced them. Tornadoes and
waterspouts were well known to the ancients; virtu-
ally all of the classical philosophers have provided
explanations of the phenomena. The Greek philosopher
Aristotle (384322 BC) in Meteorologica presented
perhaps the most renowned exposition of natural extreme
phenomena: ‘So the whirlwind originates in the failure
of an incipient cyclone to escape from its cloud. It is
* Correspondence to: Dr. P. T. Nastos, Laboratory of Climatol-
ogy and Atmospheric Environment, Department of Geography and
Climatology, Faculty of Geology and Geoenvironment, University
of Athens, Panepistimiopolis GR 157 84 Athens, Greece. E-mail:
nastos@geol.uoa.gr
due to the resistance the eddy generates and emerges
when the spiral descends to the earth dragging along the
cloud that cannot shake off. When blowing in a straight
line it carries along whatever comes by in a circular
motion and overturns and snatches up whatever it meets’
(Meteorologica, 371a915).
Later on, the Roman poet and philosopher Titus
Lucretius Carus (9050 BC) in De Rerum Natura (On
the Nature of the Universe) postulated two formation
mechanisms for whirlwinds and waterspouts. Gaius Plin-
ius Secundus, better known as Pliny the Elder (2379
AD), a Roman natural philosopher in his Natural History
(Book II, 131134) recalled more of the earlier ideas.
Tornadoes occur in many parts of the world (Fujita,
1973): Several publications during the last several
decades have presented historical records concerning tor-
nadic activity (e.g. Meaden, 1976; Peterson, 1982, 1992,
1995, 1998; Grazulis, 1993; Tarand, 1995; Tooming and
Peterson, 1995; Reynolds, 1999a, 1999b; Tyrrel, 2000;
Marcinoniene, 2003; Setv´
ak et al., 2003; Nastos and Mat-
sangouras, 2010; Gay`
aet al., 2011; Br´
azdil et al., 2012;
Rauhala et al., 2012). In addition to these historical
publications, several researchers have presented or
2013 Royal Meteorological Society
I.T. MATSANGOURAS et al.
analysed tornado occurrences and climatology for many
European countries (Dessens, 1984; Dessens and Snow,
1987; Paul, 1999; Dotzek, 2000, 2003; Gay`
aet al.,
2000; Holzer, 2000; Bertato et al., 2003; Sioutas, 2003,
2011; Gianfreda et al., 2005; Giaiotti et al., 2007;
Keul et al., 2009). Tornadoes have been recorded in all
continents with the exception of Antarctica (Goliger and
Milford, 1998; Kounkou et al., 2009). The frequent
occurrence of tornadoes associated with damage caused
in the United States is of great concern (e.g. Snow
and Wyatt, 1997, 1998; Bluestein, 1999; Forbes and
Bluestein, 2001; McCarthy and Schaefer, 2004a, 2004b;
Verbout et al., 2006; Clark, 2008; Brotzge and Erickson,
2010; Hall, 2011). Brooks and Doswell (2002) men-
tion almost 20 000 deaths associated with over 3600
tornadoes that have occurred in the United States
since 1680 AD. There is a vast literature on tornado
climatology in the United States (e.g. Kelly et al.,
1978; McCarthy and Schaefer, 2004a, 2004b; Walker,
2007; Kis and Straka, 2010; Moore and Dixon, 2011).
Besides, analysis of tornado frequencies in Canada
with large-scale atmospheric oscillations, such as El
Nino-Southern Oscillation (ENSO) events suggests that
the cooler La Nina events tend to suppress tornadic
activity, while El Nino events tend to enhance it (Etkin
et al., 2001). In addition, most of what is known in fluid
dynamics about tornadoes comes from laboratory and
numerical experiments, simulating vortices that share
similarities with what can be observed in a tornado in
nature (e.g. Rotunno, 1977, 1978, 2013, 1979, 1984,
1980; Fiedler and Rotunno, 1986; Wilson and Rotunno,
1986; Klemp, 1987; Howells et al., 1988; Fiedler, 2009).
The complex inland terrain of Greece along with the
Ionian Sea at the west and the Aegean Sea at the east,
appears to be a vulnerable area for violent phenomena
such as tornadoes, waterspouts and funnel clouds. How-
ever, tornado literature for Greece was very limited up
to 2000, after which a systematic recording and investi-
gation of tornado activity and the first national tornado
database was initiated by the Hellenic National Agri-
cultural Insurance Organization (ELGA) (Sioutas, 2002,
2006). Significant research has been carried out dur-
ing the last decade, including tornado and waterspout
overviews and the first climatology for Greece (Sioutas,
2003; Matsangouras et al., 2011b), waterspout occur-
rences and forecasting (Sioutas and Keul, 2007; Keul
et al., 2009) and a 20th century climatology (Nastos
and Matsangouras, 2010). In addition to these publica-
tions, reports and analyses of some important tornado and
waterspout case studies have been carried out (Sioutas,
2001, 2002; Sioutas et al., 2006; Matsangouras and Nas-
tos, 2010; Matsangouras et al., 2010, 2011a, 2011b, 2012;
Nastos and Matsangouras, 2012).
In this study, we present historical records of tor-
nadoes, waterspouts and funnel clouds over Greece
from the 18th century up to 2012, including their
spatial and temporal variability. The Laboratory of
Climatology and Atmospheric Environment (LACAE,
http://lacae.geol.uoa.gr) of the University of Athens has
undertaken a systematic effort in recording tornadoes,
waterspouts, and funnel clouds in Greece beginning in
2007. LACAE developed in 2009 an open-ended online
tornado report database web system (http://tornado.
geol.uoa.gr), contributing to the compilation of a clima-
tology of these extreme weather events. All database’s
records have been assembled as photographs, videos, eye-
witness reports, literature reports and synoptic reports by
the HNMS.
The goal of this article is to discuss the historical
records and to create an updated climatology for Greece.
Data and the methodology to collect reports are dis-
cussed in Section 2. In Section 3. historical records of
tornadic events in Greece are presented. Sections 4. and
5 summarize the geographical and temporal (seasonal,
monthly and diurnal) distributions of tornadoes. The syn-
optic meteorological and sea surface temperature (SST)
conditions associated with the development of tornadic
events are summarized in Section 6. Section 7 summa-
rizes the impact of the tornadoes in Greece. Conclusions
are presented in Section 8.
2. Data and methodology
Owing to changes in the methods of data collection
and data resources over time, the tornado climatology
of Greece discussed here is comprised of two different
datasets: (a) an inhomogeneous time series (17091999)
of events, sparsely distributed in time and (b) a homo-
geneous time series (20002012) of events uniformly
distributed over all the years.
The historical dataset covering the period 17091999
was synthesized after a methodical search through very
large historical digital archives of the National Libraries
and Greek newspapers from the 19th and 20th centuries.
We examined newspaper archives with daily and weekly
publication of the following Greek historical news-
papers: Ermis (18751880), Acropolis (18831884),
Skript (18931963), Empros (18961969), Macedonia
(19111981), Eleftheria (1944 1967) and Taxydromos
(19581977).
Historical data were also retrieved from the academic
literature, and administration and HNMS records. It is
noteworthy that we came across significant reports about
tornadic activity in Greece in historicalarchaeological
and philological reviews (Politis, 1880) in Parnassus,an
historical Greek journal.
The recent dataset covers the period 20002012,
beginning when systematic tornado climatology research
was started. A systematic effort in recording tornadoes,
waterspouts and funnel clouds in Greece was started in
2007 by the Laboratory of Climatology and Atmospheric
Environment (LACAE) of Athens’ University. Since then
LACAE has been actively collecting information on
tornadoes in Greece.
An open-ended online tornado report database
(http://tornado.geol.uoa.gr) web system was launched in
2009 by LACAE in order to record any tornadic activity
2013 Royal Meteorological Society Int. J. Climatol. (2013)
A CLIMATOLOGY OF TORNADIC ACTIVITY OVER GREECE
Figure 1. LACAE’s flowchart presenting tornado report system.
in Greece. With the simplicity of a Google Map interface
(as the majority of internet users have been familiarized
with) and within three steps, the user is able to report any
tornado, waterspout or funnel cloud in Greece. The user
could send any addition information concerning damage,
fatalities and attach any media file (photograph or video)
to support his/her report. The submitted tornado report,
after passing a plausibility check by a quality control
manager based on confirmation from other sources
(newspaper, HNMS, etc.) or other tornadoes’ reports,
confirmation of the attached media files, and finally
the liability of the user, it is stored in the database. A
flow chart of the LACAE tornado report data stream
system is presented in Figure 1. More than 100 events
have been reported since 2009, but only 90 reports have
passed the plausibility check. In this article, only the
90 events that have been fully verified have been taken
into account. Moreover, the last 10 years’ database was
also enriched by reports from HNMS’s forecasters, as
several waterspouts were spotted over the Ionian and
Aegean Sea by those forecasters themselves who were
on duty when the waterspouts occurred. Furthermore,
our data base has been crosschecked with the data from
the European Severe Weather Database (ESWD, Dotzek,
2009) and we have verified our records.
With the use of the geographical information system
(GIS) software ArcGis 9.3 we developed a geo-database
with more than 612 tornadic events from February 1709
to December 2012. The application of GIS software pro-
vides not only the ability to add significant information
such as disasters, fatalities, dates and other information,
but also the ability to produce several thematic layers that
could assist the research for specific regions.
In this climatological study, the Greek territory was
divided in 12 subregions – domains (Figure 2) in order
to reveal the spatial and temporal variability of tor-
nadic events. A regular grid with steps every 0.2lat-
itude/longitude was used to draw the 12 domains in
order to obtain a more useful and objective analysis
for the purpose of calculating frequencies of spatial dis-
tribution. The subdomains do not cover all the Greek
territory (especially the interior of Thessaly) but cover
the most favourable regions of Greece with respect to
tornado, waterspout and funnel clouds development. As
far as domain D9 is concerned, we did not follow the
abovementioned methodology (grid with steps of 0.2lat-
itude/longitude), because there were scattered events over
few islands or along eastern Aegean Sea region. Instead
of the mentioned grid a free non rectangular schema was
drawn to merge all the tornadic events and thus it does not
count in the objective analysis of the spatial distribution.
The Thessaly interior area was not taken into account in
spatial analysis as there was an inaccuracy of location for
several events. Regarding temporal analysis, Thessaly’s
events were taking into account only in annual and sea-
sonal variability. The spatial analysis of these subdomains
is presented in Section 4., but in Section 5 (temporal dis-
tribution) all tornadic events recorded in Greek territory
(in and out of the subdomains) were considered in our
analysis, accompanied with specific subdomain temporal
distribution (Table 3). With the term of ‘tornadic events’
(hereinafter TE) we have included any event that could
2013 Royal Meteorological Society Int. J. Climatol. (2013)
I.T. MATSANGOURAS et al.
Figure 2. The constructed 12 domains over Greek territory using GIS.
D1 subregion is located over north Ionian Sea including Corfu Island;
D2 area is located on central Ionian Sea enclosing the NW part of
Peloponnesus; D3 area concerns the Crete Island; D4 is limited over
Attica (enclosing Athens city). D5 concerns the complex islands of
Sporades, over NW Aegean Sea, D6 the central Macedonia, D7 the
south areas of Peloponnesus, D8 the Cyclades’ islands complex and
D9 the east part of Aegean Sea. At D9 domain we did not follow the
above mention methodology (grid with steps of 0.2latitude/longitude),
because there were scattered events over few islands or along eastern
Aegean Sea region. Instead of the mentioned grid a free non rectangular
schema was drawn to merge all the tornadic events and thus it does not
count in the objective analysis of the spatial distribution. Finally, we
added three more domains: D10 covering the west parts of Greece, D11
over the north part of D3 and D12 over north Crete Island (Iraklion
city).
be characterized as a tornado or waterspout or funnel
cloud.
3. Historical records of tornadoes in Greece
3.1. Tornado references in ancient and traditional
documents
During our research we came across, in historical news-
papers, several synonyms for tornadoes and waterspouts,
which were derived from local traditions and mytholo-
gies (e.g. in Gortynia, Arkadia’s prefecture in west
Greece, was known as anemogazou; in the Macedo-
nia region, in northern Greece, were called anemikes;
in Corinth, central Greece, was called saganakia;in
Corfu Island were called trompa; in western Greece were
known as roufoulas).
The twisting winds according to tradition (synonymous
with ‘strovilos=vortex in ancient Greek), were called
anemostrovilas,anemostrouvilos,anemostrofoulas,
anemoroufoulas and anemostroufoulas. These synonyms
were compounded words of anemo (=wind) +strovilas
or roufoulas or ... (=vortex) and the meaning of these
words in Modern Greek traditions reveals that twisting
winds ‘suck’ or they ‘were violently rotated’. Charles
du Fresne, sieur du Cange or Ducange (16101688),
a distinguished philologist and historian of the Middle
Ages and Byzantium, in his important work Glossarium
mediae et infimae Latinitatis (Glossary of medieval and
late Latin, 1678) used the synonym anemostrofilon:
‘severe storm developed accompanied with gale winds
and anemostrofilon crushed their ships’.
In Parnassus (Politis, 1880), a historical Greek
journal comprising mainly philological, historical
archaeological and also scientific essays, it is indicated
that during medieval times tornadoes, not only in Greece
but also in several European countries, were commonly
believed to come from demonic forces. This superstition
in Greece came from the noise of the air motions
associated with the swirling cloud. As a tornado formed,
people near the tornado’s path lay down in order to be
protected and whispered prayers. Salvanos and Salvanu-
Papavlasopulu (1930) quote a prayer that citizens of
south Corfu (NW Greece D1 in Figure 2) used to
whisper in order to escape from tornado’s force impact.
Aristophanes (446 BC386 BC), a comedy playwright
and poet of ancient Athens, in his work ‘Batrakhoi’ (The
Frogs, 405 BC) supports this superstition and states that
tornadoes are matched with demons: ‘
, ,
,
’ (the horrible snake-like demon, who fought
against Gods, got associated with the most powerful
whirl wind, and was called ‘efyros’). Although it was
believed in many places in Greece that tornadoes came
from evil spirits, there were several places especially in
north, west and central Greece where people believed
that tornadoes came from fairies. Similarly, in these
places when a tornado formed, the people asked the poets
to satisfy the fairies, so the fairies would not destroy
their crops or lift them in the air. An ancient proverb
states: ‘ ’(surrendered to the winds)
and it was believed that twisting winds are able to lift
you up and throw you far away. The epic poet Homer
used this ancient curse in his epic poem Iliad; he states
that, Helen cursed herself, wishing a windstorm to lift
her up. Moreover, in his epic poem The Odyssey, Homer
stated that Penelope wished a storm had lifted her up, as
happened to Pandarus’ daughter.
In Meteorological Phenomena and Climate in Byzan-
tium (Telelis, 2004a, 2004b, 2005) we came across
with significant reports relating to severe weather within
4th15th century AD that occurred in the Byzantine
Empire. Based on damage description and eyewitness
reports more than 20 events could be possible charac-
terized as tornado impact all over SE Europe and Middle
East (the Byzantine Empire).
Tomkinson (2006) in Travellers’ Greece: Memories of
an enchanted land presents a broad selection from rich
body of travel literature which has been inspired by those
lands which are now part of Greek state. Tomkinson
presents eyewitness reports from 1870 that describe the
formation and dissipation of several waterspouts over the
2013 Royal Meteorological Society Int. J. Climatol. (2013)
A CLIMATOLOGY OF TORNADIC ACTIVITY OVER GREECE
Ionian Sea. The spectacular view of waterspout over the
Ionian Sea inspired Joseph Cartwright, an artist of 17th
century to capture this nature phenomenon near Lefkada
island (Tomkinson, 2006).
On the basis of these ancient and traditional reports, it
is evident that tornadoes were not an unknown event, and
the difficulty to understand these violent events forced
people to believe that tornadoes came from evil spirits.
Moreover, the various names given to tornadoes prove
that in specific subregions of Greece tornadoes were not
as rare as they were in other places.
3.2. Inventory of historical tornadic events in Greece
A number of rare events have been documented in several
subregions in Greece, such as among others the deadly
tornado in Astakos of Aitoloakarnania, central west
coast of Greece (D2 domain in Figure 2), on October 18,
1934; it initially began as a waterspout and then came
onshore causing three fatalities, injuring forty people and
damaging many houses (Kanellopoulou, 1977). Further-
more, ‘fish-rain’ events associated with waterspouts that
sucked small fish from lakes occurred in October, 1951
in Alona, Florina, western Macedonia, Greece (Livadas,
1954), and in December 2002 in Korona, Kilkis, central
Macedonia, Greece (Sioutas, 2011).
The first tornadic event in our historical records
refers to a waterspout in Gasitsa of Corfu Island (D1
domain in Figure 2) on 13 February 1709, causing a
fisherman casualty. The next day (14 February 1709) a
tornado developed over the same area (Garitsa), inducing
structural damages to a monastery. On 23 September
1843, a tornado developed in Athens during evening with
no significant damage in contrast to a severe tornado
in Athens in 1852 that caused significant damage. Mr
Mackenzie (a traveler, who recorded the interaction
of atmospheric conditions with crops productivity in
Corfu Island) in a letter to Dr. George Lawsen (a
Scottish-Canadian botanist and Professor of Chemistry
and Natural History at Queen’s University) stated that a
severe hailstorm occurred over Corfu Island, followed by
a waterspout having the appearance of a huge inverted
funnel (Mackenzie, 1858). A more pronounced event
happened on 11 August 1902, when, a tornado caused
the derailment of Athens local train and more than 50
passengers were injured as the nine wagons were derailed
in Heraklio suburban area (D4 domain in Figure 2).
The most deadly tornado event in Greek history took
place in Megdova Lake of Karditsa (central Greece)
on the morning on 17 December 1959. The Elefthe-
ria (EEYEPIA) newspaper reported on 18 December
1959 that, 21 workers lost their lives as their boat was
destroyed by a waterspout. An eyewitness ranger offi-
cer reported that the boat was suddenly lifted up by
the waterspout and was lost inside the risen column of
water. These fatalities triggered by waterspouts were not
the only ones as on 8 September 1963 a waterspout in
Kalamata’s port (D4 domain in Figure 2) caused the
fatality of one sailor. Another remarkable event took
place on 8 August 1902, when a waterspout sank a 7-
ton ship 810 miles offshore of Avlona port in Albania
(north from D1 in Figure 2). On 8 July 1963, a tornado
formed in the evening near to Kalamata (D7 domain in
Figure 2) causing significant damage near the villages of
Manesi and Sterna. The tornadoes not only caused dam-
age to structures, but also to crops as more than two
thousand olive oil trees were uprooted and three men
were injured. On 19 October 1976, a tornado formed in
Rhodes island (south island in D9 domain, Figure 2),
causing significant damages while six people were
injured.
Studying the impact of tornadic activity on humans
from 1709 to 1999, we found that 27 people were killed
and more than 103 people were injured due to tornadoes
and waterspouts. Table 1 presents all the historical (before
2000) tornadic events, accompanied with information
about Greek subregions and fatalities.
It should be emphasized that the 300-year database of
tornadic events, includes decades in which a number of
tornadoes and waterspouts most likely remained unre-
ported, as they occurred in unpopulated areas or they
didn’t cause significant damage to be considered note-
worthy. Table 1 presents a detailed summary of historical
events per type for every subregion; accompanied with
the number of fatalities and injuries.
4. Geographical distribution
Our analysis concerns the period 17092012, as the first
event was recorded on 13 February 1709 at Corfu Island
(NW Greece, D1 domain in Figure 2) and the latest event
on 31 December 2012 at Rhodes island (SW Greece, D9
domain in Figure 2).
A total of 612 events were recorded and catalogued on
405 days, as there were several days with multiple events.
During this period (17092012), the 612 recorded events
consist of 171 tornadoes, 374 waterspouts and 67 funnel
clouds (Figure 3). Tornadoes and waterspouts during this
period caused more than 29 fatalities and more than 114
people were injured, while in numerous cases tornadic
activity had great impact on the local society as numerous
damage to crops and structures was recorded (see tornado
impact in next section).
The complex inland terrain of central and north Greece
along with the surrounding Seas (Ionian Sea at the west
and Aegean Sea at the east) appears to be a favourable
area for intense vortices such as tornadoes, waterspouts
and funnel clouds. Although tornadoes can occur any-
where over Greece’s mainland, Figure 3 (based on 300-
year dataset) shows that they are more frequent along
west Greece, Attica region, Thessaly and over north
Greece (central Macedonia). A common characteristic
for these regions is the flat topography; the absence of
complex topography allows intense vortices to be spot-
ted from several kilometres far away. In addition these
regions are characterised as the most density populated
areas (more than 70 citizens/km2) based on 2011 census,
2013 Royal Meteorological Society Int. J. Climatol. (2013)
I.T. MATSANGOURAS et al.
Table 1. Historical records of tornadoes (TR), waterspouts (WS) and Funnel Clouds (FC) over Greece from 1709 to 1999 along
with the location and fatalities/injuries information.
Date Type Location/Domain Injuries/Fatalities
13 February 1709 WS Garitsa Corfu/D1 0/1
14 February 1709 TR Garitsa Corfu/D1
23 September 1843 TR Athens/D4
1852 TR Athens/D4
17 October 1857 WS Corfu/D1
2 December 1895 TR Athens/D4
11 August 1902 TR Athens/D4 50/0
17 December 1910 TR Argostoli Cephalonia/D2
21 December 1910 TR Argostoli, Cephalonia/D2
14 April 1922 WS Hrakleio Crete/D12
4 September 1923 TR Chania Crete/D11
5 July 1924 TR Athens/D4 3/0
13 November 1930 TR Chania Crete/D11
18 October 1934 TR Astakos/D2 40/3
26 December 1934 TR and WS Chania Crete/D11
27 November 1937 TR Rethymno Crete/D11 0/1
15 October 1951 TR Alona, Florinas
12 August 1953 TR Zante/D2
9 October 1953 TR Mytilini Lesvos/D9
1 February 1954 TR Patra/D2
2 February 1954 TR Rhodes/D9
15 November 1955 TR Athens/D4
3 November 1956 TR Athens/D4
1957 Jul 13 FC Rethymno Crete/D11
5 September 1957 TR Rhodes/D9 0/1
26 October 1959 TR Iraklion/D13
26 October 1959 TR Ymitos-Vironas/D4
7 December 1959 WS Megdova Lake, Karditsa 0/21
24 September 1960 FC Chania Crete/D11
8 July 1963 TR Manesi and Sterna/D7 3/0
9 September 1963 WS Kalamata/D7 1/0
31 October 1963 FC Araxos/D2
17 June 1964 FC Florina
20 November 1964 TR Chania Crete/D11
17 July 1966 FC Larisa
25 November 1966 FC Andravida/D2
19 November 1966 TR Chania Crete/D11
9 December 1967 FC Araxos/D2
26 November 1968 FC Souda Chania Crete/D11
2 January 1971 TR Avlonas Attica/D4
1 September 1973 TR and WS Corfu/D1
22 July 1975 TR Thessaloniki/D6
1975 WS Sporades/D5
19 Oct 1976 TR Floraki Rhodes/D9 6/0
25 March 1977 TR Kourouta Ilias/D2
28 November 1978 TR Sounio, Xanthi
2 January 1980 TR Garitsa Corfu/D1
13 June 1984 WS Corfu/D1
15 October 1984 TR Vlasi, Kalamata/D7
July 1989 WS Chalkidiki/D6
15 August 1995 TR Corfu/D1
14 March 1996 TR Corfu/D1
24 November 1996 TR Corfu/D1
13 September 1996 TR Platanos, Nafpakto/D2
20 November 1998 TR Zante/D2
8 August 1999 WS Corfu/D1
conducted by the Hellenic Statistical Authority. Simi-
larly, waterspouts can occur anywhere over Ionian and
Aegean Sea, but they occur more frequently around Crete
and the Corfu Islands (domains D3 and D1, respectively
in Figure 2). The majority of waterspouts events were
reported close to the coasts in contrast with the small
number of waterspouts in open seas, reported by sailors.
Fewer waterspouts have been reported over the open
seas than near the coast owing to the higher population
density near the coasts. In addition, numerous observed
2013 Royal Meteorological Society Int. J. Climatol. (2013)
A CLIMATOLOGY OF TORNADIC ACTIVITY OVER GREECE
Figure 3. Spatial variability of tornadoes, waterspouts and funnel clouds
over Greece for the period 1709–2012. (Numerous tornadic events lie
under the tornadic type symbols, due to low resolution of the image).
waterspouts came offshore and caused significant dam-
age. On 20 September 2011, a waterspout came offshore
in Vlychos bay, marine of Lefkada Island (Ionian Sea),
causing significant damage to several moored boats and
one fatality.
The geographical distribution of historical tornadic
events (17091999), indicates that they are favourable
over north Crete, Corfu Island (north Ionian Sea), NW
Peloponnesus and Attica. In addition significant tornadic
occurrences were spotted over south Peloponnesus, Spo-
rades complex islands and north parts of Rhodes island.
Sparse events were also recorded over Thessaly, central
Macedonia and eastern Aegean Sea.
The geographical distribution of tornadic events during
the last 13 years (2000–2012) reveals that they may occur
all over the Greek territory. The most frequent occurrence
of tornadic events are over north Crete and Corfu Island
(D12 and D1, respectively). Significant tornadic density
(events/year) contouring (not shown) is evidence along
west Greece with maxima over NW Peloponnesus; sec-
ondary maxima are located over the central Ionian Sea
and south Peloponnesus (D2 and D7, respectively). Attica
and Thessaly could also be characterized as favourable
areas for tornadic development. More specifically, tor-
nadic cases over Attica are related to tornadoes, funnel
clouds (mainly over the southern parts) and waterspouts
over Saronikos Gulf. However, high values of tornadic
density over Thessaly concern tornado events. The occur-
rence is lowest in north Greece and in some inland areas
of central Greece, where topography is higher and den-
sity population is lower; Figure 4 illustrates these areas
by the number of tornadic events normalized by popula-
tion per year per Greek prefecture, based on 2011 census
of the Hellenic Statistical Authority.
Figure 4. A thematic map of tornadic events (20012011) normalized
by population per year per Greek prefecture based on 2011 census by
the Hellenic Statistical Authority.
Tornadoes are more frequent (not shown) over NW
Peloponnesus followed by central Macedonia, south parts
of Corfu Island, north parts of Rhodes island and
Thessaly area. Tornadoes also occurred over Attica (south
mainland), south Peloponnesus (as numerous waterspouts
came onshore). Waterspout frequency (not shown) is
highest over north Crete (D12 in Figure 2) followed by
Corfu Island (D1). Waterspouts were also recorded along
west Greece (D10 and D7), Sporades complex islands,
central Aegean Sea and eastern Aegean Sea.
Significant results were found concerning the distribu-
tion of tornadoes and waterspouts for the aforementioned
domains during recent 13 years dataset (20002012).
D10 (west Greece) domain is the largest Greek subregion,
with 19.09 tornadic events/year, 9.69 waterspouts/year
and 5.08 tornadoes/year (Table 2). Almost the half of the
mean annual tornadic events of D10 region took place
over D1 domain (8.53 events/year). Tornadoes are more
frequent over D2 domain with a mean annual value of
2.85 tornadoes/year, followed by D6 (central Macedonia)
and D1 (Corfu Island) domains. Waterspouts are more
frequent around Crete Island as 11.23 waterspouts/year
took place in D3 domain (thus, the 1/3 of mean annual
events). More specifically, 10 waterspouts/year occurred
north of Crete (D11 domain) and 8 waterspouts/year
developed north of Heraclion city (D12), a number almost
approaching the annual mean waterspouts over the Ionian
Sea (D10, approximately 9.69 waterspouts/year). The sec-
ond domain where waterspouts occurred is Corfu Island,
with an annual mean of 6.92 events (Table 2), a num-
ber that represents 1/4 of the total annual events in
Greece and its nearby territories. Based on a statis-
tical analysis of the recent dataset (20002012) it is
seen that Greece faces an annual mean of 42.15 tornadic
events per year which is broken down into 10 for TR
2013 Royal Meteorological Society Int. J. Climatol. (2013)
I.T. MATSANGOURAS et al.
Table 2. Annual mean tornadic events and frequency (events/year/10 000 km2) for each domain considered in the analysis for the
period 20002012.
Domain Period 20002012
Annual mean
Tornadic events TR WS Tornadic events frequency (events/year/10 000 km2)
Greek territory 42.15 10 27.69 3.19
D1 (Corfu) 8.53 0.92 6.92 9.52
D2 (north west Peloponnesus) 5.54 2.85 2.38 2.68
D3 (Crete) 12.92 0.46 11.23 4.64
D4 (Attica) 2.38 0.69 1 2.54
D5 (Sporades) 1.85 0.15 1.62 1.32
D6 (central Macedonia) 2.23 1.38 0.77 0.85
D7 (south Peloponnesus) 1.54 0.85 0.77 1.27
D8 (Cyclades) 1.31 0.07 1.23 0.46
D10 (west Greece) 19.09 5.08 9.69 3.89
D11 (north Crete) 11.38 0.31 10 7.71
D12 (Heraklion area) 9.08 0.15 8 16.65
In addition the annual mean tornadic events, tornadoes (TR) and waterspouts (WS) for each domain considered in the analysis is shown for the
period 2000– 2012 (past 13 years).
and 27.69 WS events per year. The frequency of tor-
nadic events in absolute values of events/year/10 000 km2
is 3.19 (Table 2), pointing out to a rather high fre-
quency of tornado events as compared with that in other
European countries. Dotzek (2000) from a conserva-
tive estimate of tornadic activity in Germany deduced a
recurrence density of about 0.2 (events/year/10 000 km2).
Holzer (2000) estimated an overall average for Aus-
tria at 0.3 tornadoes per 10 000 km2a year. Paul
(2000) calculated the annual risk probability of signif-
icant tornadoes in France to 0.66 ×105and Tyrrell
(2003) gave an annual tornado density for Ireland of
1.2 tornadoes/10 000 km2. In Table 2 (column 4) we
present the frequency of tornadic events per geograph-
ical domain (events/year/10 000 km2). The maximum
frequency (16.65) is depicted at D12 (Heraklion area),
followed by D1 (Corfu) (9.52) and D11 (north Crete area)
(7.71). Besides, West Greece (D10) experiences a fre-
quency of 3.89 tornadic events/year/10 000 km2against
2.68 at NW Peloponnesus, favourable regions for tornado
genesis.
The spatial distribution of tornadic cases from 1709 to
1999 (historical dataset) is in agreement with the geo-
graphical distribution within 2000 to 2012. Similarities
concern tornadic occurrences over north Crete, Corfu,
NW Peloponnesus, Attica, south Peloponnesus, Sporades
complex islands and Corfu Island. The geographical dis-
tribution presented in this section is in agreement with the
preliminary results of Nastos and Matsangouras (2010)
and Sioutas’ (2011) geographical distribution of annual
mean days of tornadoes and waterspouts.
5. Temporal variability
A discussion of tornadic events with detailed information
about their seasonal, monthly and diurnal occurrence is
presented in this section. Any presentation of the annual
variability of tornado events based on the historical
300-year period (17091999) would be inappropriate
because the historical dataset consists of discrete events
distributed inhomogeneously within the examined period.
However, from the inter annual variability based on
the recent dataset 20002012 (not shown), a continuous
sample of events that occurred during the past 13 years,
it is found that 2010 was the year with the most recorded
tornadic events (76 TE events) followed by 2006 (72 TE
events) and 2009 (67 TE events). In order to explain
the tornadic activity during the aforementioned years, we
examined the influence of atmospheric circulation with
respect to North Atlantic Oscillation Index (NAOI) and
Mediterranean Oscillation Index (MOI). NAOI is defined
as a difference of standardized sea level pressure (SLP)
time series from a station close to the centre of the
Azores High (Gibraltar) and a station close to the centre
of Islandic Low (Reykjavik) (Jones et al., 1997). Posi-
tive NAOI phase (stronger dipole) results in prevailing
westerly winds that are strengthened and moved north-
wards causing increased precipitation and temperatures
over the northern Europe and drier and cooler anoma-
lies in the Mediterranean region, while opposite condi-
tions occur during the negative (weaker dipole) NAOI
phase (Hurrell and Van Loon, 1997). MOI has been sug-
gested by Conte et al. (1989) and Palutikof et al. (1996),
and concerns the existence of a teleconnection pattern in
the annual geopotential height fields at 500 hPa between
Western and Eastern Mediterranean. The authors defined
the Mediterranean Oscillation index (MOI) as the normal-
ized pressure difference between Algiers (36.4N, 3.1E)
and Cairo (30.1N, 31.4E). Several studies (Corte-Real
et al., 1995; Maheras et al., 1999; Dunkeloh and Jacobeit,
2003; Nastos et al., 2011) have revealed the Mediter-
ranean Oscillation as the most important regional circu-
lation mode affecting the time series of temperature and
rainfall between Western and Eastern region. The MO
is related to the activity of cyclogenesis in the Mediter-
ranean, mainly in the bay of Genoa. In the positive phase
2013 Royal Meteorological Society Int. J. Climatol. (2013)
A CLIMATOLOGY OF TORNADIC ACTIVITY OVER GREECE
Figure 5. Seasonal variability of tornadoes, waterspouts, funnel clouds
and total events over Greece for the period 2000–2012.
of MO, the cyclogenesis is anomalously intense while in
the negative phase it is anomalously weak. The MOI and
NAOI are highly correlated among themselves, as the
passages of the cold fronts from the Atlantic, described
by NAOI variability, are one of the triggers for the
Mediterranean cyclogenesis.
We found that during the rain period of the year
(OctoberMarch) NAOI values were 1.83, 0.47 and
0.54 in 2010, 2006 and 2009, respectively. Similar
results were revealed with respect to MOI (0.78,
0.12, 0.48 in 2010, 2006 and 2009, respectively).
These findings explain in some extent the higher tornado
activity in 2010, due to more cyclone passages through
the Mediterranean, affecting mainly western Greece.
The most tornado events (TR) were recorded during
2009 (16 TR), 2005 (16 TR), 2002 (16 TR) followed by
2010 (15 TR) and 2008 (15 TR). The most waterspout
events were observed in 2006 (66 WS) followed by 2010
(53 WS), 2009 (44 WS) and 2012 (42 WS). Funnel clouds
were reported most in 2007 (13 FC) followed by 2008
(12 FC).
During the analysis of data we found a large increase
of documented tornadic events during the last 13 years,
but this increase does not reveal or imply any clima-
tologically anomaly or trend. It is very likely due to the
new technologies that entered in our life the last 13 years,
such as the use of social networks based on the internet,
mobiles phones equipped with cameras, and finally the
increased interest of the media or the general public for
such extreme phenomena.
5.1. Seasonal variability
Tornadoes, waterspouts and funnel clouds occur during
all four seasons in Greece. The seasonal variability, based
on 20002012 dataset, is shown in Figure 5. Autumn
is the dominant period as from the seasonal variability
in the recent dataset it is seen that almost 47% of the
events were recorded during autumn (265 TE), followed
by winter (107 TE), spring (93 TE) and summer (92 TE)
season.
Tornadoes’ seasonal distribution shows that tornadoes
were more frequent during autumn (68 TR), followed by
winter (26 TR), summer (25 TR) and spring season (16
TR). Waterspouts’ frequency exhibits a maximum during
autumn (180 WS) followed by summer (64 WS), spring
(59 WS) and winter season (59 WS). Funnel clouds
are reported with higher frequencies during winter (21
FC), spring (18 FC) and autumn (17 FC), against lower
frequency during summer (3 FC); these frequencies are
lower than those of tornadoes and waterspouts.
The seasonal variability of tornado, waterspout and
funnel cloud days for 20002012 period (not shown)
depicted that autumn (155 days) is the most favourable
season followed by winter (81 days), summer (55 days)
and spring (54 days). The seasonal pattern is in agreement
with the aforementioned seasonal variability of tornado
events and with the recent research by Sioutas (2011).
Tornado days are more frequent during autumn (57 TR
days) followed by winter (25 TR days), summer (19 TR
days) and spring (8 TR days). The seasonal mean number
of days with at least one tornado is 4.75 days during
autumn, 2.08 days during winter 1.58 during summer and
0.67 during spring season. Waterspout days are more
frequent during autumn season (86 WS days) followed by
winter (41 WS days), spring (34 WS days) and summer
(34 WS days). The seasonal mean number of days is
7.17 days during autumn, 3.42 during winter, and 2.83
during summer and spring season. Funnel cloud days are
commonly observed during winter (15 FC days) followed
by autumn (12 FC days) and spring (12 FC days), while
the seasonal mean number of days is 1.25 days during
winter and 1 day during autumn and spring season.
5.2. Monthly distribution
It was found that tornadoes, waterspouts and funnel
clouds occur during all months over Greece (Figure 6).
The monthly variability of tornadic events from 2000 to
2012, displays a maximum during October (104 TE),
followed by September (95 TE), November (57 TE),
December (53 TE), May (35 TE) and July (35 TE).
October (30 TR) is the month with the highest tornado
frequency (based on events), followed by November (28
TR), July (12 TR), February (11 TR) and December (9
TR). This monthly tornado distribution over Greece with
a maximum during autumn is much different from the
respective distribution over central European countries.
The continental region of central Europe has a maximum
monthly tornado frequency during the warm season
(from June to August), as already stated by Wegener
(1917), Dotzek (2000) and Rauhala et al. (2012). This
kind of distribution, however, is evident over the north
part of Greece, the region of central Macedonia (D6
domain), where the maximum frequency is found during
the summer months (July and June), as the greatest
potential instability occurs then as a result of strong
surface heating. The monthly distribution of tornado days
(not shown) presents a maximum during October (27 TR
days) followed by November (20 TR days), February
2013 Royal Meteorological Society Int. J. Climatol. (2013)
I.T. MATSANGOURAS et al.
Figure 6. Monthly variability of tornadoes, waterspouts, funnel clouds
and total events over Greece for the period 2000–2012.
(11 TR days), December (9 TR days) and July (8 TR
days). Taking into consideration the period 2000– 2012,
the monthly mean number of days with at least one
tornado is 2.1 days during October, followed by 1.54 days
during September, 0.85 days during February, 0.69 days
during December and 0.62 days within July.
The majority of waterspouts (24% of all waterspouts)
took place during September (87 WS) (Figure 6). The
second highest waterspout month is October (64 WS) fol-
lowed by December (30 WS), June (26 WS), November
(23 WS), July (22 WS), April (22WS) and May (21WS)
during June and July there were multiple waterspouts
events during just a few days). The monthly distribu-
tion of waterspout days (not shown) shows a maximum
of waterspout days during October (37 days), followed
by September (31 days), November (19 days), December
(17 days) and June (17 days). Based on the past 13 years,
the monthly mean number of days with at least one water-
spout is 2.85 days in October, 2.4 days in September, 1.46
in November, 1.31 in December and June.
Taking into account the seasonal and monthly variabil-
ity within the described domains in Section 4. (Table 3),
it is clear that tornadic events are most frequent during
the autumn. Exceptions to this rule were the tornadic
events that developed over D5 and D6 domains, where
the most favourable season was during the summer. Also,
tornado events for D5 and D3 domain are most common
during the winter against summer regarding D6 domain.
Autumn was the most favourable season for the formation
of waterspouts except for the D5 and D6 domains (max-
ima during summer) and over the D9 domain (maxima
during spring). The unusual maximum of waterspouts in
spring over D9 is very likely due to small number of
waterspouts recorded. More specifically, 10 out of 19
WS within D9 were occurred in 3 days during spring
season. So, there is not a reliable pattern of the compos-
ite mean synoptic conditions based on only these three
WS days, in order to derive clear conclusions for the
Figure 7. Diurnal distribution of tornadoes, waterspouts and funnel
clouds events over Greece for the period 2000–2012. The numbers
above the orange bars illustrate the number of total events.
synoptic conditions favour WS in spring. In Section 2, we
have remarked that there were scattered events over few
islands or along eastern Aegean Sea region (D9 domain).
Matsangouras et al. (2013) presented a seasonal objec-
tive synoptic classification for west Greece (D10) per
tornadic type and identified specific synoptic weather pat-
terns favourable for tornado and waterspout occurrence.
Details about monthly variability per type (TE, TR and
WS) for each domain are presented in Table 3.
5.3. Diurnal variability
The diurnal variability of tornadoes, waterspouts and fun-
nel clouds that developed from 2000 to 2012 over Greece
is depicted in Figure 7. It exhibits a broad maximum in
the distribution between 08:00 and 15:00 hours UTC time
period. However, only 39% of recorded cases have been
classified with an exact time of incidence. Nevertheless,
taking into consideration the available cases, a distinct
maximum of tornadic activity is obvious at 09:00 hours
UTC, which is in agreement with Sioutas’ (2011) diur-
nal distribution. More than 75% of all cases occur from
08:00 to 15:00 hours UTC. Tornado diurnal distribution
displays a maximum frequency from 11:00 to 13:00 hours
UTC, while a uniform distribution appears for the rest
of the hours of the day (Figure 7). This maximum fre-
quency is related to tornado occurrence over D6 during
summer season, as convective and unstable weather con-
ditions are the main characteristic over central Macedonia
during that season.
Waterspout diurnal variability in contrast to the tor-
nado diurnal variability presents two diurnal peaks
(Figure 7): the first is early in the morning (07:00
09:00 hours UTC) and the second one after noon time
(14:00– 15:00 hours UTC). These two maxima of diurnal
distribution reveal a diurnal pattern of activity broadly
similar to the behaviour in Key West Florida vicinity
(Golden, 1973), in Nassau Bahamas (Peterson, 1978)
and around Catalonia during 19502009 (Gay`
aet al.,
2011). Golden (1974a, 1974b) and Simpson et al. (1986)
demonstrated that waterspouts are often initiated by
2013 Royal Meteorological Society Int. J. Climatol. (2013)
A CLIMATOLOGY OF TORNADIC ACTIVITY OVER GREECE
Table 3. Seasonal and monthly temporal variability of total events (TE), tornadoes (TR) and waterspouts (WS) for each domain
considered in the analysis, during the period of 20002012.
Domain Maximum frequency
of events within season
Maximum frequency
of events within month
TE TR WS TE TR WS
Greece Autumn Autumn Autumn September November September
D1 (Corfu) Autumn Autumn Autumn November November November
D2 (north west Peloponnesus) Autumn Autumn Autumn November November September
D3 (Crete) Autumn Winter Autumn September January September
D4 (Attica) Autumn Autumn Autumn October/November November October
D5 (Sporades) Summer Winter Summer June December June
D6 (central Macedonia) Summer Summer Summer July July June
D7 (south Peloponnesus) Autumn Autumn Autumn November October/November November
D8 (Cyclades) Autumn Autumn Autumn October October November
D9 (east Aegean Sea) Autumn Autumn Spring September October/December May
D10 (west Greece) Autumn Autumn Autumn October November September
D11 (north Crete) Autumn Autumn Autumn September November September
D12 (Heraklion) Autumn Autumn/Winter Autumn September January September
fair-weather cumuli or cumulus congestus and not nec-
essarily by thunderstorms, as a pre-existing vertical vor-
ticity within the boundary layer is amplified by vortex
stretching below and within the cumulus up-draft.
Sources for vertical vorticity near the ground may be
convergence lines, outflow boundaries from advancing
cold pools or sea breezes. Keul et al. (2009) enclosed
numerous waterspouts in mesoscale situation of sea-land
breeze over Ionian and Aegean Seas, by implemented
Szilagyi’s waterspout nomogram (Szilagyi, 2009) within
20022006. In our research all waterspout events were
associated with fair-weather cumuli or cumulus congestus
clouds. The aforementioned two diurnal peaks of water-
spouts are in agreement with the timing of the sea-land
breeze circulations regarding waterspout events with a
lack of significant convective instability or synoptic forc-
ing. The funnel cloud temporal distribution (Figure 7) is
similar to the waterspout distribution, with two maxima
at 09:00 and 14:00 hours. The rest of funnel cloud events
are scattered in all other hours.
6. Synoptic-scale meteorological and SST conditions
The synoptic-scale conditions associated with the devel-
opment of tornadic events, based on seasonal and
geographical composite analyses, are discussed in this
section. Synoptic-scale composites were synthesized for
the autumn season for the domains D1, D2 and D11,
where tornadoes and waterspouts are most frequent; tor-
nadoes are most frequent over D2 while waterspouts are
most frequent over D1 and D11 domains.
The daily composite mean and anomaly of the
synoptic-scale height/pressure fields at two levels in the
lower half of the troposphere during tornadic days were
constructed from National Centers for Environmental
PredictionNational Center for Atmospheric Research
(NCEPNCAR) reanalyses (Kalnay et al., 1996) plot-
ted from the Web page at the Physical Sciences Division
(PSD) of the NOAA/Earth System Research Laboratory
for the period 1 September 1948 to 31 December 2012.
The daily mean anomaly is calculated by the subtraction
of daily mean composite during tornadic days (1 Septem-
ber 1948 to 31 December 2012) from the daily mean
composite of 30 years (1981 2011) climatology (clino).
Figure 8 shows the daily mean composite and daily mean
anomaly of geopotential heights (gpm) at 500 hPa, along
with the mean sea-level pressure at the surface (SFC) for
tornado days during autumn period (the most active sea-
son) for D2 domain. Figures 9 and 10 show the daily
mean composites and daily mean anomalies of synoptic
conditions for the D1 and D11 domains during tornado
days, respectively.
For the D2 domain, the daily composite mean synoptic
conditions at 500 hPa for tornado days is characterized
by a broad trough along central and southern Italy, and
is associated with a SW upper-air stream over the area
of interest (Figure 8). The daily mean composite of air
temperature (not shown) is characterized by a short-wave
thermal trough along Italy; the relatively cold isotherm of
20 C crosses central Italy. The air temperature over the
central Ionian Sea and NW Peloponnesus ranges between
17.5 Cand15.5 C. The centre of the cyclonic
circulation at the surface (SFC) is positioned over the
Gulf of Taranto. The composite daily anomaly at 500hPa
during this season appears to be more than 130 gpm
centred over the Gulf of Taranto The SFC daily mean
anomaly of tornado days is centred over the same vicinity
and it is calculated to 11 hPa.
For the D1 domain, the composite mean synoptic
conditions at 500 hPa during autumn are characterized
by a long-wave trough from central Italy to Sicily;
this pattern is associated with a SW upper air flow
over the area of interest (Figure 9). It is seen from
the composite mean of SFC that a shallow cyclonic
circulation is centred over Italy, forcing a S-SW surface
air flow. The composite daily anomaly at this barometric
pressure level during autumn presents a deepening of
geopotential height at least equal to 90 gpm over central
2013 Royal Meteorological Society Int. J. Climatol. (2013)
I.T. MATSANGOURAS et al.
Figure 8. Composite mean (upper graphs) and anomaly (lower graphs) of tornadoes days during autumn for the geopotential height (gpm) at
500 hPa (left graphs) and mean sea-level pressure (hPa) (right graphs) for D2 domain for the period 1 September 1948 to 31 December 2012.
Mediterranean Sea (between Corsica and central Italy).
The daily composite anomaly of SFC during this season
reveals a pressure decrease more than 5 hPa over an
extensive region at SE Europe (Balkans).
For the D11 domain, a very broad through at 500 hPa
is located along Aegean Sea, suggesting W-NW upper-
air flow over Crete (Figure 10). Finally, the combi-
nation of high pressure over central Europe and low
pressure over the eastern Mediterranean Sea, estab-
lishes N-NW air flow at the surface over the area of
interest.
Long-term means (clino) during autumn season (not
shown); based on NCEPNCAR reanalysis 1981 2010
period, at 500 hPa reveals a zonal flow over south
Europe. Based on this long-term climatology at SFC
the combination of high pressure over Balkans and low
pressure over east Mediterranean Sea establishes a NE
and N-NW air flow over north and south Aegean Sea,
respectively. Over central Mediterranean Sea (Corsica,
Italy) a shallow cyclonic circulation forcing a SW-S
surface air flow over Ionian Sea.
Our daily composite mean synoptic results for D1,
D2 and D11 within 19482012, are in agreement
with Sioutas (2011) favourable synoptic types for
tornadoes and waterspouts development in Greece within
20002009.
Apart from the synoptic conditions that favour tornadic
events, topographic factors such as the varying depth of
the sea, the sea currents, varying shapes of the coastline
and SST (Carapiperis, 1952), are all likely to contribute
to the conditions related to the development of tornadoes
or waterspouts. The Ionian Sea and the coasts of west
Greece are the most favourable regions in the appearance
of tornadic events. The Ionian water body and the sea
near the northern coasts of Crete Island have lower
refresh rates of waters as compared with the north Aegean
and even to the southeast Aegean; the former receives the
colder Black Sea waters while the latter is affected by the
influx of Levantine waters mainly through the Rhodes
passage. This Levantine water mass travels northwards
along the eastern Aegean Sea, close to the coasts of Asia
Minor, having no impact in the central Aegean Sea. The
Levantine waters, when reaching the Black Sea outflow
in the vicinity of the island of Limnos, get subducted
below the very light Black Sea water layer. This layer
is formed by waters flowing out from the Dardanelles
and then moving westwards and eventually southwards. It
covers the north Aegean and moves in a cyclonic motion
westwards and southwards along the east coast of the
Hellenic Peninsula.
Additionally, the role of surface currents is significant
in configuring the spatial distribution of SSTs. This is
2013 Royal Meteorological Society Int. J. Climatol. (2013)
A CLIMATOLOGY OF TORNADIC ACTIVITY OVER GREECE
Figure 9. Composite mean (upper graphs) and anomaly (lower graphs) of waterspout days during autumn for the geopotential height (gpm) at
500 hPa (left graphs) and mean sea-level pressure (hPa) (right graphs) for D1 domain for the period 1 September 1948 to 31 December 2012.
the case of the current along the west coast of Pelopon-
nese that moves northwards relatively warmer waters [i.e.
Levantine Intermediate water (LIW)], while through the
Otrando straits fresher Adriatic waters move southwest-
wards (Malanotte-Rizzoli et al., 1997). Aegean’s SST
fluctuate spatially and seasonally, with mean monthly
SSTs varying from 8 C, in the north during winter, up
to 26 C, in the south (D3 domain in Figure 2) during
late summer (Poulos et al., 1997). The overall seasonal
spatial distribution pattern of Aegean’s Sea SST values
shows that the surface temperature changes is dependent
upon: (a) the distribution of the colder Black Sea Waters;
(b) the advection of the warmer Levantine Waters, influ-
encing only the eastern part of the south and central
Aegean; (c) coastal upwelling induced by the Etesians;
and (d) to a lesser extent but locally important (espe-
cially at the north) freshwater riverine outflows (Poulos
et al., 1997). Moreover, Skliris et al. (2011) stated that the
largest part of the Aegean Sea surface is cooling between
1985 and 1992; whilst a general warming is obtained
throughout the basin of Aegean Sea after 1992. Max-
imum SST warming trends are observed in the Cretan
Sea (D3); based on satellite-derived SST analysis higher
warming rates obtained during summer (0.068 C/year)
and autumn (0.050 C/year) than those obtained during
spring (0.037 C/year) and winter (0.026 C/year) for the
Aegean Sea. The highest positive seasonal SST anomaly
is obtained in summer 2002 with the obtained mean sum-
mer SST being about 1 C higher than the long-term
(19852008) summer mean (Skliris et al., 2011). It is
noteworthy that we came across significant number of
waterspouts (more than 40 WS) during 2002 over the
Aegean Sea.
Several publications during the last years are asso-
ciated with tornadic formation over fresh water bodies
(lakes and lagoons) and introduced some forecasting tech-
niques. Keltieka (1987) developed a forecast technique
about forecasting waterspout based on limited sample of
data over Lake Erie (one of the Great Lakes of North
America). Szilagyi (2009) using a large dataset of water-
spouts investigated 14 parameters as possible correlators
to waterspout formation over the Great Lakes. Out of
these, two instability parameters (Water-850 hPa tempera-
ture difference and convective cloud depth) and one wind
constraint (850 hPa wind speed, less than 40 knots) were
judged to be the strongest correlators. Matsangouras et al.
(2010), using remote sensing data (Andravida’s Doppler
radar) over the water bodies (scattered lagoons and small
lakes) of NW Peloponnesus, detected a shallow cyclonic
circulation accompanied with wind shear at low-levels
(from SFC up to 850 hPa) ahead the cold front prior
to tornado formation in D2 domain. King et al. (2003)
2013 Royal Meteorological Society Int. J. Climatol. (2013)
I.T. MATSANGOURAS et al.
Figure 10. Composite mean (upper graphs) and anomaly (lower graphs) of waterspout days during autumn for the geopotential height (gpm) at
500 hPa (left graphs) and mean sea-level pressure (hPa) (right graphs) for D11 domain for the period 1 September 1948 to 31 December 2012.
suggested that lake breezes may produce low level bound-
aries on which tornado may be form during SW air flow
ahead of a cold front in southern Ontario Lake area.
The difference of heating energy process by the solar
radiation between shallow waters and the adjoining allu-
vium soil (alluvium soils are heating faster and retaining
higher temperature) during synoptic days with SW air
flow may produce a shallow convergence line. As the
cold front encounters that shallow convergence line, the
additional moisture convergence and low-level vorticity
may provide a trigger for tornado formation. In these
cases, the near ground convergence alone can amplify
the vertical vorticity to tornado intensity and a down-
draft such as the rear-flank downdraft is not needed for
tornadogenesis.
7. Tornadoes activity impacts
Analysing the record of more than 300-year tornadic
activity (17092012), we found that at least 114 people
were injured and 29 lost their lives. As mentioned before,
tornadoes (or waterspouts that came offshore) caused
numerous fatalities and significant damage to structures
and crops in Greece. The geographical distribution of
injured people (not shown) during 17091999 is such
that west and central communities of Greece have been
affected, mainly NW Peloponnesus, Attica and south
Peloponnesus (D2, D4 and D7 domains, respectively in
Figure 2) accompanied with scattered cases in Crete,
Corfu and north Greece (D3, D1 and D6, respectively,
in Figure 2).
An intensity analysis has been carried out with respect
to the 300-year tornado database, based on the Torro or
T-scale (Meaden, 1976), broadly used in Europe against
the Enhanced Fujita or EF-scale implemented in United
States. The Enhanced Fujita or EF-scale (McDonald,
2002; McDonald et al., 2003, 2004; WSEC, 2004)
was implemented in USA by NOAA on 1 February
2007. It classifies the damage observed for typical
structures in the USA and then assigns, based upon an
expert elicitation, velocities claimed sufficient to explain
the degree of damage. EF-scale damage description is
endemic to the USA and it is not readily applicable
internationally. Any resulting variations in threshold
values from country to country could severely affect the
climatological consistency of worldwide tornado records.
Dotzek (2009) related several wind scales (T-scale,
F-scale, EF and Beaufort scale) by proposing a class
of new wind speed scales; they are called Energy- or
E-scales, in which the relevant scaling factors are derived
from physical quantities like mass flux density, energy
density (pressure), or energy flux density.
2013 Royal Meteorological Society Int. J. Climatol. (2013)
A CLIMATOLOGY OF TORNADIC ACTIVITY OVER GREECE
Figure 11. The geographical distribution of damage to crops and
structures over Greece for the period 1709–2012.
For 137 out of the 171 tornado events recorded in
Greece, sufficient information of tornado impacts to
buildings and crops is known in order to classify these
events using T-scale. It is found that 48 out of 137 tor-
nado events caused only structural damage, 17 out of
137 tornadoes caused only damage to crops and 72 out
of 137 tornadoes caused damage to both buildings and
crops. The geographical distribution of both damage to
crops and structural damage (Figure 11) is such that weak
tornadoes in north Greece did not cause significant dam-
age to crops with an exception of few strong tornadoes.
Nevertheless, tornadoes over NW Peloponnesus caused
significant damage not only to crops but also to struc-
tures. Attica experienced a significant impact in structural
damage during 17091950. A remarkable impact in both
destruction (to both crops and buildings) was also noted
in Corfu Island (Figure 11).
The dominant (27.73% of total cases) intensity of tor-
nadoes in Greece (Figure 12) is of T4 degree in T-scale.
The percentage of incidence of weak tornado events
(T0T3) is 51% versus 49% for strong tornado events
(T4T7). The combination of both datasets exhibits a
normal distribution for tornado intensity (Figure 12). T4
events are the dominant events as they occurred with a
frequency of 27.5 and 31.76% of total events, respec-
tively for both the historical and recent datasets.
From the geographical distribution of tornado intensity
(Figure 13) based on the 17092012 dataset, it is seen
that strong tornadoes (T4T7) occurred mainly along
west Greece (D10 in Figure 2). NW Peloponnesus (D2
in Figure 2) experienced strong tornadoes; the major-
ity of them were classified as T4T5. However, weak
tornadoes developed in Thessaly and central Macedonia
(D6 in Figure 2); the majority of them were classified as
weak (T0T3) and a few tornado events were noted in
Figure 12. Tornado intensity distribution over Greece for the period
1709– 2012.
Figure 13. Spatial tornado intensity distribution based on T-scale over
Greece for the period 1709–2012.
central Macedonia as T4 and T5. Attica experienced tor-
nado intensity from weak to strong scale with significant
impact in structures during 17091950.
8. Conclusions
Ancient and traditional documents give evidence that
tornadic activity over Greece is not as rare as considered
a couple decades ago. In this study, several unknown
historical tornadic events were found across Greece after
1709. Moreover, deadly tornadoes and waterspouts have
been documented, in which 29 fatalities and 114 injuries
from 1709 to December 2012 were reported.
The geographical distribution of historical dataset
17091999 is such that tornadic events are most com-
mon over north Crete, Corfu Island (north Ionian Sea)
and Attica. The geographical distribution of tornadic
events during the past 13 years (20002012) is such that
tornadic events occur all over the Greek territory with
relatively high values of tornadic occurrence over west
2013 Royal Meteorological Society Int. J. Climatol. (2013)
I.T. MATSANGOURAS et al.
Greece and north Crete island (D10 and D11, respec-
tively). Tornadoes are more frequent to occur over NW
Peloponnesus followed by south parts of Corfu Island,
north parts of Rhodes island and Thessaly area. Water-
spouts are more frequently observed over north Crete
(D12), followed by Corfu Island (D1). Waterspouts were
also recorded along west Greece (D10 and D7), the Spo-
rades complex islands, central Aegean Sea and eastern
Aegean Sea.
Taking into consideration the impacts of tornado activ-
ity over Greece, the most common intensity of torna-
does is estimated as T4. The geographical distribution of
injured people during 17091999 is such that west and
central communities of Greece were affected, mainly in
NW Peloponnesus, Attica and south Peloponnesus. The
geographical distribution of tornado intensity (not shown)
based on the 17092012 dataset reveals that strong torna-
does (T4T7) occurred mainly along west Greece (D10).
However, weak tornadoes developed in Thessaly and cen-
tral Macedonia (D6 in Figure 2), where the majority of
them were classified as only T0T3 intensity.
During the past 13 years, (20002012), 2010 was the
year with the most recorded tornadic events (76 TE
events) followed by 2006 (72 TE events) and 2009 (67
TE events). Negative values of the atmospheric circula-
tion indices NAOI and MOI within the rain period could
explain in some extent the higher frequency in the afore-
mentioned years, due to more cyclones passages through
the Mediterranean, affecting mainly western Greece. 50%
of the total tornadic events were recorded during autumn,
followed by the winter and summer seasons. Tornado’s
seasonal distribution shows that tornadoes were more fre-
quent during autumn (68 TR), followed by winter (26
TR), summer (25 TR) and spring season (16 TR). Water-
spout frequency exhibits a maximum during autumn (180
WS) followed by summer (64 WS), spring (59 WS) and
winter season (59 WS). Tornadic (TE) monthly variability
during 20002012, depicts a maximum during October,
followed by September, November, December, May and
July. October (30 TR) is the month with the highest tor-
nado frequency (based on events), followed by November
(28 TR), July (12 TR), February (11 TR) and December
(9 TR). Tornadoes most commonly develop during the
warm time of the day, as more than 75% of all cases
occur from 08:0015:00 hours UTC and the maximum
is at 12:00 hours UTC. Waterspout frequency of occur-
rence has two maxima during the day, the first early in
the morning (07:0009:00 hours UTC) and the second
after the noon time period (14:0015:00 hours UTC).
As far as the favourable synoptic conditions are
concerned for tornado events prevailing over west Greece
(NW Peloponnesus), a trough over south and central Italy
at 500 hPa accompanied with closed cyclonic circulation
over Taranto’s Gulf are the dominant pattern. Regarding
waterspouts’ synoptic conditions that occurred north
of Crete Island, a trough is also eminent over the
eastern Aegean Sea. The aforementioned conclusions
indentify regional features of tornadic activity in Greece,
constituting a useful contribution to the Greek tornadic
climatology.
Authors’ future work is oriented to estimate
spatialtemporal thermodynamic parameters analy-
sis using proximity soundings analysis based on NCEP
reanalysis and Weather Research Forecast (WRF-ARW)
atmospheric model simulations, as an effort to find an
optimum set of parameters to determine which events
were probably associated with supercells and which
were not. The reanalysis system has shown a great deal
of promise as a source of environmental information,
and proximity sounding analysis has long been a tool
to determine environmental conditions associated with
different kinds of weather events and to discriminate
between them, although it has been limited, necessarily,
by the spatial and temporal distribution of soundings
(Brooks et al., 2003). Apart from spatial-temporal anal-
ysis, further research of weather types classification with
respect to tornadic activity over Greece would interpret
the different seasonal and monthly variability of TE over
the examined domains. To this direction, Matsangouras
et al. (2013) suggested a seasonal objective synoptic
classification for west Greece (D10) per tornadic type,
and recognized specific synoptic weather patterns
favourable for tornado and waterspout occurrence.
Acknowledgements
The authors thank the Hellenic National Meteorological
Service (HNMS), Hellenic Agricultural Insurance Orga-
nization (ELGA), Hellenic National Library, Aristotle
University of Thessaloniki (Psifiothiki), state authori-
ties, media (blogs, news site, TV and radio stations),
the Hellasweather group and all anonymous people, who
contributed to accomplish this extensive dataset. Daily
composite images were provided by the NOAA/ESRL
Physical Sciences Division, Boulder Colorado. The
authors have no conflict of interest to declare.
References
Bertato M, Giaiotti DB, Manzato A, Stel F. 2003. An interesting case
of tornado in Friuli-North-eastern Italy. Atmos. Res. 67 68: 3 21.
Bluestein HB. 1999. Tornado Alley, Monster Storms of the Great Plains.
Oxford University Press: New York, Oxford.
Br´
azdil R, Chrom´
a K, Dobrovoln´
yP, ˇ
Cernoch Z. 2012. The tornado
history of the Czech Lands, AD 1119–2010. Atmos. Res. 118:
193– 204.
Brooks HE, Doswell CA III. 2002. Deaths in the 3 May 1999 Oklahoma
City tornado from a historical perspective. Weather Forecast. 17:
354– 361.
Brooks HE, Lee JW, Craven JP. 2003. The spatial distribution of severe
thunderstorm and tornado environments from global reanalysis data.
Atmos. Res. 67 68: 7394.
Brotzge J, Erickson S. 2010. Tornadoes without NWS warning. Weather
Forecast. 25(1): 159– 172.
Carapiperis LN. 1952. On the surface temperature of Greek waters.
Pure Appl. Geophys. 23(1): 153–161.
Clark M. 2008. The ‘Super Tuesday’ tornado outbreak of 5– 6 February
2008 over the Southern US. Int. J. Meteorol. 33: 127–140.
Conte M, Giuffrida S, Tedesco S. 1989. The Mediterranean oscillation
impact on precipitation and hydrology in Italy. In Proceedings of the
Conference on Climate and Water, Vol. 1 . Publications of Academy
of Finland, Helsinki, 121– 137.
2013 Royal Meteorological Society Int. J. Climatol. (2013)
A CLIMATOLOGY OF TORNADIC ACTIVITY OVER GREECE
Corte-Real J, Zhang X, Wang X. 1995. Large-scale circulation regimes
and surface climatic anomalies over the Mediterranean. Int. J.
Climatol. 15: 1135– 1150.
Dessens J. 1984. Les Trombes en France. Climatologie et Car-
act´
eristiques Physiques: Lannemezan, France.
Dessens J, Snow JT. 1987. Tornadoes in France. Weather Forecast. 4:
110– 132.
Dotzek N. 2000. Tornadoes in Germany. Atmos. Res. 56: 233– 251.
Dotzek N. 2003. An updated estimate of tornado occurrence in Europe.
Atmos. Res. 67 68: 153161.
Dotzek N. 2009. Derivation of physically motivated wind speed scales.
Atmos. Res. 93: 564 574.
Dunkeloh A, Jacobeit J. 2003. Circulation dynamics of Mediterranean
precipitation variability 1948– 1998. Int. J. Climatol. 23: 1843– 1866.
Etkin D, Brun SE, Shabbar A, Joe P. 2001. Tornado climatology of
Canada revisited: Tornado activity during different phases of ENSO.
Int. J. Climatol. 21: 915 938.
Fiedler BH. 2009. Suction vortices and spiral breakdown in numerical
simulations of tornado-like vortices. Atmos. Sci. Lett. 10: 109 14.
Fiedler BH, Rotunno R. 1986. A theory for the maximum windspeeds
in tornado-like vortices. J. Atmos. Sci. 43: 2328 40.
Forbes GS, Bluestein HB. 2001. Tornadoes, tornadic thunderstorms,
and photogrammetry: a review of the contributions by T.T. Fujita.
Bull. Am. Meteorol. Soc. 82: 73–96.
Fujita TT. 1973. Tornadoes around the world. Weatherwise 26: 56– 83.
Gay`
a M, Homar V, Romero R, Ramis C. 2000. Tornadoes and
waterspouts in the Balearic Islands: phenomena and environment
characterization. Atmos. Res. 56: 253 267.
Gay`
a M, Lasat M, Ar´
us J. 2011. Tornadoes and waterspouts in Catalo-
nia (1950– 2009). Nat. Hazards Earth Syst. Sci. 11(7): 1875 1883.
Giaiotti DB, Giovannoni M, Pucillo A, Stel F. 2007. The climatology
of tornadoes and waterspouts in Italy. Atmos. Res. 83: 534 541.
Gianfreda F, Miglietta MM, Sans`
o P. 2005. Tornadoes in Southern
Apulia (Italy). Nat. Hazards 34(1): 71 89.
Golden JH. 1973. Some statistical aspects of waterspout formation.
Weatherwise 26: 108117.
Golden JH. 1974a. The life cycle of Florida Keys’ waterspouts I.
J. Appl. Meteorol. 13: 676–692.
Golden JH. 1974b. Scale-interaction implications for the waterspout
life cycle II. J. Appl. Meteorol. 13: 693–709.
Goliger AM, Milford RV. 1998. A review of worldwide occurrence of
tornadoes. J. Wind Eng. Ind. Aerodyn. 74 76: 111 121.
Grazulis TP. 1993. Significant Tornadoes, 1680 –1991 . Environmental
Films: St. Johnsbury, Vermont.
Hall SJA. 2011. The Joplin tornado, Missouri 22 May 2011: a report
from a storm chaser’s perspective. Int. J. Meteorol. 36: 161–176.
Holzer AM. 2000. Tornado climatology of Austria. Atmos. Res. 56:
203– 211.
Howells PA, Rotunno R, Smith RK. 1988. A comparative study
of atmospheric and laboratory-analogue numerical tornado-vortex
models. Q. J. Roy. Meteorol. Soc. 114: 801–22.
Hurrell JW, Van Loon H. 1997. Decadal variations in climate associated
with the North Atlantic oscillation. Clim. Change 36: 301 326.
Jones PD, J´
onsson T, Wheeler D. 1997. Extension to the North Atlantic
Oscillation using early instrumental pressure observations from
Gibraltar and South-West Iceland. Int. J. Climatol. 17: 1433– 1450.
Kalnay E, Kanamitsu M, Kistler R, Collins W, Deaven D, Gandin L,
Iredell M, Saha S, White G, Woollen J, Zhu Y, Chelliah M, Ebisuzaki
W, Higgins W, Janowiak J, Mo KC, Ropelewski C, Wang J, Leetmaa
A, Reynolds R, Jenne R, Joseph D. 1996. The NCEP/NCAR 40-year
reanalysis project. Bull. Amer. Meteor. Soc. 77: 437 471.
Kanellopoulou H. 1977. The waterspout of 1st September 1973 in
Corfu (in Greek). Bull. Greek Meteorol. Soc. 5(5): 41– 50.
Kelly DL, Schaefer JT, McNulty RP, Doswell CA, Abbey RF.
1978. An augmented tornado climatology. Mon. Weather Rev. 106:
1172– 1183.
Keltieka F. 1987. Internal report. National Oceanic and Atmospheric
Administration: Cleveland, Ohio.
Keul AG, Sioutas MV, Szilagyi W. 2009. Prognosis of central-eastern
Mediterranean waterspouts. Atmos. Res. 93(1 3): 426436.
King P, Leduc JM, Sills D, Donaldson RN, Hudak RD, Joe P, Murphy
PB. 2003. Lake breezes in southern Ontario and their relation to
tornado climatology. Weather Forecast. 18: 795807.
Kis KA, Straka MJ. 2010. Nocturnal tornado climatology. Weather
Forecast. 25: 545– 561.
Klemp JB. 1987. Dynamics of tornadic thunderstorms. Annu. Rev. Fluid
Mech. 19: 369– 402.
Kounkou R, Mills G, Timbal B. 2009. A reanalysis climatology of
cool-season tornado environments over southern Australia. Int. J.
Climatol. 29: 2079– 2090.
Livadas G. 1954. Fish rain in western Macedonia. Meteorologica No.
2 (Publications of the Meteorological Institute of the University of
Thessaloniki).
Mackenzie. 1858. Notice of the produce of the olive crop in the Island
of Corfu during the past season, in a Letter from Mr Mackenzie,
Corfu, to Dr George Lawson, dated January 11, 1858. Trans. Bot.
Soc. Edinb. 1–4: 42– 43, DOI: 10.1080/03746606009467693.
Maheras P, Xoplaki E, Davies T, Martin-Vide J, Bariendos M,
Elkoforado MJ. 1999. Warm and cold monthly anomalies across
the Mediterranean Basin and their relationship with circulation:
1860– 1990. Int. J. Climatol. 19: 1697 1715.
Malanotte-Rizzoli P, Manca BB, D’Alcal`
a MR, Theocharis A, Berga-
masco A, Bregant D, Budillon G, Civitarese G, Georgopoulos D,
Michelato A, Sansone E, Scarazzato P, Souvermezoglou E. 1997. A
synthesis of the Ionian Sea, hydrography, circulation and water mass
pathways during POEM-Phase I. Prog. Oceanogr. 39(3): 153– 204.
Marcinoniene I. 2003. Tornadoes in Lithuania in the period of
1950– 2002 including analysis of the strongest tornado of 29 may
1981. Atmos. Res. 67 68: 475484.
Matsangouras IT, Nastos PT. 2010. The 27 July 2002 tornado event in
Athens, Greece. Adv. Sci. Res. 4: 9 13.
Matsangouras IT, Nastos PT, Nikolakis D. 2010. Study of meteorolog-
ical conditions related to the tornado activity on 25-3-2009 over
NW Peloponnesus, Greece. In Proceedings of 10th International
Conference on Meteorology, Climatology and Atmospheric Physics
(COMECAP2010). 25– 38 May 2010, Patras, Greece, 417 425 (in
Greek).
Matsangouras IT, Nastos PT, Pytharoulis I. 2011a. Synoptic-mesoscale
analysis and numerical modeling of a tornado event on 12 February
2010 in northern Greece. Adv. Sci. Res. 6: 187 194.
Matsangouras IT, Nastos PT, Sioutas MV. 2011b. 300 Years historical
records of tornadoes, waterspouts and funnel clouds over Greece.
In 6th European Conference on Severe Storms (ECSS 2011), 3–7
October 2011, Palma de Mallorca, Balearic Islands, Spain.
Matsangouras IT, Nastos PT, Pytharoulis I. 2012. Numerical Inves-
tigation of the Role of Topography in Tornado Events in Greece.
In Advances in Meteorology, Climatology and Atmospheric Physics,
Helmis CG, Nastos PT (eds). Springer Berlin: Heidelberg, Germany;
209– 215, DOI: 10.1007/978-3-642-29172-2_30.
Matsangouras IT, Nastos PT, Smith R, Blair D, Dahni R. 2013. An
objective synoptic classification of tornadic days over Greece. In
13th Annual Meeting of the European Meteorological Society (EMS)
and the 11th European Conference on Applied Climatology (ECAC),
Vo l . 29: 9–13 September 2013, EMS2013-514, Reading, UK.
McCarthy D, Schaefer J. 2004a. Tornado trends over the past thirty
years. Bull. Am. Meteorol. Soc. 4879–4883.
McCarthy D, Schaefer J. 2004b. Tornado trends over the past thirty
years. In 14th Conference on Applied Meteorology, Preprints, Seattle,
WA, American Meteorological Society 3.4.
McDonald JR, 2002. Development of an enhanced Fujita scale for
estimating tornado intensity. In Preprints, 21st Conference on Severe
Local Storms. Austin, Boston, American Meteorological Society,
174– 177.
McDonald JR, Mehta KC, Mani S, 2003. F-scale modification process
and proposed revisions. In Preprints, 83rd AMS Annual Meeting:
Symposium on the F-Scale and Severe-Weather Damage Assessment.
Long Beach, American Meteorological Society, Boston, 5 pp.
McDonald JR, Forbes GS, Marshall TP, 2004. The enhanced Fujita
(EF) scale. In Preprints 22nd Conference on Severe Local Storms.
Hyannis, Boston, American Meteorological Society, 7 pp.
Meaden GT. 1976. Tornadoes in Britain: their intensities and distribu-
tion in space and time. J. Meteorol. 1(8): 242– 251.
Moore TW, Dixon RW. 2011. Climatology of tornadoes associated with
gulf coast-landfalling hurricanes. Geogr. Rev. 101(3): 371 395.
Nastos PT, Matsangouras IT. 2010. Tornado activity in Greece within
the 20th Century. Adv. Geosci. 26: 49 51.
Nastos PT, Matsangouras IT. 2012. Composite mean and anomaly of
synoptic conditions for tornadic days over North Ionian Sea (NW
Greece). In Advances in Meteorology, Climatology and Atmospheric
Physics, Helmis CG, Nastos PT (eds). Springer Berlin: Heidelberg,
Germany; 639–645, DOI: 10.1007/978-3-642-29172-2_91.
Nastos PT, Philandras CM, Founda D, Zerefos CS. 2011. Air tempera-
ture trends related to changes in atmospheric circulation in the wider
area of Greece. Int. J. Remote Sens. 32(3): 737 750.
Palutikof JP, Conte M, Casimiro Mendes J, Goodess CM, Espirito SF.
1996. Climate and Climate change. In Mediterranean Desertification
2013 Royal Meteorological Society Int. J. Climatol. (2013)
I.T. MATSANGOURAS et al.
and Land Use, Brandt CJ, Thornes JB (eds). John Wiley and Sons:
London, England.
Paul JF. 1999. An inventory of tornadoes in France. Weather 54:
217– 219.
Paul F. 2000. A developing inventory of tornadoes in France. Atmos.
Res. 56: 269– 280.
Peterson RE. 1978. Waterspout statistics for Nassau, Bahamas. J. Appl.
Meteorol. 17: 444– 448.
Peterson RE. 1982. Tornadic activity in Europe – the last half-century.
In Paper presented at the Conference on Severe Local Storms,
63– 66.
Peterson RE. 1992. Letzmann and Koschmieder’s “Guidelines for
research on funnels, tornadoes, waterspouts and whirlwinds”. Bull.
Am. Meteorol. Soc. 73: 597– 611.
Peterson RE. 1995: Johannes Peter Letzmann: pioneer tornado
researcher. Meteorology in Estonia in Johannes Letzmann’s Times
and Today, H. Eelsalu and H. Tooming. Estonian Academy Publish-
ers: Tallinn, Estonia; 9–43.
Peterson RE. 1998. A historical review of tornadoes in Italy. J. Wind
Eng. Ind. Aerodyn. 74– 76: 123 130.
Politis NG. 1880. Modern Greek mythology. Folk legends. Parnassus
4(10): 762– 773.
Poulos SE, Drakopoulos P, Collins MB. 1997. Seasonal variability in
the oceanographic conditions in the Aegean Sea (eastern Mediter-
ranean): an overview. J. Mar. Syst. 3(1 4): 225 244.
Rauhala J, Brooks EH, Schultz MD. 2012. Tornado climatology of
Finland. Mon. Weather Rev. 140: 1446 1456.
Reynolds DJ. 1999a. A revised U.K. tornado climatology, 1960–1989.
J. Meteorol. 24(242): 290– 321.
Reynolds DJ. 1999b. European tornado climatology 1960–1989.
J. Meteorol. 24: 376– 403.
Rotunno R. 1977. Numerical simulation of a laboratory vortex.
J. Atmos. Sci. 34: 1942 56.
Rotunno R. 1978. A note on the stability of a cylindrical vortex sheet.
J. Fluid Mech. 87: 761 71.
Rotunno R. 1979. A study in tornado-like vortex dynamics. J. Atmos.
Sci. 36: 140– 55.
Rotunno R. 1980. Vorticity dynamics of a convective swirling boundary
layer. J. Fluid Mech. 97: 623 40.
Rotunno R. 1984. An investigation of a three-dimensional asymmetric
vortex. J. Atmos. Sci. 41: 283 98.
Rotunno R. 2013. The fluid dynamics of tornadoes. Annu. Rev. Fluid
Mech. 45: 59– 84.
Salvanos G, Salvanu-Papavlasopulu M. 1930. Folklore material from
Corfu: Folklore collection of Argiradon Corfu, Athens, 97 pp.
Setv´
ak M, ˇ
S´
alek M, Munzar J. 2003. Tornadoes within the Czech
Republic: from early medieval chronicles to the “internet society”.
Atmos. Res. 67 68: 589605.
Simpson JS, Morton BR, McCumber MC, Penc RS. 1986. Observations
and mechanisms of GATE waterspouts. J. Atmos. Sci. 43: 753 782.
Sioutas MV. 2001. Tornado activity in Greece, November 2000.
J. Meteorol. 26(256): 69– 70.
Sioutas MV. 2002. Damaging tornadoes in Greece: July and September
2001. J. Meteorol. 27(265): 17– 22.
Sioutas MV. 2003. Tornadoes and waterspouts in Greece. Atmos. Res.
67– 68: 645 656.
Sioutas MV. 2006: Tornado and waterspout characteristics in Greece.
In Proceedings of 8th Pan-Hellenic conference on Meteorology,
Climatology and Atmospheric Physics. Athens, Greece, 23 26 May
2006, 121– 128 (in Greek).
Sioutas MV. 2011. A tornado and waterspout climatology for Greece.
Atmos. Res. 100(4): 344 356.
Sioutas MV, Keul AG. 2007. Waterspouts of the Adriatic, Ionian and
Aegean Sea and their meteorological environment. Atmos. Res. 83:
542– 557.
Sioutas MV, Doe R, Michaelides S, Christodoulou M, Robins R. 2006.
Meteorological conditions contributing to the development of severe
tornadoes in southern Cyprus. Weather 61(1): 1016.
Skliris N, Sofianos S, Gkanasos A, Axaopoulos P, Mantziafou A,
Vervatis V. 2011. Long-term sea surface temperature variability in
the Aegean Sea. Adv. Oceanogr. Limnol. 2(2): 125 139.
Snow JT, Wyatt AL. 1997. The tornado, Nature’s most violent wind:
part 1 – world-wide occurrence and categorization. Weather 52:
298– 304.
Snow JT, Wyatt AL. 1998. The tornado, nature’s most violent wind:
part 2 – formation and current research. Weather 53: 6672.
Szilagyi W. 2009. A waterspout forecasting technique. In 5th European
Conference on Severe Storms, Landshut, Germany, 12–16 October
2009, ECSS-129.
Tarand A. 1995. How often do tornadoes occur in Estonia?. In
Meteorology in Estonia in Johannes Letzmann’s Times and Today,
Eelsalu H, Tooming H (eds). Estonian Academy Publishers: Tallinn,
Estonia; 132– 138.
Telelis I. 2004a. Meteorological phenomena and climate in Byzantium.
Acad. Athens 1: 1 494 (in Greek).
Telelis I. 2004b. Meteorological phenomena and climate in Byzantium.
Acad. Athens (2): 495 923 (in Greek).
Telelis I. 2005. Historical-climatological information from the time of
the Byzantine Empire (4th–15th centuries AD). Hist. Meteorol. 2:
41– 50.
Tomkinson J. 2006. Travellers’ Greece, Memories of an Enchanted
Land. Anagnosis Books: Athens, Greece; 1–604.
Tooming HK, Peterson RE. 1995. Vigorous tornadoes and waterspouts
during the last 35 years in Estonia. In Meteorology in Estonia in
Johannes Letzmann’s Times and Today, Eelsalu H, Tooming H (eds).
Estonian Academy Publishers: Tallinn, Estonia; 168–179.
Tyrrell J. 2000. Tornadoes in Ireland: an historical and empirical
approach. Atmos. Res. 56: 281 290.
Tyrrell J. 2003. A tornado climatology for Ireland. Atmos. Res. 67– 68:
671– 684.
Verbout SM, Brooks HE, Leslie LM, Schultz DM. 2006. Evolution
of the U. S. tornado database: 1954–2003. Weather Forecast. 21:
86– 93.
Walker SA. 2007. Spatial and temporal analysis of tornado fatalities in
the United States: 1880–2005. Weather Forecast. 22: 1214– 1228.
Wegener A. 1917. Wind-und Wasserhosen in Europa. Vieweg: Braun-
schweig.
Wilson T, Rotunno R. 1986. Numerical simulation of a laminar end-
wall vortex and boundary layer. Phys. Fluids 29: 3993 4005.
WSEC. 2004. A Recommendation for an Enhanced Fujita Scale
(EF-scale). Wind Science and Engineering Center, Texas Tech
University, Lubbock, Texas, 95 pp.
2013 Royal Meteorological Society Int. J. Climatol. (2013)
... Taszarek et al. (2018) used the ERA-Interim reanalysis from ECMWF, comparing these data with over 1 million sounding measurements in the European region, whereas Ingrosso et al. (2020) applied both ERA-Interim and ERA-5, the recently issued higher resolution reanalysis from ECMWF (Hersbach et al., 2020), to the Italian area. Renko et al. (2013) and Matsangouras et al. (2014) characterised the synoptic environment where tornadoes develop by means of ERA-INTERIM for the Eastern Adriatic basin and NCEP-NCAR reanalysis for Greece, respectively. ERA-5 was also recently used by Rodríguez and Bech (2020) in their analysis of tornadic environments over the Iberian peninsula. ...
... In the present article, we investigate synoptic patterns associated with tornadoes in different Italian regions. The literature about southern Europe tornadoes focused on the synoptic-scale conditions associated with the tornado events is very limited (Sioutas and Keul, 2007;Keul et al., 2009;Renko et al., 2013;Matsangouras et al., 2014). Sioutas and Keul (2007) detected synoptic patterns favourable to waterspouts development in the Adriatic, Ionian and Aegean Sea. ...
... Results consistent with Sioutas and Keul (2007) were shown by Renko et al. (2013) for the Eastern Adriatic basin. Matsangouras et al. (2014) analysed three different Greek sub-regions in autumn, when tornadoes and waterspouts are most frequent. They found a typical synoptic configuration for tornadoes over west Greece, i.e., a trough over southern and central Italy at 500 hPa accompanied by a closed cyclonic circulation over the Gulf of Taranto, while a trough over the eastern Aegean Sea is the synoptic pattern for waterspouts North of Crete. ...
Article
An analysis of synoptic patterns and mesoscale precursors is produced for Italian tornadoes for the period 2000–2018. Anomaly maps of different parameters are extracted from ERA-5 reanalysis. To highlight typical large-scale configurations, a clustering analysis is applied to define different regional clusters, representative of areas affected by a large number of tornadoes. The analysis shows that: a)Significant anomalies of synoptic variables and mesoscale precursors are generally present over and nearby the region where tornadoes occur, in particular for significant tornadoes (EF2+). b) Differences among clusters suggest that the environmental conditions favourable to tornado development differ between southern and northern Italian regions. In the Ionian region and Sicily, tornadoes occur when the northern end of a warm/moist tongue is advected toward an area of strong low-level shear. In the northern regions, high-shear cold air intrusions from continental Europe, combined with the high humidity content due to intense evaporation from the Mediterranean during summer, favour supercell triggering. c) tornadoes in southern regions are characterised by the highest anomaly values in wind shear, storm relative helicity and CAPE; d) significant SST positive anomalies are observed for southern tornadoes.
... The main goal of this article is to present a homogeneous and robust dataset of tornado and waterspout reports registered from 2000 to 2019 in Catalonia, analysing their main spatial and temporal features. This 20-year period is comparable to other recent tornado climatologies carried out in Europe (Matsangouras et al., 2014b;Miglietta and Matsangouras, 2018;Sioutas and Doe, 2019;Leitão and Pinto, 2020) and gathers results from technological advances on severe weather observations, decreasing the number of unreported events and providing a comprehensive updated analysis of one of the European tornado hotspots. ...
... Similarly, the density of waterspouts normalized per year and 100 km of coastal line in Catalonia is 5.0, which is higher than in Croatia (3.0, Renko et al., 2016), in Greece (2.1, Matsangouras et al., 2014b), in Germany (2.1, Kühne et al., 2017), in Italy (0.9, Miglietta and Matsangouras, 2018) and in Portugal (0.2 in mainland, Leitão and Pinto, 2020). Nevertheless, if only near-coastal waterspouts are considered (i.e., located less than 10 km offshore), the density decreases to 2.7, which is similar to some of the previous-mentioned countries. ...
... This result is similar to that previously found by Gayà et al. (2011), although the maximum has been shifted towards autumn. The monthly distribution presented here is similar to other Mediterranean countries such as Greece (Matsangouras et al., 2014b) or Italy (Miglietta and Matsangouras, 2018), but differs from central and northern Europe, where tornadoes are usually observed between June and August (Groenemeijer and Kühne, 2014;Antonescu et al., 2016). Differences between central and northern Europe and the region of study can be explained by the influence of the Mediterranean Sea shifting the maximum convective activity towards autumn and the more frequent presence of the jet stream later in the season compared to higher latitudes. ...
Article
Tornado climatologies are essential to enhance our knowledge about their frequency and spatial distribution, together with their damage path characteristics. Due to their small spatiotemporal scale, tornado detection is strongly linked to visual observations. Therefore, social networks have become one of the most important sources of severe weather reports during the present century, including tornadoes, improving their detection and increasing the number of observed cases compared to previous decades. This article presents an analysis of tornado and waterspout events reported between 2000 and 2019 in Catalonia (NE Iberian Peninsula), one of the southern European regions where these phenomena are most frequent. The study includes 105 tornadoes and 329 waterspouts reported in a 32,000 km2 area, and therefore, it presents one of the highest tornado and waterspout densities in the Mediterranean basin according to recent climatologies. Remote-sensing tools such as weather radar, lightning detection and satellite imagery have been used to validate each event. Moreover, fieldwork performed after several tornado cases and the use of high-resolution aerial imagery have provided information to characterise damage paths, including the length and width of the damage swath and the maximum intensity according to the Enhanced Fujita Scale. Finally, a list of the analysed events is provided for further research.
... Several publications during the last several decades have presented historical records concerning tornadic activity (e.g., Meaden, 1976;Tooming and Peterson, 1995;Peterson, 1998;Reynolds, 1999;Tyrrell, 2003;Macrinoniene, 2003;Dotzek, 2003;Nastos and Matsangouras, 2010;Gayà et al., 2000;Brázdil et al., 2012;Rauhala et al., 2012;Haghroosta et al., 2014). As far as tornadic activity over Greece is concerned, a comprehensive spatial distribution of a total of 612 events (171 tornadoes, 374 waterspouts and 67 funnel clouds) recorded in 405 d was presented (Matsangouras et al., 2014a, b: Nastos and as there were several days with multiplied events within the period 1709-2012 (Fig. 5). This study gives evidence that even in an eastern Mediterranean region these extreme atmospheric phenomena are abundant, causing catastrophic impacts on infrastructures and in many cases loss of life. ...
... Spatial variability of tornadoes, waterspouts and funnel clouds over Greece for the period 1709-2012. Numerous tornadic events lie under the tornadic type symbols due to low resolution of the image (adapted fromMatsangouras et al., 2014a). ...
Article
Full-text available
Risk assessment constitutes the first part within the risk management framework and involves evaluating the importance of a risk, either quantitatively or qualitatively. Risk assessment consists of three steps, namely risk identification, risk estimation and risk evaluation. Nevertheless, the risk management framework also includes a fourth step, i.e., the need for feedback on all the risk assessment undertakings. However, there is a lack of such feedback, which constitutes a serious deficiency in the reduction of environmental hazards at the present time. Risk identification of local or regional hazards involves hazard quantification, event monitoring including early warning systems and statistical inference. Risk identification also involves the development of a database where historical hazard information and hazard effects are included. Similarly, risk estimation involves magnitude–frequency relationships and hazard economic costs. Furthermore, risk evaluation consists of the social consequences of the derived risk and involves cost-benefit analysis and community policy. The objective of this review paper is twofold. On the one hand, it is to address meteorological hazards and extremes within the risk management framework. Analysis results and case studies over Mediterranean ecosystems with emphasis on the wider area of Greece, in the eastern Mediterranean, are presented for each of the three steps of risk assessment for several environmental hazards. The results indicate that the risk management framework constitutes an integrated approach for environmental planning and decision-making. On the other hand, it sheds light on advances and current trends in the considered meteorological and environmental hazards and extreme events, such as tornadoes, waterspouts, hailstorms, heat waves, droughts, floods, heavy convective precipitation, landslides and wildfires, using recorded datasets, model simulations and innovative methodologies.
... The studied areas of the Ionian and Aegean Seas are characterized by irregular coastlines combined with numerous small and larger islands. These geographic configurations can often create low-level differential heating and local wind convergence, which tends to initiate waterspout occurrences [10,13,16,17]. In this study, the twelve greatest waterspout outbreak cases occurred in the Greek Seas over the last two decades (2000-May 2023) with 10 or more individual waterspouts observed are investigated and their synoptic and mesoscale environments and a forecasting method are presented. ...
... At surface level, south and southwesterly winds are usually present; these are promoted by a surface low located in southern Italy. These findings are consistent with those of Matsangouras et al. (2014). Moreover, Kahraman (2021) focused on characterising weather configurations that trigger mesocyclonic tornadoes in Turkey. ...
Article
The present article analyses the synoptic framework associated with tornadic events on the Iberian Peninsula and the Balearic Islands. We discussed the spatial and temporal distribution of tornadoes in the study area, subsequently performing a principal component analysis using 500 hPa geopotential height and mean sea level pressure data from ERA5 reanalysis to classify into 12 synoptic patterns the 465 tornadic events (comprising 608 individual tornadoes) reported between 1950 and 2021. Furthermore, we employed 900 hPa specific humidity and temperature, 300 hPa and 900 hPa zonal and meridional wind, CAPE, medium-layer shear (MLS) and the product of MLS and the square root of two times CAPE (WMAXSHEAR03) variables to create the mean composites related to each weather type. We also analysed the influence of several teleconnection patterns on the occurrence of tornadoes. Results showed a wide range of configurations prone to tornadogenesis, although these are usually characterised by a surface low and/or a trough whose axis is oriented north-south close to the Iberian Peninsula. Tornadoes are typically related to warm, moist low-level advections, i.e. maritime 900 hPa wind together with 900 hPa specific humidity relative maxima. The variable WMAXSHEAR03 presents a pattern similar to the tornado distribution for each weather type throughout the study area, whereas CAPE and MLS separately do not (Mediterranean events are mainly associated with high CAPE and moderate MLS environments, whereas Atlantic tornadoes tend to be promoted in small CAPE and strong MLS conditions). Tornadoes in western Iberia are usually formed under the negative phase of the NAO and the AO, whereas in the Mediterranean area they are typically associated with the positive phase of the WeMO and the MO, which contrasts with torrential rainfall events in this subregion.
... Regarding the climate of Athens, according to the Koppen climate classification (2018) [7], it is defined as Hot Summer Mediterranean (Csa), which means mild and rainy winter and dry summer periods (see appendix 1). Nevertheless, the urbanization of Athens city resulted in enhanced urban heat island [8] and increasing rain intensity, as well as the appearance of extreme weather phenomena ( [9], [10], [11]). Mechanical turbulence from increased surface roughness, sensible heat from the urban warm air and the anthropogenic condensation nuclei floating in the urban air are the synergistic factors, which cause urban induced changes in precipitation [12]. ...
Article
Full-text available
Climate change is an ongoing fact with evident impacts on earth, humanity and our heritage. The management and protection of heritage from the effects of climate change should be taken into consideration, so as to act dynamically and immediately, since the effects of climate change are rapidly evident in all aspects of life. The objective of this research is to study the vulnerability of the ancient Greek Theatre of Dionysus, as it forms a monument with embedded heritage values exposed to the climate change. This case study is the tangible result of a particular cultural and historic research, bearing historic knowledge, cultural meanings via a recognizable architectural structure and it reflects the conjunction of culture with society and nature. Materiality and its pathology combined with the condition of the natural landscape and the altering pattern of cultural tourism can lead to the study of climate change imprint on this kind of heritage, which should be examined as a wholeness of culture and nature. The uniqueness in the Theatre’s identity and historic path deserves interpretation of the effects of climate change, so as to manage adaptation, proactive planning, mitigation and dissemination of the discovered results.
... These data gave the opportunity to elaborate pan-European climatologies of tornadoes and of other severe weather events [18,19] and to identify the mesoscale conditions favorable to their development [20][21][22]. ESWD also represented the basis for the development of tornado climatology in Mediterranean countries [23,24], including Italy, where Miglietta and Matsangouras (2018) [25] (MM18 hereafter) developed a new climatology for tornadoes and waterspouts covering the 10-year period 2007-2016. ...
Article
Full-text available
Characteristics of extratropical cyclones that cause tornadoes in Italy are investigated. Tornadoes between 2007 and 2016 are analyzed, and statistical analysis of the associated cyclone structures and environments is performed using the JRA-55 reanalysis. Tornadoes are distributed sporadically around the cyclone location within a window of 10° × 10°. The difference in the cyclone tracks partially explains the seasonal variability in the distribution of tornadoes. The highest number of tornadoes occur south of the cyclone centers, mainly in the warm sector, while a few are observed along the cold front. Composite mesoscale parameters are examined to identify the environmental conditions associated with tornadoes in different seasons. Potential instability is favorable to tornado development in autumn. The highest convective available potential energy (CAPE) in this season is associated with relatively high-temperature and humidity at low-levels, mainly due to the strong evaporation over the warm Mediterranean Sea. Upper-level potential vorticity (PV) anomalies and the associated cold air reduce the static stability above the cyclone center, mainly in spring and winter. On average, the values of CAPE are lower than for US tornadoes and comparable with those occurring in Japan, while storm relative helicity (SREH) is comparable with US tornadoes and higher than Japanese tornadoes, indicating that the environmental conditions for Italian tornadoes have peculiar characteristics. Overall, the conditions emerging in this study are close to the high-shear, low-CAPE environments typical of cool-season tornadoes in the Southeastern US.
... Several publications during the last several decades have presented historical records concerning tornadic activity (e.g., Meaden, 1976;Tomming et al., 1995;Peterson, 1998;Reynolds, 1999;Tyrrel, 2003;Macrinoniene, 2003;Dotzek, 2003;Nastos and Matsangouras, 2010;280 Gayà et al., 2000;Brázdil et al., 2012;Rahuala et al., 2012;Haghroosta, et al., 2014). As far as tornadic activity over Greece is concerned, a comprehensive spatial distribution of a total of 612 events (171 tornadoes, 374 waterspouts and 67 funnel clouds), recorded on 405 days was presented (Matsangouras et al., 2014a;, as there were several days with multiplied events, within the period 1709-2012 ( Figure 5). This study give evidence that even in an https://doi.org/10.5194/nhess-2020-155 ...
Preprint
Full-text available
Risk assessment constitutes the first part within the risk management framework and involves evaluating the importance of a risk, either quantitatively, or qualitatively. Risk assessment consists of three steps, namely risk identification, risk estimation and risk evaluation. Nevertheless, the risk management framework also includes a fourth step, i.e. the need for a feedback of all the risk assessment undertakings. However, there is a lack of such feedback, which constitutes a serious deficiency in the reduction of environmental hazards at the present time. Risk identification of local or regional hazards involves hazard quantification, event monitoring including early warning systems and statistical inference. Risk identification also involves the development of a database, where historical hazard information and their effects is included. Similarly, risk estimation involves magnitude/frequency relationships and hazard economic costs. Furthermore, risk evaluation consists of the social consequences of the derived risk and involves cost-benefit analysis and community policy. The objective of this review paper is twofold: On one hand, to address meteorological hazards and extremes within the risk management framework. Analyses results and case studies over Mediterranean ecosystems with emphasis on the wider area of Greece, in the eastern Mediterranean, are presented for each of the three steps of risk assessment for several environmental hazards. The results indicate that the risk management framework constitutes an integrated approach for environmental planning and decision making. On the other hand, it shed light to advances and current trends in the considered meteorological and environmental hazards and extreme events, such as tornadoes, waterspouts, hailstorms, heat waves, droughts, floods, heavy convective precipitation, landslides and wildfires, using recorded datasets, model simulations and innovative methodologies.
Article
Full-text available
Severe convective windstorms and tornadoes regularly hit the territory of Russia causing substantial damage and fatalities. An analysis of the climatology and formation environments of these events is essential for risk assessments, forecast improvements and identifying of links with the observed climate change. In this paper, we present an analysis of severe convective windstorms, i.e., squalls and tornadoes reported between 1984 and 2020 in the Perm region (northeast of European Russia), where a local maximum in the frequency of such events was previously found. The analysed database consists of 165 events and includes 100 squalls (convective windstorms), 59 tornadoes, and six cases with both tornadoes and squalls. We used various information to compile the database including weather station reports, damage surveys, media reports, previously presented databases, and satellite images for windthrow. We found that the satellite images of damaged forests are the main data source on tornadoes, but their role is substantially lower for windstorm events due to the larger spatial and temporal scale of such events. Synoptic-scale environments and associated values of convective indices were determined for each event with a known date and time. Similarities and differences for the formation conditions of tornadoes and windstorms were revealed. Both squalls and tornadoes occur mostly on rapidly moving cold fronts or on waving quasi-stationary fronts, associated with low-pressure systems. Analyses of 72-h air parcel backward trajectories shows that the Caspian and Aral Seas are important sources of near-surface moisture for the formation of both squalls and tornadoes. Most of these events are formed within high CAPE and high shear environments, but tornadic storms are generally characterised by a higher wind shear and helicity. We also differentiated convective storms that caused forest damage and those did not. We found the composite parameter WMAXSHEAR is the best discriminator between these two groups. In general, storm events causing windthrow mainly occur under conditions more favourable for deep well-organised convection. Thus, forest damage can be considered as an indicator of the storm severity in the Perm region and in adjacent regions with forest-covered area exceeding 50%.
Article
Full-text available
In the months of November 2012, 2014 and 2018, four tornadic events affected southeastern Italy (Calabria and Apulia regions). Tornado-spawning supercells hit the same regions, had similar trajectories, and were characterized by common synoptic conditions. The events are analysed through the use of multiple observational tools; large-scale and mesoscale analysis are performed, in addition to high-resolution numerical simulations and sensitivity tests with the WRF model. In particular, the effect of changes in Sea Surface Temperature on the intensity and tracking of tornadic supercells in the Mediterranean is investigated, finding that small variations for SST may cause significant changes on instability parameters, consequently on the supercell intensity. Mesoscale-model simulations show that the structure and the track of the convective cells can be correctly reproduced, as well as the triggering of convection due to interaction of the flows with the topographic features such as the land-sea interface and orography, within a relatively small temporal delay. The mesoscale analysis also reveals, in all cases, an environment highly favourable for tornado formation, with most simulated instability parameters/indices exceeding their critical thresholds.
Conference Paper
Full-text available
Although the Fujita Scale has been in use for 30 years, the limitations of the scale are well known to its users. The primary limitations include a lack of damage indicators, no account of construction quality and variability, and no definitive correlation between damage and wind speed. These limitations have led to inconsistent ratings of tornado damage and, in some cases, overestimates of tornado wind speeds. Thus, there is a need to revisit the concept of the Fujita Scale and to improve and eliminate some of the limitations. Recognizing the need to address these limitations, Texas Tech University (TTU) Wind Science and Engineering (WISE) Center personnel proposed a project to examine the limitations, revise or enhance the Fujita Scale, and attempt to gain a consensus from the meteorological and engineering communities. A steering committee first was organized to initiate the project. The next step involved assembling a forum of users to identify the issues and develop strategies to improve the Fujita scale. A panel of wind damage experts met and assigned failure wind speed values to various degrees of damage (DOD) to buildings and other objects. Through this expert elicitation process, wind speeds corresponding to the DOD's were estimated. These estimated wind speeds then determined the EF (Enhanced Fujita)-scale category appropriate for the observed damage. This paper discusses the work to date in finalizing the EF-scale.
Article
A class of new wind speed scales is proposed in which the relevant scaling factors are derived from physical quantities like mass flux density, energy density (pressure), or energy flux density. Hence, they are called Energy- or E-scales, and can be applied to wind speeds of any intensity. It is shown that the Mach scale is a special case of an E-scale. Aside from its foundation in physical quantities which allow for a calibration of the scales, the E-scale concept can help to overcome the present plethora of scales for winds in the range from gale to hurricane intensity. A procedure to convert existing data based on the Fujita-scale or other scales (Saffir-Simpson, TORRO, Beaufort) to their corresponding E-scales is outlined. Even for the large US tornado record, the workload of conversion in case of an adoption of the E-scale would in principle remain manageable (if the necessary metadata to do so were available), as primarily the F5 events would have to be re-rated. Compared to damage scales like the �Enhanced Fujita� or EF-scale concept recently implemented in the USA, the E-scales are based on first principles. They can consistently be applied all over the world for the purpose of climatological homogeneity. To account for international variations in building characteristics, one should not adapt wind speed scale thresholds to certain national building characteristics. Instead, one worldwide applicable wind speed scale based on physical principles should rather be complemented by nationally-adapted damage descriptions. The E-scale concept can provide the basis for such a standardised wind speed scale.
Article
A summary of the four tornado events which occurred in Greece during July and September 2001 is given, including their fundamental characteristics and synoptic meteorological background. The tornadoes were classified from strong to severe in the categories T3 and T4 on the International TORRO Intensity Scale. No tornadoes are known for August. Two occurred during July in northern Greece, and on one occasion (11 July) a site investigation of the damage was made. Some weather radar information and hail-pad measurements are included. There was a small tornado outbreak on 5 September when two tornadoes were reported from north-eastern and south-western Greece.
Article
It has been nearly 15 years since the last tornado climatology of the United Kingdom was produced and, with new events and corrected or otherwise updated information to hand, there has been a need for such a revision for some time. This paper presents some of the results of a larger research project undertaken by the writer, which fulfil the requirement for an accurate and up-to-date British tornado climatology; consequently, this climatology supersedes all those on British tornadoes which have gone before it. The various findings are also occasionally contrasted with earlier climatological studies from both the United Kingdom and the United States. The research was undertaken at the University College of Swansea throughout the 1990s and is based primarily on the 30 year period of 1960 - 1989. It is now intended that a comprehensive tornado climatology for the United Kingdom will be produced by TORRO every decade, some five years after the relevant 30 year period has elapsed. In addition for the time required to undertake such an analysis, this also allows for late data to be included in the climatology. This first decadal climatology has taken a longer period of time to complete due to the lengthy and time-consuming process of computerising the tornado database, followed by a thorough validation and verification of the information. Now those major tasks have been completed, it is a far more straight forward and quicker task to maintain the database.
Article
Although there have been earlier attempts at collating reports of European tornadoes and waterspouts over differing time periods and with varying degrees of success, the following tornado and waterspout climatology is believed to be the first comprehensive investigation at the European continental level. Previous studies have been at national or regional levels only. Whereas TORRO was originally primarily interested in British whirlwinds, reports of any tornadoes elsewhere in Europe were also maintained in a file containing solely European reports, and this was officially established when TORRO was founded in October 1974 (Meaden, 1985). Although even in 1985 TORRO was not actively seeking European reports, the foundations of a European tornado database originated from that year when Meaden (1985) proposed that 'TORRO could in future act as a clearing house for all known European tornado events'. This paper is based on research - undertaken at the University College Swansea throughout the 1990s - which helps to fulfil that proposal, as well as extending it to include waterspouts. It has now been some 10 years since TORRO started to pursue European reports actively, a process which has accelerated in the past few years with the on-going appointment of European Representatives in each nation. Such people act as points of contact who are in a better position to collate and forward reports from their own nations. Already the expansion of TORRO from a British organisation into a continental-wide one is bearing fruit, with the last few years seeing a rapid increase in the number of reports reaching the organisation. The more recent events lie outside the period of under study here but, as with the United Kingdom tornado climatology, TORRO will update the European version at regular intervals. However, that the United Kingdom has vastly more reported tornadoes than the rest of Europe put together clearly indicates that there is much more work to be done - this also hints at what TORRO believes to be an exceedingly tornadic continent. Furthermore, tornadoes have not even been reported to TORRO from several nations. So while many aspects of this climatology are based on a relatively small data-set, some useful conclusions can be drawn. At the very least, the existence of this work alone will serve to bring more European tornado (and indeed other severe weather) reports to TORRO.