ArticlePDF Available

The forest fires of 1995 and 1998 on Penteli mountain

The forest fires of 1995 and 1998 on Penteli mountain
Dr. Gavriil Xanthopoulos
National Agricultural Research Foundation
Institute of Mediterranean Forest Ecosystems and Forest Products Technology
Terma Alkmanos, Ilissia, 115 28 Athens - Greece
Tel.: +3-010-7793142, Fax: +3-010-7784602, E-mail:
The 1995 fire on Penteli mountain
1995 was a better than average year, with respect to forest fires, for Greece. The area
burned was 27,203 hectares which is approximately half the area burned for each of the
three previous years and much less than later fire seasons (e.g. 95.571 ha in 1998, 167.000
ha in 2000). However, in spite of the good overall results, one large fire at the mountain of
Penteli, in Attica, a distance of a few kilometres NE of Athens, burned about 6,200 hectares
in three days (21-24 July 1995) and created, both nationally and internationally, the
impression of a catastrophic fire season (Xanthopoulos 1996). A short description of the
evolution of this fire follows.
The start of the fire
The fire started around 07:50 on the Friday, 21
of July 1995, in a thick Aleppo Pine (Pinus
halepensis) forest, under an unusually strong northerly “meltemi” wind with gusts reaching
75 km/hr. Meltemi wind is a regular phenomenon in the summer in the eastern part of
Greece, affecting the Aegean sea and its islands, but they rarely reach such intensity. This
wind had started blowing on the previous day and continued at this strength throughout the
first two days of the fire. An extreme fire danger warning had been issued on July 19
the next four days. Three other fires, at short distances from Penteli mountain, that erupted
in the sixteen hours preceding the Penteli fire (a fire at Vouliagmeni lake at the SE outskirts
of Athens on Thursday evening, a fire in the area of Kato Souli near Marathon village that
burned through the night, and a fire from a garbage dumb in the area of Keratea in the
South part of Attica) were controlled successfully in spite of the adverse conditions.
However, they drained resources from Penteli for their suppression and mop-up. These
efforts were still in progress at the time the Penteli fire started.
The fire started next to the road that connects the area of Nea Makri with the area of
Dionissos at a location called “Agios Petros” (Figure 1). There is quite heavy traffic on the
road at that time, as people drive from their summer (vacation) homes to work in Athens.
Thus the fire was reported immediately, through mobile phone, by drivers who stopped and
tried unsuccessfully to extinguish it.
Figure 1. General map (1:50.000) of the eastern slopes of Penteli mountain where the fires
took place. The extent of the urban-wildland interface areas is quite obvious.
Fire behavior
The fire accelerated quickly and in a very short time developed into a crown fire exhibiting
extreme fire behavior. This is very common in Aleppo pine forests under such conditions as
there is usually heavy understory and the trees are not very tall, leaving little gap between
the two vegetation strata. The smoke plume was so thick that it shadowed Athens and
reached well beyond Crete to the south (Figure 2).
Relative humidity was relatively high (>35%) and temperature was around 25
C in the
beginning. In spite of that, this wind-driven massive fire produced many firebrands and,
according to witnesses, there was some, but not excessive, medium range spotting later in
the day. The average observed rate of spread on the first day reached 1,7 km/hour for
lengthy periods of time, especially around noon. It was relatively slower while passing
through the area of Drafi (about 1,2 km).
Figure 2. The smoke plume from the Penteli fire reaching well beyond Crete, and making
the direction of the wind quite obvious. This NOAA-AVHRR Channel 1,
satellite image was provided by Dr. Rosa Lasaponara.
Response to the fire
Two fire trucks, returning from the previous fires, were the closest ones and they were the
first ground resources dispatched to the fire. They were there in about 20 minutes but they
were unable to control the fire which, by that time, was spreading as a crown fire. The
aerial means (amphibian Canadair CL-215 airtankers) that were already operating on two of
the other fires in the area were diverted immediately to this fire but were unable to perform
efficiently due to extreme wind-caused turbulence over the mountainous terrain. In less
than an hour it was clear that the fire was going to be a real disaster as it headed towards an
extensive urban-wildland interface area (Drafi) where individual homes were close to or
even within the thick Aleppo pine forest. It passed through that general area between 10:30
and 11:30 and continued spreading southwards towards Pallini.
Figure 3. The outline of the 1995 fire on Penteli on July 21
at 13:30, and the simulated fire
spread by a fire management information system (FMIS) developed by
ALGOSYSTEMS S.A. (Varela et al. 1999).
In the meantime, the coordinating center kept sending to Penteli all available ground and air
resources. There were some futile efforts to control the fire front, as the aerial means (both
CL-215 and CHINOOK CH-47D helicopters with 5 m
bambi-buckets) were not able to
perform effectively under such conditions (wind, turbulence, heavy smoke). The ground
forces were pushed away by the advancing front when they tried to make a stand on wide
roads or relatively open areas. It soon became evident that it was not possible to stop the
fire front and efforts by the fire trucks and crews of the Forest Service and the Fire Service
concentrated on protection of homes, monasteries, other buildings, and of course people
who were in the path of the fire.
The Firefighting Coordinating Center soon started bringing in resources from other parts of
the country. In the next two days firetrucks arrived from as far as Thessaloniki in Northern
Greece and Kalamata in the south end of Peloponnese, bringing the total number to
approximately 100. However, coordination was quite poor as this responsibility sifted from
operational people to ministers, and with the priority to save lives and homes the general
picture of the fire, expected fire behavior, etc. was lost. There was no concerted effort to
work on the flanks and the heel of the fire.
Evolution of the fire
The fire front reached the area of Pallini and the main road connecting Athens with Rafina
(Marathon avenue) around 13:30. The area of Pikermi along the same road was reached
later in the afternoon. There the vegetation continuity is broken by agricultural fields with
various cultivations, including vines, and there are many vegetation free fields where near
industrial installations homes etc. Hence, it was possible to control the fire along that road
and in the agricultural fields around it. The main fire front was controlled by that evening.
More than 2000 ha had burned by that time.
During that night the wind did not loose strength. As efforts concentrated on saving homes
and lives where the fire had already passed and on the flanks, the heel of the fire kept
widening. New “waves” originated there and aided by the effect of the topography on the
wind, through the night, reached villages (Anthousa, Penteli) on the west flank creating
more need for efforts to save lives and homes.
On Saturday, with the wind continuing at full force, the areas of Kallitechnoupoli and later
Neos Voutzas, on the eastern flank of the fire (that in the meantime kept widening) faced
destruction. The final strong firefighting effort was made on Sunday morning in the area
close to Dionissos (location called “German cemetery”) at the heel of the fire. Mop-up
efforts continued through Sunday and Monday. The fire was considered fully controlled on
Monday 24
of July. The runs at the flanks during Saturday and Sunday increased the
burned area to its final size.
The total burned area reached 6200 ha (figure 4). About 105 buildings of various types
were heavily damaged or fully destroyed. Some of them were high quality houses built with
reinforced concrete frame, clay-tile roofs etc. whereas most of them out-houses, mobile
homes, small temporary buildings, farm-barns etc. made of flammable materials.
Fortunately, and in spite of the adverse conditions, there was no loss of life.
Figure 4. The final burned area of the 1995 fire. Image Source: “Land Use Change
Interactions with Fire in Mediterranean Landscapes” (LUCIFER) project.
Firefighting resources
Ground forces
About 100 fire trucks of the Forest Service and the Fire Service were dispatched to
this fire. Some of them were sent from areas at distances exceeding 500 km from
Approximately forty tanker-trucks belonging to the municipalities were used for the
supply of the firefighting trucks with water.
More than 1150 firefighters and 1400 soldiers were also involved. They were
supported by a large number of volunteers.
Aerial firefighting means
Nearly all the heavy aerial firefighting means of the country were involved in the
suppression efforts.
Nine CL-215 amphibian water bombers, one MAFFS equipped C-130 air tanker
and two army CHINOOK CH-47D helicopters were used for class A foam, fire
retardant and water drops respectively, while two army BELL UH-1H "Huey"
helicopters were used for reconnaissance and coordination.
This fleet was augmented, after the first day, by two Canadair CL-415 water
bombers that were sent by the Italian government and one FOKKER-27 airtanker
plus one BE-20 lead-plane sent by the French government
The 1998 fire on Penteli mountain
1998 was a very special year for forest firefighting in Greece. In May 1998, a law
transferred the responsibility of forest firefighting from the Forest Service to the Fire
Service (Xanthopoulos 1999, Eftichidis and Varela 2000). The later historically had been
involved in forest firefighting but on a secondary role, mainly firefighting from roads,
protecting housing areas in the Wildland urban interface etc. It lacked appropriate
equipment (clothing, long and small-diameter hoses, etc.) and most important it did not
have the previous experience of full responsibility for coordinating a large forest fire. Given
the late transfer of the responsibility, just before the start of the fire season, it did not have
time to work on these deficiencies and to correctly incorporate the material it received
from the Forest Service (e.g. service the fire trucks it received, work on the forest maps,
etc.). Given all these, and the development of a hostile environment between the personnel
of the two Services, the scene was set for disaster. The problems started in the beginning of
July and kept worsening as the fire season turned out to be windy combined with very low
relative humidity. In August, the fire on Penteli mountain became one of the worse fire
disasters of that year that added 7.500 ha to the burned area of that year, bringing the total
up to the final figure of 95.571 ha.
The start of the fire
The fire started at 22:00 on 2 August near the village of Stamata northeast of Athens and
north of the area of Dionissos under relatively calm conditions (light NE wind at less than 2
BF, air temperature at 27
C, and relative humidity around 30% at that time).
Unfortunately, as it was not close to a road and access to it was difficult, it was not attacked
aggressively and effectively through the night, in spite a meteorological forecast that
predicted strong winds for the next day.
Early next morning the firefighting forces were still waiting the fire on a road. The CL-215
did not appear with the first morning light as requested. When they did, around 08:00, it
was too late. The wind had started picking up, the fire accelerated, it crowned in thick
Aleppo pine (Pinus halepensis) forest and started spotting making all efforts to stop it along
roads totally ineffective.
Fire behavior
Initial spread, during the night, was quite slow due to the low wind. However, in that dry
fire season the fuels were very dry and the night air relative humidity at 30% was a bad
omen for the next day.
On the 3
of August, once the meltemi wind picked-up to roughly 7 BF, the fire crowned
quickly. As relative humidity dropped to less than 15% during the day the fire soon started
producing heavy medium and some long-range spotting. The wind speed did not preclude
aerial firefighting as in 1995, but the spotting made all efforts to stop the fires along roads
Coordination tried but failed to stop the fire from entering the area burned in 1995. When
the fire entered this area, which by this time was covered by some regeneration of small
pines, resprouting shrubs and mainly heavy grass growth, fire spread rate became very fast.
The fire quickly reburned practically all the area burned in 1995.
Average observed rate of spread, including the runs in the pine forest, was about 1,2 km/hr.
Runs in the light fuels of the reburned area at times exhibited higher rate of spread but the
average was also affected by firefighting efforts. This fire, as it was fanned by a strong
wind, can clearly be described as a wind-dominated fire, the same as the 1995 one
(Rothermel 1991). This is common in most fires in Greece occurring under full-strength
meltemi wind.
Response to the fire
The most important event characterizing initial response to the fire was the half-hearted
effort to reach the fire during the night. A bulldozer was used to develop an access road to
the fire. It did not complete the job due to the rugged topography, stopping a few hundred
meters from it. Then the effort was practically abandoned, waiting for the fire on the road.
The firefighting forces attacking the fire were relatively inexperienced in forest firefighting,
especially in using hand tools away from roads.
The second event was that the fire was not attacked from the air immediately after sunrise.
By 08:00 of August 3
, the wind speed increased sharply and the fire accelerated, and
reached quickly the roads where the firefighters were waiting for it. They were unable to
stop it there as well as on the asphalt road connecting Dionissos to Nea Makri a few km
further down its path. There 15 fire trucks lined-up hoping to stop the fire but the
advancing crown fire front quickly pushed them away.
Evolution of the fire
The fire front reached practically the same line where the previous fire was stopped due to
favorable fuel conditions.
After the initial burn of August 3
and the mop-up efforts of August 4
, in the morning of
August 5
, the fire restarted under windy conditions, probably due to incomplete mop-up.
This time it affected unburned forest in the area of Penteli village at the base of Penteli
mountain, entering deep in the wildland-urban interface, reaching even the central square of
the village and burning many buildings.
The fire was finally controlled on the 5
of August, while mop-up efforts continued for two
more days.
The final burned area exceeded 7.500 ha. This fire destroyed the forest that had remained on
the mountain of Penteli after the fire of 1995, and reburned most of the previously burned
area diminishing the probability for natural regeneration of pine there due to lack of seed.
Hundreds of houses and other buildings (hospitals, restaurants, factories, a school, etc.)
were destroyed or suffered significant damage. A 67 year old man who lived in the village
of Penteli was caught by the fire and died as he was fleeing his home.
Firefighting resources
All the available firefighting resources of the Fire Service in Attica and from other parts of
the country were assigned to this fire. In total there were more than 100 firetrucks as well as
all the available water-trucks of the municipalities. Also, the majority of the fleet of 15
Canadair CL-215s made drops on this fire, as well as Army CHINOOK CH-47D
helicopters. After the disaster in the wildland-urban interface area, local people volunteered
to develop patrols trying to prevent further arson. However, although the actual cause of the
fire was never identified, evidence (area, distance from Athens, time and conditions of
start) shows that the fire was not started to create the havoc it did. The disaster can clearly
be attributed to ineffective initial attack due to lack of appropriate action and not due to
inadequate dispatched forces.
Literature cited
Varela, V., G. Eftichidis, G. Xanthopoulos, A. Sfiris, and I Perros. 1999. Application of the
FMIS system for the simulation of the conditions and the evolution of forest fires in
two test cases: the Penteli fires of 1995 and of 1998. Presented at the International
Scientific Conference on “Fires in the Mediterranean Forests: Prevention -
Suppression - Soil Erosion - Reforestation” held in Athens, February 3-6, 1999.
Organized by the Hellenic National Commission for UNESCO.
Eftichidis, G. and V. Varela. 2000. The conflagration of 1998 in the Wildland-urban
interface of Athens. Pp. 68-76. In proceedings of the research special session of the
Euromediterranean wildfire meetings held in Hyeres-les-Palmiers, Var-France,
October 24
, 2000. CEREN, Domaine de Valabre, F-13120, Gardanne, France. 198
Rothermel, R. C. 1991. Predicting behavior and size of crown fires in the Northern Rocky
Mountains. USDA For. Serv. Res. Pap. INT-438. 46 p.
Xanthopoulos, G. 1996. Greece: the 1995 forest fire season. International Forest Fire News
(ECE/FAO) 14: 23-26.
Xanthopoulos, G. 1999. The 1998 forest fire season in Greece: A forest fire expert’s
account. International Forest Fire News (ECE/FAO) 20: 00-00 (available on the
Internet on the site:
... In a fire behavior simulation study, the fire rate of spread and fireline intensity in Aleppo pine forests reached a ROS of 61 m min À1 and a FI of 102,000 kW m À1 , respectively (Mitsopoulos and Dimitrakopoulos, 2007a). Xanthopoulos (2002) using real time observations reported a mean ROS of 29 m min À1 during the large fire on Penteli mountain in 1995, which propagated in an area dominated mainly by Aleppo pine forest. ...
In this study we examined the effect of stand, topographic, and climatic factors on the fuel complex characteristics of Aleppo (Pinus halepensis Mill.) and Calabrian (Pinus brutia Ten.) pine forests of Greece, based on measurements in fifty six (56) plots dominated by Aleppo pine (28 plots) and Calabrian pine (28 plots) in four different forest sites. The data were analyzed through Random Forest regression, linear regression, and quantile regression. The results showed that for all the examined factors, basal area, stand density index and stand height were the most important factors explaining fuel complex characteristics in the region for both species. Topography and climate factors were not found to be significant factors for canopy fuels. However, elevation and annual precipitation showed a minor effect mainly on the shrub fuel layer of Calabrian pine. The analysis and the conclusions of the study can be important inputs for forest management decisions (thinning, fuel treatments, stand regeneration planning) quantifying the effect that certain fuel management actions can have on the fuel hazard that the stands of these two important Mediterranean pines represent.
... Στην ευρύτερη περιοχή της βορειοανατολικής Πεντέλης (π.χ. Άγιος Πέτρος) η πυρκαγιά μεταδόθηκε καίγοντας φρύγανα, κατά τόπους λίγους χαμηλούς θάμνους αείφυλλων πλατύφυλλων και νεαρά δέντρα χαλεπίου πεύκης (11 έως 14 ετών) που είχαν προέρθει από φυσική αναγέννηση ή αναδάσωση σε περιοχές που είχαν καεί για δεύτερη ή και τρίτη φορά κατά τα τελευταία 14 χρόνια (Xanthopoulos 2002). ...
Conference Paper
Full-text available
This paper describes the evolution of the large and destructive fire of NE Attica, on 21-24 August 2009 and makes reference to the firefighting efforts, based on in situ observations during the fire and subsequent collection of data from visits to burned areas, testimonials of residents and individuals who were involved in the firefighting effort, satellite data and data collected from recording of images and information broadcasted by the media in those days. Utilization of all this information allowed a complete documentation of the behavior of the fire and a description of the weaknesses of the firefighting mechanism that manifested themselves during its four-day evolution.
... Fifteen years later, a large fire in July 1995, on Penteli mountain near Athens, burned 6500 ha and also burned about 100 buildings, many of them homes. A second large fire on Penteli, in August 1998, burned 7500 ha, re-burning most of the area that burned in 1995, and resulted in the destruction of even more houses and the death of one civilian (Xanthopoulos 2002). . Damage to property due to wildfires is not limited to buildings. ...
Full-text available
One of the most recent significant wildfires in a wildland-urban interface in Greece occurred in Northern Attica, north of Athens, Greece, that includes the towns of Sikaminos, Milesi, Oropos, Markopoulo, and Kalamos. The fire erupted in the afternoon of June 4, 2001 under strong winds associated with the passage of a cold front. It burned in a similar pattern to previous fires in the same area, in many parts burning Pinus halepensis forest under regeneration, while olive groves and other agricultural cultivations were also found in its path. Its fast rate of spread in combination with the need to protect homes, which absorbed a lot of firefighting effort, increased the burned area quickly. The fire slowed down with the onset of the night as the wind speed decreased and the fuels started picking-up moisture. It was controlled completely in the morning of the next day under favorable conditions by the strong firefighting forces that in the meantime had arrived in the area. In this paper the behavior of the fire and the firefighting efforts are described, discussed and compared to past similar fires in the same area. Special difficulties encountered in firefighting due to the extensive wildland-urban interface are also pointed-out and discussed with the help of images that show the landscape where the fire spread. An IKONOS satellite image from the year 2000 is used, among others, for this purpose. The extent and nature of damages to structures are documented and presented. Finally, there is a short discussion about the post-fire problems that followed.
... On August 5, 1998, a 7500 ha fire that had started on August 2, killed a 67-year old man who was trying to fill his car with some of his belongings and evacuate (Xanthopoulos 2002). He was found on the ground next to the car, victim of inhaling smoke and hot gases. ...
Full-text available
Forest fires are dangerous phenomena that can cause loss of life of firefighters and other citizens. Their death toll is relatively small, comparing for example to the number of lives lost in traffic accidents. However, fire related deaths are always dramatic and create strong emotions to the public. More important, they are a concern to all those working in fire management because, as a rule, most of these deaths can be avoided if the necessary knowledge exists and the necessary care is exercised. This paper presents a number of forest fire related deaths of both firefighters and citizens that occurred in Greece in the last decades. The incidents are described briefly, providing the necessary information that leads to the basic conclusion of this work: the causes of each fatal accident are not unique to that case. They have been identified before in some other place and time, and they fall into broad categories for which warnings have been established in international literature. The question that has to be answered then is if the victims posessed, at the time of the accident, the necessary knowledge on how to avoid putting their life in danger. A negative answer to this question necessarily points to the ever-existing need to educate firefighters and the public, especially when the latter lives in wildland-urban interface areas, on what they should do, what they shouldn't, and which are the situations that shout "watch- out.
Describes methods for approximating behavior and size of a wind-driven crown fire in mountainous terrain. Covers estimation of average rate of spread, energy release from tree crowns and surface fuel, fireline intensity, flame length, and unit area power of the fire and ambient wind. Plume-dominated fires, which may produce unexpectedly fact spread rates even with low ambient windspeeds, are covered and supplemental methods suggested for estimating their occurrence. The spread information can be used to estimate and map fire area and perimeter.
Greece: the 1995 forest fire season
  • G Xanthopoulos
Xanthopoulos, G. 1996. Greece: the 1995 forest fire season. International Forest Fire News (ECE/FAO) 14: 23-26.
The conflagration of 1998 in the Wildland-urban interface of Athens
  • G Eftichidis
  • V Varela
Eftichidis, G. and V. Varela. 2000. The conflagration of 1998 in the Wildland-urban interface of Athens. Pp. 68-76. In proceedings of the research special session of the Euromediterranean wildfire meetings held in Hyeres-les-Palmiers, Var-France, October 24 th, 2000. CEREN, Domaine de Valabre, F-13120, Gardanne, France. 198 p.
The 1998 forest fire season in Greece: A forest fire expert's account
  • G Xanthopoulos
Xanthopoulos, G. 1999. The 1998 forest fire season in Greece: A forest fire expert's account. International Forest Fire News (ECE/FAO) 20: 00-00 (available on the Internet on the site: