Arid Lands Newsletter (link)No. 54, November/December 2003
Fire Ecology I

Fire in the European Mediterranean

by Oliver Rackham

"Big fires are a predictable result of rural depopulation, land abandonment, increase of wild vegetation, modern forestry, legislation against fire, and the growth of a fire-fighting industry."


(Back to top)
The Mediterranean region of southern Europe forms a fringe along the north shore of the Mediterranean Sea, from Portugal to Turkey, including Corsica, Sardinia, Sicily, Crete and many lesser islands. It is defined by climate: warm wet winters and hot dry summers, often with no rain between May and September. It is a very varied region, with a violent tectonic history: many mountain ranges are over 2000 m in height, most of them near the sea and some on islands. Annual rainfall varies from over 1000 mm to less than 200 mm; even the wettest areas have a dry season. Rainfall varies widely from year to year, especially in the more arid parts. High altitudes experience cold winters as well as dry summers, and the growing season can be very short.

The region has been markedly affected by glacial cycles, although there has been little actual ice cover. The present climate is only about 5000 years old, and has fluctuated within that period (Grove and Rackham 2001, chapters 7-9).

Natural vegetation varies, according to climate and geology, from forest to savanna to shrubland to desert. Pasturage ("rangeland" in American terms) often consists not of grassland but of trees bitten down into the form of shrubs, which can persist indefinitely in that state.

The Mediterranean was the main seat of development of European human culture. Agriculture has existed here for some 7000 years. At times, notably during the Roman Empire and the nineteenth century CE, there has been a dense rural population and extensive and intensive land use, especially in the form of terrace cultivation. In all countries there was severe rural depopulation in the 20th century: cultivation has retreated from the mountains and natural vegetation has increased.

The Mediterranean basin extends into Asia and Africa, where the vegetation is broadly similar (especially at the arid end of the scale). The human history is also broadly similar, until the last hundred years: depopulation set in later, and in much of the region the population is still rising. A similar climate exists in the "mediterraneoid" regions of California, middle Chile, the southwestern tip of Africa, and small parts of Australia. The first two of these regions share the mountains and the violent tectonics. However, the plants and animals of the mediterraneoids are dissimilar or utterly different from the Mediterranean and from each other; the human histories are utterly different (Grove and Rackham 2001, chapter 1).

link to Rackham Figure 1
Link to Fig. 1, ~36K

In the Mediterranean, as in the mediterraneoids, fire occurs where vegetation is flammable. Combustibility is not a misfortune but an adaptation: plants that burn do so because they are fire-adapted. They make fire-promoting resins and other chemicals, or they have structural adaptations, such as producing a loose, airy litter of dead leaves and twigs which dries out and burns. Their ecology involves catching fire from time to time and burning up competitors.

Link to Rackham Figs. 2, 3
Link to Figs. 2&3, ~59K

Some trees survive fire by an insulating bark, the cork oak being a notable example. Others are killed to the ground but sprout, as with holm oak and many other evergreens. Others are killed outright but regenerate from seed, as with Aleppo pine. Although these mechanisms are not as well investigated in the Mediterranean as in California or southwestern Australia, it is known that many species will not germinate readily, if at all, without fire. However, there are fire-sensitive plants, such as junipers, cypress, and firs (Abies spp), which grow in places protected from fire.

Characteristics of Fires

Link to Rackham Fig. 4
Link to Fig. 4, ~25K

(Back to top)
Fires are of three types:

  1. Natural fires, ignited by lightning (or rarely volcanoes).
  2. Occupational fires, used deliberately to improve or control natural vegetation.
  3. Wildfires, ignited by human malice or accident, including occupational fires that get out of control.

Most fire damage is caused by a few very big conflagrations in the dry season. Conflagrations are promoted by hot weather, high winds, or preceding drought. A big late-season fire, driven by a hot sirocco wind off the Sahara, leaps over firebreaks, highways, and gorges, and defies all fire-fighting technology. However, no very close relation between fire and weather has been found. The most fiery regions are not the driest, but relatively rainy areas such as Corsica, northwestern Spain, and northwestern Crete. Only in smallish areas of desert or semi-desert is the vegetation too sparse to support a fire.

Link to Rackham Fig. 5
Link to Fig. 5, ~35K

Fires can occur, indeed probably have occurred, in most places where there is flammable vegetation. They are nearly always cyclical: for example in Greece Pinus halepensis, although killed by fire, can support a pine-and-fire cycle of only 15 years. South-facing sites tend towards a regime of frequent small fires. On north-facing sites, unless the vegetation is incombustible (e.g. beech forest), fires are rare but probably catastrophic when they occur. On Mount Athos, where there was a conflagration in 1990, Dr Philip Oswald and I found evidence of previous fires, but too long ago for anyone now living to remember them. It is unusual for fires to occur in vegetation that has not previously burnt, except on abandoned farmland where the vegetation has only recently become combustible.

Effects of fire

Link to Rackham Figs. 6 & 7
Link to Figs. 6 & 7, ~56K

(Back to top)
Fire is an important and positive factor in biodiversity (e.g. de Lillis and Testi 1992): hundreds of herbaceous plant species appear in recently burnt areas, some of them apparently from seed dormant since the previous fire. Promoting such plants, for livestock to feed on, is a main reason for occupational burning. In a study of the Greek island of Kythera Dr Jennifer Moody and I find that the biggest concentrations of plant species per square kilometer are in (1) gorges; (2) the sites of medieval towns; and (3) areas burnt in the previous year or two.

Link to Rackham Fig. 8
Link to Fig 8, ~38K

Some ecosystems, however, are threatened by fire, especially where recent changes have made fire possible in fire-sensitive but previously incombustible vegetation. An example is the Samariá National Park in Crete, a first-class historic cultural landscape with its buildings, ancient cypress trees, and cultivation remains; the park is now much invaded by pine which has made it flammable.

Among the objections advanced to fire is that it accelerates erosion. There appear to be some well-authenticated examples (e.g. Ballais 1992), but in my experience it is more common for fires to have no observable effect. There seems not yet to have been a rigorous comparison of burnt and unburnt catchments over one or more whole fire cycles. Still less is it known whether more erosion results from frequent small fires than from a single conflagration.

link to Rackham Fig. 9
Link to Fig. 9, ~47K

In theory, if fires were suppressed for long enough, suppression might become permanent because fire-promoting trees and plants would be ousted by fire-sensitive competitors, as happens (for example) at the boundary between savanna and rain-forest in Australia. In the Mediterranean I have observed firs spreading down into flammable forests at lower altitudes, and chestnuts invading and shading out fiery grassland, but only on a small scale. Such changes towards incombustibility are probably rare: most of the more invasive trees and plants in the European Mediterranean are fire-promoting.

People and fire

(Back to top)
Most Southern European pasturage is occupationally burnt on a cycle ranging down to five years or even three: the landscape is a mosaic of triangular patches of different ages of regrowth. Burning is often a skilled art. In Liguria (northwestern Italy) the flower-rich grasslands of lower and middle mountains are traditionally burnt every year late in the rainy season, when fires are cool; this pre-empts an unscheduled, uncontrollable, hot fire in the dry season. Occupational burning encourages the growth of grasses and the young shoots of trees and shrubs, which are good forage for sheep and goats. It also sets back the less palatable and nutritious undershrubs, which are somewhat fire-sensitive.

Link to Rackham Figs. 10, 11a, 11b
Link to Figs. 10, 11, ~70K

In the historic Mediterranean, fires seldom reached houses or crops. To judge by recent experience, gardens and orchards (often irrigated) round villages, hamlets, and monasteries usually act as fire-breaks. Olive groves, however, were at risk: the tradition of plowing under olives is doubtless partly in order to reduce the fuel.

History of fire

(Back to top)
The existence of elaborate fire adaptations, such as the fire-resisting bark of cork oak, implies that fire has been a factor over many generations of trees and perennial plants, and therefore for much longer than human history or prehistory. There is as yet no evidence on the frequency or other characteristics of fires in the Pleistocene.

By analogy with other parts of the world, one might expect Paleolithic and Mesolithic peoples to have used occupational fire to manipulate flammable vegetation in favor of their hunting and gathering activities. In a general way, finds of charcoal in pollen deposits have been taken to indicate widespread occupational burning in late prehistory, but suitable deposits are rare and the idea as yet is not well developed (e.g. Atherden and Hall 1999). It is not possible to say what was burnt, how often, or at what time of year.

Ancient Hebrew, Greek, and Roman writers were aware of fire, but usually in a general way: they were not much interested in the uncultivated roughlands where fires occur.

From the Middle Ages onwards there is a stream of evidence for occupational burning, from statutes forbidding or legitimatizing it, and from accounts of travellers, for whom fire was often a new and terrifying experience. In Crete in 1414, for example, a decree by the Senate of Venice was prompted by the reported decline of cypress, a fire-sensitive tree. The Carta de Logu, the 14th-century law code of Sardinia, has a clause concerning fire-breaks to keep wildfires away from towns. The amount of evidence increases into the 18th and 19th centuries, but this is not evidence for increase of fires; as with earthquakes and floods, the farther back in time we go the bigger an event has to be to be put on record (Grove and Rackham 2001, p.229-30).

Is fire increasing?

(Back to top)
Mediterranean governments publish annual reports of fires. Most official statistics are difficult to interpret, and these raise especially difficult problems of definition. Does occupational burning (such as burning a grain field after harvest) count as a fire? Where "forest fires" are recorded, what counts as a forest? Are fires in savanna (heathland with scattered trees) or maquis (trees reduced to the stature of shrubs) counted? Some countries claim to record the number of fires (how big does a small fire have to be to be counted?) or the source of ignition (where the fire itself often destroys the evidence).

From this evidence, one can tentatively infer the following (Grove and Rackham 2001, p. 231-2):

  1. Lightning is not a negligible cause, even though the modern landscapes of the Mediterranean contain few of the dead or hollow trees where lightning fires often start.
  2. Most of the burning is accounted for by a few very large fires.
  3. Some years are more fiery than others, by a factor of at least tenfold.
  4. Smoking and matches are not such important sources of ignition as might be expected.
  5. In most countries the area reported as burnt each year has increased, typically by about fivefold between the 1960s and 1990s. It is difficult to say how much of this is due to more fires and how much to more complete recording.

My own experience and that of my colleagues, for example the 35 years for which I have known Crete, is that the increase in fires is real, though less than the statistics make it.

Link to Rackham Figs. 12 and 13
Link to Figs. 12 & 13, ~44K

There is abundant reason why fire should have increased. There is now much more to burn than when the countryside was densely populated and used. Abandoned farmland or neglected pasture turn into forest or shrubland. They are invaded especially by pines and Cistus, which happen to be among the most fire-promoting genera in the European flora. Moreover, natural vegetation used to be interspersed with fields and vineyards which acted as fire-breaks. When these are overgrown, fire can spread much farther before it meets an obstacle.

Link to Rackham Figs. 14a, 14b, 14c
Link to Fig. 14, ~67K

The area around Myrtos (Crete), an arid part of the Mediterranean, was first studied by me in 1968, some twenty years after a decline in browsing by sheep and goats. Vegetation was beginning slowly to increase, and was still increasing when I revisited the area in 1988. Following that visit I predicted that it was "liable to change from the grazing-dominated landscape of the past to a fire-dominated landscape in the future" (Rackham 1990). A conflagration duly happened in 1996.

Link to Rackham Figs. 15a, 15b
Link to Fig. 15, ~56K

Exotic invasive species are less of a factor than in some other regions. Fire-promoting eucalypts from Australia, once established, are almost impossible to get rid of, but with one exception (see below) do not spread. The nearest that the Mediterranean has to an "elephant-grass" problem - where giant grasses invade forest or savanna and result in more frequent and hotter fires - is in Spain and the Balearics, where the native giant grass Ampelodesmus tenax is often a factor in fires.

Forestry and fire

(Back to top)
Most Mediterranean countries regard themselves as ruined landscapes, "degraded" through thousands of years of misuse of the land, which might be "restored" to the forests supposed to have existed in an idealized past. Modern forestry, usually on the German model of planting trees on non-forested hillsides, has been extensively attempted, especially in Spain and Italy.

Foresters are not content with the natural fire risk. Most of the trees planted are pines, which are fire-dependent and fire-promoting. Much of the Pinus radiata planted in Italy in the "reafforestation" campaigns of the 1930s was consumed by fire. P. halepensis and P. pinaster, among the most widely planted in Spain, turn out to burn, on average, once in 20-30 years. After 1975 it was claimed that the tree'd area burnt in Spain exceeded the area planted (Martínez Hermosilla 1990); this was before 1994, the most fiery year yet known in that country.

Link to Rackham Fig. 16
Link to Fig. 16, ~37K

In Portugal and other countries, planting eucalypts has had a similar effect, except that they are not killed by fire but sprout, grow up, and generate the fuel for an indefinite succession of future fires. On the Greek island of Kythera, whose people are familiar with eucalypts through their Australian colony, plantations of Pinus halepensis and Eucalyptus camaldulensis, established around 1970, have already created a fire cycle; both pines and eucalypts re-establish from seed after a fire and prepare fuel for the next fire.

Link to Rackham Fig. 17
Link to Fig. 17, ~44K

Mount Vesuvius, if left alone, has its own peculiar shrubby and herbaceous vegetation on the unstable surfaces resulting from successive volcanic eruptions. Since the last eruption in 1944 foresters have been energetic in creating a forested landscape of pines and the giant broom Genista ætnensis, introduced from Mount Etna, which has proved to be very invasive on Vesuvius. This landscape is much more flammable than the native oak and chestnut forest surviving on areas not reached by recent eruptions; it sets the scene for a conflagration the next time the volcano erupts.

Link to Rackham Fig. 18
Link to Fig. 18, ~28K

Foresters promote fire in another way, when they exclude browsing animals and thus allow native vegetation (especially grasses and shrubs) to grow up and accumulate fuel. As at Myrtos, suppression of goats and sheep predictably results in a fire-dominated landscape.

Human attitudes to fire

(Back to top)
Those who do not have daily contact with fire regard it as a disaster and a tragedy, the result of human carelessness or greed. They assume that humanity has the power to prevent fire: a recent Prime Minister of Greece was called upon to resign for failing to stop the pines around Athens from burning, and promised never to neglect this 'duty' again. Attention focuses on the source of ignition, as if fire damage could be reduced merely by having fewer fires.

Legislators, in their vanity, think they will prevent fire by making it illegal. In many countries occupational burning is a crime, even the preemptive burning of the grass in Liguria. This does not prevent it from happening, but prevents the lawmakers from having control over it. Responsible people no longer stay to watch and contain the fire: if they do they get arrested. If fires are suppressed the fuel accumulates, to feed a conflagration later in the season or in a later year. Forbidding fire also antagonizes shepherds and others for whom occupational burning is part of their livelihood.

There is, however, some sign of more practical ways of controlling fire, for example by removing flammable shrubs alongside roads. In the south of France large areas of fiery vegetation have been broken up by "green firebreaks", strips of land that are grazed or even cultivated (Challot 1993); this however involves finding animals and shepherds, and conflicts with the European Union's policy of abandoning agriculture and pasturage. Preemptive burning is little used in countries where ownerships are small and nobody dares take responsibility.


(Back to top)
Most of what I have said will be familiar to readers who have studied fire in other flammable parts of the world. Fire is an inescapable fact of life in Mediterranean vegetation, and many ecosystems could not function without it. Compared to Australia and America, Europe is backward in understanding the effects of different frequencies, intensities, and seasonalities of fire.

Fire was believed to play a part in the desertification which in the 1990s was supposed to threaten southern Europe. In reality I know of no instance where fire is known to have created a new desert. If desertification is progressive, fire could play a part only in the early stages, for deserts and semi-deserts will not burn. In practice, as far as I am aware, fire is always followed by regrowth. The erosion-promoting effects of fire, such as they are, are probably not significant. Forest and shrubland fires, by definition, do not affect land that is cultivable by present standards.

Human problems with fire result from people ignorantly putting themselves and their property in dangerous places. Instead of living in villages surrounded by relatively fireproof gardens, they build houses in the bush and even in pinewoods. They banish sheep and goats, and shepherds' fires, from their environment, and live surrounded by masses of uneaten fuel. Planning regulations in many countries are notoriously lax, and the death penalty (by fire) is not a sufficient deterrent to enforce them.

Europe has been slow to learn from the fire experience of North America and Australia. Administrators still believe that they have it in their power to prevent fire. For centuries laws have been passed against fire, and for decades the public has been exhorted against sources of ignition; neither of these activities has had any detectable effect on reducing fire, and yet they are forever being repeated. A huge investment in fire-fighting, especially in Spain, has been associated with an increase in conflagrations. There seems to be no interest in what pattern of fire it would be desirable to achieve. Big fires are a predictable result of rural depopulation, land abandonment, increase of wild vegetation, modern forestry, legislation against fire, and the growth of a fire-fighting industry. The general effect of human reactions to fire is probably to lengthen the fire cycle, resulting in less frequent but more intense and more destructive fires in any one place.


(Back to top)
This work draws on many projects in which I have participated over the years, notably at Myrtos in 1968, in the European Union's MEDALUS 2 project in the 1990s, and on the archaeological surveys of Laconia, Sphakiá, and Kythera. Among other colleagues I acknowledge the help and friendship of Professor Peter Warren, Professor Diego Moreno, Mr A.T. Grove, Dr Jennifer Moody, and Dr Cyprian Broodbank.


(Back to top)
Atherden, M.A. and J.A. Hall. 1999. Human impact on vegetation in the White Mountains of Crete since AD 500. The Holocene 9:183-93.

Ballais, J.L. 1992. L'érosion consécutive à l'incendie d'août 1989 sur la Montagne Sainte-Victoire. Trois années d'observations (1989-92). Bulletin de l'Association des Géographes Français 5:423-7.

Challot, A. 1993. La place des grandes coupures agricoles et pastorales dans la prévention des incendies de forêt. Forêt Méditerranénne 14:130-40.

de Lillis, M. and M. Testi. 1992. Fire disturbance and vegetation dynamics in a mediterranean maquis of central Italy. Ecologia Mediterranea 18:55-68.

Grove, A.T. and O. Rackham. 2001. The nature of Mediterranean Europe: An ecological history. New Haven, CT: Yale University Press.

Martínez Hermisilla, P. 1990. Enfoque histórico de los trabajos de restauración. Ecologia fuera de serie 1:367-71.

Rackham, O. 1990. The greening of Myrtos. In Man's role in the shaping of the eastern Mediterranean landscape, ed. S. Bottema et. al, 341-8. Rotterdam: Balkema.


bar denoting end of article text

Author information

(Back to top)
Dr. Oliver Rackham, a historical ecologist, is a Fellow at Corpus Christi College, Cambridge University. He can be reached for comment at

About the Arid Lands Newsletter

Link to ALN home page Link to index page for back web issues Link to index page for pre-web issue archive Link to ALN #54 table of contents