Fire is an important recurrent phenomenon in all forested and non-forested regions of the globe. In some ecosystems fire plays an ecologically significant role in biogeochemical cycles and disturbance dynamics. In other ecosystems fire may lead to the destruction of forests or to long-term site degradation. As a consequence of demographic and land use changes and the cumulative effects of anthropogenic disturbances many forest types adapted to fire, are becoming more vulnerable to high-intensity wildfire. Ironically, this is often due to the absence of periodic low-intensity fire. In other forest types, however, as well as many non-forest ecosystems e.g. in savannas and grasslands, fire plays an important role in maintaining their dynamic equilibrium productivity and carrying capacity (Goldammer, 1990; Goldammer and Furyaev 1996; van Wilgen et al. 1997).
In most areas of the world wildfires burning under extreme weather conditions will have detrimental impacts on economies, human health and safety, with consequences that are comparable to the severity of other natural hazards. In all ecosystem fire needs to be managed to balance the benefits derived from burning with the potential losses from uncontrolled fires.
Fires in forests and other vegetation produce gaseous and particle emissions that have impacts on the composition and functioning of the global atmosphere (Crutzen and Goldammer, 1993; Levine, 1991, 1996; van Wilgen et al. 1997). These emissions interact with those from fossil-fuel burning and other technological sources which are the major cause for anthropogenic climate forcing. Smoke emissions from wildland fires also cause visibility problems which may result in accidents and economic losses. Smoke generated by wildland fires also affect human health and in some cases contribute to the loss of human lives.
Fire risk modelling in expected climate change scenarios indicate that within a relatively short period of the next three to four decades, the destructiveness of human-caused and natural wildfires will increase. Fire management strategies which include preparedness and early warning cannot be generalized due to the multi-dimensional effects of fire in the different vegetation zones and ecosystems and the manifold cultural, social, and economic factors involved.
However, unlike the majority of the geological and hydrometeorological hazards included in the IDNDR Early Warning Programme, wildland fires represent a natural hazard which can be predicted, controlled and, in many cases, prevented.
Comprehensive reports with final data on losses caused by forest and other vegetation fires (wildland fires) are only occasionally available. The main reason for the lack of reliable data is that the majority of both the benefits and losses from wildland fires involve intangible non-use values or non-market outputs which do not have a common base for comparison, i.e. biodiversity, ecosystem functioning, erosion, etc. (González-Cabán, 1996).
Market values such as loss of timber or tourism activity have been calculated in some cases. The large wildfires in Borneo during the drought of 1982-83, which was caused by the El Niño Southern Oscillation (ENSO), affected a total of more that 5 million hectares of forest and agricultural lands (Goldammer et al., 1986). It resulted in the loss of timber values of ca. US$8.3 billion, and a total of timber and non-timber values and rehabilitation costs of US$ 9.075 billion (Schindele et al., 1989). The damages caused by the fire episode of 1997 in Indonesia are not yet known at the time of writing this report.
The 1988 fires in the Yellowstone area of the United States cost around US$ 160 million to suppress and caused an estimated loss of US$ 60 million in tourist revenues between 1988 and 1990 (Polzin et al., 1988). In the longer term, however, the increased biodiversity created by the fires in Yellowstone National Park may well yield benefits that outweigh these losses.
Australia’s Ash Wednesday Fires of 1983, which were also linked to the ENSO drought of 1982-83, resulted in a human death toll of 75, the loss of 2539 houses and nearly 300,000 sheep and cattle. In South Australia alone the estimated direct losses of agricultural output (sheep, wool, lambs, cattle, pasture, horticulture) of the Ash Wednesday fires were estimated AUS$ 5.7 million (on the basis of 1976-77 prices), and the estimated value of the net costs to the Government Sector of South Australia of the 1983 bushfires were AUS$ 33 million (Healey et al., 1985).
Wildfire damage to agricultural lands, particularly in the tropics, may have tremendous impact on local and regional famine. In 1982-83 the West African country Côte d’Ivoire was swept by wildfires over a total area of ca. 12 million ha (Goldammer, 1993). The burning of ca. 40,000 ha of coffee plantations, 60,000 ha of cocoa plantations, and some 10,000 ha of other cultivated plantations had detrimental impacts on the local economy. More than 100 people died during this devastating fire period.
The “Great Black Dragon Fire” of 1987 in the People’s Republic of China burned a total of 1.3 million hectares of boreal mountain forest, the houses of 50,000 inhabitants and resulted in a human death toll of 221, mostly caused by high carbon monoxide concentrations in the forest villages. The long-term statistics in China reveal that between 1950 and 1990 a total of 4,137 people were killed in forest fires (Goldammer, 1994).
The last large fire event occurred in Mongolia between February and June 1996. A total of 386 forest and steppe fires burned over an area of 2.3 million ha of forest and 7.8 million ha of pasture land, involving the loss of 25 human lives, more than 7000 livestock, 210 houses, 560 communication facilities, and 576 facilities for livestock; the preliminary damage assessment was ca. US$ 2 billion (Naidansuren, 1996).
Reliable statistical data on occurrence of wildland fires, areas burned and losses are available for only a limited number of nations and regions. Within the northern hemisphere the most complete data set on forest fires is periodically collected and published for the member states of the Economic Commission for Europe (ECE). It includes all Western and Eastern European countries, countries of the former Soviet Union, the U.S.A. and Canada. The last data set covers the period 1993-95 (ECE/FAO 1996). In the European Union a Community Information System on Forest Fires has been created on the basis of information collected on every fire in national databases. The collection of data on forest fires (the common core) has become systematic with the adoption of a Commission Regulation in 1994. The Community Information System on Forest Fires currently covers 319 provinces (departments, states) of Portugal, Spain, France, Italy, Germany and Greece (European Commission, 1996; Lemasson, 1997). It contains information on 460,000 fires recorded between 1 January 1985 and 31 December 1995 involving a total of six million hectares. Other countries from outside the ECE/EU region report fire statistics in the pages of International Forest Fire News or are included in the FAO report on global wildland fires (FAO, 1992).
In many countries (e.g. Australia) where fire is used as a management tool by the indigenous population, graziers and managers of forests and natural areas it is impossible to discriminate between management fires and wildfires. Statistics for wildfires are usually available only for production forest and national park lands.
A global data set has been developed on the basis of active fires detected by the NOAA AVHRR sensor. The “Global Fire Product” of the International Geosphere-Biosphere Programme Data and Information System (IGBP-DIS; further details are provided below in the section on Global Fire Monitoring).
From the perspective of the IDNDR wildland fires may affect two basic environmental problem areas, (1) atmospheric pollution (direct impact of smoke on human health and economies; influence of gaseous and particle emissions on the composition and functioning of the atmosphere), and (2) biodiversity, ecosystem functioning, and landscape stability. These both can have deleterious consequences for the severity of other hazards.
Human fatalities and health
Smoke pollution generated by wildland fires occasionally creates situations during which human lives and local economies are affected. Fatalities in the general public caused by excessive carbon monoxide concentrations have been reported from various fire events, e.g. the large forest fires in China in 1987. Firefighters who are regularly subjected to smoke are generally at higher health risk.
The use of fire in forest conversion and other forms of land clearing and wildfires spreading beyond these activities are very common in tropical countries. In the 1980’s and 1990’s most serious pollution problems were noted in the Amazon Basin and in the South East Asian region. The most recent large smog episodes in the South East Asian region were in 1991, 1994 and 1997 when land-use fires and uncontrolled wildfires in Indonesia and neighbour countries created a regional smog layer which lasted for several weeks. In 1994 the smoke plumes of fires burning in Sumatra (Indonesia) reduced the average daily minimum horizontal visibility over Singapore to less than 2 km; by the end of September 1994 the visibility in Singapore dropped to as low as 500 metres. In the same time the visibility in Malaysia dropped to 1 km in some parts of the country. A study on asthma attacks among children revealed a high concentration of fire-generated carbon monoxide (CO), nitrogen dioxide (NO2) and inhalable suspended particulate matter (PM10) was responsible for the health problems (ASEAN, 1995a). The smog situation in September 1997 caused the worst smoke pollution in the region, reflected by a value of 839 of the Pollutant Standard Index (PSI, for further details see the Section below on Atmospheric Pollution Warning) in the city of Kuching (Sarawak Province, Malaysia); the government was close to evacuating the 400,000 inhabitants of the city.
In the same regions, the smoke from fires caused disruption of local and international air traffic. In 1982-83, 1991, 1994 and 1997 the smog episodes in South East Asia resulted in the closure of airports and marine traffic, e.g. in the Strait of Malacca and along the coast and on rivers of Borneo. Smoke-related marine and aircraft accidents occurred during September 1997. The loss of an airplane and 234 human lives in September 1997 in Sumatra was partially attributed to air traffic control problems during the smog episode.
Wildfires burning in radioactively contaminated vegetation lead to uncontrollable redistribution of radionuclides, e.g. the long-living radionuclides caesium (137Cs), strontium (90Sr) and plutonium (239Pu). In the most contaminated Regions of the Ukraine, Belarus and the Russian Federation (the Kiev, Zhitomir, Rovmo, Gomel, Mogilev and Bryansk Regions), the prevailing forests are young and middle-aged pine and pine-hardwood stands with high fire danger classes. In 1992 severe wildfires burned in the Gomel Region (Belorussia) and spread into the 30-km radius zone of the Chernobyl Power Plant. Research reveals that in 1990 most of the 137Cs radionuclides were concentrated in the forest litter and upper mineral layer of the soil. In the fires of 1992 the radionuclides were lifted into the atmosphere. Within the 30-km zone the level of radioactive caesium in aerosols increased 10 times (for more details on resuspension of radioactive matter from forest fires, see Dusha-Gudym, 1996).
Fire emissions, atmosphere and climate
In recent years increasing attention has been given to the role of vegetation fires in biogeochemical cycles and in the chemistry of the atmosphere (Crutzen and Goldammer, 1993). According to recent estimates some 1.8-4.7 billion tons of carbon stored in vegetation may be released annually by wildland fires and other biomass burning (Crutzen and Andreae, 1990). It must be noted that not all of the biomass burned represents a net source of carbon in the atmosphere. The net flux of carbon into the atmosphere is due to deforestation (forest conversion with and without involving the use of fire) and has been estimated by Houghton (1991) to be in the range of 1.1-3.6 billion tons per year. Important contributions to the total worldwide biomass burning, which are included in the numbers mentioned above, are fires in savannas, shifting agriculture, agricultural waste burning and firewood consumption (Andreae and Goldammer, 1992).
Although the emissions from tropical vegetation fires are dominated by carbon dioxide [CO2], many products of incomplete combustion that play important roles in atmospheric chemistry and climate are emitted as well, e.g., a number of gases that influence the concentrations of ozone and hydroxyl radicals and thus the oxidation efficiency of the atmosphere, in particular NO, CO, CH4 and reactive hydrocarbons. The influence of these emissions affects especially the southern hemisphere during the dry (winter) season, i.e. during August – November, and manifests itself in strongly enhanced tropospheric ozone concentrations, extending from the regions regularly affected by biomass burning in Brazil and southern Africa across the Atlantic and the Indian Ocean all the way down to Tasmania (Andreae et al., 1993; Journal of Geophysical Research, 1996; van Wilgen et al., 1997). Other gases whose atmospheric concentrations are strongly dominated by biomass burning are CH3Cl and CH3Br, which together with CH4 play a significant role in stratospheric ozone chemistry (Manö and Andreae, 1994).
Biodiversity, ecosystem functioning, and landscape stability
The impacts of wildfires on the functioning and stability of ecosystems has been described widely in numerous publications, covering the full range of geographical, ecological, socio-cultural and economic conditions of the globe. The magnitude of phenomena resulting from wildfires prohibits any detailed review in the context of this report.
On the one hand fire is an integrated element which contributes to the stability, sustainability, high productivity and carrying capacity of many ecosystems. On the other hand wildfire, in conjunction or interaction with land use systems and exploitation of natural resources, leads to the loss of forest and agricultural products and can have negative impacts on biodiversity, ecosystem function and land stability. For example, in the dry forests of Australia low-intensity fire is regularly applied to maintain understorey plant species and habitat for native fauna as well as to reduce surface fuels to mitigate against the impacts of high-intensity wildfires. During the dangerous summer period all fires are suppressed as quickly as possible both to reduce damage to forest values and to reduce the chance of wildfire burning out of the forest and causing severe losses to houses and structures in the built environment.
Many plant and animal species, e.g. in the tropical lowland rain forest ecosystems and elsewhere, are susceptible to fire influence and are easily destroyed by fire and replaced by less species-rich communities. Human-induced fire regimes in tropical rain forests result in degraded vegetation types (grasslands, brushlands) which are less stable and productive, both from an ecological and economic point of view. Fires may also lead to the depletion of soil cover, resulting in increased runoff and erosion, with severe downstream consequences, e.g., mudflows, landslides, flooding or siltation of reservoirs.
Fires often interact with other disturbances, e.g. extreme storm events (hurricanes) or insect outbreaks. The extended rain forest fires of 1989 in Yucatan (Mexico) represent a typical example because they were a result of a chain of disturbance events. Hurricane “Gilbert” in 1987 opened the closed forests and increased the availability of unusual amounts of fuels. The downed woody fuels were then desiccated by the subsequent drought of 1988-89, and the whole of the forest area was finally ignited by escaped land clearing fires. None of these single three factors, the cyclonic storm, the drought, or the ignition sources, if occurring alone, would have caused a disturbance of such severity and magnitude on an area of 90,000 hectares (Goldammer, 1992).
In Krasnoyarsk Region, Russian Federation, a mass outbreak of the Siberian Gipsy Moth (Dendrolimus superans sibiricus) going on since 1989 has meanwhile affected a total of 1 million hectares of boreal forest (Baramchikov, 1997). It is expected that large wildfires will occur in the partially or completely killed stands within the next years.
Early Warning Systems in Fire Management and Smoke Management
Early warning (fire intelligence) systems are essential components of fire and smoke management . They rely on
evaluation of vegetation dryness and weather;
detection and monitoring of active fires;
integrating and processing of these data in fire information systems with other relevant information, e.g. vegetation cover and values at risk;
modelling capabilities of fire occurrence and behaviour; and
dissemination of information.
Early warning of fire and atmospheric pollution hazard may involve locally generated indicators, such as local fire-weather forecasts and assessment of vegetation dryness. Advanced technologies, however, which rely on remotely sensed data, evaluation of synoptic weather information and international communication systems (e.g., Internet) are now also available for remote locations.
In this report the large variety of standards, methods and technologies of fire and smoke management which are used in national programs cannot be described in detail. Generally speaking, however, it is obvious that, due to the lack of resources, fire management systems are disproportionately less available in developing countries.
In some industrialized countries, e.g. in Central and Northern Europe, wildfires have been largely eliminated due to high-intensity land use, improved accessibility of potentially threatened land and the availability of infrastructures and advanced fire management technologies. Regions with less developed infrastructures are found in densely populated lands (e.g., in the tropics and subtropics) and in sparsely inhabited regions (e.g., in the northern boreal forests) as well. They are equally subjected to high wildfire risk because of the abundance of human fire sources or the lack of human resources to control fires respectively.
Relation to Other IDNDR Early Warning Working Group Reports
Some of the issues described in this report are closely related to other IDNDR Early Warning Working Group reports, e.g. the reports on hydrometeorological hazards, on technological opportunities, and on local perspectives. The cross-cutting issues show that there are areas of potential common activities and programmes.
The conclusions of the most recent global wildland fire forum, the “Second International Wildland Fire Conference” (Vancouver, Canada, May 1997), clearly underscored the fact that unlike other natural disasters, fire is one of the few natural disturbances that can be forecast and mitigated (Anonymous, 1997). This fact may explain why forecasting fire events and the potential of mitigating fire impacts are comparably better developed as compared to other natural disasters. The description of the early warning systems for wildfires, which are available, in the development stage or proposed, may therefore serve as examples for other local, regional and international mechanisms of cooperation in disaster early warning and management.