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

Fire management in the Kruger National Park

by Navashni Govender

"...the KNP is now implementing a fire management policy that simulates as closely as possible the intensity, spatial and temporal heterogeneity of a natural fire regime, taking into account the importance and reality of anthropogenic fires in this system and thereby achieving its mandate of maintaining biodiversity in all its facets and fluxes."


(Back to top)
Fire management practices within conservation areas have been debated for decades. Fire is an obvious ecosystem driver and as such extremely important in shaping the savanna landscape. However, there are important gaps in our understanding about the effects of fire, the role it has played in savannas and the interactive effects with factors such as herbivory (Bond 1997). As a result there are numerous practical fire management options and methods of application available. Conservation managers realize the importance and necessity of fire, feel the need to choose between various options to manage it, but acknowledge that their decisions are currently based on imperfect knowledge. In order to effectively manage, learn and apply fire as a usable tool in conservation, ongoing research into fire management strategies and the effects thereafter is extremely important, as is the learning-while-doing operating principle.

Africa is often referred to as the "Fire Continent" due to the widespread occurrence of biomass burning, particularly in the savanna biome. Covering approximately fifty percent of the land surface in Africa, savannas are a tropical vegetation type characterized by the coexistences of both grasses and trees (Scholes and Walker 1993). Fires are a major disturbance in many biomes across the world (Parr and Chown 2003), comparable to herbivore activities in their ability to consume up to 80% of the aboveground primary production in some systems (Bond and van Wilgen 1996). The landscapes characterizing the savanna biome today have developed with fire playing a key role in determining the structure, function and dynamics of the system.

The macro-level requirement for natural fires is a fire climate which comprises dry and wet periods so that fires can burn the plant fuels during the dry period that have been produced and accumulated during the wet period. Within Africa these climatic conditions are characteristic of the grasslands and savannas, with lightning as the so-called main "natural" ignition source. There is consensus that fires have been occurring in savannas for thousands of years, shaping the landscape and selecting for fire resistant/tolerant flora and fauna. It has been accepted that ancient fires were caused by natural systemic events such as lightning, but that as humans came to gather in social groups, they became the dominant source of ignition for wildfires (Goldammer and Crutzenan 1993). This is well illustrated in the savanna areas comprising the Kruger National Park (KNP) in South Africa, where anthropogenic activities have now become the dominant ignition source of fires (Trollope 1993).

Description of the KNP

Link to Govender Fig. 1
Link to Fig. 1, ~29K

(Back to top)
The Kruger National Park is one of the largest proclaimed and therefore officially protected natural areas in the world. Established in 1926, the park covers approximately 1 898 458 ha, occupying almost 2.5 percent of the total land surface area in South Africa (Figure 1). It is situated in the northeastern region of South Africa and is separated from adjoining Mozambique by the Lebombo mountain range in the east and from Zimbabwe by the Limpopo valley in the north. Elevations range from 260 m above sea level on the eastern plains to 839 m above sea level in the southwestern areas of the park. Although the mean average rainfall for the entire park is approximately 500mm, it varies from around 350 mm in the north to around 750 mm in the south. The rainfall regime within an annual climatic season can be described as highly erratic and usually confined to the summer months (September to March), and over a longer period can be described by extended wet and dry periods. These wet and dry periods have marked effects on the dynamics of the ecosystem and on the occurrence of fires. The park is also distinctively divided in two by its geology, with granitic sandy soils on the western half of the park and basaltic clay soils on the eastern half.

The vegetation of the park is a well-wooded savanna, dominated by trees in the genera Acacia, Combretum, Sclerocarya and Colophospermum. The flora of the park comprises +/- 2000 taxa, including over 400 tree and shrub species, and over 220 grasses. The fauna of the park includes 148 mammal and +/- 500 bird species. Megaherbivores are an important component of the fauna, and include elephant (Loxodonta africana), white and black rhinoceros (Ceratotherium simum and Diceros bicornis, respectively), and hippopotamus (Hippopotamus amphibious).

Fire history in KNP

(Back to top)
Very little is known of the fire regime in the KNP before the 1900s. It is assumed that before the arrival of humans to the area the main ignition source of fires was lightning. The first anthropogenic users of fires in the area are thought to have been the hunter-gathers, who used fire to entice game into their hunting area by encouraging the growth of palatable fodder. Later Bantu tribes and eventually European settlers used fires for agricultural and stock farming purposes (Brynard 1971). This period within the park's history, with very low numbers of staff in the early years and large tracts of land with no firebreaks, is often described as an era of uncontrolled and sporadic burning by local inhabitants and lightning.

During the early part of the period after establishment of the Kruger National Park in 1926 there was no official policy for burning range (also referred to as "veld" in South Africa). Fires were often regarded as having a detrimental effect on the vegetation. Seen as an unavoidable evil the range was also burnt, often during February, March and April, to provide short green growth during the winter months for game and sheep of farmers who held grazing rights within the area.

During the period 1926 to 1947, annual autumn burning of areas that had escaped accidental fires was applied mainly to provide new fodder for grazing. In an attempt to prevent fires burning outside the park's boundaries from crossing into the park, a network of firebreaks along the boundary was established in 1934. In 1937, after a "disastrous" fire, KNP warden Colonel James Stevenson-Hamilton remarked that firebreaks alone would not be able to prevent large fires and hence the range should be burnt in autumn at least every second year (Brynard 1971). Safety was not the only factor that influenced his decision. The rationale was also given that burning long old grass was necessary, as such grass provided no nourishment for animals but did harbor ticks and provide sites for ambush of grazers by predators.

In 1946, Col. Stevenson-Hamilton was succeeded by Col. J. A. B. Sandenbergh, who opposed range burning in all forms and "felt that deliberate burning in an area which should be kept in its natural state, must upset the natural balance" (Brynard 1971); hence, between 1948 and 1956 there was an attempt to reduce the impacts of fire, by ensuring wherever possible that no area burned more often than once every five years and then only after good spring rains.

Early in the 1950s, a new era of fire management in the park dawned and the involvement of management in the ecology of the system began to increase. This led to the appointment of a professional biologist and assistant in 1951, followed by the establishment of a long-term experimental burning trial in 1954. The initial objective of this experiment was to investigate the effects of season and frequency of burning on vegetation in the four major vegetation communities in the park (Van der Schijff 1958). To date the treatments in this long-term burning experiment have been judiciously applied (Biggs et al. 2003) and numerous projects, scientific papers and research questions and answers have emanated from the experiment (Govender, Potgieter and Biggs 2003).

During the period 1957 to 1980, regular prescribed burning was practiced every 3 years, in spring after the first rains, on fixed areas of around 4000 ha ("burn blocks"). The burn blocks (approximately 480) were delimited by a network of firebreaks and tourist roads established for that purpose. The reasoning behind this frequency and seasonal application of fires was that it would result in "cool" fires causing little damage to the flora and none to the fauna.

Between 1981 and 1991 a more flexible prescribed burning policy was applied, where fires in burning blocks were timed to take fuel loads, post-fire age and mean annual rainfall into account. Managers aimed to burn up to 50% of the park following high rainfall years, but only 20% following low rainfall years. The seasonal distribution was also varied to allow for some burns before the first spring rains, and in midsummer during high rainfall periods. The approach was intended to result in a more variable fire return period, a wider seasonal distribution of fires, and longer fire return periods in drier areas.

During the nearly forty-year implementation of a rotational burning program, where management set fires and suppressed lightning and "refugee" fires, phenomena were observed such as the growing dominance of grass species characteristic (in agricultural terms) of very poorly managed range, a decline in large tree densities, the continued practice of "ring burns" that allegedly led to large areas being burnt by very homogenous and intensely hot fires, and the lack of a flexible fire regime. These observations resulted in the park's adaptation of a so-called "natural" fire management policy in 1993. This new policy stipulated that all lightning ignited fires were to burn freely while other fires were to be suppressed. The reasoning for this change was due to the wilderness fire philosophy and addressed concerns about the negative impacts of perimeter fires (or "ring burns") on the Park's vegetation. The reasoning was that lighting ignitions (point fires) would yield a more variable, less intense fire regime and that the extent of areas burnt would vary widely, allowing for a patch mosaic burn pattern with different intensities over different areas and under different conditions. An intended rule of the policy was that all lightning fires should burn to their fullest natural extent and that rangers should carry and light the fires across man-made barriers, such as roads that would have stopped the fire spread. In practice this did not always prove to be very successful; analysis of the fire patterns during this era demonstrated that boundaries of most fire scars were either roads or management firebreaks.

By 2001, after an almost 10-year implementation of the lightning fire management policy in the KNP, it was clear that the objectives of the policy were not being achieved. Reasons for the collapse of the lightning policy can be summarized as follows. Firstly, the majority of fires that burnt during this era were usually ignited by illegal transmigrants traversing the park. Those fires that could have been deemed "desirable" were usually large runaway fires (due to the removal of many firebreaks), were often lit under hazardous conditions and according to the policy had to be combated. Secondly, the predominant lightning season in the park is during the wet season (September to March). The fact that the range is burnable from April onwards, whereas lightning ignitions were only expected in early spring, meant that other ignition sources were dominant during the dry season. Finally it was only during one year, 1996, that large accumulations of fuel in the late wet season of 1995 led to massive fires in the central and southern KNP; and it was only in this year that lightning fires actually burnt a larger surface area than anthropogenic fires. The dominance of anthropogenic ignitions over lightning constituted an exceedance of a threshold of potential concern (TPC) (van Wilgen, Biggs and Potgieter 1998), with illegal transmigrants driving the fire policy and not lightning fires. This led to one of two scenarios, either the recalibration of the present TPC or a rethinking of the fire management policy.

A new mission statement for the KNP included philosophical ideals such as "the maintenance of biodiversity in all its facets and fluxes" (Braack 1997) and this led to management re-evaluating the burning policy during mid-1997. Hence, in 2002 a new fire policy was established. This current policy combines point ignitions and wildfires in areas where fire is considered necessary, according to ecological criteria (grass sward composition and fuel biomass). In these areas, an annual fire target is established; this target is based on rainfall and distributed throughout the burning season as monthly targets. The policy further stipulates the suppression of fires in areas where fire is not considered necessary, and tolerance of all lightning fires. It is envisioned that this fire system will:

  1. result in the maintenance of biodiversity;
  2. provide a practical solution to the occurrence of large, extensive and highly dangerous unplanned wildfires by pre-empting many transmigrant fires and breaking up the fuel load, while managing such unplanned burns as do occur (whether from anthropogenic or other causes); and
  3. allow lightning fires to occur in a controlled way.

Integrated Fire Management Policy

Link to Govender Fig. 2
Link to Fig. 2, ~18K

(Back to top)
The Park is divided into three distinctive classes of FMUs (fire management units), where the application of fire within each unit will vary according to specific criteria. First, there are at least 22 large fire management units or LFMUs which are usually based on similar edaphic conditions. In addition, there are 12 designated pristine wilderness areas (PWAs) and six functioning concession areas (that is, areas of land within KNP that are leased on a long-term basis for private entrepreneurship). Each of these concession areas will constitute a totally independent small fire management unit (SMFU) (Figure 2).

Link to Govender Fig. 3
Link to Fig. 3, ~8K

At the end of the growing season (March/April), the veld condition assessments (VCA) will provide the baseline data (the botanical composition and the standing grass fuel load present) for estimating the percentage of surface area to be burnt in each of the FMUs by the end of the fire season (Figure 3). Under specific conditions (possible poor representativity of VCA sites, expectation of a drier season ahead etc.) the annual target to be burnt is allowed to be modified slightly according to the judgment of the ranger in charge.

Link to Govender Fig. 4
Link to Fig. 4, ~24K

The annual target percentage to be achieved by the end of the fire season is then translated into a month-by-month cumulative burn target for each FMU. The family of curves for the monthly distribution of percentage area to be burnt will vary for a wet (Figure 4a) or dry (Figure 4b) year. In combination with transmigrant fires, rangers aim for their monthly target, by applying patch fires from early on in the fire season (April) to the end of September. In order to keep somewhere near the target, adaptive "catching-up" or "slackening-off" is practiced as the season progresses. The application of patch burns by management will stop at the beginning of the lightning season (September) to then allow lightning a chance to contribute to the area burnt as a "natural" ignition source. During this period all anthropogenic fires will be combated as far as practically possible and lightning fires that occur will be allowed to burn to their greatest possible extent, with a ceiling of 50% of the FMU to be burnt by a lightning fire.

By setting the fire as point ignition (patch burns) and not along the perimeter of sites (ring burns), a wide range of fire types will result. The application of fires throughout the fire season, beginning from early autumn through to winter and spring, is designed to allow for significant variation in the range of fire intensities due to differences in the degree of curing and moisture content of the grass fuel. The frequency of burning under this system is also varied because the total area burnt during a season will depend upon the condition of the grass sward, which will vary naturally from season to season in response to normal variations in rainfall.

Details of the range condition criteria used to select areas to be considered for burning in the LFMUs are as follows:

  • The grass sward (outside of wilderness areas) must not be dominated by (i.e. must not consist of more than 50%) Increaser II grass species, which are pioneer species that increase with heavy utilization and are characterized by low grass species diversity.
  • The grass fuel load must be greater than approximately 4000 kg/ha because under these conditions the grass sward is in a moribund and unpalatable state requiring some form of defoliation to restore its vigor and grazing potential (Trollope and Potgieter 1985).
  • Point ignitions will be placed at random within areas of the LFMU that meet the above two range condition criteria.

Details of the burning condition criteria used to select areas to be considered for burning in the PWAs are as follows:

  • Only the target percentage to be burnt within PWAs is based on the grass fuel load, which must be greater than 4000 kg/ha.
  • Random ignition within PWAs will be set in order to achieve the monthly and annual targets.

Because of the special requirements and various safety factors within concession areas (SFMUs) in the park, a special hybrid "block-patch" procedure has been developed for these areas. This method of burning will be applied by the ranger if he identifies that the patch fire system is extremely risky to implement according to the current policy within SFMUs.

Major criticisms of the previous fire policies were that the application of fire to the system was too rigid and did not allow for climatic variations and that fire managers were not allowed to make any on-site decisions and were forced to combat all non-lightning fires. The new policy is flexible; targets are influenced by ecological and climatic variations in the ecosystem and fire managers are given considerable freedom of judgment and fire decision support tools to assist their implementation of the fire policy. These new methods of determining fire application for each type of FMU within the KNP should be also seen as a multi-scaled field experiment that will aim to test the effects of these three burning systems (range condition criteria, lightning "natural" criteria and an intensively management system) on the vegetation. The success or failure of the integrated burning system will be evaluated according to a series of ecological criteria that have been developed for the Kruger National Park and are referred to as the Threshold of Potential Concern (TPCs).

Analysis of the fifty-year fire historical data set maintained by the KNP has indicated that only particular factors (spatial heterogeneity, season and intensity) that influence the fire regime in the park can be affected by management strategies; the frequency of fires and total area burnt in a particular fire season is largely driven by climatic variables, such as rainfall, not the fire management strategy practiced at that particular time (van Wilgen et al., in press).

The new KNP fire objectives, which have evolved over decades of fire research will lead to a range of fire types, intensities and effects that will best maintain biodiversity. In addition, a few interesting and important programs of research in the park will provide crucial information and further our understanding of the role that fire plays within the savanna ecosystem. A team from the University of Cape Town, are investigating the role of fire and its influence on tree/grass dynamics; researchers from several South African universities (in particular the Universities of Witwatersrand and Stellenbosch) are working on the effect of fire on soil nutrients and nutrient cycling; and the U.S. National Science Foundation's biocomplexity program is working to develop a conceptual and model-based understanding of savanna ecosystems. Other major initiatives for fire research in the pipeline include a group of scientists from the Konza Prairie Biological Station attached to Kansas State University in the USA who are planning work on comparative grassland dynamics under herbivore influence. Such research both rests on and continues to build an extremely well documented, maintained and invaluable fire management history and large-scale fire experimental data set in the KNP. A recent account of the role fire plays in the KNP ecosystem is contained in van Wilgen et al. (2003).

After a long fire management history in the park and through many years of experience and learning, the KNP is now implementing a fire management policy that simulates as closely as possible the intensity, spatial and temporal heterogeneity of a natural fire regime, taking into account the importance and reality of anthropogenic fires in this system and thereby achieving its mandate of maintaining biodiversity in all its facets and fluxes. Together with international and local fire ecologists, the Kruger National Park is taking the lead in answering many fire-related questions with sound research, application and management of fires within savanna conservation areas.


(Back to top)
I would like to thank Harry Biggs and Andre Potgieter for their helpful comments on the article and Don Ntsala for the maps.


(Back to top)
Biggs, R., H.C. Biggs, T. Dunne, N. Govender, and A.L.F. Potgieter. 2003. What is the design of the Experimental Burning Plot (EBP) trial in the Kruger National Park? Koedoe 46: 1 - 15.

Bond, W. J. and B.W. van Wilgen. 1996. Fire and plants. London: Chapman and Hall.

Bond, W.J. 1997. Fire. In Vegetation of Southern Africa, ed. R.M. Cowling, D. M. Richardson and S.M. Pierce, 421-446. Cambridge: Cambridge University Press.

Braack, L.E.O. 1997. A revision of parts of the management plan for the Kruger National Park. Volume VIII: Policy proposals regarding issues relating to biodiversity maintenance, maintenance of wilderness qualities, and provision of human benefits. Skukuza: South African National Parks. Available online: (Scientific Services, Management Plan).

Brockett, B.H., H.C.Biggs and B.W. Van Wilgen. 2001. A patch-mosaic burning system for conservation areas in southern African savannas. International Journal of Wildland Fire 10: 169-183.

Brynard, A. M. 1971. Control burning in the Kruger National Park: History and development of the range burning policy. Proceedings: Tall Timbers Fire Ecology Conference 11: 219-231.

Goldammer, J.G. and Crutzen, P.J. 1993. Fire in the environment: Scientific rationale and summary of results of the Dahlem Workshop. In Fire in the environment: The ecological, atmospheric and climatic importance of vegetation fires, ed. P.J. Crutzen and J.G. Goldammer. John Wiley and Sons.

Govender N., A.L.F. Potgieter and H.C. Biggs. 2003. Scientific value of research conducted on the Experimental Burn Plots in the Kruger National Park. Proceeding: Fifth International Rangeland Conference, Durban.

Parr, C. L. and S.L. Chown. 2003. Burning issues for conservation: A critique of faunal fire research in Southern Africa. Austral Ecology 28: 384-395.

Scholes, R. J A. and B.H. Walker. 1993. An African savanna: Synthesis of the Nylsvely study. Cambridge: Cambridge University Press.

Trollope, W.S.W. and A.L.F. Potgieter., 1985. Fire behaviour in the Kruger National Park. Journal of the Grassland Society of Southern Africa 2: 17-22.

Trollope, W.S.W. 1993. Fire regime of the Kruger National Park for the period 1980-1992. Koedoe 36: 45-52.

van der Schijff, H. P. 1958. Inleidende verslag oor rangebrandnavorsing in die Nasionale Krugerwildtuin. Koedoe 1:60-93.

van Wilgen, B. W., N. Govender, H.C. Biggs, D. Ntsala, D. and X.N. Funda. In press. Manipulating savanna fire regimes to meet conservation objectives: Lessons from a large African national park. Conservation Biology.

van Wilgen, B.W., W.S.W Trollope, H.C. Biggs, A.L.F. Potgieter and B.H. Brockett. 2003. Fire as a driver of ecosystem variability. In The Kruger experience: Ecology and management of savanna heterogeneity, ed. J. du Toit, K.H. Rogers and H.C. Biggs, 149-170. Washington, D.C.: Island Press.

van Wilgen, B.W., H.C. Biggs, and A.L.F. Pogieter. 1998. Fire management and research in the Kruger National Park, with suggestions on the detection of thresholds of potential concern. Koedoe 41: 69-87.


bar denoting end of article text

Author information

(Back to top)
Navashni Govender is the Fire Ecologist at Kruger National Park. You can reach her for comment at

Additional web resources

(Back to top)
Kruger National Park web site

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