A Structured versus a Laissez Faire Approach to Controlled Burning
in the Kruger National Park in South Africa

(IFFN No. 19 - September 1998,p. 70-75)

The Kruger National Park is situated in the north eastern portion of South Africa where it is contiguous to Mozambique in the east and Zimbabwe in the north. It comprises primarily arid savanna (473 mm p.a.) and is 1.9 million hectares in extent. The Park is one of the premier ecotourism centres in southern Africa and is the second oldest national park in Africa tracing its origin back to 1898. It has one of the greatest diversities of animal and plant life in the world and has as one of its management objectives the maintenance of species and structural diversity of the ecosystem. The attainment of this objective is achieved through active managerial involvement but based on the principle of management with minimum interference. One of the most important management practices that has been applied in the Kruger National Park since its establishment is controlled burning. This is because fire is recognised as a natural factor of the environment in the Park where it has occurred since time immemorial and has and still is one of the most important factors influencing the composition and structure of the savanna vegetation. Systematic controlled burning in different forms has been applied since 1954 to provide palatable, nutritious grazing for wildlife and maintain an optimum balance between grass and bush vegetation. A program for monitoring the condition of the rangelands in the Park since 1989 has indicated a steady decline in the forage production potential and a decrease in the diversity of perennial grass species. Excessively frequent burning has been identified as one of the likely causes for these changes and debate has developed on strategies to decrease this rate. One option is a laissez-faire approach which has been adopted in the Park and where fires caused by lightning are permitted to burn (but anthropogenic fires as far as possible are extinguished) believing that this will result in the development of a natural fire regime. Another option is a structured approach where controlled burning is applied following a decision-support system using ecological criteria. In both cases it is believed that this will result in a decrease in the frequency of burning in the Park and lead to sustainable and maximum biodiversity.


32797 Byte

Fig.1. Experimental prescribed burn in Kruger National Park during the Southern Africa Fire-Atmosphere Research Initiative (SAFARI-92). Photo: J.G.Goldammer


Analysis of the records of the fire regime based on the area burnt during the period 1985 to 1992 showed that the most important ignition source in the Park was controlled burning (47 %), which was applied on a rotational basis to remove moribund and/or unpalatable grass material and to maintain an optimum balance between grass and bush vegetation. Fires lit by refugees fleeing from Mozambique via the Park into South Africa accounted for 24 %. These homeless people light fires for protection, cooking and warmth which then spread either by accident or when left unattended. The next most important causes of fires have been grouped under a general heading others (20 %). This includes wildfires caused by poachers, tourists, arsonists, accidents and reasons unknown. Surprisingly lightning (10 %) was not a very significant ignition source and the probable reason for this was that ignition from other sources had a greater probability of causing fires under the form of management in the Park and political conditions in Mozambique.

The frequency of burning in the Kruger National Park is highly variable and is influenced by the annual rainfall and the grazing pressure exerted by the wildlife populations that fluctuate from year to year. The most commonly occurring types of fires in the Kruger National Park are surface fires burning in the grass sward either as head fires with the wind or back fires against the wind. Crown fires do occur when the aerial portions of trees and shrubs occasionally ignite during fierce high intensity head fires but these are the exception rather than the rule. Fire intensities vary primarily according to fuel load consequently the intensity of the fires will be influenced by the rate of accumulation of grass fuel as a product of the annual rainfall. Suffice it to say that fire intensities recorded during experimental burns applied during 1992 in the Park ranged from 8845 to 22 kJ/s/m for head fires and 160 to 20 kJ/s/m for back fires illustrating the wide range of intensities that can occur within and between different types of fires.

As a result of the development of simplified techniques for estimating the standing crop of grass material and assessing the condition of the grass sward in the Kruger National Park a program to monitor the condition of the rangelands was initiated in 1989. This comprises recording the standing crop of grass with a disc pasture meter and the presence and absence of 18 key grass species at 532 sampling sites located in the 35 landscape units. These data are used to monitor annually the grass fuel loads, the potential of the rangeland to produce grass forage and fuel and trends in the grazing intensity of the grass sward. An assessment of the condition of the rangelands in 1989 led to the conclusion that the majority of the rangelands were dominated by Increaser II grass species which indicates that the grass sward was being heavily utilised by grazing animals. Analysis of the data collected during 1994 indicates that Increaser II grass species increased by approximately 40%. This increase between 1989 and 1992 suggests that the rangelands are being overutilized and that the grazing intensity is non-sustainable in the long term. In the absence of any other objective measure of the biodiversity of the grass sward this would indicate that the current management of the rangelands is not achieving one of the primary objectives of biological management in the Kruger National Park viz. the maintenance of species diversity of the grass sward. The condition of this component of the vegetation is of particular importance as it is the primary source of forage for a large proportion of the ungulates and other wildlife species in the Park.

There is evidence to suggest that too frequent burning has also caused an increase in the proportion of Increaser II grass species in the rangelands. This is because an analysis of the grass fuel loads recorded in the Park during the period 1989 to 1992 indicates that the majority of the areas (73%) subjected to controlled burns during this period were in a non-moribund condition (<4000 kg/ha) and need not have been burnt. These data would suggest that the rangelands were being burnt too frequently during the burning policies that were implemented since 1980.

One of the solutions to the declining condition of the rangelands in the Kruger National Park is to modify the controlled burning program, particularly reducing the frequency of burning.

This was addressed in 1994 when a laissez faire burning policy was introduced to replace the previous subjectively structured burning program. This comprises a "hands off" policy where all fires ignited by lightning in the Park are permitted to burn freely whereas all anthropogenic fires must be extinguished as far as is possible. The basic philosophy of this laissez faire burning policy is that lightning should be regarded as the only natural ignition source and if left alone will result in the development of a natural fire regime as regards types and intensities of fires and the season and frequency of burning. It is reasoned that human interference in the functioning of one component of a natural ecosystem will eventually lead to interference in the other components with unforeseen consequences. The laissez faire burning policy is not based on any practical examples from elsewhere in the savannas of Africa but rather on the basic and sincere belief that if a natural area is large enough, as the Kruger National Park is assumed to be, the ecosystem will function normally in response to natural variations in the environment, for example seasonal variations in the rainfall.

Another option to modifying the burning program in the Kruger National Park in order to achieve the management objectives of the Park is a structured approach where controlled burning is applied following a decision support system using ecological criteria. The basic philosophy of the proposed structured approach to controlled burning is that the use of fire to achieve specific management objectives must be based on the condition of the vegetation and its known reaction to the different components of the fire regime i.e. type and intensity of fire and season and frequency of burning. Based on research results obtained in the Kruger National Park and elsewhere in the savanna areas of southern Africa the following ecological criteria can be used to apply a burning program that will achieve the management objective of maximising species and habitat diversity in the Park. These criteria can be applied using the infrastructure that was in existence up until 1994 and which comprised the Park being divided into 88 burn units and grouped into 23 management units. The burn units should be burnt on a rotational basis that varies according to seasonal rainfall and the condition of the grass sward as described by the botanical composition and the standing crop of grass during the annual monitoring of the condition of the rangeland in the different burn units. The following criteria can be used to identify the burn units that should be considered for burning.

i) If more than 50% (sourveld areas/high rainfall areas) or 33% (sweetveld areas/low rainfall areas) of the management unit has been accidentally burnt by either lightning, poachers or refugees then no burn units should be considered for burning. The 50% and 33% limits are based on field experience that in sourveld the maximum frequency of burning should not be more than biennial and in sweetveld triennial.

ii) If more than 10% of a burn unit has been burnt accidentally then that unit must become a preferred area for a controlled burn provided that it fulfils the other criteria for burning. It is generally undesirable to have a limited area burnt in a large unburnt area because it leads to excessively heavy utilization of the rangeland in the burnt area and a decline in the species diversity as reported on earlier.

iii) If Increaser II grass species are dominant in the burn unit the rangeland must not be burnt in order to allow the grassland to develop towards the Decreaser stage which will have a greater species diversity as reported on earlier.

iv) If the fuel load of grass in the burn unit is less than 4000 kg/ha then the rangeland should not be burnt because it is not in a moribund state and will not generate an intense enough fire to control bush encroachment if necessary.

v) Ideally the size and density of the bush should also be assessed when considering whether a burn unit should be burnt or not. If the bush is becoming too dense then a high intensity fire is necessary whereas if the bush is in optimum condition for habitat requirements, a cool fire is preferable.

vi) Finally all wildfires caused by either lightning, poachers, refugees or other causes should be allowed to burn freely if the affected burn units fulfil the aforementioned ecological criteria otherwise they must be controlled as far as is possible. This is because the effect of fire on rangeland will be similar irrespective of the source of ignition if burnt under the same conditions.

It is proposed that the burn units be ignited around the perimeters as in the past both for security reasons and increasing the range of types of fires burning the rangeland. It is concluded that the burn units are large enough to result in the fire front fragmenting into individual fires and spreading through the burn unit as a mosaic of different types of fires in response to changes in the direction of the wind, air temperature and relative humidity during the extended burning period. This fire mosaic will ensure a range of fire effects which will result in the maximization of habitat diversity both in the grassland and bush components of the vegetation.

It must be acknowledged at the outset that an objective assessment of the proposed structured versus laissez faire burning policy for the Kruger National Park is very difficult as it is fraught with subjectivity from both points of view. For example regarding lightning as being the only natural and therefore permissible source of ignition in the Park, will be logistically difficult to implement. This is because during the period 1985 to 1992 fires caused by lightning burnt only 10% of the Park whereas fires caused by poachers and Mozambique refugees burnt 44% of the Park despite active attempts by the rangers to control such fires. It is therefore highly likely that with the cessation of controlled burning in 1994 the dominant sources of ignition will still be anthropogenic rather than lightning because the probability of fires being caused by poachers and refugees will continue to be much greater than the occurrence of dry lightning storms in view of the enormous human populations surrounding the Park. It is for this reason that it is proposed in the structured approach fires ignited by any form of ignition be permitted to burn provided the condition of the rangeland fulfils the ecological criteria. However in order to ensure the achievement of the management objectives of the Park controlled burns applied according to the proposed ecological criteria should be the dominant ignition source.

One of the assumptions in the laissez faire burning policy is that it will automatically achieve the management objectives of the park because this is the perceived conditions of the rangelands prior to settlement in South Africa when there was no or very limited pro-active management such as controlled burning. It is believed that such an assumption could be valid if the Kruger National Park could be assumed to be a completely natural operating ecosystem. This unfortunately is not because approximately 96% of the Park is sweetveld which disqualifies it as being an ecological unit where major migrations of grazing ungulates between sweetveld and sourveld would naturally occur and provide periods of rest for the rangeland to maintain its condition. The boundary fence around the park also severely limits permanent presence of animals in sweetveld areas which were previously occupied on an intermittent basis depending upon the availability of drinking water. Also the extensive network of roads has created completely unnatural barriers that limit the natural spread of fires that would have occurred prior to settlement. It is therefore believed that proactive management such as controlled burning is unavoidable if the biological management objectives of the Park are to be achieved in a situation where some of the important natural factors of the environment no longer exist.

Finally there are no examples of a laissez faire burning policy based on lightning as the primary ignition being successfully applied in any of the savanna areas of Africa which can be used as a source of reference. Therefore such a policy is an hypothesis at this stage which must be tested over an extended period of time (20-30 years) before any valid conclusions as to its efficacy can be drawn. In contrast the proposed structured burning program for the Kruger National Park is based on research findings from long term fire experiments conducted in the Park and elsewhere in the savanna areas of southern Africa. Furthermore the structured program has been tested in terms of its possible effects on the frequency of burning by applying the ecological criteria retrospectively to the range condition data for the different burn units for the period 1989 to 1992. The results showed that if the ecological criteria had been applied to the burn units that were rotationally burnt during this period the extent of the area burnt would have been reduced by 52%. Interestingly enough on presenting these results to the management staff of the Kruger National Park their reaction from a management perspective to a reduction in the burning frequency was very positive. In conclusion there are examples of successful burning programs based on scientifically tested cause and effect research results in southern Africa which demonstrate that a structured approach is both logistically possible and ecologically acceptable for managing wildlife areas. Examples of these are to be found in the Mpofu and Tsolwana Game Reserves in the Eastern Cape Province and in the game reserves in the Drakensberg administered by the Natal Parks Board where controlled burning is an important management practice.


Gordon, R. 1983. The National Parks of South Africa. Struik Publishers, Cape Town, South Africa

Joubert, S.G.J. 1988. Master plan for the management of the Kruger National Park. Vol 1-6. Unpubl. policy document. National Parks Board of South Africa.

Paynter, D., and W.Nussey. 1986. Kruger - Portrait of a National Park. Macmillan South Africa (Publishers) Pty, Ltd, Braamfontein, South Africa.

Trollope, W.S.W. 1993. Fire regime of the Kruger National Park for the period 1980-1992. Koedoe 36 (2), 45-52.

Trollope, W.S.W, and A.L.F.Potgieter. 1985. Fire behaviour in the Kruger National Park. J.Grassland Soc. South Afr. 2 (2), 17-22.

Trollope, W.S.W., W.P.D.Gertenbach, A.L.F.Potgieter, and N.Zambatis. 1991. Monitoring range condition in the Kruger National Park. Fourth International Rangeland Congress. Montpelier, France, 336-339.

Van Wilgen, B.W., C.S.Everson, and W.S.W.Trollope. 1990. Fire management in southern Africa: Some examples of current objectives, practices and problems. In: Fire in the tropical biota. Ecosystem processes and global challenges (J.G.Goldammer, ed.), 179-215. Ecological Studies 84, Springer-Verlag, Berlin-Heidelberg-New York, 497 p.



Winston S.W.Trollope


Department Livestock & Pasture Science
Faculty of Agriculture, University Fort Hare
P Bag X1314
Alice, 5700
South Africa

IFFN No. 19
Country Notes