Arid Lands Newsletter (link)No. 55, May/June 2004
Fire Ecology II

Prescribed burning in African grasslands and savannas for wildlife management

by Winston S.W. Trollope and Lynne A. Trollope

"The author's experience is that the range condition burning system is the simplest and most practical system for both large conservation areas and smaller, more intensively managed game ranches."



Introduction

Fig. 1 - thumbnail link
Link to Figure 1, ~43K

(Back to top)
Africa is known as the "Fire Continent" (Komarek 1965) and prescribed burning is a widely recognized and essential ecological factor in for managing its grassland and savanna ecosystems. Research investigating fire regime effects on the biotic and abiotic components of the ecosystem has been conducted in these regions since the early twentieth century. This has led to a general understanding of the effects of type and intensity of fire and season and frequency of burning on the grass and tree components of the vegetation. This in turn has clarified the use of fire as a range management practice, and viable prescribed burning programs have been developed for the grassland and savanna areas used for both livestock production and wildlife management. Experience gained through research on the effects and use of fire in southern and east African grasslands and savannas (Trollope 1983, 1989; Trollope and Potgieter 1986; Trollope and Trollope 1999, 2001; Trollope et al. 2000; van Wilgen et al. 1990) has led to the conclusion that the broad groups of grasses and trees generally react similarly to the different fire regime components and, therefore, general guidelines can be formulated for prescribed burning. This article focuses on the use of fire for wildlife management, a form of land use for which Africa is famous (Figure 1).


In African grassland and savanna areas used for nature conservation and game ranching there is general consensus that fire has occurred naturally since time immemorial and that it is often essential for the ecological well-being of these ecosystems (Bothma 1996; Thomson 1992; Trollope 1990). Nevertheless, views on the most appropriate burning system for wildlife areas vary widely. Initially, ideas on the use of fire were based on ecological equilibrium theory and burning was applied at a fixed return period. However, with the development of non-equilibrium theory of savanna dynamics, prescribed burning is now applied under more diverse conditions (van Wilgen et. al. 2003). Even within this new paradigm burning systems vary from so-called "natural" systems based entirely on lightning as the ignition source, to actively applied burning systems based on the condition of the rangelands. The literature indicates that fire management is best developed in southern Africa.

This paper considers the lightning burning system, the patch mosaic burning system and the range condition burning system. The integrated fire management system recently developed in the Kruger National Park in South Africa has also been recently considered in this publication; suffice it to say here that the Kruger burning system is a product of these three burning systems, specifically developed for the unique conditions and circumstances that exist in the Kruger National Park.

Lightning burning system

(Back to top)
This system assumes that lightning is the only natural ignition source and that the biotic and abiotic components of African grassland and savanna ecosystems have evolved in the presence of lightning-caused natural fires. Therefore, in extensive wildlife areas like national parks, a natural fire regime should be driven by lightning as the primary ignition source. 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 (Trollope and Potgieter 1985).

In southern Africa, lightning burning systems have been implemented in Namibia's Etosha National Park and in South Africa's Kruger National Park. In the Etosha National Park a burning program was initiated in 1981 to simulate a lightning burning system (van Wilgen et al. 1990). This program was based on rainfall data, the period since the last burn and the natural incidence of fires caused by lightning. The objective was to remove moribund grass material and combat bush encroachment for the benefit of the wildlife in the Park, as follows:

  • The Park was divided into 24 blocks using existing roads and firebreaks;
  • Approximately 150 rain gauges were installed for uniform coverage of the Park;
  • Resultant rainfall data were used to select blocks for burning, i.e.:
    • Blocks were considered for burning when their seasonal rainfall exceeded the 20 year mean annual rainfall;
    • From this subset, blocks with the longest period since the last burn were further selected. Where the period was similar, priority was given to blocks which had received the highest rainfall;
  • Blocks were selected until not more than 12% of the Park was designated for prescribed burning;
  • Lightning-caused fires were allowed to burn but were confined to the block in which they occurred, assistance being given to them to spread if they were limited by unnatural barriers like roads;
  • Conversely, unplanned anthropogenic fires were combated and limited to the smallest area possible (van Wilgen et al. 1990).

Fig. 2 - thumbnail link
Link to Figure 2, ~45K

The following additional criteria were implemented in 1992 by du Plessis (1997):

  • With the aid of the disc pasture meter, calibrated by Kannenberg (1992), the degree of moribundness of the grass sward in a block is also assessed:
    • If the degree of moribundness is >50% this increases the priority for a block to be burnt.
    • Conversely, if the degree of moribundness is <20% this reduces the priority for burning (Figure 2);
  • The grazing pressure of wildlife in selected or adjacent blocks for burning is also considered. If this pressure is high in both cases, the priority for burning the selected block is reduced. This prevents potential shortages in available forage for wildlife.

The following fire regime was used to implement the burning system:

  • Season of burn
    • Based on the natural occurrence of lightning fires during the onset of spring rains, prescribed burns are applied in spring during September and October (van Wilgen et al. 1990).
  • Frequency of burning
    • Using the aforementioned procedure, the estimated frequency of burning will be between 5.5 and 9.3 years depending upon seasonal rainfall (van Wilgen et al. 1990). This approximates Siegfried's conclusion (1981) that burning frequency in the Park should be at least once every 10 years.
  • Type of fire
    • Since the Park is partially burnt using blocks and lightning fires occurring as point ignitions, a wide diversity of fire types occur with this burning system.
  • Fire intensity
    • No mention is made of the fuel and atmospheric conditions under which the prescribed burns and lightning fires occur. However, as Etosha National Park is in an arid environment, the grass fuel loads would not be excessive; also, as burns are applied after the first spring rains, the resultant fire intensity would be moderate.

In a lightning burning system initiated in the Kruger National Park in 1992, all lightning-caused fires were allowed to burn freely throughout the Park. The previous policy of prescribed "block burning" (i.e. applying perimeter ignitions to burn blocks) was discontinued and all unplanned anthropogenic fires were extinguished as far as possible. This policy resulted from the belief that lightning should be the primary ignition source in a natural savanna ecosystem. It could also be justified by results from the Park's range condition monitoring program,which showed that pioneer grass species had increased by 40% from 1989 to 1994, indicating a significant decline in the biodiversity of perennial grass species. The pioneer species increase was largely attributed to too-frequent burning and the over-development of watering points, resulting in excessively high grazing intensities. Results from a long-term burning trial conducted in the Park since 1954 showed that with increasing burning frequencies, pioneer grass species also increased. This was ascribed to the more frequently burnt plots being more nutritionally attractive to, and therefore more heavily grazed by, herbivores like zebra and wildebeest (Trollope and Potgieter 1985).

However, between the system's 1992 inception and 2000, more than 75% of the area burnt in the Park was ignited by anthropogenic sources. Therefore, for all practical purposes lightning could not be regarded as the Park's dominant ignition source under current socio-economic conditions. In addition, the prevalence of unplanned fires led to management staff's having almost no control over the fires occurring in the Park. This led in turn to a request by the field staff for the lightning burning system to be modified in order to cope with the realities of different unplanned ignition sources and the need to apply prescribed burns for both ecological and fire control purposes. This request resulted in the development of the Integrated Fire Management System applied in the Park since 2002.

Patch mosaic burning system

(Back to top)
The patch mosaic burning system is being successfully used in the summer rainfall areas of the moist savannas in South Africa's Pilansberg National Park. Patch burning originated in Australia, where it has been applied in Uluru and Kakadu National Parks (Brockett et al. 2001). The basic assumption here is that the burning patterns of fires are an effective surrogate measure of biodiversity; that is, the more diverse the burning patterns, the higher the biodiversity. Diverse burning patterns are developed by applying a range of numerous burns from late autumn to spring (April to November), thereby attempting to develop a heterogeneous mosaic of species and structural diversity in the vegetation. It is assumed that this is an effective and efficient means of developing and maintaining biodiversity (Brockett et al. 2001; van Wilgen et al. 1998).

The system is implemented as follows:

Fig. 3 - thumbnail link
Link to Figure 3, ~16K

Fig. 4 - Thumbnail link
Link to Figure 4, ~25K

Fig. 5 - Thumbnail link
Link to Figure 5, ~16K

  • The total area to be burnt is determined as a function of the estimated standing grass fuel load at the end of the growing season in autumn (April). This can be estimated subjectively, or objectively using a disc pasture meter as developed by Bransby and Tainton (1977), and expressed in kilograms per hectare. (Figure 3). This relationship was established in the Kruger National Park by van Wilgen et al (2000).
  • The area to be burnt monthly during the burning window is estimated as a function of the total area to be burnt (Figure 4). The actual area burnt is continually monitored against the desired cumulative monthly burn target but is allowed to vary by 10% above or below the target area thereby providing an element of flexibility in applying the patch burns.
  • Patch burns are applied at random as point ignitions in the 50,000 ha (123,553 ac) area of the Park. The number of fires ignited is a function of the total area to be burnt during any one burning season. This ranges from a target of approximately 6 fires when 1% of the Park's total area is estimated to be burnt to approximately 71 fires when 50% of the total area is estimated to be burnt (Figure 5).
  • The timing of burns over the season depends upon the grass fuel load. When this is high fires should commence in April, peak in June and continue until November. With moderate grass fuel loads fires will commence in April, peak in June/July and continue until September. With low grass fuel loads, fires are applied rarely and are skewed to the latter part of the dormant season, commencing in June, peaking in August and ending in September (Brockett et al. 2001).

The following fire regime is associated with the application of this system:

  • Type of fire
    • Numerous types of fires ranging from head and back surface fires to crown fires occur with this burning system because fires are applied as point ignitions.
  • Fire intensity
    • Fire intensities range widely with this burning system because it is applied from autumn to spring, allowing for significant variations in the moisture content of the grass fuel and its consequent effects on fire intensity. Fire intensities also vary in response to different fuel loads associated with the application of this burning system.
  • Season of burn
    • Most patch burns are applied when the grass sward is dormant, except at the beginning of the burning season in autumn and at the end in spring; at these times, fires are applied while the grass sward is still physiologically active.
  • Frequency of burning
    • The patch mosaic burning system results in a variable burning frequency because the total area burnt during a season depends upon the grass fuel load, which varies annually in response to rainfall.

Post-fire range management, patch mosaic burning system

Applying a series of patch burns at regular intervals during the burning season has the effect of attracting grazing animals to a different, newly burnt area after each fire. This minimizes the potentially detrimental impact of heavy, continuous grazing in any one area after burning (Brockett et al. 2001).

An assessment of the burning patterns associated with eight years of applying this system has showed an increase in spatial heterogeneity of the fire patterns over the years. This was generally reflected as a greater number of smaller fires occurring during any one year. Besides the assumed desirable effects on biodiversity, this burning system has proven to be very cost-effective to apply and has significantly reduced the hazard of large-scale wildfires. Experience has also shown that this burning system is best suited to conservation areas larger than 20,000 ha (49,420 ac), which would generally exceed the potential size of large-scale fires occurring in practice in Africa (Brockett et al. 2001).


Range condition burning system

(Back to top)
This system was developed from a fire research program initiated in South Africa's Eastern Cape Province in 1968; it was extended to the Kruger National Park for consideration in 1982 and to East Africa in 1992 (Trollope 1971; Trollope and Potgieter 1985; Trollope and Trollope 1999). The system is is used both for domestic livestock systems and for wildlife management. It is based on empirical results and is appropriate for use in wildlife areas as it provides a practical means of improving and maintaining the species and habitat diversity of natural grassland and savanna ecosystems (Trollope 1971; Trollope et al. 1995).

In this system fire is used to remove moribund and/or unacceptable grass material, to create or maintain an optimum relationship between herbaceous and woody vegetation where necessary, and to encourage wildlife to move to less preferred areas in order to minimize the overuse of preferred areas. The basic philosophy is that the use of fire to achieve specific management objectives must be based on the vegetation's condition and known reaction to the different components of the fire regime i.e. type and intensity of fire and season and frequency of burning. The following ecological criteria are used to achieve the overall management objectives of improving and maintaining species and habitat diversity:

  • If the grass sward is dominated by pioneer (Increaser II) species from genera like Aristida, Eragrostis, Sporobolus, Tragus and Enneapogon, burning should not be considered as it will maintain the grass sward in a pioneer condition characterized by low diversity of perennial grasses. Conversely, if the grass sward is dominated by either climax or sub-climax (Increaser I & Decreaser) species from genera like Hyperthelia, Hyparrhenia, Themeda, Andropogon and Pennisetum, burning should be considered provided the target grassland or savanna is in a moribund condition. These criteria are intended to maintain a high diversity of perennial grass species.
  • If the grass fuel load is greater than 4000 kg/ha (3572 lb/ac), the grass sward is in a moribund condition and burning should be considered. The fuel load can easily be determined with a disc pasture meter calibrated for use in African grasslands and savannas. When the grass sward is in this condition in savanna areas it can also generate an intense enough fire to create or maintain an optimum balance between herbaceous and woody vegetation where necessary. This criterion is intended to maintain the grass sward in a vigorous and palatable condition that will ensure adequate forage for grazing animals and provide adequate resistance to accelerated soil erosion. It also allows the manipulation of savanna vegetation structures to promote or maintain habitat diversity.
  • All unplanned wildfires initiated by lightning or other causes can be allowed to burn freely if the condition of the vegetation fulfils the aforementioned criteria; otherwise, such fires must be extinguished as far as possible. This is because fire's effects on grassland and savanna vegetation are similar when burning conditions are similar, irrespective of the source of ignition.
  • Finally, limits must be set to the area burnt, to help ensure adequate supplies of forage for herbivorous wildlife. It is therefore recommended that not more than 50% of the area be burnt in moist grassland and savanna ecosystems (>700 mm [27.6 in] rainfall per year) and not more than 33% in arid grassland and savanna ecosystems (<500 mm [19.7 in] rainfall per year).

Fig. 6 - thumbnail link
Link to Figs. 6a and 6b, ~60K

The following fire regime is recommended when using the range condition burning system:

  • Type of fire
    • Designated burn areas can be ignited either as point ignitions or as normal perimeter ignitions involving the "block burning" of selected vegetation.
      • Point ignitions are used to develop a mosaic of different fire types due to changes in wind direction, air temperature and relative humidity during the burning period. The ensuing range of fire effects is intended to maximize habitat diversity in both the vegetation's grassland and woody components alike. Point ignitions are best suited to large conservation areas (generally greater than 20,000 ha [49,400 ac]) where intensive range management is not critical. It is recommended that the areas to be burnt be divided into thirds, with a point ignition being applied in the center of each third. This is to ensure that the majority of the designated area is burnt by a mosaic of different fire types in order to promote and maintain species and habitat diversity.
      • In smaller conservation areas and game ranches, point ignitions are too dangerous and difficult to control and perimeter ignitions are recommended. These can also be used when high intensity surface head and crown fires are needed to restore the optimum balance between herbaceous and woody vegetation (i.e. where trees and shrubs have become too dense) (Figure 6).

 

  • Fire intensity
    • Fire intensity is an important component of the fire regime and needs to be varied according to the reason for burning:
      • When burning to remove moribund and/or unacceptable grass material a cool fire of <1000 kJ/s/m is recommended; this can be achieved by burning when the air temperature is <20° C (<68° F) and the relative humidity >50%.
      • When burning to reverse an imbalance between herbaceous and woody vegetation caused by overly dense trees and shrubs, an intense fire of >2000 kJ/s/m is required. This can be achieved when the grass fuel load is >4000 kg/ha (>3572 lb/ac), the air temperature is >25° C (>77° F) and the relative humidity is <30%. This will cause a significant topkill of stems and branches of trees and shrubs up to a height of 3 m (~10 ft), making the vegetation more available for shorter browsing animal species.
      • In all cases the wind speed should not exceed 20 km/h (~12.5 mph) for safety reasons (Trollope 1999).
      • These detailed prescriptions are better suited to smaller, more intensively managed conservation areas and game ranches. In these situations prescribed burns are applied and completed in a matter of hours, so it is possible to burn under specific atmospheric conditions. Conversely, in large conservation areas it is more difficult to vary the fire intensity relative to the reason for burning, as the fires can burn for long periods. At best, different burning times can be chosen throughout the dormant season depending on the moisture content of the grass fuel and the suitability of predicted weather conditions.
    • Season of burning
      • In wildlife areas it is also recommended that prescribed burns be applied when the grass sward is dormant, to avoid any detrimental effects on its regrowth and basal cover. The burning window can extend over the entire dry season, with the actual timing of fires varying according to the reasons for burning. When burning to remove moribund grass material prescribed burns can be applied at any time during the dormant season. When burning to reduce the density and size of trees and shrubs, fires should be applied when the grass fuel is at its driest, to ensure a high-intensity fire.
    • Frequency of burning
      • When burning to remove moribund and/or unacceptable grass material the frequency of burning will depend upon the rate at which the grass sward becomes moribund (Trollope 1989), i.e. accumulated grass fuel load >4000 kg/ha (>3572 lb/ac). Therefore the frequency of burning will vary as a function of the stocking rate of grazing animals and the amount of rainfall an area receives.
      • When burning to reduce the density and size of trees and shrubs the frequency of burning cannot be prescribed; it will depend upon the stocking rate of browsing animals and/or the rate of regrowth of the trees and shrubs.

    Post-fire range management, range condition burning system

    Grazing after burning in wildlife areas is difficult to control. To prevent overgrazing the burnt area must exceed the short-term forage requirements of the grazing animals attracted to the highly palatable and nutritious post-burn regrowth; thus, relatively large areas should be burnt at any one time (Trollope 1992). Another strategy successfully used in southern Africa is to apply a series of patch burns at regular intervals throughout the dormant season . This ensures the grazing animals are attracted to each recently burnt area in turn, spreading the impact of grazing over the entire burnt area and avoiding the detrimental effects of heavy continuous grazing in any one part of the area (Brockett et al. 2001).

    The author's experience is that the range condition burning system is the simplest and most practical system for both large conservation areas and smaller, more intensively managed game ranches. It has the advantage of being based on empirical research results and large-scale experience in both southern and eastern Africa. It is therefore strongly recommended that this burning system be considered for use whenever prescribed burning is necessary in Africa's grasslands and savannas.

    References

    (Back to top)
    Bothma, J. du P. 1996. Game ranch management. Pretoria: Van Schaik Publishers.

    Bransby, D. I. and N. M. Tainton. 1977. The disc pasture meter: Possible applications in grazing management. Proceedings of the Grassland Society of Southern Africa 12:115-118.

    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(2):169-183.

    du Plessis, W. P. 1997. Refinements to the burning strategy in the Etosha National Park, Namibia. Koedoe 40(1): 63-76.

    Kannenberg, N. 1992. Grass-Biomasse in Savannenbiomen des Etosha National Park: Anwendung des Disc Pasture Meter (DPM). In Feuer in der Umwelt, ed. K. Weiss and J. Goldammer, 76-91. Freiburg, Germany: Max-Planck-Institut fur Chemie.

    Komarek, E. V. 1965. Fire ecology: Grasslands and man. Proceedings, Tall Timbers Fire Ecology Conference 4:169-220.

    Siegfried, W. R. 1981. The incidence of veld-fire in the Etosha National Park, 1970-1979. Madoqua 12:225-230.

    Thomson, R. 1992. The wildlife game. Westville, So. Africa: Nyala Wildlife Publications Trust.

    Trollope, W. S. W. 1971. Fire as a method of eradicating macchia vegetation in the Amatola Mountains of South Africa: Experiment and field scale results. Proceedings, Tall Timbers Fire Ecology Conference 10: 99-120

    _____ 1983. Control of bush encroachment with fire in the arid savannas of southeastern Africa. PhD thesis, University of Natal, Pietermaritzburg.

    _____ 1989. Veld burning as a management practice in livestock production. In Veld management in the Eastern Cape, ed. J.E. Danckwerts and W.R. Teague, 67-73. Pretoria: Government Printer.

    _____ 1990. Veld management with specific reference to game ranching in the grassland and savanna areas of South Africa. Koedoe 33(2):77-86.

    _____ 1992. Chapter 3: Veld management in grassland and savanna areas. In Guide to grasses of South Africa, ed. F. P. van Oudtshoorn, 45-56. Pretoria: Briza Publications.

    _____ 1999. The use of fire as a management tool. In Veld management in South Africa, ed. N. M. Tainton, 240-242. Pietermaritzburg, South Africa: University of Natal 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(2):17-22.

    _____, 1986. Estimating grass fuel loads with a disc pasture meter in the Kruger National Park. Journal of the Grassland Society of Southern Africa 4:148-152.

    Trollope, W. S. W. and L. A. Trollope. 1999. Report on the assessment of range condition and the fire ecology of the savanna vegetation on the Lewa Wildlife Conservancy in Kenya - 1998. Final Report. Alice, So. Africa: Univ. of Fort Hare, Dept. of Livestock & Pasture Science.

    _____ 2001. Fire management program for the National Reserve of Gile in Zambesia Province. Final Report. Alice, So. Africa: Univ. of Fort Hare, Dept. of Livestock & Pasture Science.

    Trollope, W. S. W., H. Biggs, A. L. F. Potgieter and N. Zambatis. 1995. A structured versus a wilderness approach to burning in the Kruger National Park in South Africa. In Rangelands in a sustainable biosphere, ed. N. E. West, 574-575. Denver, CO: Society for Range Management.

    Trollope, W. S. W., C. J. H. Hines and L. A. Trollope. 2000. Simplified techniques for assessing range condition in the East Caprivi region of Namibia. Final Report. Windhoek, Namibia: Directorate of Forestry, Namibia-Finland Forestry Programme.

    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, ed. J. G. Goldammer. Ecological Studies 84. Berlin: Springer-Verlag.

    van Wilgen, B. W., H. C. Biggs and A. L. F. Potgieter. 1998. Fire management and research in the Kruger National Park, with suggestions on the detection of thresholds of potential concern. Koedoe 41(1):69-87.

    van Wilgen, B.W., H. C. Biggs, S. P. O'Reagan and N. Mare'. 2000. A fire history of the savanna systems in the Kruger National Park, South Africa between 1941 and 1996. South African Journal of Science 96:167-178.

    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. T. du Toit, K. H. Rogers, and H. C. Biggs, 149-170. Washington, DC: Island Press.

 

bar denoting end of article text

Author information

(Back to top)
Dr. Winston S. W. Trollope is a professor in the Department of Livestock and Pasture Science at the University of Fort Hare, South Africa. Dr. Trollope can be reached for comment at winfire@procomp.co.za. Dr. Trollope's wife, Lynne A. Trollope, has collaborated with him on numerous ecological projects in southern and eastern Africa since 1970.

Additional web resources

(Back to top)
Global Fire Monitoring Center, University of Freiburg, Germany
http://www.uni-freiburg.de/fireglobe

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 #55 table of contents