«VELD BURNING IN THE KRUGER NATIONAL PARK J SOUTH AFRICA-AN ADAPTIVE APPROACH Willem P. D. Gertenbach SUMMARY A rigid rotational burning programme was ...»
VELD BURNING IN THE KRUGER NATIONAL PARK J
SOUTH AFRICA-AN ADAPTIVE APPROACH
Willem P. D. Gertenbach
A rigid rotational burning programme was applied for at least 70 years in
the Kruger National Park. Research results from the experimental plots for
the last 35 years, as well as research into a natural fire regime, revealed a lack
of flexibility in the programme to allow for change in frequency of veld fires based on accumulation of fuel. An adaptive approach was recommended to ensure that frequency of veldfires and season of burn should try to simulate a' natural fire regime more closely.
INTRODUCTIONSite description The Kruger National Park (KNP) is situated in the eastern Transvaal Lowveld of South Africa on the boundary of Mocambique. It covers an area of ± 2 million hectares. The KNP experiences a subtropical climate with warm wet summers, with temperatures of above 40°C as a general phenomenon, and mild winters with frost as an exception. The annual rainfall ranges from 700 mm in the south to 450 mm in the northern section of the KNP.
The diversity of geological formations occurring in the KNP, together with diversity in climatic conditions and slopes, result in a variety of soil types.
The variety of soil types support a variety of vegetation types, varying from open savanna to dense woodland. These differences in vegetation offer different habitats for animal species. The KNP can therefore be described as a diverse system of interdependent abiotic and biotic components functioning in a near-natural manner which was protected against human exploitation for the last 87 years.
The diverse nature of the KNP can be seen in the more than 2000 vascular plant species, 50 species of freshwater fish, 33 amphibian species, 112 reptile species, and 137 species of mammals ranging from small mammals like the fieldmouse (Mastomys natalensis) to large mammals like the African elephant (Loxodonta ajricana) and 490 species of birds. The 2000 vascular plant species consist of 325 species of grasses, 200 species of trees, 500 shrub species, 900 species of forbs, and a few species of creepers and ferns.
History of veldfires Fire played a role in the development and migration of plant communities in South Africa long before human settlement (West 1965). Because of this age-old influence of fire on the ecosystems, it, like climate, must be regarded as a natural phenomenon which contributed toward the establishment of a climax vegetation. Fire occurring under such circumstances in the absence of man and initiated by lightning can be regarded as a natural fire regime (1l'ollope 1987). In the conservation and management of natural areas like KNP knowledge of a natural fire regime is of utmost importance and it will be highlighted later in this paper. Such a natural fire regime prevailed in the area currently known as KNP for centuries before any human settlement (Brynard 1971).
When man migrated to different parts of the world, he took With him the ability to manipulate his environment. West (1965) puts it as follows: "Fires resulting from such causes (lightning) probably played their part in moulding the earth's vegetation and animal life long before the advent of Man, but the full potency of fire in its ability to affect vegetation by producing and maintaining fire sub-climax types was probably not reached until Man discovered how to make fire and to use it in furthering his activities:' Proof of the first people that used fire is scarce but Maggs (1976) claims that it was during the Middle Stone Age (150,000-180,000 BC). Manmade fires in the area of the present KNP date back for at least 100,000 years (Meyer 1986).
The fire regime during early black occupation can be divided into two eras.
Firstly, the period between 100,000 BC to 1500 AD was cfiaracterized by sporadic occupation by primitive peaceful people of the hunter/gatherer type which lived in harmony with their environment also as far as a fire regime is concerned. Secondly, the shorter period from 1500 AD to 1800 AD was characterized by more permanent occupation by aggressive stock farmers, hunters, and agriculturalists who also used fire in warfare. They had a greater influence 'on their environment also as far as the fire regime is concerned.
With the migration of white men to the Lowveld the status quo was almost the same as with the black stock farmers except for the period 1896-97 when almost all the game and domestic stock in the Lowveld was destroyed in a runderpest outbreak, grazing pressure diminished and more frequent fires occurred. " With the proclamation of the Sabie Game Reserve (later to be KNP) in 1902 a policy of annual burning during autumn was adopted (Brynard 1964).
The period from 1926 to 1934 was characterized by the absence of any fire because of severe drought (Stevenson-Hamilton 1943). After 1934 a policy of biennial burns in autumn was adopted. Colonel Sandenberg became Warden of KNP in 1946 and immediately changed the policy to a burn every five years and in spring. This change of the burning season to spring was a change toward a more natural fire regime.
Rotational Burning Programme In 1952 a fire ecologist was appointed to the staff of KNP and he initiated a series of experiments to test the influence of frequency and season of fire in different vegetation types. In 1954 the triennial rotational burning programme was implemented which stated: "Until it is proved to be wrong, it be laid down by the Board as an interim policy that the whole of the KNP be. divided into sections, separated by properly constructed firebreaks and that all grass which has become long and rank be burnt every three years on the understanding that only one third of each section be burnt annually and as late as possible in spring after the first rains:' Hundreds of kilometers of firebreaks were constructed and the area was divided into ± 400 blocks. Each of these blocks was burnt every three years in spring in such a way that the burnt areas were scattered over the whole KNP. This triennial rotational burning programme lasted until 1975.
During 1975 the burning programme was slightly altered to accommodate burns before and after rains in spring and also in midsummer in cases where special habitat manipulation was necessary for specific game species (Van Wyk 1975).
Methods The fire experimental plots were laid out in 1954. The objective of the experiment was to provide knowledge on the most appropriate burning season and frequency to be applied in the dominant vegetation types'in the KNP (Van der Schijff 1959). A series of twelve different treatments was applied to each of four replicates in four of the most dominant vegetation types. The treatments were: an annual burn in August; a biennial burn in August, October, December, February and April; a triennial burn in August, October, December, February and April; and a control plot with no burn at all.
The experimental plots were used to compare the effect of different fire treatments on the vegetation structure and composition, the composition of small mammal species, as well as the soil characteristics. Treatments were applied for the last 35 years.
Data on the frequency, season, and range of lightning fires and any other fires have been collected since 1946. This data can be useful in assessing the functioning of a natural fire regime for the area. Responses of the vegetation to the current triennial rotational burning programme were also carefully monitored using remote sensing techniques. An aerial survey of KNP is conducted every year during the dry season (Joubert 1983). Among other aspects, the surveys include the numbers and location of all larger game species, distribution of water, tree and shrub phenology, an estimate of grass-cover, grass greenness, accumulated litter, and areas burned during the previous season. This data is stored in a computer for every burning block of KNP.
When the survey is completed a map can be printed of all the areas in KNP with a high accumulation,of litter. An annual veld condition aSSessment at the end of the growing season (Trollqpe et al. 1988) also focuses on the buildup of fuel in the veld. This data provides additional input to facilitate the decision to burn or not to burn during the next season.
Results Results of these surveys and trials have been published by several researchers (Van der Schijff 1958, Brynard & Pienaar 1960, Brynard 1964 & 1971, Van Wyk 1971 & 1975, Potgieter 1974, Gertenbach & Potgieter 1975 & 1979, Webber 1979, Fraser 1983, Kern 1981, Trollope & Potgieter1985, GerteIlbach 1987).
A summary of the results obtained from the experimental plots is presented below for the herbaceous and woody layers.
The Herbaceous Layer
1. In unburnt veld, the grass layer becomes moribund and eventually dies out with a simultaneous increase in the occurrence of forbs (Gertenbach & Potgieter 1979).
2. Grass on an unburnt plot is essentially unutilized (Van Wyk 1971, Van der Schijff 1958).
3. Unpalatable grass material can be removed by regular burning.
4. Regular burning can be useful in preventing selective grazing (Van Wyk 1971).
5. Veld burnt too frequently shows signs of deterioration. There is an increase in unpalatable species and a simultaneous decrease in palatable species (Fig: 1).
Figure 1. Percentage basal cover of Bothriochloa radicans, Digitaria eriantha, and Themeda triandra on the fire experimental plots of Satara, Kruger National Park (Gertenbach & Potgieter 1975).
6. Veldfires in August and October result in an unnaturally high grazing pressure after the burn with detrimental effects to the herbaceous layer.
7. Fireduring the dormant phase is less detrimental to the grass tuft than fires when the grass is actively growing (ltollope 1987).
8. The effect of grazing after a burn is as important as the effect of the burn itself.
9. The desired frequency of burning is very closely related to the rainfall and to the speed at which undecomposedlitter accumulates. This depends on rainfall and utili~ation (Gertenbach & Potgieter 1979).
The Woody Vegetation
1. Unburnt veld becomes dense and unacceptable for most plains-loving game.
2. Too frequent burns can also result in bush encroachment. Better control of bush occurs with biennial and triennial burns.
3. Regular burns in August, October, and December are more capable of controlling bush encroachment than burns in February or April.
4. Most of the woody species in the KNP are adapted to and cannot be eradicated by fire. Regular fires only influence the horizontal structure of the woody vegetation.
5. Burning is more effective in regions of high rainfan than in more arid regions (West 1965, Scott 1971, Trollope 1987). Therefore lower burning. frequency under higher rainfall.
6. The effect of burning on a specific vegetation type differs from wet to dry climatic cycles (Gertenbach 1980).
7. The effect of veldburning differs from one vegetation type to another (Brynard 1971, Gertenbach & Potgieter 1979).
8. No significant change in species composition of the woody.layer occurred since the commencement of the experiment (Van Wyk 1971).
Because no sound conclusions could be drawn from the experimental plots and because the experimental plots only give account for four of the vegetation types in KNP, research efforts were directed towards an understanding of the functioning of a natural fire regime and the drafting of an adaptive burning programme to simulate a natural regime as far as possible. By adopting this approach the chances of making erroneous management decisions are minimized.
Research on the geological, climatological, pedological, vegetational, and zoological aspects of KNP advanced to such a point that it was possible to divide KNP into landscape units based on these different attributes. (Gertenbach 1983). These landscapes are also used as management units for the purpose of a burning programme. Management objectives were drawn up for each ofthese landscapes and a new burning programme tried to meet the requirements of these objectives.
Analysis of the rainfall data for the KNP revealed a cyclic pattern which implies a period of about ten years of above average rainfall, succeeded by a period of about ten years of below average rainfall (Gertenbach 1980). A close relationship was also found between the accumulation of litter in the.
wet spells and the frequency of lightning fires (Fig. 2). This indicated that a regular burriing frequency which does not consider the climatic cycles and therefore the accumulation of litter, is unnatural and that the frequency of.
veldfires should be adapted to correspond with the climatic cycles and the buildup of grass fuel in the veld.
Figure 3 shows a graph of the buildup of grass fuel during a wet and dry. cycle. If 3000 kg/ha is the level at which a lightning fire would occur, then this level will be reached in two years in a wet cycle, but it may take up to four years to reach this level in a dry cycle. This once again implies that burning frequency must be adapted according to rainfall or climatic cycles and the buildup of grass fuel.
Research showed that the ground lightning density (OLD) (Fig. 4) increases during the summer months. The moisture content of the grass layer also increases during the summer months. The result is that the success of a lightning strike to set fire to veld decreases during the summer months and also decreases in the winter as a result of lack of thunderstorms and the associated Figure 4.. Average ground lightning density (GLD) and percentage grass moisture for Kruger National Park (Gertenbach 1989)
lightning. The only season of the 'year when the OLD is high enough and the grass moisture is low enough to start a fire is during early spring. In wet' spells with a high buildup of fuel, lightning fires occurred almost any time of the year with a peak in early spring. The incidence of these fires in dry spells was restricted to early spring. These observations are substantiated by figures from the KNP and Etosha (Thble 1). A burning programme should therefore be adaptive to accommodate such a phenomenon.