Zimbabwe: Fire and the Forest (IFFN No. 9 – 1993)


Fire and the Forest

(IFFN No. 9 – 1993, p. 23-27)

The state forest boundaries on the Kalahari Sands (KS) of northern Matebeleland, Zimbabwe, encompass an obviously very variable vegetation cover extending from Victoria Falls in the west to Mafungabusi in the east and to Chesa in the south east. Forest Management attention has been concentrated upon the commercially exploited timber species mkusi (Zambezi Teak, Baikiaea plurijuga), mtshibi (Guibourtia coleosperma), and in the last 30 years, mukwa (kiaat, Pterocarpus angolensis). Areas of mkusi in particular were the prime purpose for placing extensive areas under State control. Variation in topography, geomorphology, rainfall and frost are major determinants of the vegetation growing on a given area, vegetation then modified in structure and composition by man’s logging and other activities such as hunting, ranching and farming, and by the timing and frequency of dry-season fire.

Vegetation maps in varying detail are available for each forest. These should be used to delineate acceptably homogenous areas for which specific management objectives can be defined, and thus to which appropriate treatments can be applied. A generalized summary of the vegetation comprising the Forest Estate is presented in Figure 1. Potential fire management is then immediately apparent.


There is more than enough visual and experimental evidence of detrimental effects of fire in general upon woodland and forest intended for harvestable mkusi and its associates. Even one hot season fire, with an adequate weight and depth of fuel, may destroy perhaps 100 to 200 years’ accumulated growth in commercial terms: not the forest vegetation as such, for most of this will coppice, and in a century or two will be restored. Thus it is essential that mkusi be protected from uncontrolled or wildfire. 

On the other hand, natural (lightning-caused) fire has been a component of the veld for millennia, interspersed with the accidental or deliberate hunting fires of the original nomadic residents. There is a belief, to which I subscribe, that the rare to occasional fire in the late hot to early rainy season reaching aerially destructive proportions only in relatively small areas or patches of, say, 0.01 ha to a hectare or two before being extinguished by accompanying rain, is essential to the regeneration of the woodland or forest by thinning out competing vegetation. This will effectively reduce moisture competition and possibly increase light levels at or near ground level, thereby allowing either seedlings to establish and/or enabling established but effectively dormant rootstocks to grow up into the canopy. This fire I term a silvicultural fire.

Large-scale natural fires, apparently as devastating as those now caused frequently by man, have also undoubtedly occurred but scrutiny of available meteorological records suggests that the coincidence of heavy rainfall seasons to provide the fuel loadings, “black” radiation frosts to provide fuel depth, and dry lightning storms to provide ignition, occur at intervals of two or more decades. Small “patch” fires suggest an uneven-aged (or uneven-structured) forest, and large fires an even-aged (or even-structured) forest.

Management Implications

In summary, the Objects of Forest Management set out by F.B. Armitage in 1961 are directed towards the principal object of maintaining a tree cover over all forest or potential forest sites to protect the soil and ameliorate the climate. While placing the main emphasis upon the production of commercially exploitable species on a sustainable yield basis, increased total productivity by developing the use of minor forest products, applying multiple land use principles where feasible, and maintaining or increasing the soil and water conservation and amenity value of the forests, are also stipulated. The prevention, or specified application, of fire is a major tool in the hands of the forest manager.

The risk of wildfire can be reduced greatly by a carefully planned, vigorous and consistently applied extension programme such as that designed for the now-defunct Agricultural Unit. This implies active sincere adherence to a long-term programme of holistic resource management such as that propounded by Alan Savory. Today, forest occupants and neighbours – the major risk – are as integral a part of the forest ecosystem as the plants and animals. 

The hazard can be reduced in two ways – by burning or by physical removal. To use fire protection in mkusi woodland is a contradiction. For mkusi, fire is for silviculture, not for protection. Therefore in areas designated for mkusi, the fuel load needs to be physically removed. The only feasible large-scale technique is herbivory by larger mammals.

Wild herbivores, which have a valuable contribution to make but are sensitive to habitat changes, may of necessity be low in numbers and therefore in reduction effect, and difficult to manipulate. Domestic herbivores, i.e. cattle and goats, can be manipulated with relative ease and are readily harvested. 

It is obviously impossible to stock all mkusi areas with cattle at a rate to reduce the fuel hazard to tolerable levels every fire season because of the large fluctuation in food (grass growth) with varying rainfall. It is completely feasible to carry sufficient cattle to concentrate in order to render designated areas fuel-free, such as high risk zones, to create fuel breaks, or to disperse through areas for acceptable fuel reduction. Control can be by herding rather than by capital investment in fencing; watering and handling facilities can be portable not permanent. These are economic, logistical, decisions.


30899 Byte

Fig. 1. The Forest Estate: a summary of its vegetative resourcesor components, potential use and fire management


Logging is a transient but special case. Conventional practice entails dragging logs through the forest for up to 100m to a temporary access road used by trailers, which move the logs to the mill. This massive disturbance by repeated tractor movement, half of which is accompanied by dragged logs, effectively cultivates the logged area resulting in an explosive increase in dense grass and woody (largely coppice) regrowth. This may persist 10 to 20 years or more, depending on the original density and the rate of canopy closure. 

While this major hazard can be given extra attention to reduce risk, and extra precautions in terms of more intensive fuel breaks, fuel reduction (grazing and/or herding) and suppression emphasis, it would seem preferable to avoid it by adopting “Minimum Impact Logging” whereby logs are felled and loaded at stump, thereby eliminating virtually all dragging disturbance. Regrettably, a proposal to investigate the feasibility of minimum impact logging in comparison with conventional practice was vetoed in 1981/82 in the interests of “instant profit”. This investigation should be re-opened. 

Incidentally, the belief that herbaceous fuel levels in these forests build up cumulatively in proportion to the number of seasons since last burn is fallacious. Fuel levels grow to a peak at the end of the rains depending on seasonal rainfall, plant density and species composition of a given area. Fallen litter, if not entirely, breaks down in the following rainy season under soil faunal and floral activity. Perennial grasses can accumulate moribund material if not utilised, but an accumulated excess will effectively kill the plant in a few years. Fuel loadings are therefore primarily an annual event, dependent on seasonal rainfall and current grass and woody species composition and density. 

Silvicultural fire offers a range of options. Anticipating further research, I suggest that planned fires could be applied within defined limits to fuel loads, seasonal and diurnal timing, to thin out established stands and to create, temporarily, competition-free regeneration areas. 

The miombo woodlands could be grazed and/or early-burnt in the May-July period to prevent or reduce fire passage into mkusi areas. Early burning of miombo woodland proved successful in Zambia over several decades, and I know of no technical reason why it should not be equally successful here. It seems likely that fire would favour mukwa as a component of the miombo, which is economically desirable. 

Mukwa areas seem a subject lacking quantitative information and are confounded by the problems of decline and die back. Research into fire and mukwa has been needed for at least two decades; not possible yet due to staff and financial constraints. Nonetheless, I believe that periodic late (early-rainstorm) fire is beneficial to mukwa, healthy trees being apparently impervious to, for example, late-August (accidental) fire, which may also reduce moisture competition from fire-sensitive neighbouring species, by removing unhealthy specimens and making room for healthy regrowth. 

This suggestion applies particularly to the “isibabe” areas where an overwood of mukwa of varying density is found over a variable shrub understory, including “dwarf” mkusi. I cannot suggest a frequency other than a possible linkage with periodic general fruiting seasons, currently about three to ten years. 

The grasslands and vleis which intersperse the mkusi or miombo wooded ridges should be managed to complement fire protection of the mkusi, i.e. prevent the passage of unwanted fire. An advantage is to integrate vlei and grassland management into that of wild or domestic livestock. 

Again, there is a management conflict because for mkusi fire protection the grasslands should be burnt off relatively early before the hot fire season, whereas for veld management (i.e. the removal of moribund grass and/or the suppression of unwanted tree growth) fire needs to be applied relatively late at the time of new grass flush. This occurs at or before the onset of the rains, depending on residual soil moisture. 

By combining early burnt fire breaks with a rotational late burning plan, this conflict can be reconciled, as proved by the Insuza Vlei programme successfully implemented in the early 1980s. Burning must be based upon vegetation condition (senescence and density) and state of growth (at grass flush), not on the calendar. Frequency is seasonally dependent and can vary from possibly two to six years apart, or twice in four to twice in twelve years with variations between. Streambeds and sodic areas should be fire-protected to encourage water holding and to protect very sensitive topsoil respectively. 

Peat beds (Phragmites) must be protected. In a dry rainfall cycle these can, and do, catch fire and burn underground creating a sterile dust-bowl. The most prominent example is the Dete Vlei below the Hwange Safari Lodge, but the Jijima (Lugo Ranch) and Tobotobo (Eland Blocks) vleis are others. 

The basalts which fall outside the forest property and fire management must complement protection of mkusi. Complicated by forest occupancy, in general terms the riverine gallery forest should be protected from both fire and people. The skeletal rocky soils of the slopes, plus the pockets of alluvium and vertisols of the valley floors, offer sweet grazing (grasses which retain their crude protein levels in the dry season), invaluable in the winter dry season, together with mopane and other valuable browse species which should similarly be protected from wildfire.

Arable areas require little comment other than as a source of accidental fire when “burning” gardens. Efficient extension is the solution here. 


Despite the evidence of increasingly poor fire protection records from the last few years, I do not share the view that fire protection of mkusi to ecologically acceptable levels is virtually impossible. With positive commitment to the Objects of Management, careful consideration of the resources of the Forest Estate as a whole, and thoughtful planning so that these resources are integrated both ecologically and economically, I am still convinced that the indigenous State Forest can be managed (cost-) effectively for the purposes for which they were demarcated. 

People and domestic livestock are critical components of the management system. I have recently become aware of an attitude of negative economic pessimism which seems to have influenced indigenous State Forest management for at least the last 20 years. Time appears to be a major mental stumbling block in adopting a more positive attitude. That the indigenous forests are a slow growing community in which change and growth are measured in centuries or half centuries rather than decades, or even years, must be clearly understood and accepted. Unless a positive approach incorporating the patience entailed by these time scales is adopted, I believe effective management is doomed to failure. 


There is little, if any, originality in the above and I acknowledge the many colleagues and friends who have contributed their thoughts and experience over the last thirty years: Denys Fanshawe, the late Paddy O’Dare, Jimmy Sharp, the late Ian Farquhar, John Rushworth, the late Bill Rainford and Peter Butler, come immediately to mind, but there are many others. Also, my sincere appreciation of Jonathan Timberlake’s editorial help and criticism. The interpretation (and errors!) are mine.



From: G.M. Calvert
Matopos Research
Ministry of Agriculture, Zimbabwe
Private Bag K 5137

Country Notes


Print Friendly, PDF & Email
WP-Backgrounds Lite by InoPlugs Web Design and Juwelier Schönmann 1010 Wien