Managing fire in natural grasslands

January 30, 2017

Photo 1: Cymbopogon grass species are often seen along forest edges.

Managing fire to maintain healthy conservation areas and protect plantation trees....

By Dr Liza Joubert-van der Merwe, Mondi Ecological Wetland Programme, Stellenbosch University

Ecological Networks (ENs) among commercial timber compartments can perform equally well as nature reserves in supporting ecosystem services and conserving biodiversity when they are managed and designed well. Here, the term ‘management’ refers specifically to the management of disturbances in grasslands.

Disturbances can be divided into two categories: natural (e.g. burning and grazing) and anthropogenic disturbances (e.g. mowing and ploughing). While natural disturbances played a role in the evolutionary development of grasslands long before human settlement, anthropogenic disturbances were introduced by humans. However, the distinction is not always this clear (see Box 2 at the bottom of this post). This article is specifically about fire management aimed at conserving biodiversity, and incorporates aspects of management for optimizing grazing conditions in grasslands.

Yes, burn grassland
Burning is an essential part of management for grasslands in the eastern part of the country with >650 mm mean annual precipitation (MAP). If fire is excluded for an extended period of time, these grasslands will gradually develop into forest (Bond 2005). This succession is marked by (1) the accumulation of a dense mat of plant biomass, followed by (2) a change in grass composition (palatable Themeda triandra and Tristachya leucothrix → impalatable Aristida junciformis and Cymbopogon species) (Uys et al. 2004; Kirkman et al. 2014) (Photo 2 below). One reason for the invasion of A. junciformis into grassland left unburned for too long is because this species has thin, pointy leaves that allows it to penetrate through the thick mat of accumulated biomass. Tiller position (at or near the soil surface), and susceptibility to shading also play important roles in determining which grass species dominate plant assemblages in different successional stages.

After the shift in grass composition comes:  (3) a change in vegetation structure, as forest pioneer species (Leucosidea sericea and Halleria lucida) establish to pave the way for gradual development into forest (Photo 1 above). The change in vegetation structure, as grassland develops into woodland, can be relatively quick (10-20 years). Such succession over a large spatial scale is normally not encouraged, as this intermediate woody vegetation state is of little value for grassland biodiversity and grazing.

However, to maintain regional biodiversity, fire can be excluded from small, demarcated areas, if the overall aim is to protect or link up nearby indigenous forest patches, especially if they are dominated by old Yellowwood (Podocarpus fulcatus) trees.

Recommendation: Burn grassland routinely to maintain good quality grassland dominated by palatable grass species (e.g. T. triandra) and a wide variety of other plant and invertebrate species. The timing depends on the state of the grassland (see below). Only if there is a specific reason, such as the protection of old-growth indigenous forest patches, should fire be excluded from demarcated patches of grassland with the aim of establishing woodland.


Photo 2: A river at Good Hope Forestry Estate separates two vegetation types (grassland and woodland) created by differences in fire management. Fire exclusion on the right hand side of the river resulted in the influx of woody shrub and tree species.

Burning too frequently
Annual burning is probably the opposite of fire exclusion, and takes place predominantly in firebreaks (Photo 3 below). When comparing firebreaks to nearby grasslands with longer fire-return intervals (2-3 years), annual burning homogenizes and shifts the composition of plant and grasshopper assemblages (Joubert et al. 2014; Joubert et al. 2016). (Here, plants and grasshoppers are used as surrogates of biodiversity, while acknowledging that grasshoppers can eat much of the grass layer.) This means that all the firebreaks in different parts of the landscape provide habitat for the same plant and grasshopper species, which is not ideal. We want different species to inhabit the different local areas. Nevertheless, annual burning of firebreaks provide habitat for early-successional species. Hence, if confined to specific areas, this management practice can have value.

However, when annual burning is combined with cattle grazing, plant species richness declines and the abundance of many common grasshopper species increases. Therefore, we caution against the large-scale application of this fire regime.

Recommendation: Annual burning should only be allowed in firebreaks used for the protection of property (e.g. farms or forestry estates). Where possible, land managers should consider shifting firebreaks, e.g. burn either sides of a fence line in alternate years. Where annual burning is unavoidable, land managers should control livestock grazing diligently to prevent bare patches from appearing in firebreaks, which precedes invasion by alien bramble species and local extinction of plant species.


Photo 3: Annual burning homogenizes the landscape, but also creates habitat for early-successional species. When limited to only firebreaks, this management practice can have value for regional conservation efforts. However, when coupled with uncontrolled cattle grazing, bare patches start to appear leading to loss of plant species.

Manage veld condition, not fire
To avoid annual burning and fire exclusion are not very helpful management guidelines. However, these are the only two fire frequencies that have negative effects on biodiversity. Local biota are able to deal with any intermediate fire frequency, although this resilience might be further influenced by local weather conditions and grazing regimes.

If we want a more specific answer, we need to approach this question from a different angle. We actually know much more about how fire influences grazing conditions than we know about how fire influences biodiversity. It is for this grazing perspective that we consult Prof. Winston Trollope for further recommendations (2011). Prof. Trollope is an internationally-renowned fire specialist with decades of experience, especially in African grassland and savanna. First, he offers a change in perspective from managing fire to managing grassland. In other words, decisions to burn or not should be based upon the condition of grassland, rather than a burning program. Then, he follows this change in perspective with the recommendation to burn when plant biomass reaches 4 000 kg/ha. Using this guideline as the basis for burning accounts for the effect of periods of drought, historical differences in disturbances, and current consumption of biomass by grazers.

To test whether this recommendation works, we turn to the Ukulinga Research Station near Pietermaritzburg. The long-term above-ground production of plant biomass at this station is 423 ±102g/m2 (or 4 230kg/ha) per annum. (The variation is caused by differences in fire management and periods of drought.) Importantly, there are no domestic grazers at this research station, only the occasional buck. Hence, there is practically no consumption of plant biomass by herbivores. So what happens if we follow the recommendation to burn when there is a standing dry biomass of 4 tons / ha i.e. one year’s worth of production? A grass layer dominated by palatable T. triandra is maintained (Kirkman et al. 2014). If fire frequency is reduced to burning every 2 or 3 years, T. triandra is replaced by other grasses, such as unpalatable A. junciformis. Importantly, this means that T. triandra dominance is maintained by managing the moribund plant biomass, and not primarily by this specific fire frequency.

In the rest of KwaZulu-Natal with domestic grazers, this recommendation will translate into longer fire-return intervals.

Recommendation: Base your decision to burn on grassland condition, and not primarily on a burning program. It is good to burn when there is ~ 4 tons of dry plant biomass per hectare. Depending on stocking density, climatic conditions and disturbance history, this could imply burning every 2-6 years.


Photo 4: Decisions to burn should be based upon grassland condition, and not primarily by a burning programme. In one plot at the Ukulinga Research Farm, dominance of Themeda triandra (Red Grass) is maintained, because it is burned when dry plant biomass reaches -4 tons/ha. Towards the back of this landscape, in a plot with longer fire-return intervals, Aristida junciformis (white tussocks) dominate.

Patch mosaic burning
Differences in fire frequency allow us to assess the effect of time since last fire, which can be used as a measurement of the ‘age’ in grassland. Fire frequency sets a maximum limit to the age of grassland. For example, the age of annually-burned firebreaks is always less than 1 year. For biennially-burned grasslands, it cannot be more than 2 years, and so forth. During the first year after fire, grasshopper and plant species richness is greatest, most rare geophytes are flowering, and many indigenous game species graze on the flush of new growth. Thereafter, the accumulation of moribund grasses causes self-shading, with cooler and low-light conditions leading to a decline in plant diversity. Therefore, it appears as if it might be best to burn as often as possible to maintain species-rich, one-year-old grasslands. However, this does not do the complexity of grasslands and its associated biodiversity justice.

Although most plants and many animals use young grasslands (<1 year old), young grasslands cannot provide all the resources that species need to complete their life cycles. Hence, young grasslands cannot function as habitat on their own, and they need older grasslands to support them. (A habitat is a physical space in the landscape containing all the resources necessary for a species to complete its life cycle.) For example, Oribi (Ourebia ourebi) graze in recently-burned short grass, but they hide in patches of tall grass. The age of grassland is also pivotal for grassland specialist birds, such as the endemic Yellow-breasted Pipit (Anthus chloris) that needs unburned grassland patches to breed successfully (Little et al. 2013). Where the successful conservation of an invertebrate species, such as the Karkloof Blue Butterfly (Orachrysops ariadne), depends on the simultaneous protection of other invertebrate species, a patch mosaic burning pattern may also be of value (Armstrong and Louw 2013).

Recommendation: Maintain a mosaic of grassland patches of different ages to provide the habitat requirements of a diverse range of plant, invertebrate, bird and mammal species.


Photo 5: Burning mosaics of patches differing in age provide habitat for a wider range of species than if the entire landscape was burned in exactly the same way. Here, a gravel road separated two grassland patches that are burned biennially, but in alternate years. The bright green sheen at the top of the hill indicates that it was recently-burned, while the grassland in the foreground was not burned this year. Unburned grassland patches are critical for the conservation of grassland-nesting birds (e.g. Yellow-breasted Pipit) and fire-intolerant invertebrates.

Burning season
It does not hurt to cut your fingernails, except when you cut into the nail bed. In the same way, it does not hurt grassland to burn, except if you burn into the live parts. In areas with strong seasonal cycles of wet summers and dry winters, it is best to burn when the grass layer is moribund and dry. In summer rainfall areas, such as the KwaZulu-Natal Midlands, the dormant season is from autumn after the first frost (April / May) to a few weeks after the first spring rains (September/October).

Plant assemblages do not respond to burning at different times (autumn vs. winter vs. spring) within this dormant period.

However, the abundance and species richness of various ground living, flying and foliar arthropods were greater in sites burned in winter (August) than in sites burned in spring after the first rains (Chambers and Samways 1998; Uys and Hamer 2007). This is probably because invertebrate life cycles are strongly synchronized with seasonal cycles. Burning in spring (after eggs have hatched) can easily kill off a large proportion of juvenile invertebrates, including pollinators that perform a valuable ecosystem service. Therefore, if the aim is to conserve invertebrates associated with grasslands, it might be best to burn in winter.

However, there are legal constraints to consider when deciding on a suitable burning season. Hence, where it is not possible to burn in winter, I reiterate the importance of maintaining burning mosaics.

Recommendation: Burn in the dormant season when the grass layer is moribund and dry. If it is not possible to burn in winter due to legal constraints, take special care to leave unburned grassland patches that can act as habitat remnants for sensitive invertebrate species.


Photo 6: Burning in summer damages the grass sward, which leads to poor post-burn regrowth. Especially in areas renowned for their high-intensity summer thunderstorms, this management practice drastically increase grassland susceptibility to soil erosion and top soil loss.

Type of fire
Burning with and against the wind results in vastly different fires. Head fires burn upwards (i.e. into the sky), they are driven by the wind, and they pass quickly through the grass sward. Consequently, high temperatures are not maintained at grass root level for an extended period of time. In contrast, back fires burn against the wind, they take longer to burn through any specific patch of grass, and they transfer more heat energy downwards into the grass tussock. Consequently, sensitive buds inside the grass tussock suffer greater damage, and post-burn regrowth of grasses is poorer after a back burn.

Back burns are easier to control than head fires, which explain their popularity in the forestry sector. However, back burns may jeopardize the health of grass swards in ecological networks. A healthy and robust grass sward is desirable for biodiversity conservation and timber production purposes, as it limits top soil loss due to sheet erosion, especially in areas with steep slopes.

Therefore, it is recommended that management steer away from the practice of back burning, wherever possible, but without jeopardizing the economic viability of the plantation. Doing so could aid long-term sustainability of remaining natural grasslands.

Recommendation: Look for opportunities to burn with the wind without jeopardizing profitability of commercial land uses. For example, in a wide conservation corridor with forestry compartments on either side, a perimeter can be burned with back fires after which the remaining interior of the corridor can be burned with a head fire.


Photo 7: Back fires burn against the wind, and transfer more heat energy downwards into the grass tussock. The resultant damage to grass buds results in poorer regrowth than after a head fire.

Fire intensity
Bush encroachment is something that we often associate with grasslands that are overgrazed. However, especially in Zululand, a similar phenomenon is seen in many narrow corridors among commercial forestry compartments, which tend to be overgrown with woody shrub and tree species that established in recent decades. A very hot crown fire that is driven by a wind can be very severe i.e. it can kill off many of these woody species (Trollope 2011). However, this kind of fire presents a serious risk to the safety of firefighters and the profitability of forestry plantations and is, therefore, avoided at all cost. Most controlled burns in plantations are on cool days with little wind and moderate to high relative humidity, resulting in low fire intensity.

However, the current shift to mechanical harvesting of entire plantations (instead of individual compartments) represents an opportunity to diversify the fire regime. After mechanical harvesting, the remaining slush in compartments is burned in any case. Is it feasible to explore the option of very hot fires in narrow, overgrown corridors if they can only spread to compartments that have already been harvested and burned? If this option is feasible, we have to ask whether it is desirable. How important is the conservation of grasslands as opposed to woodland in the forestry context? Dr. René Gaigher of the Mondi Ecological Network Programme (MENP) is currently exploring the biodiversity value of these unmanaged, overgrown corridors in a forestry context. Hopefully, her findings will shed some light on this issue.

Recommendation: Maintain an open mind about the use of high intensity head fires to control bush encroachment in narrow conservation corridors, especially if the majority of the plantation has already been harvested.

Fire is only one disturbance used for the management of grasslands in ecological networks among forestry compartments. There are a few other disturbance, which I have not addressed in this article. I will discuss the effect of anthropogenic disturbances (grazing intensity and mowing of firebreaks) in the next article.


Photo 8: High intensity fires can be used to control influx of woody shrub and tree species in narrow corridors. Historically, these kinds of fires posed a serious risk to the safety of fire fighters, and the profitability of plantations.

Box 1: Definitions

Management practices: They are active interventions by land managers for a specific purpose e.g. crop production or pest control. Here, we talk about the deliberate application of fire for lending protection to properties (e.g. firebreaks), for removing moribund plant material to rejuvenate grass growth, and for controlling bush encroachment (or woody succession).

Fire regime: A term that describes 1) fire frequency, 2) season of a burn, 3) type of fire (head vs. back burn), 4) fire intensity (energy output, sometimes measured as flame length), 4) fire severity (i.e. measure of ecosystem impact), and 5) spatial arrangement of burns (Bond and Keeley 2005).

Box 2: History of changing fire regimes

Originally, before human influence, the fire regime was incredibly heterogeneous. Fires were ignited by lightning strikes and rock falls when there was enough dry fuel to catch fire. Fires burned hot and fast when driven by a dry berg wind on a hot day. But they were cool and their spread slow if there was little wind or if temperatures were low and relative humidity high. They stopped when it started raining, or when there was an interruption in flammable fuel, e.g. at forest edges, rocky cliffs, or at sites heavily-grazed and trampled by indigenous game. Only natural phenomena controlled how fast and far the burn extended, and how frequently fire returned. But because natural phenomena change all the time, the fire regime probably changed too.

Adaptation of local biota to variability in the historical fire regime probably accounts for much of the resilience in grassland ecosystems that we see today.

During the next period, which started about 100 000 years ago, San hunter-gatherers started to play a role in the fire regime. It is likely that the original fire regime from the first era persisted to a large extent due to low human population density and the nomadic nature of human life. However, humans were able to start fire; hence, an additional source of ignition was introduced. The ways in which early humans used fire were more diverse than today. There is the familiar example of using fire to attract indigenous game to the new flush of grass in burned patches. However, there is also a myriad of other examples in the book ‘Human beginnings in South Africa’ by Deacon and Deacon (1999) that illustrate the importance of fire as tool for human survival. The routine use of fire in caves enabled humans to compete with large cats for shelter in dark places. It was really a different world than the one we live in today. Furthermore, evidence is presented that geophytes with potato-like corms and bulbs, particularly Watsonia species, were ‘fire-managed’ to stimulate flowering, after which each plant divided to produce two new corms. These were dug up for food with fire-hardened stakes. Contrasting with the burning of medium to large patches to attract game and farm indigenous bulbous plants, there are also indications that fire was used on a micro-scale i.e. single plants.

Changing fire regime
The next major change in the fire regime was probably due to the start of pastoralism. Husbandry practices spread from North and East Africa southwards across the equator to Botswana, Zimbabwe and South Africa over a period of 5 000 years (Deacon and Deacon 1999). In South Africa, the earliest signs of pastoralism was about 2 000 years ago. Indigenous people had herds of sheep and cattle, as is evident from rock paintings of these animals throughout South Africa. The Khoikhoi herders and siNtu-speaking African farmers were nomadic, and migrated with their animals to make use of available water resources and pastures. The working assumption is that they used fire to improve forage quality for when they return with their animals to the same general area sometime in future. When the Europeans arrived in the 17th century, they reported seeing herds numbering thousands of head of cattle in the Western Cape. Also, there were different breeds of cattle kept in different parts of the country. In KwaZulu-Natal, the siNtu-speaking farmers had Sanga cattle, which is extremely adaptable to available forage (graze vs. browse) allowing them to do well in grassland and woodland. Coincidentally, their adaptability renders these cattle a very good fit for these dynamic ecosystems, especially if we consider how woody and grass biomass fluctuates in response to the local fire regime.

European settlers
There was again a change in fire regime with the influx of European settlers in KwaZulu-Natal between 1824 and 1857. No brief summary of history can do this 150-200 years justice. The changes in political power, human population size, human movement patterns (nomadic à permanent resident), land ownership, land use (livestock grazing à crop production and forestry), demographic structure, industry, and economy of scales were enormous. This reflected in the rate of change in the fire regime, which became more homogeneous (i.e. less variable) over time, primarily due to one factor: the ability to stop fire. Although fire is a natural disturbance, the current fire regime is largely anthropogenic and under human control. Official burning guidelines were recently published to balance the ecological benefit of burning against the threat of damage to infrastructure and natural resources (SANBI 2014).

*First published in SA Forestry magazine, Dec 2016

*This article is one of a series by Dr By Dr Liza Joubert-van der Merwe. Read the previous article: Designing functional ecological networks.

Notify of
1 Comment
Newest Most Voted
Inline Feedbacks
View all comments
linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram