Fire plays a fundamental role in maintaining ecosystem functioning in many South African vegetation types (Bond & Keeley 2005). It is a common and regular driver of change in many ecosystems and biomes across the globe, from boreal forests to tropical grasslands. Fire influences species composition, vegetation structure and dynamics (Bond & Van Wilgen 1996), particularly in grassland and savanna systems. In grassland and savanna biomes, fire frequency, extent and intensity can be a major determinant factor in the abundance of woody and herbaceous species (Bond, Woodward & Midgley 2005; Cavender-Bares & Reich 2012). Fire can be used as a management tool in order to promote ecosystem benefits.

Grasslands are the most primitive evolving systems of diverse plant communities (Little, Hockey & Jansen 2015). South African grasslands are being increasingly degraded through the increasing influence of overgrazing (Neke & Du Plessis 2004; O’Connor 1985), extensive and frequent burning (Uys, Bond & Everson 2004), plantation forestry (Lipsey & Hockey 2010) and invasion by alien plant species (Le Maitre et al. 1996). It is estimated that 60% of the grassland biome has been permanently transformed, while as little as 15% remains as natural grassland with only 2% formally conserved in South Africa (Carbutt et al. 2011; Macdonald 1989). What is alarming is that the majority of the natural grassland is vastly fragmented and most are ill-managed (Mucina & Rutherford 2006).

Mountain Zebra National Park (MZNP) is a unique conservation area occurring at the interface of three biomes, that is, grassland, Nama Karoo and thicket. Fire has recently become more common in the Nama Karoo and other arid ecosystems, potentially threatening the persistence of native species (Du Toit, O’Connor & Van den Berg 2015; Syphard, Keeley & Abatzoglou 2017). Climate change may be the reason why there is an increase in fire-conducive weather conditions (Kraaij, Cowling & Van Wilgen 2012; Wilson et al. 2010), changes in the structural composition of vegetation affecting fuel attributes of ecosystems (D’Antonio & Vitousek 1992; Syphard et al. 2017). Just like other grassland vegetation types, MZNP mountainous grassland is prone to occasional lightning-caused fires (Bezuidenhou & Brown 2016). Studies from the eastern Karoo suggest that fire can significantly modify vegetation structure and composition, to a similar degree as grazing and herbivory (Du Toit et al. 2015). The complexity of the interaction between grazing, herbivory and climate in the region has been broadly acknowledged (Kraaij & Milton 2006). Fire plays a huge role in shaping the structure and composition of the vegetation (Veen et al. 2008). It provides a green flush of nutrient-rich grass by removing moribund grass materials (Bond & Van Wilgen 1996). Although South African grassland systems are naturally maintained by winter and spring fires (Mucina & Rutherford 2006), there is concern over the possible detrimental effect of unnaturally frequent fires on plant diversity (Collins & Calabrese 2012). It has been suggested that controlled burning should be based on the rate of litter accumulation and that grazing should not start until sward height reaches 250 mm (Mentis 1981). Some of the palatable grass species require frequent burning, for example, Themeda triandra, and in the absence of fire, species such as Merxmuellera disticha and Merxmuellera stricta increase (Owen-Smith & Danckwerts 1997). Although there are studies carried out on grassland ecology in South Africa, limited studies exist in MZNP mountainous grassland which to some extent hinder the development of a suitable fire management plan for the park. Hence, this study aims to determine the influence of fire on: (1) grass species richness and diversity and (2) grass relative abundance cover. We hypothesise that: (1) burned sites will have high species richness and species diversity because of fire reducing moribund and making it easier for plants to get sunlight and other nutrients, and (2) burned sites will have more palatable grass compared to unburned sites because fire encourages the growth of palatable grasses.

Even though fire is known to cause replacement or addition in a habitat which results in both species richness and species diversity, there was no significant difference in terms of species richness and species diversity. There was no significant difference in species richness and diversity in burned and unburned areas in this study. This is similar to a study done by Nepolo and Mapaure (2012) where there was no significant difference in burned and unburned sites species richness in semi-arid savanna woodlands in Namibia.

The results have shown the importance of fire in grassland ecosystems as an ecological factor that determines the plant species composition (Bond & Van Wilgen 1996). The grass with high moribund such as M. disticha and M. stricta were dominant in unburned (control) sites for both Site A (2002 and 2010) and Site B (2010 and 2019). The presence of M. disticha and M. stricta is a strong signal of habitat degradation or disturbance, as these two species are associated with moribund material which inhibits the growth of other species (Owen-Smith & Danckwerts 1997). These are big tussock, wiry grasses that easily outcompete other grasses under conditions of selective grazing and the absence of fire. With fire suppression, these grasses create a huge amount of moribund which limits sunlight and other resources for other plant species. One of the highly palatable grasses, that is, T. triandra, was high in burned sites compared to unburned sites in both fire frequencies. The absence of fire seems to suppress the palatable grass and promotes grass that does well under shady conditions. Themeda triandra is one of the most significant grasses in MZNP as it is the most palatable grass available for grazing animals (Zacharias 1990). Consequently, the grasslands dominated by T. triandra are of significant ecological importance at MZNP. Themeda triandra is considered to be a climax or subclimax species (Smit, Bredenkamp & Van Rooyen 1992) and an indicator of grassland in good condition and it declines in abundance when overgrazed, underutilised or in the absence of fire (also commonly referred to as a ‘Decreaser’ species) (Van der Westhuizen et al. 2001; Wiegand et al. 2004). In South Africa, Australia and Southeast Asia, T. triandra dominate mainly in frequently burnt grasslands and their populations may decline to local extinction if fires are excluded for longer than a decade (Snyman, Ingram & Kirkman 2013). This is similar to this study as T. triandra was low in terms of composition in unburned areas. Manipulation of the fire interval is a key tool for influencing the biodiversity of vegetation stands (Bond & Keeley 2005). Grass and shrubs species composition was found to differ significantly between burned and unburned areas.

The overall grass cover differed between burned and unburned sites; however, the difference was not statistically significant. Grass cover was noted to be higher in the burned area in comparison to an unburned area in the 2002 and 2010 fires (Site A). Fire occurrence removes the moribund and the litter build up on the soil surface and this changes the microclimate and nutrient levels in the surface soil (Lavorel, McIntyre & Grigulis 1999). Therefore, the removal of these growth inhibitors by fire in the burned area has resulted in higher grass cover observed for this site. According to a study carried out by Snyman (2003), it was conveyed that fire increases plant nutrient availability in the soil and acts as a natural selective force in the development of grass species. Fires play an important role in driving species composition in MZNP.

Although there was no significant difference in grass species richness and species diversity in burned and unburned sites, there was a significant difference in grass species composition in burned and unburned sites species composition. One of the limitations of this study is the absence of prior data collected on the sites before the 2002 fire which was going to bring a good comparison. Fire plays an important role in species composition; hence, there was a significant difference in Site B where the fire burned less than 3 years ago and no significant difference in Site A where the fire has burned more than 10 years ago. This does show that the fire return interval should not be more than 10 years as it affects grass species composition and a site that burned more than 10 years ago is almost similar to a site that has not burned in 20 years or more. One of the highly palatable grass T. triandra was low in unburned sites compared to burned sites, and a decline in abundance of T. triandra in grassland is usually coupled with a decline in grazing value, species richness and cover. Consequently, we recommend intermittent fire application lesser than 10 years as it could increase the abundance of palatable species. This is done to maintain healthy functional grassland composition. If any natural fire occurs in a veld that has not burned in more than 8 years, then they must not suppress it, unless there is a threat to property or lives.