Fire Trials of Wild Rooibos sites in the Suid Bokkeveld

Wild Rooibos grows naturally in highly bio-diverse sites, along with a large number of other species of fynbos plants. All of these plants are adapted to fire, and indeed many species will become moribund, cease to reproduce and die if not exposed to fire at least every few decades.

In the Suid Bokkeveld fires occur naturally during summer thunder storms, which occur most frequently in the late summer months (January – April). Since settlement in the mid-1700s, farmers have also used fire as a management tool to promote regeneration of plants and stimulate the production of grasses and other palatable species.

Rooibos has become a traded product in the past century. Most rooibos is produced in plantations, which farmers protect from fires so as to avoid financial loss. Wild harvest has remained a source of income for some farmers, but changes in the fire regime of the area due to human influence have made it difficult to assess what fire regime will benefit rooibos harvesters and also retain the biodiversity of the area. In order to address this question, EMG and members of the Heiveld Co-operative undertook to experimentally burn some trial plots, and to compare the regenerating vegetation on these plots to that of adjoining plots that would be spared from burning.

Nick Helme was commissioned to assess the vegetation within plots, which were set out on two properties in the Suid Bokkeveld. The assessment took place in 2009 prior and post burning of half of the plots with the specific intention of observing the impact of fire upon all of the plant species present in the trial plots so as to draw general conclusions and recommendations relating to the management of fire in areas of the Suid Bokkeveld where wild rooibos occurs.

The initial survey identified all species present in the autumn of 2009, and follow up surveys each spring have monitored plant populations using the Braun Blanquet scoring system to quantify the relative cover provided by each species.

Each site (Dobbelaarskop & Blomfontein) had two 10 by 10m control plots, and two 10 by 10m experimental plots (controlled fires/burns). The experimental plots were burned in April 2009, and plot data were collected in September 2009, September 2010 and September 2011. The following section reflects Nick Helme’s observations and interpretations of the data.

Overall plant cover

Dobbelaarskop: overall cover increased rather significantly in the two control plots over the three years, whilst this was less dramatic in the burn plots, due mainly due to the relatively coarse BB scale used and the subjective nature of making these estimates. Essentially it could be said that overall cover increased in all plots, with cover remaining significantly higher in the control plots, as three years is nowhere near enough for full vegetation recovery in this relatively arid area. It is estimated that the burn plots will take between seven and ten years to recover to a pre burn state of overall cover.

Blomfontein: Increasing cover was less pronounced at this site, but was still evident, although this is not clear from the coarse scale data.

Aspalathus linearis cover

Dobbelaarskop: Control plots showed insignificant cover gain, but burn plot B showed excellent cover growth, whilst burn plot D showed relatively little gain. I am not sure what would account for this difference, and it may be useful to check the seedling data for these plots to see if there were significant differences in post fire recruitment. It would appear that fire is certainly not negative for wild rooibos in this area.

Blomfontein: There was a dramatic increase in cover in both burn plots, but also in one of the control plots, presumably due to the good rainfall experienced in 2009 and 2010. Quite clearly fire is not negative for wild rooibos, at least in this area!

Species diversity

Dobbelaarskop and Blomfontein: The pre and post burn 2009 data cannot be compared in terms of annuals and geophytes (eg. Schizodium flexuosum), as most of these taxa would not be evident in April (end of dry season). This is evident from the control plot data as well (many species absent from the April 09 surveys) – which is why surveys do need to be conducted at the same time of year (although not all years are equal either!). In this regard a notable difference in annual taxa abundance was found between 2010 (low, due to poor rains) and 2011 (high, presumably due to slightly better rains), but partly balanced out by the usual effect of annual numbers gradually decreasing in the years after a fire (high in first few years, and then declining).

The relatively coarse BB scale was again a problem when looking at total cover of small species, such as annuals and geophytes – although they may have been common (hundreds of plants) their cover often did not add up to more than 1m2, and certainly to not more than 5m2. A number of species only became evident in 2010 or even 2011, which is probably mainly a function of observer experience and a better species list for the plots as time progressed – it is easy to miss relatively cryptic species. Wahlenbergia sp nov was curiously absent from all plots in 2011, for unknown reasons – presumably it was earlier or later in flowering than usual.

Some species, such as the reseeding (vs. resprouting) restios, suffered a notable drop in cover after the fire, and due to relatively slow growth rates were not identifiable in the first season after a fire, and were thus only noted again in 2010 or 2011, or were not yet identifiable even in 2011 (e.g. Thamnochortus platypterus at Dobbelaarskop). Short fire intervals (<5 or 6yrs) could thus potentially remove reseeding restios from the veld, which would have a dramatic structural impact, as they are dominant in many Fynbos situations. In general most geophytes and annuals are present in both the control and burn plots, although they may be marginally to substantially more common in burn plots (but the data do not show this clearly, as noted). Succulent perennials such as Ruschia dichroa and Ruschiella lunulata not surprisingly did not respond well to the fire. Those that evaded it survived, but there does not seem to have been any post fire recruitment of these species. One of the striking things about the plot data is that species composition and dominance varies hugely from plot to plot, a typical Fynbos phenomenon which helps drive the high diversity figures in the biome. This unfortunately makes direct comparisons for this sort of study very difficult, as it is almost impossible to find identical plots. Non-fire related mortality was also evident (eg. Leucadendron sheilae and Erica rigidula), although on a very limited scale, and quite possibly associated with summer drought.

Threatened species

The plots themselves did not include many threatened species, but there were some, most of which are regional or local endemics. Blomfontein included Leucadendron sheilae (Red Listed as Vulnerable; the single plant died, but with a larger population nearby), Babiana engysiphon (Endangered), Athanasia leptocephala (Rare) and Wahlenbergia sp nov. (a regional endemic). Dobbelaarskop plots included only Wahlenbergia sp nov. (a regional endemic), although the local endemic Babiana rigidifolia (LC) is common in nearby rocks.
Analysis of the plot data regarding the impacts of fire on these species is not particularly useful, although it can be said that Babiana engysiphon was consistently present in both control and burn plots at Blomfontein.

Want to know more? Email us