Last update: February 7, 2013 03:50:38 PM E-mail Print





L. van den Berg# & J.C.O. du Toit

Grootfontein Agricultural Development Institute, Private Bag X529, Middelburg (EC), 5900

#E-mail:  Loraine van den Berg



Fire shapes the structure and composition of natural rangelands, and has been the focus of considerable research (e.g. Trollope & Trollope, 2002).  While planned fires can be used to induce managerially desirable shifts in composition and structure, accidental fires often cause significant damage to rangelands (SCFPA, 2009), such as by destroying standing grazing reserves.  Accidental fires also affect local economic sustainability and productivity, and can endanger property and human lives.  Unplanned fires can negatively impact biodiversity, especially in environments that are not adapted to being burnt.  It is important to manage areas that have been burnt in unplanned fires to avoid or minimise declines in veld condition (Everson, 1989).


Karoo veld, when burnt, may take up to 30 years to recover (SABONET, 2011), primarily because of the immediate and drastic decrease in the abundance of dwarf shrub species (Donaldson & Mostert, 1958).  Additionally, fires reduce the short-term productivity of Karoo ecosystems because of the immediate loss of forage, and the reduced regrowth potential of plants during the season following a fire (Tainton, 1999; Nkomo & Sassi, 2009). 


Fire can alter the cycling of nutrients and the biotic, physical, moisture and temperature characteristics of soil (Bruhjell & Tegart, 2012).  Physical impacts of fire on soil can include the breakdown or loss of soil structure, and reduced moisture retention and capacity, which can increase susceptibility to erosion.  Recovery of soil nutrient levels after fires can be fairly slow in some ecosystems, particularly those with limited nitrogen, and in semi-arid regions where decomposition rates are slow (Neary et al., 1999).  Chemically, fire-impacted soils can experience changes in nutrient pool cycling rates, loss of elements to the atmosphere, increased pH and loss of organic matter.  Under high burning severity, carbon and nitrogen levels are usually strongly affected (Neary, 2004).  Bruhjell & Tegart (2012) stated that numerous exchangeable elements including P, Mg, K, Na and Ca typically increase following a fire and as a result nutrient cycling is increased. 


In the past, veld fires in the Karoo were rare occurrences, except in the Karoo mountain areas where it is sometimes used as a veld management tool (Du Preez, 1983; Tainton, 1999).  However, above-average rainfall and a shift towards early-season rains in recent years in parts of the Karoo (Du Toit, 2010), may have benefitted the growth of the herbaceous (grassy) component in the Eastern Upper Karoo.  Because grass is highly flammable, this may have resulted in an increased likelihood of high-intensity, and hence damaging, veld fires.


Research on the effects of fire and the subsequent response of plants to it has been conducted throughout the grassland and savanna areas of Africa (West, 1965; Trollope & Trollope, 1999; Snyman, 2006), but little work has been conducted in this respect in the drier Karoo area of South Africa.  The effect of fire on the vegetation composition and recovery processes of rangelands in the Eastern Upper Karoo is not well understood, but will be of notable importance if the current trends towards an increased grassiness continue.  The aim of this project therefore is to evaluate the response and recovery of the natural veld in the Eastern Upper Karoo after a veld fire.



Study site location

The study was undertaken on the farm The Mills, which is situated 12 km from Hanover towards Richmond on the N1 (S31°10.502; E24° 18.721).  The farm is situated in the Eastern Upper Karoo (Mucina & Rutherford, 2006) within an extensive ecotone between the Nama Karoo Biome in the west and the Grassland Biome to the east.  A complex mix of grass- and shrub-dominated vegetation types, which are subject to dynamic changes in species composition, dependent on seasonal rainfall events, occurs within this vegetation type (Low & Rebelo, 1996).  On 24 October 2011, 652 ha of natural vegetation were burnt during an uncontrolled accidental fire (Figure 1). 



Figure 1.  Map showing the extent of the accidental fire on The Mills, Hanover 


Experimental layout

The accidental fire offered the opportunity to establish a ‘natural experiment’ to explore the effects of fire on Karoo vegetation.  Because vegetation composition has significant spatial variation, and the effect of the fire on composition is of interest here, sampling effort was minimised by conducting species composition comparisons across the fire line.  This excluded the problem of running the risk of sampling communities that were different before the fire, and burnt or did not burn based on for example the amount of grass present.  Four sample sites were identified in the study area and at each sample site the burnt treatment was compared with the unburnt treatment.  Data was collected within the four identified sites. 


Data collection

Photomonitoring provides a photographic record of sites by creating a panoramic snapshot of the vegetation from carefully selected vantage points over time (Hall, 2001; Hurford & Schneider, 2006).  While photographs cannot tell the entire story about an area, much information can be gathered by comparing photographs taken of the same area over a number of years (McDougald et al., 2003).  Digital photographs were taken of the burnt and unburnt areas at the four identified sites from the burn perimeter during the survey and a photo series for each sampling site was compiled.


A soil auger was used to collect five soil samples per treatment per sample site (to a depth of 15 cm).  The five samples per treatment were combined to give one sample per treatment per site, equalling eight samples in total.  The sampling was conducted at the onset of the project (three months after the burn).  The soil samples were analysed for pH, calcium (Ca), manganese (Mn), sulphur (S), carbon (C), ammonium nitrogen (N), oxygen (O), total cations, phosphorous (P), sodium (Na) and the proportions of clay, silt and sand.  The treatment effects on the various cations and pH were tested with a t-test.


The relative abundance of life forms (grass, dwarf shrub and ephemerals) was measured in all the sites by using the descending point technique.  A point was lowered at one meter intervals and the life form of rooted plants within a radius of 300 mm were recorded (Sutherland, 1996).  If no plant occurred within a radius of 300 mm, it was noted as bare ground.  A total of 200 points, repeated three times, were recorded in each treatment (burnt and unburnt) at each of the four sample sites and the percentage abundance of each life form per treatment determined.  Detrended Correspondence Analysis (DCA), using the CANOCO computer program (Ter Braak, 1992) was used to describe the compositional similarities between the different sampling sites as well as their similarities with the various life forms. 





From the photographs taken during the first survey it is clear that there were visual differences between the burnt and unburnt veld in all four sites (Figure 2). These photographs will be retaken during the forthcoming surveys at the end of the growing season. 


Figure 2.  Unburnt (left) versus burnt (right) treatments at Site 2 (31-01-2012)


Results of soil analyses

Results from soil analysis showed that soil pH ranged from being mildly acidic to mildly alkaline across the sites (Figure 3).  Soil carbon levels were generally low (up to c 1.5%), as was total organic matter (up to c 2.5%) (Figure 4).  There was no treatment effect on pH (KCl), Carbon (C), or organic matter contents (p>0.05).


Figure 3.  pH (KCl) levels at the four sites


Figure 4. Soil carbon and organic matter levels at the four sites


Botanical composition

The Karoo is characterised by a huge floral diversity and dynamic interaction between different vegetation types such as grasses and dwarf shrubs.  Fires however reduce the abundance of woody plant species and favour herbaceous species (Perkin et al., 2012).  It is expected that the herbaceous layer would recover faster than the woody layer as the growth cycles of grasses are faster than woody species.  The results from the surveys conducted three months after the fire showed a strong similarity between the four Unburnt sites, and the four Burnt sites with regards to abundance of plant life forms.  As expected, the Unburnt sites were correlated with the dwarf shrub life form, while the Burnt sites were closely correlated with the grass life form.  The Burnt sites were associated with the pioneer life form along the first axis of the DCA.  Dwarf shrub species were largely absent in the Burnt sites three months after the fire occurred.



Results of surveys conducted three months after the fire showed that the short-term effects of the fire were restricted to the relative abundance of life forms, with soil chemical and physical characteristics appearing unaffected. Monitoring will continue over the next ten years to document changes. It is anticipated that the experiment will have important implications for understanding vegetation dynamics after burning, in the context of global climate change.



Figure 5.  Detrended Correspondence Analysis (DCA) showing similarities between survey sites and life forms (Closed circles – Burnt; Open circles – Unburnt)




Bruhjell, D. & Tegart, G.  2012.  Fire effects on soil.  Factsheet 2 of 6 in the Fire effects on rangeland factsheet series.  British Columbia:  Ministry of Agriculture.  Website:

Donaldson, C.H. & Mostert, J.W.C.,  1958.  Alarming encroachment of bitterbos in the Orange Free State.  Farming in South Africa.  34: 53 – 56.

Du Preez, G.,  1983.  Riglyne vir die brand van veld in die Karoo-berggebied.  Karoo Agric.  3(2): 8 – 11.

Du Toit, J.C.O.,  2010.  An analysis of long-term daily rainfall data from Grootfontein, 1916 to 2008.  Grootfontein Agric 10(1): 24-36.

Everson, C.S., George, W.J. & Schulze, R.E.,  1989.  Fire regime effects on canopy cover and sediment yield in the Montane Grasslands of Natal South.  African Journal of Science.  85: 113 – 116.

Hall, F.C., 2001.  Photo point monitoring handbook: Part C – concepts and analysis.  Gen. Tech. Rep. PNW-GTR-526.  Portland.  USA.  48p.

Hurford, C. & Schneider, M., 2006.  Monitoring nature conservation in cultural habitats: A practical guide and case studies.  Springer Netherlands.  ISBN 9781402037566.

Neary, D.G.  2004.  An overview of fire effects on soils.  Southwest Hydrology.  September/October 2004.

Neary, D.G., Klopatek, C.C., DeBano, L.F. & Ffolliott, P.F.,  1999.  Fire effects on belowground sustainability:  a review and synthesis.  Forest Ecology and Management.  122: 51 – 71.

Low, A.B. & Rebelo, A.G., 1996.  Vegetation of South Africa, Lesotho and Swaziland. DEAT.  Pretoria.

McDougald, N., Frost, B. & Dudley, D., 2003.  Photomonitoring for better land use planning and assessment.  Rangeland Monitoring Series.  Publication 8067.  University of California, Division of Agiculture and Natural Resources.

Mucina, L. & Rutherford, M.C.,  2006.  The vegetation of South Africa, Lesotho and Swaziland, Strelitzia 19,  SANBI,  Pretoria.

Nkomo, G.V. & Sassi, M.,  2009.  Impact of veld fires on land on smallholder farmers in Cashel Valley in Zimbabwe.  Natural Resources, Agricultural Development and Food Security (NAF):  International Working Paper Series.  Website:

Perkin, B.K., Wittkuhn, R.S., Boer, M.M., Macfarlane, C. & Grierson, P.F., 2012.  Response of plant species and life form diversity to variable fire histories and biomass in the Jarrah forest of South Western Australia.  Austral Ecology.  37(3): 330 – 338.

SABONET (Southern African Botanical Diversity Network).,  2011.  Karoo Desert National Botanical Garden.  Website:

SCFPA (Southern Cape Fire Protection Agency).,  2009.  Introduction.  Website:

Snyman, H.A.,  2006.  Estimating grassland production loss due to fire for a semi-arid climate.  South African Journal of Animal Science.  36 (Issue 5, Supplement 1): 38 – 41.

Sutherland, W.J., 1996. Ecological Census Techniques, a handbook, Cambridge University Press, Cambridge.

Tainton, N.M.,  1999.  Veld burning:  Karoo.  In:  Tainton, N.M. 1999.  Veld management in South Africa.  University of Natal Press.  Pietermaritzburg.  ISBN0 86980 947 4.

Ter Braak, C.J.F., 1992. CANOCO - A FORTRAN program for canonical community ordination. Ithaca:  Microcomputer Power.

Trollope, W.S.W. & Trollope, L.A.,  2002.  Fire behaviour a key factor in the fire ecology of African grasslands and savannas.  In:  Viegas (ed.)  2002.  Forest Fire Research & Wildland Fire Safety.  Millpress, Rotterdam.  ISBN 90 77017 72 0.

West, O.,  1965.  Fire in vegetation and its use in pasture management with special reference to tropical and subtropical Africa.  Mem. Pub. Commomw. Agric. Bur., Farnham Royal, Bucks.  England No. 1.




Grootfontein Agric 13 (1)