In this post Izak Smit discusses his recent paper ‘An examination of the efficacy of high-intensity fires for reversing woody encroachment in savannas‘.

Woody densification in savannas

Many studies have documented how grasslands and open savannas are being invaded by woody plants. This phenomenon is predicted to increase as atmospheric CO2 levels increase, favouring woody plants at the expense of grasses. Woody encroachment can have many negative effects, including loss of agricultural production (↑cover ≈ ↓livestock), changed herbivore assemblages (↑cover ≈ ↓grazers) and reduced eco-tourism opportunities (↑cover ≈ ↓game viewing opportunities ≈↓tourism). Herbicides or mechanical clearing can reverse woody densification, but are expensive and impractical over large scales, and often have unintended and negative ecological consequences. Frequent fire can also be used to combat woody densification, but few studies have explored whether high intensity fire can be used as a large, infrequent disturbance to reverse densification. In our study we examined the effectiveness of high intensity fires for reducing woody cover in the Kruger National Park (South Africa) (Fig. 1).

Fig 1
Fig. 1: Large-scale fire experiment in Kruger National Park, South Africa, used to test the effectiveness of high intensity fires for reducing the cover of woody shrubs. Photo credit: Izak Smit

Fire intensity effects on woody cover

We compared woody vegetation structure (cover and height) in areas exposed to repeated high intensity and low intensity fires, as well as protection from fire. We collected pre- and post-fire 3D information on woody vegetation using a LiDAR sensor mounted on an aeroplane (Fig. 2). We found that two successive high intensity fires opened up the landscape and reduced woody cover, as intended. In comparison the areas burnt by two successive low intensity fires became even denser over the 4 year study period (Fig. 3). The high intensity fire was successful, at least in the short term, in reducing woody cover, but we were surprised at the number of tall trees (>10m tall,  with a canopy radius of >3m) that collapsed following high intensity fires (Fig. 4). Tall trees are normally considered to be unaffected by fires because their canopies are beyond the flame zone, but more than a third of tall trees collapsed after two successive high intensity fires. This was much higher than the rate of collapse following low intensity fires (6%), or protection from fire (3%). Trees are apparently vulnerable to high-intensity fires because they have been compromised by having their bark removed by elephants (often also leading to wood borer infestation, Fig. 5).

Fig 2
Fig. 2: Pre- and post-fire LiDAR of the same area, showing how 3D vegetation structure was affected by high intensity fire (black depicts ground cover, and grey-to-white scale indicates increasing height of woody vegetation; blue lines show examples of large-canopied trees observable both before and after fire treatments, whereas the red circle highlight an area where fire reduced cover of small shrubs). Images derived from Carnegie Airborne Observatory (CAO) imagery collected during the study.


Fig 3
Fig 3: LiDAR derived woody cover in 2010 (before first fire application) and 2014 (after fire applications in 2010 and 2013) for four fire intensity treatments [no fire (control), two successive low intensity fires, two successive high intensity fires, and a high intensity fire followed by a moderate intensity fire]. Fires were two years apart. Woody cover was reduced after high intensity fires, but increased under the no-burn and low intensity fire treatments. Image credit: Izak Smit.
High intensity fires as management tool

We were hoping our results would indicate that high intensity fires could be used to reduce cover of encroaching shrubs. Although we confirmed this, at least in the short term, the accompanying damage to tall trees creates a management conundrum. High-intensity fires can be used to reduce bush encroachment, but only at considerable cost to the tall tree component. Considering that top-killed small shrubs can potentially recover rapidly after fires  (Fig. 6), and that recovery of tall trees will be slow (Fig. 4), it is clear that a regime of infrequent, high-intensity fires can simultaneously produce a positive and a negative outcome. Trees are important in the landscape, inter alia for nutrient cycling, shade for large mammals, habitat for small fauna and nesting sites for raptors. Therefore, managers will need to make trade-offs when contemplating the manipulation  of fire intensity to achieve specific goals. One solution may be to repeatedly apply high-intensity treatments to some areas, and not to others. This could generate a heterogeneous landscape where grasses become dominant and tall trees become scarce in some places, but in others tall trees persist (or at least decline at slower rates), and shorter woody shrubs increase in dominance. Whether this would be acceptable, or practical, remains to be tested. We predict that in areas without elephants or with lower elephant densities, high intensity fires may have less of an impact on tall trees.

Fig 4
Fig. 4: Collapse of tall mature trees following high intensity fire, probably exacerbated by pre-fire damage (e.g. bark removal by elephants). Recovery of the tree layer would take a long time. Photo credit: Izak Smit.
Fig 5
Fig. 5: Elephant damage to mature tall trees – elephants remove bark within the “flame-zone” height (a), allowing wood to dry and become vulnerable to wood borer infestation (b). These exposed scars create “weak” points, enabling high intensity fires to set the main stem alight, leading to eventual toppling of the tree (c). Photo credit: Brian van Wilgen (a) & Izak Smit (b & c).
Fig 6
Fig. 6: Basal resprouting from the roots (a-b) or coppicing from aerial parts (c) allows small shrubs to recover previous dimensions much faster after fires, as opposed to tall mature trees (see Figs. 4-5). These photos were taken within weeks of fire. Photo credit: Izak Smit (a) & Navashni Govender (b & c).