Project topics and materials
Get Complete Project Material File(s) Now! »
Discussion
We aimed to identify the influence of elephants for vegetation in an African savanna.
To achieve this we searched for changes in sub-canopy vegetation with changes in canopy shapes induced by elephants. We distinguished between open and closed woodlands as differences in their structures and tree densities may influence the response of sub-canopy vegetation to disturbances. At our study site, like elsewhere, elephants changed tree canopies (Barnes 1982, Jachmann & Croes 1991, Lewis 1986), thereby affecting the microhabitat (Belsky et al. 1993, Belsky & Canham 1994). We compared sub-canopy community variables of areas exposed to elephants (inside the Tembe Elephant Park) with those of areas protected from elephants (local communal lands adjoining the Park). We recognised that browsers and grazers other than elephants living inside the Park may influence the responses we recorded. These responses thus may be considered as a “park effect” rather than an “elephant effect”. Nearly 60% of the trees in closed and open woodlands inside the Park had altered canopies, whereas the canopies of most trees (80%) on communal lands were intact. The altered canopies outside the Park may be ascribed to natural tree mortality, or remnants of elephant browsing before the Park were fenced. Independent of canopy shape and woodland type indices for grasses were lower at trees than at controls. The opposite was true for seedlings and saplings. In open woodlands, trees within intact canopies were associated with more seedlings and saplings than control sites. Here the perennial grass species Panicum maximum Jacq. dominated while controlled sites supported a variety of pioneer grass species (Ludwig et al. 2004). Grass and woody sub-canopy species in closed woodlands did not respond to the small changes (<10% totally removed) in tree canopies. As elsewhere, (e.g. Favier et al. 2004) gaps that developed in response to elephant feeding in closed woodlands may favour grass growth (Norton-Griffiths 1979). This could result from variability and increase in sunlight (Naumburg & de Wald 1999; Ludwig et al. 2004). The relatively low community indices for saplings inside the Tembe elephant Park may not only be ascribed to elephants, but also to browsing by species such as kudu Tragelaphus strepsiceros Pallas, nyala Tragelaphus angasii Gray, impala Aepyceros melampus Lichtenstein, red duiker Cephalophus natalensis A. Smith and suni Neotragus moschatus von Dueben). The decrease in the sapling evenness may be result from selective browsing that reduces dominant species (exploiter-mediated coexistence; Begon, Harper & Townsend 1996). Furthermore, elephants may have facilitated (sensu Connell & Slatyer 1977; van de Koppel & Prins 1998) the growth of saplings by reducing the inhibitory effect of shading by canopies. This may also hold for open woodlands.
In the open woodlands of Tembe, elephants reduced tree canopies. This and frequent hot fires may have reduced the grasses and woody saplings. Fire is unselective (Bond et al. 2005) and may reduced dominant grasses, thus explaining inducing increased species richness and evenness (see Pimm 1991). This and changes in tree canopies may have generated similar recruitment and establishment conditions for grasses at both our treatment and control sites. Grasses accordingly increased in diversity and evenness. Browsing and fire may keep woody elements within the flame zone, and stop saplings from further development. This situation may be similar to the fire-mediated recruitment bottleneck referred to by Higgins et al. (2000).
In our study area seedlings, on the other hand, were not influenced by canopy changes. This differs from other studies (e.g. Barnes 2001; Jachmann & Croes 1991) where elephants destroy seedlings. Canopies may provide microhabitat conditions for the establishment of seedlings (e.g. through establishing fertility patches; Anderson et al. 2001), but changes in the canopy clearly have little consequence for seedlings once they have been established (see Caylor et al. 2005).
To summarise, our study suggests that elephants enhance the structural heterogeneity of closed woodlands but homogenise that of woodlands. This is supported by the significant interactive term between the overall “park effect” and reduced canopy structures for grass and saplings indices in the open woodland (see Table 3.1). Inside the Park, there is a higher similarity in the respective grass and saplings indices between tree samples (with their canopies modified by elephants) and their local controls.
The study area
The study was conducted in Tembe Elephant Park (27°01’S 32°24’E) (300 km2) and adjacent communal land (200 km2) situated within the Maputaland region of northern KwaZulu-Natal, South Africa. Elephants always occurred in Maputaland, but have been confined to the Park since 1989 following the fencing of its northern boundary, which borders southern Mozambique. During 2001 an estimated 179 elephants (95% CI of 136 to 233) resided in the Park, and the population is presently increasing at a rate of 4.64±0.06% per annum (Morley 2005).
From 1959/60 until 2001/02 (corrected for the June-July rainfall season) the area received a mean (±SD) annual rainfall of 748±388 mm. Sand forests and mixed woodlands dominate the landscape (Matthews et al. 2001). Van Wyk (1996) describes sand forests as a very dense and dry semi-deciduous to deciduous forest type. Based on tree and shrub densities, we divided the mixed woodlands into closed and open woodland types (adapted from Edwards 1983; One-tailed t-test t257=13.45, P<0.0001).
Dense stands (mean ± SD; 2,423.3±873.1 / ha-1) of trees, shrubs and undergrowth, with an enclosed and layered canopy cover characterise the closed woodland. Grass swards and sparsely spaced mature trees and shrubs (1,060.9±728.9 / ha-1) dominate the open woodland.
Chapter 1
General introduction
Chapter 2
Study area
Chapter 3
The effect of elephant-modified tree canopies on sub-canopy savanna plant communities
Chapter 4
The impact of elephants on plant community variables of the Tembe Elephant Park
Chapter 5
Range constriction and landscape use of elephants in Maputaland, southern Africa
Chapter 6
A meta-analysis of elephant impact
Chapter 7
Synthesis
References
Appendices
