While densities of species such as elephant are being manipulated through population reductions (for example hunting, or contraception), removal or introduction, these management decisions are likely to have implications for other species or plants.
An example of this comes from the USA where the introduction of wolves in Yellowstone National Park caused the numbers of elk to decline, which in turn increased recruitment of aspen due to the release in browsing pressure from elk. These subsequent effects are also known as trophic cascades.
To test for these indirect cascading effects of human intervention in herbivore populations on the different species of trees that occur together (tree assemblages) and the diet choices of other herbivores, an existing exclosure experiment (where one or more species is artificially removed) was used in the sand forest in Phinda Private Game Reserve. This reserve is 180 km2 in size, and is situated in north-eastern KwaZulu-Natal.
The research is a joint collaboration between And Beyond (previously known as CC Africa) and the University of KwaZulu-Natal, Durban and was initiated to investigate the effects of elephant and nyala on the sand forest vegetation.
Sand forest is a deciduous dry forest with a high biodiversity and endemic species, and is only found in north-eastern KwaZulu-Natal and southern Mozambique. This critically endangered forest generally occurs in a mosaic of distinct patches enclosed by savanna woodland.
Threats to sand forest include fire and selective species utilisation by both man and herbivores. In addition the forest has a very low resilience to any of these disturbance factors and is known to have poor recruitment (germination and growth of new seedlings and saplings) rates.
To protect the forest from elephant impact an electrified elephant exclosure fence was constructed in 2005. Within this enclosed area, smaller sized exclosures were set up to exclude nyala (a medium-sized browser). Unfenced control plots were situated outside the elephant exclosed area.
With the help of students from the University of KwaZulu-Natal, all trees were monitored in 36 plots of 400 m2 each (12 per treatment), during four weeks of field work in 2007. Data gathered included species, height of tree, canopy dimensions and the recording of any browsing damage by any herbivore.
After counting over 2800 stems, the researchers found over 140 tree species in the sand forest. Most surprisingly many seedlings and saplings were encountered. However, seedlings have to escape a "browsing trap" caused by nyala, and once these recruits survive this they potentially face herbivory from both elephant and nyala.
In the short term that the experiment was running, the study found tree species assemblages to be affected by both elephant and nyala together, and by each species on their own. Browsing activities by only suni and duiker did not cause any changes in species assemblages.
It was also discovered that nyala and smaller sized herbivores changed their diet (resource) utilisation in absence of their larger counterparts (which would be elephant and nyala respectively). This implies that there might normally be some competition between these different sized herbivores, when all species are present, and that species might change their feeding strategies when others are removed.
From this short term experiment it can be concluded that when species are removed from the system this will have indirect cascading effects on for instance vegetation dynamics, tree composition and even on feeding strategies of other herbivore species. The effects on vegetation are most important to acknowledge, as this may affect biodiversity (in the worst case scenario even a loss of species).
For example, while it is well known that elephant can have a destructive effect on tall trees, there remains a problem when recruitment is impeded by other herbivores, and no small trees recruit into the taller height classes to compensate for (natural) die-offs.
This implies that every form of manipulation needs to be carefully considered before implementation, and needs to be monitored in order to detect any (undesirable) cascading effects, to conserve our habitats and biodiversity.
The University of KwaZulu-Natal would like to acknowledge the joint collaboration with And Beyond (Phinda Private Game Reserve - www.andbeyond.com), and especially thank them for putting up all the fencing used in this study.
By Georgette Lagendijk