Herbivore distribution patterns mapped on SA Game Reserve
Data gathered on a South African game reserve shows the importance of terrain type for herbivores within the reserves and the accuracy of using line-transect sampling as opposed to aerial surveys to gather the data. The results will aid conservation planning on the reserves.
Private game reserves in South Africa are central to the maintenance and enhancement of its herbivore biodiversity. These reserves are (usually) fenced and as such require intensive land and livestock management to maximize their ‘game’ carrying capacity for wildlife tourism purposes.
At the same time reserve managers must manage the herbivore population so as to prevent habitat degradation from over-grazing and to utilise their dunging/feeding activities and patterns to maintain optimum nutrient conditions. Understanding animal habitat requirements can enable game reserve managers to predict animal distribution patterns and consequently their impact on vegetation.
Estimating game distribution is therefore an important part of game reserve management. One method for monitoring large game stocks is via aerial surveys in which all visible game are counted.
An alternative method is to monitor game using line-transect sampling. Line transect sampling has been a popular and widely used technique for estimating abundance of biological populations in a variety of habitats and situations. Animals can be surveyed from game vehicles and, provided that habitat is recorded along with each animal sighting, population density models can be corrected for variation in detectability according to habitat type.
The field work related to this study took place in Welgevonden Game Reserve. The reserve is situated in the Waterberg plateau in the Limpopo Province of South Africa where it forms part of the Waterberg Biosphere Reserve declared in 2001 by UNESCO, which in its entirety covers in excess of 14,500km2.
During this study, large mammal populations on the reserve were monitored from game vehicles using distance sampling along six 10km line transects that coincide with the reserve road network. The transect lines in total incorporated all of the seven different terrain types previously identified as present on the reserve.
Researchers travelled the transect lines and road systems and recorded their encounters with herbivores in accordance with accepted protocols. Each time an animal was seen, the species was identified, aged and sexed, the number of individuals recorded, the distance along the transect line they were encountered and the Geographical Positioning System (GPS) location of the animal(s). The large mammal species commonly encountered during surveys were elephant, white rhino, blue wildebeest, Burchell’s zebra, giraffe, impala, warthog, kudu, waterbuck and red hartebeest and their distribution forms the basis of all subsequent analysis.
GPS and terrain data obtained from the transect surveys were combined with existing GIS maps of the reserve to investigate ranging and habitat use of the species encountered. Population estimates were then formulated for each species using DISTANCE© V.6.0, a software programme which estimates density and abundance from transect data.
Most species positively selected plains except impala and warthog
These data were added to monthly game counts that incorporated the entire road network in the reserve. In addition annual game count data were collected via helicopter and these data were compared against the estimates obtained from the game transects.
Data from the ten herbivore species were used to determine habitat preferences with respect to terrain class. We used Jacobs’ (1974) modification of the Ivlev selection index to determine significant preference and avoidance for each terrain class. This index measures numbers of individuals of a species selecting each terrain class relative to the availability of the terrain class in the reserve. All species avoided crest summits and hill slopes as indicated by the negative Ivlev index. All species for which significance could be assessed, prefer ‘old land’ (land previously in agricultural production) and plateau, with white rhino showing a particularly strong preference for old land.
Most species positively selected plains except impala and warthog, whilst blue wildebeest, red hartebeest and white rhino showed notable preferences for this terrain type. With the exception of impala, most species prefer riparian areas, with elephant showing an extremely strong preference for this terrain type. Finally most species, except wildebeest, prefer valley bottom. These selection and avoidance relationships may reflect grazing quality, perhaps mediated by drainage, and predation risk.
Herbivore abundance estimates calculated from distance sampling were compared to abundance estimates calculated from aerial game counts. In almost all cases the aerial count estimate falls within the errors of the distance estimate. These initial results indicate that if DISTANCE© estimates continue to compare satisfactorily against aerial census data, the frequency of these aerial census could be decreased, and perhaps discontinued.
Vehicle-based transects therefore not only create significant cost savings, but also provide more accurate spatial data, which increases the application of data in other areas of conservation planning on the reserve.