Dave J. Druce

Ecological Thresholds in the Savanna Landscape: Developing a Protocol for Monitoring the Change in Composition and Utilisation of Large Trees

Background Acquiring greater understanding of the factors causing changes in vegetation structure - particularly with the potential to cause regime shifts - is important in adaptively managed conservation areas. Large trees (≥5 m in height) play an important ecosystem function, and are associated with a stable ecological state in the African savanna. There is concern that large tree densities are declining in a number of protected areas, including the Kruger National Park, South Africa. In this paper the results of a field study designed to monitor change in a savanna system are presented and discussed. Methodology/Principal Findings Developing the first phase of a monitoring protocol to measure the change in tree species composition, density and size distribution, whilst also identifying factors driving change. A central issue is the discrete spatial distribution of large trees in the landscape, making point sampling approaches relatively ineffective. Accordingly, fourteen 10 m wide transects were aligned perpendicular to large rivers (3.0–6.6 km in length) and eight transects were located at fixed-point photographic locations (1.0–1.6 km in length). Using accumulation curves, we established that the majority of tree species were sampled within 3 km. Furthermore, the key ecological drivers (e.g. fire, herbivory, drought and disease) which influence large tree use and impact were also recorded within 3 km. Conclusions/Significance The technique presented provides an effective method for monitoring changes in large tree abundance, size distribution and use by the main ecological drivers across the savanna landscape. However, the monitoring of rare tree species would require individual marking approaches due to their low densities and specific habitat requirements. Repeat sampling intervals would vary depending on the factor of concern and proposed management mitigation. Once a monitoring protocol has been identified and evaluated, the next stage is to integrate that protocol into a decision-making system, which highlights potential leading indicators of change. Frequent monitoring would be required to establish the rate and direction of change. This approach may be useful in generating monitoring protocols for other dynamic systems.

Sampling strategies for millipedes (Diplopoda), centipedes (Chilopoda) and scorpions (Scorpionida) in savanna habitats

At present considerable effort is being made to document and describe invertebrate diversity as part of numerous biodiversity conservation research projects. In order to determine diversity, rapid and effective sampling and estimation procedures are required and these need to be standardized for a particular group of organisms to allow for comparisons between studies and habitat types. The savanna biome is one of the largest and most important in Africa; however, it is also one of the most poorly studied, especially in terms of invertebrates. This study was undertaken in the Greater Makalali Conservancy, Limpopo Province, South Africa. The effectiveness of six sampling methods (pitfall traps, active searching one 25 m2 nested quadrat, active searching ten 2.25 m2 random quadrats, cryptozoan traps, wet cloths and drive transects) was tested to determine which method, or combination of methods, was the best for sampling millipedes, centipedes and scorpions in the savanna environment. Active searching in a 9 m 2 area was the most effective way to sample millipede species, while the drive transect method was important for sampling larger millipede species. Both the active searching of the nested quadrats and the random quadrats proved to be the most effective methods to sample centipedes. Scorpions were most effectively sampled using pitfall traps. Efficiency for all methods was calculated as the number of species collected per hour. The most effective method was not always the most efficient one, and this needs to be considered when designing a sampling strategy. Sampling more than one period in the summer is important for determining species richness. The number of random quadrats required to sample the fauna is likely to vary in different habitats, and there is a large amount of variation in the number of species collected between samples, which is probably related to high levels of habitat heterogeneity on a small scale. These points, as well as the size, mobility and other biological features of taxa need to be considered in designing a sampling strategy for invertebrates.

Spatial and temporal changes in group dynamics and range use enable anti-predator responses in African buffalo

The reintroduction of large predators provides a framework to investigate responses by prey species to predators. Considerable research has been directed at the impact that reintroduced wolves (Canis lupus) have on cervids, and to a lesser degree, bovids, in northern temperate regions. Generally, these impacts alter feeding, activity, and ranging behavior, or combinations of these. However, there are few studies on the response of African bovids to reintroduced predators, and thus, there is limited data to compare responses by tropical and temperate ungulates to predator reintroductions. Using the reintroduction of lion (Panthera leo) into the Addo Elephant National Park (AENP) Main Camp Section, South Africa, we show that Cape buffalo (Syncerus caffer) responses differ from northern temperate ungulates. Following lion reintroduction, buffalo herds amalgamated into larger, more defendable units; this corresponded with an increase in the survival of juvenile buffalo. Current habitat preference of buffalo breeding herds is for open habitats, especially during the night and morning, when lion are active. The increase in group size and habitat preference countered initial high levels of predation on juvenile buffalo, resulting in a return in the proportion of juveniles in breeding herds to pre-lion levels. Our results show that buffalo responses to reintroduced large predators in southern Africa differ to those of northern temperate bovids or cervids in the face of wolf predation. We predict that the nature of the prey response to predator reintroduction is likely to reflect the trade-off between the predator selection and hunting strategy of predators against the life history and foraging strategies of each prey species.

Prey selection by a reintroduced lion population in the Greater Makalali Conservancy, South Africa

Lion prey selection was studied on the Greater Makalali Conservancy (140 km2), Limpopo Province, South Africa, in order to assist with management strategies. Monitoring was carried out between February 1998 and December 2001. Lion killed 15 species, with warthog, blue wildebeest, Burchell’s zebra, kudu and waterbuck constituting approximately 75% of their diet. Between 2.2% and 3.1% of the available prey biomass was killed yearly, while each female equivalent unit (FEQ) killed between 3 kg and 3.2 kg daily. Lion predation was greater for warthog, wildebeest and waterbuck and less for impala than expected. When male lion were present, a greater number of warthog and giraffe were killed, while number of females had a significant effect on medium sized prey species and total prey species killed. Significantly more warthog, wildebeest and kudu were killed in winter than summer. More prey than expected was killed in open habitats and less than expected in thickets. Managers of small, enclosed reserves need to constantly monitor prey populations, especially medium sized prey and may be able to reduce predation on large prey species by manipulating male lion numbers. Reserves also need to contain adequate open habitats for lion to make use of these areas for hunting.

Patterns of millipede (Diplopoda), centipede (Chilopoda) and scorpion (Scorpionida) diversity in savanna habitats within the Greater Makalali Conservancy, South Africa

ABSTRACT Although savanna is one of the most important biomes in southern Africa, it is, apart from the botanical component, one of the least studied. This study aimed to document the distribution patterns, richness and diversity of selected ground-dwelling, flightless arthropods (millipedes, centipedes and scorpions) within savanna habitats to improve invertebrate conservation planning. Five habitat types (white sand bushveld, brown sand bushveld, general mixed bushveld, rocky outcrops and mopane woodland) within the Greater Makalali Conservancy, Limpopo Province, South Africa, were studied using four successful sampling methods (active searching of two types of quadrats, pitfall traps and drive transects) during three sampling periods between February 1999 and March 2000. Millipedes were consistently the most species rich and abundant taxon. Species were not uniformly distributed across all habitat types; some species were unique to certain habitats. Generally, the more heterogeneous habitats supported the greatest millipede, centipede and scorpion diversity, richness and density, and in the case of the millipedes contained the highest number of regional endemics and habitat specialists. There was no significant difference in scorpion diversity among habitats.