Large herbivores in space. Resource partitioning among savanna grazers in a heterogeneous environment.

Large mammalian grazers are among the most conspicuous species groups of the animal kingdom. The richest assemblages of large grazers can be found on the African continent. Diverse large grazer assemblages and the grazing systems they live in have been ascribed great socio-economic as well as ecological value. The impact of wild large grazers on humans has been enormous throughout the evolutionary history of mankind. Domesticated grazers have taken over most of the socio-economic role of their wild counterparts in industrialized societies but in many African countries, where domesticated animals were introduced relatively late, the socio-economic impact of wild ungulates is still strong, forming an important protein source and driving the fast-growing tourism industry. Next to this socio-economic role, the potential influence of large grazers on terrestrial ecosystems can be enormous, from driving large-scale changes in vegetation structure to influencing system nutrient cycling. By shaping the systems they inhabit large grazers influence communities of many other taxa that depend on these systems (from arthropods, to birds and large carnivores). Moreover, an increasing number of studies discuss the importance of grazer species diversity for the functioning of grazing systems as species differ in the way they shape their environment. Alarmingly, these diverse grazer communities and their ecological and socio-economic role are increasingly threatened. Free-roaming large grazers have disappeared from large parts of Africa and the remaining areas with high species richness seem to coincide with regions that have the highest human population growth, so the human-wildlife conflict is due to increase. To conserve these diverse assemblages and their functional role we need to understand what factors shape the large grazer communities in time and space. In other words we need to understand how these different large grazer species can locally coexist. Though other aspects such as differences in predation pressure and disease susceptibility have been mentioned, partitioning of the food resource is generally accepted to be the basis of large grazer niche differentiation and ultimately coexistence. In the past several studies have emphasized the importance of body size to explain the separation of feeding niches amongst large African grazers along resource quality and quantity axes, based on the hypothesis that larger grazers can tolerate a lower-quality diet than smaller ones. Prerequisite is that there is sufficient variation in food quality and quantity (i.e. resource heterogeneity) available to large grazer species to be able to coexist. Up to now, in the African context resource heterogeneity has mostly been defined in terms of variation in plant species and in vegetation structure, such as grass height or leaf-stem ratio. Furthermore, several studies argued that different-sized grazers partition resources over time, where species use the same areas and plant species but at different moments in time during large-scale migrations, exploiting different vegetation growth stages that vary in resource quality and quantity. In many areas, however, these large scale migrations do not (or no longer) occur (such as in many South African reserves). Still these same areas sustain species rich and abundant grazer assemblages without clear evidence of competition for resources. Clear empirical evidence how these species partition their resources is still lacking. In this thesis I argue that this is partly due to the fact that the spatial dimension has not been well incorporated into our thinking on large grazer resource partitioning. Spatial variation in resource quality and quantity can be defined at different spatial scales, from the feeding patch level to landscape gradients. Past studies have linked spatial variation in resource quality and quantity to single species foraging behavior and some studies have modeled the effect of resource heterogeneity on the stability of grazer population numbers. We, however, lack empirical studies that specifically relate resource heterogeneity to spatial resource partitioning and ultimately coexistence patterns in species rich grazing systems. The aim of this thesis was, therefore, to explore how naturally coexisting large African grazers might spatially partition resources by defining variation in resource quality and quantity at different spatial scales using experimental as well as observational techniques. All studies were carried out in Hluhluwe-iMfolozi Park, a 90.000 ha reserve on the South African east coast, just south of Swaziland and were mainly focused on the 6 commonly occurring large grazer species in the park; impala, warthog, wildebeest, zebra, buffalo and white rhino. In the thesis I start with describing how the body mass of these grazer species influences their distribution over the landscape and how this distribution is linked to landscape variation in habitat type and quality. The larger species were more evenly distributed over the landscape than the smaller species and had a more diverse habitat use, though digestive physiology influenced this relationship. By differently distributing over the landscape, large grazer species might partition their resources at a coarse landscape scale (chapter 3). At a finer scale (hectares) differences in soil type cause spatial variation in grassland types. I describe how grazer species partition these grassland types that differ in resource quality and availability and how fire interacts with grassland type to affect grazer community composition (chapter 4). In general, grazing lawn sites, situated on relatively sandy soils, were nutrient hotspots and attracted all grazer species. On the relatively clayey soils bunch grasslands attracted the larger grazer species (wildebeest and larger), especially in the months after a fire. At an even finer scale (meters), most grassland in Hluhluwe-iMfolozi is characterized by a high spatial heterogeneity at the patch level, with alternating patches of short and tall grass. I showed how this within-grassland heterogeneity might increase opportunities for resource partitioning amongst savanna grazers and ultimately mediate their coexistence (chapter 5). In a field experiment where I manipulated variation in short grass patch size and resource quality I found that species differ in their selection of patch size and quality. Moreover, at this scale grazers do not only respond to heterogeneity, but they can also shape vegetation heterogeneity. I describe and experimentally test a scale-dependent mechanism that might drive short-tall grass patch dynamics in savanna grasslands (chapter 6). This mechanism is based on an interaction between very large (mega) grazers, such as the white rhino, and smaller grazers. The mega grazers cause disturbances above a certain scale (such as wallows and middens) which attracts smaller grazers. The intense localized grazing of the smaller grazers hence promotes lawn development. These results support the suggestion that the diversity of grazer communities is a significant aspect of savanna ecosystem functioning. In the concluding chapter I synthesize the results and discuss how resource partitioning among large grazers might be nested across different spatial scales (chapter 7). Results from this thesis suggest that resource partitioning does not simply happen at one spatial scale but that large grazers actually partition a complex resource landscape where a hierarchical set of factors determines the spatial heterogeneity of resource quality and quantity at different spatial scales. However, many studies in the past were actually limited to one spatial scale. Our future challenge is that we need to design studies where we really integrate spatial scale into our thinking about how herbivores partition this complex resource landscape. These new studies might benefit from newly available techniques, such as remote-sensing of vegetation quality and satellite tracking of grazers that allow us to better integrate observed patterns of spatial resource use across spatial scales. From a practical conservation perspective the results of this study emphasize the need to take spatial resource heterogeneity into account when managing savanna systems. These results support the shift towards a new paradigm in the management of grazing systems. This so-called heterogeneity paradigm states that management should promote grassland heterogeneity to maintain biologically diverse communities in these systems. At the end of the thesis (chapter 7) I philosophize how fine scale heterogeneity might counterbalance a lack of reserve extent, because optimal and suboptimal resources become available in home ranges of sedentary species. In contrast coarse-grained systems, where there are only large-scale gradients that affect resource quality and quantity, have to be sufficiently large for animals to find the right balance of optimal and suboptimal resources. This means that focusing on reserve size would still be necessary in some cases (e.g. Serengeti system) but in places where the abiotic template is heterogeneous small reserves might provide a good alternative supporting diverse and abundant large grazer populations. This offers opportunities for areas where many different stakeholders compete for space, such as the province of Kwazulu-Natal in South Africa where Hluhluwe-iMfolozi Park is situated. In other words next to reserve size, the available scale of heterogeneity of e.g. topography or soil types should be an essential criterion for reserve planning.