Linking long-term patterns of landscape heterogeneity to changing ecosystem processes in the Kruger National Park, South Africa
Thesis (PhDAgric)--Stellenbosch University, 2018.
ENGLISH ABSTRACT: Biodiversity loss is a global threat to ecosystem function and human well-being. Environmental heterogeneity is a recognised driver of biodiversity under a niche-based view of available species habitats. As such, an increase in environmental heterogeneity is expected to promote species coexistence, persistence and diversification. Loss of environmental heterogeneity is therefore considered proximal evidence of biodiversity loss. At a landscape scale, this heterogeneity is defined as the degree of difference between landscape elements and is often described as landscape heterogeneity. Patterns of landscape heterogeneity are generated and maintained by the physical landscape template or abiotic environment (e.g. topography, geology and climate), upon which complex adaptive interactions between landscape pattern (structure and composition) and ecological processes (function) occur. Landscape pattern can therefore be described as the self-organising expression of landscape function which varies not only across space but also through time. Accordingly, observable variations in landscape pattern are conjectured to signify divergence in landscape function. This thesis explores this relationship further within the Kruger National Park (Kruger): a large (~ 20,000 km2 ), longestablished (proclaimed 1898) protected area in South Africa’s semi-arid savanna. Results therefore describe landscape heterogeneity, in terms of the abiotic and biotic components (environmental drivers) that generate and maintain landscape pattern in Kruger, to inform strategic biodiversity planning. Chapter 1 introduces the reader to landscape heterogeneity and its relevance to protected area management and biodiversity conservation. Chapter 2 begins by isolating the effects of ‘stationary’ landscape properties on environmental heterogeneity through their relationship with Landsat spectral variance. Results show this relationship is sensitive to season and rainfall with the effects of dynamic ecosystem processes dominating many areas. Thereafter, Chapters 3 and 4 examine in more detail the nature of selected dynamic drivers in Kruger, namely rainfall and elephants. Results demonstrate the existence of longterm spatiotemporal changes in both rainfall and elephant density and distribution patterns in Kruger from 1985-2015. Together these results feed into chapter 5, where a Structural Equation Model (SEM) is used to investigate the causal structure of landscape heterogeneity with stable landscape properties, rainfall, herbivory and fire. Results are presented as path coefficients and long-term driver dominance maps showing the magnitude and direction of the different cause and effect relationships between heterogeneity, the physical landscape template, rainfall, herbivory and fire return interval. Finally the nature of the environmental-heterogeneity theory is operationalised in Chapter 6 using R, Shiny and Leaflet to provide an interactive web interface for protected area managers to explore heterogeneity differences in context with park specific research questions. Chapter 7 concludes the thesis with a brief synthesis of results in context with current literature and highlights future research opportunities and possible directions.
AFRIKAANSE OPSOMMING: Geen opsomming