Abstract:
Landscape connectivity, the degree to which landscape structures and elements facilitate or inhibit movement between locations, is widely used for conservation management and planning. It is, however, difficult to measure landscape connectivity directly, so it is commonly estimated using a multi-step process. It is likely that uncertainties, such as selection of an inappropriate spatial grain, may be introduced at multiple points during this process, reducing the accuracy and usefulness of any connectivity estimate produced. The goal of this thesis was to develop an improved understanding of the effects that uncertainties at a number of key steps in the connectivity estimation process have on landscape connectivity estimates. As landscape connectivity is difficult to study empirically a virtual ecology approach was adopted throughout the thesis. In this virtual approach a simulation of the ecosystem of interest was developed, from which samples were taken and statistically analysed, as would occur empirically; these results were then evaluated against the known ‘true’ data in the virtual model. As a first step, the effects of a number of potential sources of uncertainty in the creation of cost-surfaces were investigated. I then examined the performance of three commonly used movement data types; relocation, pathway and genetic data. These data were used to assess the accuracy of the cost values assigned to a cost-surface to identify the most reliable data type. The ability of the most commonly used connectivity models and their associated metrics to accurately estimate landscape connectivity assuming no uncertainty, was then assessed. Finally, I explored how the accuracy of landscape connectivity estimates may alter over time due to environmental variability. The outcomes of the virtual ecology approach I adopted show that at each step examined the introduction of uncertainty could result in large inaccuracies in landscape connectivity estimates, making these estimates less useful. A number of important factors that influenced the inaccuracies were highlighted and methods suggested for how to increase the accuracy estimates of landscape connectivity. The thesis also demonstrated the effectiveness of virtual ecology for assessing complex multi-step processes that are difficult to study empirically.