A systematic approach to planning, decision-making and management has become best-practice in conservation over the past two of decades. The field of ‘systematic conservation planning’ is concerned with identifying cost-effective places and actions to protect biological diversity. Past research has focused on static assessments. However, given the fact that biological diversity and processes that threaten its persistence vary in space and time, conservation assessments might need to be made in a dynamic context. In addition, we must explicitly account for the trade-offs associated with implementing conservation actions and investing in improved knowledge and learning to reduce uncertainty on where, how and when to act. The aim of this thesis was to develop novel approaches for accounting for both ecosystem dynamics and uncertainty in conservation planning. Ecosystems are generally treated as static in conservation planning despite many being spatially and temporally dynamic. For example, pelagic marine ecosystems are quite dynamic because ecological processes, such as eddies, that produce resources that many species depend on can be erratic. In chapter two we explored the issue of developing a system of fixed protected areas that consider the physical and biological dynamics typical of the pelagic realm. The approach was to maximize the representation of key fisheries species and species of conservation concern due to significant declines in their abundance, within a network of protected areas. We also ensured that protected area design reflected system dynamics and this was achieved by representing key oceanographic process (such as upwellings and eddies), and biological processes (such as the abundance of small pelagic fish) in protected areas. To account for the variability where these processes occur, we used time series data to find both predictable areas and anomalies, assuming that their past location was somewhat reflective of their future locations. Implementing conservation actions that are fixed in space and time are probably not the most effective strategy in ecosystems that are dynamic. This is because of the movements of particular species. For example, many species have distributions and abundances that change seasonally and might only require temporary management in particular areas. In chapter three, we tested the utility of three approaches to implementing fisheries closures to reduce bycatch in the South African Longline Fishery; 1) time closures, 2) permanent spatial closures and 3) episodic spatial closures. In chapter three, we identified these closures using an existing database containing catch and bycatch data from 1998 to 2005. There was variation where and when different species were caught as bycatch, and it was determined seasonal area closures were the best strategy. This was because it achieved the same conservation objectives for bycatch species as the other types of closures, but impacted less on the long-lining industry. While this result is intuitive, it demonstrated quantitatively, how much more effective moveable management can be. Decisions on where conservation actions are implemented are always based on incomplete knowledge about biological diversity. It is generally assumed that gathering more data is a good investment for conservation planning. However, data can take time and incur costs to collect and given habitat loss, there are both costs and benefits associated with different levels of investments in knowledge versus conservation implementation. In chapter four, the aim was to determine the return on investment from spending different amounts on survey data before undertaking a program of implementing new protected areas. We found that, after an investment of only US$100,000, there was little increase in the effectiveness of conservation actions, despite the full species dataset costing at least 25 times that amount. Surveying can take time because of expertise limitations, logistics and funding shortfalls. Biological diversity may be lost while data collection occurs conversely, not collecting enough data can lead to erroneous decisions. Additionally, resources spent on learning may be better spent on other actions. In chapter five, in a series of retrospective simulations, we compared the impact of spending different amounts of time collecting biological data prior to the implementation of new protected areas. The aim was to find the optimal survey period given the trade-off between gaining knowledge to improve conservation decisions while there is concurrent loss of habitat. We discovered that surveying beyond two years rarely increased the effectiveness of conservation decisions, despite a substantial increase in the knowledge of species distributions. Often there are choices between different actions and uncertainty as to which are the most effective. In chapter six, we discuss how the principles of adaptive management might be applied to conservation planning. Improving future management decisions through learning should be viewed as essential in all conservation plans but such learning is often included as a minor step, or is completely ignored. In this chapter we provide a brief overview of an adaptive framework for conservation planning and ideas for future research.

A systematic approach to planning, decision-making and management has become best-practice in conservation over the past two of decades. The field of ‘systematic conservation planning’ is concerned with identifying cost-effective places and actions to protect biological diversity. Past research has focused on static assessments. However, given the fact that biological diversity and processes that threaten its persistence vary in space and time, conservation assessments might need to be made in a dynamic context. In addition, we must explicitly account for the trade-offs associated with implementing conservation actions and investing in improved knowledge and learning to reduce uncertainty on where, how and when to act. The aim of this thesis was to develop novel approaches for accounting for both ecosystem dynamics and uncertainty in conservation planning. Ecosystems are generally treated as static in conservation planning despite many being spatially and temporally dynamic. For example, pelagic marine ecosystems are quite dynamic because ecological processes, such as eddies, that produce resources that many species depend on can be erratic. In chapter two we explored the issue of developing a system of fixed protected areas that consider the physical and biological dynamics typical of the pelagic realm. The approach was to maximize the representation of key fisheries species and species of conservation concern due to significant declines in their abundance, within a network of protected areas. We also ensured that protected area design reflected system dynamics and this was achieved by representing key oceanographic process (such as upwellings and eddies), and biological processes (such as the abundance of small pelagic fish) in protected areas. To account for the variability where these processes occur, we used time series data to find both predictable areas and anomalies, assuming that their past location was somewhat reflective of their future locations. Implementing conservation actions that are fixed in space and time are probably not the most effective strategy in ecosystems that are dynamic. This is because of the movements of particular species. For example, many species have distributions and abundances that change seasonally and might only require temporary management in particular areas. In chapter three, we tested the utility of three approaches to implementing fisheries closures to reduce bycatch in the South African Longline Fishery; 1) time closures, 2) permanent spatial closures and 3) episodic spatial closures. In chapter three, we identified these closures using an existing database containing catch and bycatch data from 1998 to 2005. There was variation where and when different species were caught as bycatch, and it was determined seasonal area closures were the best strategy. This was because it achieved the same conservation objectives for bycatch species as the other types of closures, but impacted less on the long-lining industry. While this result is intuitive, it demonstrated quantitatively, how much more effective moveable management can be. Decisions on where conservation actions are implemented are always based on incomplete knowledge about biological diversity. It is generally assumed that gathering more data is a good investment for conservation planning. However, data can take time and incur costs to collect and given habitat loss, there are both costs and benefits associated with different levels of investments in knowledge versus conservation implementation. In chapter four, the aim was to determine the return on investment from spending different amounts on survey data before undertaking a program of implementing new protected areas. We found that, after an investment of only US$100,000, there was little increase in the effectiveness of conservation actions, despite the full species dataset costing at least 25 times that amount. Surveying can take time because of expertise limitations, logistics and funding shortfalls. Biological diversity may be lost while data collection occurs conversely, not collecting enough data can lead to erroneous decisions. Additionally, resources spent on learning may be better spent on other actions. In chapter five, in a series of retrospective simulations, we compared the impact of spending different amounts of time collecting biological data prior to the implementation of new protected areas. The aim was to find the optimal survey period given the trade-off between gaining knowledge to improve conservation decisions while there is concurrent loss of habitat. We discovered that surveying beyond two years rarely increased the effectiveness of conservation decisions, despite a substantial increase in the knowledge of species distributions. Often there are choices between different actions and uncertainty as to which are the most effective. In chapter six, we discuss how the principles of adaptive management might be applied to conservation planning. Improving future management decisions through learning should be viewed as essential in all conservation plans but such learning is often included as a minor step, or is completely ignored. In this chapter we provide a brief overview of an adaptive framework for conservation planning and ideas for future research.