New Zealand’s natural hazardscape has long posed risks for government owned water infrastructure and its stakeholder communities. In the context of climate change, there are predictions that the frequency and intensity of natural hazard events are set to increase, but there are uncertainties about when such events might occur, and the severity and impact on water infrastructure. These threats are compounded by pressures such as population growth and constrained funding. There is a need, therefore, to consider the suitability of current governance arrangements and decision processes for the management of complex, long-term risk associated with state assets. This research, therefore, asks, are current local governance arrangements and associated decision-making processes for the management of water infrastructure appropriate for long-term risk management in an environment of uncertainty? The thesis addresses this question by critically evaluating the extent to which governance arrangements enable effective risk management decisions that protect government owned infrastructure from natural hazards. A novel conceptual framework which matches the type of problem within the scales of time and complexity is developed on the basis of a literature review on risk management, adaptability and resilience, and modes of governance across multiple levels. This framework is then used to inform an exploration of risk management decision processes. The examination of these processes involved a participatory design involving local and central government officials and associated risk management professionals. A mixed-methods approach was applied to explore experiences of risk management as it relates to water infrastructure within local government, within the context of wider relationships with central government. Following a pilot workshop with sector actors and risk professionals to establish the broad research focus, the research involved 36 semi-structured interviews and a further workshop with sector officials to discuss the findings and critique the resulting decision-making model. The research found that current risk management decision making is characterised by, largely, a short-term horizon, this being connected with the tendency of elected representatives to operate within the three-year election cycle. The absence of a longer-term focus is explained in part by the autonomous environment within which councils operate, removed from central government oversight, and with little sharing of knowledge, expertise or costs of risk management. The key findings of the research indicate: a lack of consistent guidance and direction by central government to local government on adaptive risk management; a lack of coordination across the network of councils, specifically in terms of sharing knowledge, data and risk management activities; a lack of specificity in legislation regarding long term risk management which leaves risk decision making open to interpretation by councils; a culture of localised short term decision making in councils, with a low level of comprehension as to the management of long-term risks; short-term competing priorities for the financial resources required to manage risk; and a mistrust of internal guidance and advice due to the lack of in-house expertise, and a preference for external subject matter expertise. Overall, this research has identified that the complex nature of the environment of uncertainty posed by climate change has not been adequately acknowledged within local government and this is reflected in the prevailing risk management arrangements being largely reliant on annually renewed insurance arrangements based on in-house planning practices. The decision-making framework developed in this thesis acknowledges the nature of risk and the need for adaptive approaches within the constraints of time and complexity, and has potential to inform both governance arrangements and decision processes.