Abstract:
Urban stormwater presents many common problems such as flooding, pollution and ecological degradation of freshwater and coasts. Part of a potential solution to these problems is to engage urban residents in stormwater conversations and thus prompt action. One way to achieve this might be using hydrological modelling. These models portray information which is out of sight and out of mind and so are suited to the urban stormwater problem. This thesis, then, examines the technical challenges of developing stormwater models for a large urban audience.
Proof of concept flow path models are developed using relatively simple and tested methods. Stream networks and stormwater infrastructure are burned into a 1 m resolution digital elevation model, and the D8 algorithm applied to visualise stormwater flow paths. This proof of concept modelling is used to assess the feasibility and technical challenges of developing an urban flow path model on a city scale.
Models are tested to demonstrate their reliability, accuracy, and computational performance. Fundamentally, any point within the catchments produces a flow path to an outlet. Runtime on a standard desktop computer was found to be between 0.1 and 2 seconds. Using dedicated and more powerful computer infrastructure would accelerate the generation of these maps substantially. A key outcome is a formalisation of the challenges to represent stormwater infrastructure in a 2D model. These challenges include representation of buildings in a DEM, curb and channel features, assigning flow directions to underground pipes, and depicting flow paths near bridges and culverts. Several approaches to dealing with each are presented so that future research can address them.
After dealing with the technical challenges, the thesis turns its attention to the tensions involved in developing such models (a technical exercise) with a deliberate hydrosocial starting point (a social perspective). A key finding is that the disciplinary path by which the hydrosocial is approached, and where it is invoked in the modelling process, can significantly influence the models themselves. Crucially, a hydrosocial framing changes the conversation on why we model and perhaps more importantly reminds us that the value of such models is the (political) work they do and not necessarily the technical outcome.