Characterising urban catchments for explaining storm runoff and application in UK flood estimation

The impacts of urbanisation on catchment hydrology have been the focus of investigation over the last few decades, but quantifying and predicting the impacts remains an ongoing area of active research. One such area has been improving characterisation of urban land cover to predict urbanisation impacts whereby lumped catchment characterisation of urban land cover limits the ability of attribution and modelling methods to consider the spatial role of land cover in runoff response. This thesis evaluates the potential for spatially explicit characterisations of urban land cover based on landscape metrics, commonly employed in landscape ecology, to explain storm runoff in urban catchments and their application in UK flood estimation methods. Rainfall and channel flow monitoring across two towns containing 18 variably urbanised sub-catchments were used to provide high-resolution time-series of rainfall and runoff and to identify storm events which were quantified using a range of hydrological metrics. Analysing storm runoff along a rural-urban gradient showed a lumped measure of urban extent can generally explain differences in the hydrological response between rural and urban catchments but not between more urbanised catchments in which soil moisture does not play a contributing role. Using high resolution geospatial data can improve the representation of the urban environment and landscape metrics can better represent the form and function of urban land cover, improving estimates of the index flood QMED over lumped catchment descriptors. Regression analysis of hydrological metrics showed the potential of landscape metrics for explaining inter-catchment differences in rainfall-runoff and point to the importance of considering the location and connectivity of urban surfaces. Landscape metrics provide a workable means of overcoming the limitations inherent in using lumped characterisation of complex urban land cover and their ability to express connectivity, size and location of urban land cover promises potential applications in hydrological applications such as UK design flood estimation methods.