The flood hydrology of urban catchments in Greater London

• Main Author: Eyre, W. S.
• Published: University College London (University of London) 1979
• Subjects: 627
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.455050

The thesis investigated four south London catchments which drain northwards to the River Thames over superficial deposits which overlie clay in the North and chalk to the South. The catchments are densely urbanised ranging from 32.2 to 80.8 percent and range in area from 43.5 to 176.0 Km2. After evaluating several deterministic sewer and flood routing models it was decided to analyse 96 storm events by the unit hydrograph method. Testing of alternative identification techniques using objective error functions indicated that matrix inversion of response runoff and effective rainfall calculated by the loss rate curve was the most consistently accurate. Analysis of the unit hydrographs indicated that those with high peak discharges and short times to peak were caused by short duration, high intensity storms on a dry catchment, whereas unit hydrographs with a small peak discharge and a long time to peak were caused by long duration, low intensity storms on a wet catchment. The unit hydrographs from the four catchments showed no significant change through time. The mean unit hydrographs of the four catchments were not related to the degree of urbanisation, but to the physical characteristics of the catchments. The unit hydrographs were analysed and split tested using four different models. A quasi-linear, straight line approximation of the unit hydrograph proved to be the most consistently accurate. A subsidiary analysis compared the performance of seventeen linear conceptual models but was not followed up. A sensitivity analysis was conducted on the thesis's findings and quantified the significant effect of rainfall separation and profile on peak discharge and spill volume. The effect of substituting a straight line approximation of the unit hydrograph was examined and found to have a minor effect on peak discharge estimates but a more significant effect on spill volume.