Hydrometeorological characterisation of urban flash floods

Student thesis: Doctoral ThesisDoctor of Philosophy (PhD)

Abstract

Under a climate change scenario, the frequency and intensity of hydrometeorological extremes is likely to increase. Amongst current natural hazards, floods account for the largest and costliest events and these have caused significant economical and societal losses which sometimes taken years to recover from. Flash floods in particular have a greater potential for damage given their associated quick onset and inefficient response from the population. From their many causes, flash flooding from intense, localised rainfall in urban areas represents a major challenge to forecasters and this is reflected in the insufficient adaptation and mitigation strategies in terms of awareness and preparedness.

A crucial step towards flash flood risk reduction is the improvement of current numerical modelling capabilities. Given that there are many approaches to simulate and link the physical processes that lead to an urban flash flood, there is a pressing need to define strengths and weaknesses of the current numerical modelling tools to select the most efficient models and approaches that facilitate a quantitative assessment on hazard and exposure.

The research presented here introduces a methodology for the hydrometeorological characterisation of urban flash floods at sub-daily and catchment scales (i.e. less than 100 km2), which focuses on the parameterisation of cities to determine the influence of the urban canopy on atmospheric processes and in the response to intense rainfall. It aims at providing information on the capabilities and limitations of the numerical tools involved while identifying how to improve their efficiency and accuracy. It also accounts for the specific layout of a given city (thus potentially transferrable to any urban environment), proposing an advance towards the accurate numerical representation of these events.

This study presents, for the first time in the context of flash flooding, the use of a widely applied numerical weather prediction tool that has the capabilities to account for urban areas in the atmospheric processes during the origin of intense rainfall. It replaces the use of satellite-derived, remote-sensed and ground-based rainfall information (and the uncertainty associated) and instead provides information on the degree of contribution of model structure and parameters to capture the critical development and magnitude of intense rainfall. Outputs of this meteorological tool are used as climatological forcing for a hydrological model. A recent benchmark study consolidated its robustness as a highly flexible numerical tool for rainfall-runoff simulation at daily scale, so the inclusion of urban areas and the evaluation of hourly variations of river flows represents a novelty.

Two major flash flood events in the United Kingdom in the past 20 years were selected as case studies given the magnitude of damages and losses in major English cities. The proposed methodology evaluates the impact of the parameterisation of cities in the meteorological and hydrological models on the simulated rainfall and flows, respectively, and pinpoints the most suitable configuration for further applications via the quantification of the error propagation.

The results provide information on the effectiveness of the novel framework proposed and areas for its improvement, while opening the discussion on its potential to be applied to further case studies. This shows that the framework proposed contributes to the improvement of numerical tools to reproduce and map urban flash floods, therefore strengthening the basis of strategies for flash flood risk reduction.
Date of Award24 Mar 2020
Original languageEnglish
Awarding Institution
  • The University of Bristol
SupervisorPaul D Bates (Supervisor)

Keywords

  • Hydrology
  • Meteorology
  • Urban
  • Flash floods

Cite this

'