Fluvial, storm-surge and tsunami flooding represent one of the largest hazards to coastal and river population centres. For example, the low elevation “coastal zone” represents <2% of the Earth’s total land area, but supports >10% of total world GDP. The future risks posed by large flooding is normally assessed by studying gauged records of past flooding – but these are temporally limited by the length of historical records and spatially restricted by the sparsity of where gauges are located. With our knowledge of flooding only based on a few comparatively modern examples we are, therefore, ill-equipped to prepare for, or assess the likely change in, future flood risk to river or coastal infrastructure. To enhance this historical record of flooding, researchers have looked to the sedimentary record within river floodplains, deltas and coastal areas. In effect, inverting the sedimentary record to garner a history of flood events. With such studies, it is assumed that deposits of sediment correlate with large events, or periods of increased instances of flooding. However, the relationship between sediment deposition and flood magnitude is non-linear and complicated meaning that correlation is difficult.
In other fields, such as catchment hydrology, machine learning methods have been successfully used to generate good predictions of how systems respond as well as, to a lesser extent, increase our understanding of the linkages between model inputs and outputs. Therefore, this project will advance new machine-learning methods to invert, and distinguish between, past records of terrestrial flooding, storm-surges and tsunamis, that are preserved in sedimentary deposits in river flood plains or coastal zones. Inversion of these deposits will provide data on past flood events, addressing the current weakness in risk assessment of flooding geohazards. The inversion of flood risk critically depends on development of machine learning techniques, at different spatial and temporal scales, to link sedimentary deposits to forward fluid dynamic models. From this, the project will deliver a step-change tool for assessing the probabilistic likelihood of coastal flood risk and the current and future hazard they pose.
The successful candidate will be supported by an international team of PhD advisors, integrating support and expertise in forward modelling and machine learning from the University of Hull and Kyoto. Further, a core component of the project will be the co-creation of research with end-users, this will be delivered by on-going engagement with government and non-government organizations and industry.
The Living with Water: Perceptions, Processes, Responses and Mitigation of Flooding cluster
This PhD project is part of a PhD cluster focusing on flooding. The magnitude and frequency of flooding and the associated risks to infrastructure, economic activity and human life are known to be increasing due to changes in rainfall and storm intensity, frequency and seasonality, in conjunction with the pressures of increased development on floodplains and in the coastal zone. There is a need for society to adapt to Live with Water. This PhD cluster will focus on better understanding the impacts of flooding and improving the approaches to mitigating flood risk to societies. Improved understanding of the impacts of flooding requires an improved understanding of the processes leading to flooding and the processes active during floods; from the changes in fluxes of water and debris, through to emergency responders’ and the public’s behaviour and perception in response to flood events. The cluster brings together researchers in Sport, Health and Exercise Sciences, Life Sciences, and Computer Science with social scientists and physical and numerical modellers in Geography and Geology.
Applicants should have at least a 2.1 undergraduate degree in engineering, computer science, physical geography, earth sciences or related disciplines, together with an interest and ideally experience in numerical modelling and environmental systems.
For further details, please contact Prof Tom Coulthard ( [email protected]
Application deadline: Friday 13 December 2019