NSFGEO-NERC: CHANCE - Understanding compound flooding in the past, present and future for North Atlantic coastlines

Lead Research Organisation: University of Southampton
Department Name: School of Ocean and Earth Science

Abstract

Floods are the most dangerous and costly of all natural hazards. From 1980 to 2013, floods accounted for more than $1 trillion in losses and resulted in at least 220,000 fatalities globally. More than 50% of these deaths, and a large proportion of the losses, occurred in densely populated low-lying coastal regions, especially those at the coastal-river interface. Continuing to advance our understanding of flooding is therefore of utmost importance.

In coastal regions, floods are often caused by multiple factors. Floods can arise through the joint occurrence of factors such as (1) storm surges plus astronomical tides (storm-tides) and/or (2) local or remotely (swell) generated waves; but also from heavy precipitation, either through (3) increased river discharge (fluvial) and/or (4) direct runoff (pluvial). Most flood risk assessments to date have considered these four main drivers of flooding separately. However, the adverse consequences of a flood in coastal regions can be greatly exacerbated when the oceanographic (storm-tides and waves), fluvial, and/or pluvial sources of flooding occur concurrently or in close succession, a condition known as 'compound flooding', which can result in disproportionately extreme events. Despite their high impact potential, compound events remain poorly understood, in large part because of the lack of information on the inter-dependence of the driving factors, which varies considerably from place to place, and the perceived difficulty of the joint probability analysis methods required to analyse these interdepencies. This is why the World Climate Research Program Grand Challenge on Extremes has identified climatic compound events as an international research priority.

A recent example of a compound event is that associated with Hurricane Harvey in 2017. Record breaking rainfall, river discharge and runoff, combined with a moderate but long-lasting storm surge, resulted in disastrous flooding in Houston. It was the second costliest natural disaster in US history. Moreover, it is recognised that, by not considering compound flooding, the risk to Houston and elsewhere had been, and still is, greatly underestimated.

In CHANCE we will deliver a new integrated approach, incorporating all the spatial and temporal dependencies between the four main source drivers of flooding in coastal regions. This will allow us to make a step change in our understanding and prediction of the source mechanisms driving compound flood events in coastal areas around the North Atlantic basin. We will address the following key questions:

1. Where do (and where don't) compound events occur and which combinations of source-variables are most important in different regions?

2. Which weather types favour the occurrence of compound events and will the frequency of compound events increase/decrease in the future as weather patterns change?

3. What is the likelihood and spatial extent of compound events in different regions?

4. How do compounding effects from multiple flood sources exacerbate impacts to coastal communities?

We will do this through a series of methodological innovations (e.g., novel dependence analyses and state-of-the art weather typing approaches, along with inventive multivariate extreme value analysis techniques and advanced ensemble hydrodynamic modelling) that not only have relevance to the serious issue of compound flooding, but which will also be transferable to other cascading hazards in the earth and environmental sciences, such as: heat waves, drought and bush fires; extreme rainfall, landslides and cliff falls; earthquake and tsunami; and river discharge and turbidity currents. Our new methods will enable us to fully assess and predict all the source variables associated with compound flood events and their spatial extents in coastal regions (past, present and future) and will result in a major advance in the way compound flooding is understood, quantified and managed.

Planned Impact

CHANCE will develop new understanding of compound flooding. In close collaboration with our project partners, we will provide guidance on where compound flooding will be most significant and should be considered in flood forecasting, management and planning. The project's major impact will therefore be in the delivery of new capability in mapping, characterising and predicting compound flooding potential. The project will lead to a major advance in the way compound flooding is understood, quantified and managed, resulting in increased preparedness and resilience. Although our focus is on the North Atlantic, our results and policy driven impact will be relevant to other coastal regions worldwide, and in other disciplines where multi-hazards pose a risk.

The pathways to impact have been developed with, and will be delivered through, direct, regular and two-way engagement with multiple end-users in the UK (Flood Forecasting Centre, Met Office, EDF Energy, HR Wallingford, Environment Agency) and US (Army Corps of Engineers, US Geological Survey and National Oceanic & Atmospheric Administration).

Our partners at the UK Flood Forecasting Centre (FCC) use an operational forecasting tool called Coastal Decider, which is based on predefined weather patterns. A key output of our proposed project will be a database of the synoptic-scale weather patterns that led to compound events in the past. As part of the project, we will deliver a major upgrade to Coastal Decider by directly integrating our database of synoptic weather patterns into Coastal Decider. This will allow the FCC to flag early forecast events that are displaying similar meteorological characteristics to previous compound events.

Project results will feed directly into the Environment Agency's (EA) 'Risk of Flooding from Rivers and Sea' dataset. This is the standard dataset for determining categories of flood risk in the insurance industry, yet it currently ignores compounding effects. Project outputs will also inform the UK National Risk Register (NRR) which likewise does not yet consider compound flooding.

The project will also shape future iterations of the EA's 'Coastal Flood Boundary Conditions'. To date the Coastal Flood Boundary Conditions dataset just considers storm-tide and wave flood sources independently and does not include information on compound events or the spatial footprints of such events. Outputs from our project, and the innovative methods we will develop, will be used to greatly improve a planned future version of the dataset, providing a significant step forward in the reliability of information available for stakeholders.

We will work closely with EDF Energy to feed new information on compound events into their operations and safety policies for the nuclear industry; particularly in relation to their nuclear new builds. EDF are also interested in applying our methodologies to other cascading hazard problems.

Our results will also complement ongoing activities undertaken by our partners US Army Corps of Engineers, US Geological Survey and National Oceanic and Atmospheric Administration for various impact and risk mitigation studies where compounding effects are of particular relevance.

The three main ways in which we will engage with end-users are via stakeholder workshops, secondments and a policy brief. We will hold four three-day workshops throughout the project, three in the UK and one in the US, which will be attended by our project partners. At these workshops we will present methods and results on each component of the project, and the regular interaction with end-users will enable them to feedback on key findings and shape the direction and outputs of the project as it progresses. We will also prepare and disseminate a policy brief, in collaboration with the University of Southampton's dedicated Public Policy knowledge transfer team, which will summarise the main implications of our work.

Publications

10 25 50