Unlocking the potential of surface water flood nowcasting for emergency services in a changing climate

Surface water flooding affects 3.2 million properties in England, and is seen as "the biggest flood risk of all" by the Environment Agency. Average annual damage due to surface water flooding in the UK exceeds £290 million. Climate change is expected to increase the intensity of heavy bursts of rainfall, and weather is expected to become more 'uncertain' and 'unfamiliar'. The cost of associated damage under climate change could rise by 40% by the 2050s if current management approaches remain unchanged.

In the UK, Ambulance and Fire & Rescue Services are the primary emergency responders to extreme flood events, during which demands for services often peak. For example, London Fire & Rescue Service saw more than a three-fold increase in emergency 999 calls attended during flooding on the EU Referendum Day (23 June 2016). Surface water flooding such as this affects the operation of emergency responders who have to operate under flood conditions, while meeting their mandatory response time target. For example, the majority of Fire & Rescue Services in the UK aims to reach incidents within 8-10 minutes. Flood incidents compound the challenges of meeting their response time targets. For example, during the 23 June 2016 flooding, the average response time of London Fire Brigade increased from 6 minutes under normal days to 16 minutes, thereby missing the response time target. 40% of incidents were reported to be delayed due to weather, flooded roads, and congestion. To combat this and enable effective decision making, knowing where and when it could flood is important.

The operational decision making of emergency responders often involve determining when and where to allocate resources (e.g. closing flooded roads; pumping road sections to allow access; dispatching sandbags; positioning emergency vehicles; arranging cross-boundary and multi-agency operations). However, current system for predicting surface water flooding is not designed for operational purpose. The only (based on our knowledge) system exists is the UK Flood Forecast Centre's daily service, which provides 5-day outlook of flood risks including surface water flooding. A flood guidance document is issued daily and surface water flood risk is provided at the county-level, based on a pre-fun library of impact scenarios, rather than real-time analysis of risks. The spatial resolution and temporal frequency of the existing system means that such detailed decisions cannot be supported. Fundamentally, episodes of surface water flooding are typically less than 2-3 hours in the UK and associated with convective weather systems. Daily forecasts of surface water flooding are not able to capture the spatiotemporal dynamics of the fast-developing convective storms associated with such events.

Under two NERC projects (2016, 2018), we developed the first nowcasting technology for surface water flooding. This approach involves around the clock real-time modelling of surface water flooding at street-level resolution for the next 3-6 hours, updated every 2-3 hours. In this way we use the latest weather forecast from the Met Office and capture every short duration intense rainfall events that cause flooding. Our stakeholder-driven research and innovation in surface water flood nowcasting have opened up several new, untapped research and innovation opportunities. The proposed project aims to address two key research questions in order to unlock the potential of surface water flood nowcasting for emergency services to support their operational decision making in a changing climate. These include: (i) uncertainty propagation from precipitation nowcasting and forecasting products to high-resolution surface water flood predictions; and (ii) effective communication of complex surface water flood risk information to support emergency responders' operational decision making.

The immediate impact of the research outcomes will be on emergency responders in the UK, who have to respond to and operate under flood conditions, meeting their service target. We choose Fire & Rescue Service to constrain our efforts in this 12-month project. However, understanding gained through this project, visualisation tools and risk metrics co-developed with stakeholders are of great value to many Category 1 emergency responders under Civil Contingencies Act 2004 (e.g. Ambulance Service and local authorities). Emergency responders will be able to allocate their limited resources when flood strikes with more confidence whilst knowing the uncertainties in flood predictions based on the risk metrics and visualisation tools developed.

Project outcomes will be hosted by the Cabinet Office's ResilienceDirect, which is the data portal for emergency responders across England and Wales, and used by emergency responders regularly. The project team has been given access since 2014. Its national reach, flexibility and mapping capacity make it an ideal place to host the datasets to be generated from this project. Results can be accessed by all stakeholders and evaluated by them for future improvement. A dedicated surface water nowcasting module is envisaged within ResilienceDirect (see letter of support). The Cabinet Office committed £50,000 to develop this capacity within ResilienceDirect (see letter support).

An important venue for international reach is the Emergency Planning College, which is part of the Cabinet Office Civil Contingency Secretariat and regularly hosts training for emergency responders from overseas. Cabinet Office will help explore means to contribute to the college as detailed in the letter of support. Our project could potentially contribute to the training courses offered by the college, including, e.g. "Operational Emergency Management" which aims to equip operational emergency managers with information, knowledge and tools to influence the development of integrated emergency management within their own and partner organisations, and "Improving Risk Management in Civil Protection", which aims to examine risk management in civil protection, in order to improve practice in Local Resilience Forums. Deliverable 3 will aim to deliver such a training course/workshop in the Emergency Planning College.

Apart from through the Emergency Planning College, we will also disseminate the project outcomes internationally. PI Yu and Co-I Wilby have active research links internationally. PI Yu has been collaborating with a number of universities in China since 2008 and a lot of his work in Southeast Asia has been published, jointly with academics in that region. Similarly, Co-I Wilby works extensively in Africa and East Asia, providing support to policy making and generating research outputs. Both Yu and Wilby also work with universities in the US, and more recently on the resilience of emergency responding in New York City jointly with the Princeton University. We will also seek to attend an international conference (American Geophysical Union Fall Meeting 2019, see Justification of Resources) and publish in high-impact journals (a manuscript in month 9). Through these venues, the impacts of this project will be internationally far-reaching.