FRACAS: a next generation national Flood Risk Assessment under climate ChAnge Scenarios

Lead Research Organisation: NERC Centre for Ecology and Hydrology
Department Name: Boorman

Abstract

This proposed research will develop the new methodology required to make a step-change in our ability to quantify fluvial flood risk at large scales, incorporating climate change. This will combine existing and emerging technologies, to provide national and regional estimates of flood risk based on gridded models for improved assessment of flood risk to recurrence intervals in excess of 50 years. Linking gridded rainfall, runoff, flood defence performance and flood inundation models will significantly improve our ability to assess flood risk from extreme events and explore the potential impacts of climate change, including new scenarios, as they become available from UKCIPnext. This will include a spatially and temporally consistent gridded rainfall model operating over large spatial domains, a high resolution gridded runoff and flow routing model capable of modelling at the national scale and a continuous system analysis of flood inundation, taking account of defence performance. As each of these models will be run continuously in time, a continuous, linked flood risk analysis system will be developed for the first time. Each model will also be able to use derived future changes in climate to produce predictions of future in flood risk. Moreover there will be an assessment of the model and data uncertainties, as well as estimates of uncertainty due to climate change. These uncertainty assessments will include the propagation of uncertainty through the linked modelling system. The research will utilise many existing sources of data and build upon some established models and techniques, such as the Neyman-Scott Rectangular Pulses (NSRP) stochastic rainfall at the University of Newcastle, the CEH Grid-to-Grid (G2G) model, the RASP system models, and the use ensemble scenario sets to represent uncertainty. At the regional or large basin-scale analyses will include a grid-based (5km) rainfall model linked to a (1km) gridded runoff and routing model and associated knowledge of defence systems and new routines developed to translate rainfall to river levels. Such a modelling system is ultimately applicable at a national scale and this will be demonstrated for river flows. The precipitation for this demonstration will be sourced from observed rainfall datasets, or modelled time series, such as those available from RCMs driven with re-analysis data. The impact of future changes in rainfall, runoff and river levels on flood risk will be assessed within an enhanced version of the HR Wallingford RASP HLMplus model. Scenarios of climate change will be derived from a range of both global (GCMs) and regional climate models (RCMs). There will also be an analysis of the application of multi-ensemble climate scenarios and the generation of probabilistic scenarios of change in future flood risk.
 
Description Models were developed and improved by each of the partners for the purposes of developing the FRACAS system. This included the NCL spatial rainfall model, new processes within the Met Office high-resolution regional climate model, novel application of new data from these rainfall models for use in a new version of the CEH Grid-to-Grid model, and the extension of the hydraulic modeling capacity at HR-Wallingford. The project generated the first application of the outputs from the new spatial rainfall model (Newcastle University) for present and future scenarios for input to CEH Grid-to-Grid hydrological model. The CEH Grid-to-Grid (G2G) model was "coupled" with the UKCP09 Regional Climate Model (RCM) ensemble to provide, for the first time, UK-scale estimates of river flows at a 1km resolution for a set of 11 current and possible future climates. The national G2G flow estimates were used as input to the HR-Wallingford hydraulic model, RASP, which converts them to river levels, flood inundation and risk. A second catchment-scale hydraulic model, Infoworks RS, demonstrated how gridded flow estimates can drive a 2-D inundation model connected to a property database which can assess the probability of various defence failure scenarios and resulting economic damages. A newly-developed stochastic weather generator can potentially provide higher resolution estimates of weather variables than the RCM, but for a limited area. FRACAS has enabled a comparison to be made between river flow estimates driven by (i) 5km weather generator and (ii) 25km RCM output for a catchment in the North West of Britain, the Eden to Carlisle. Use of extended lengths of weather generator data (~100 years) can extend the return period estimates of river flows to higher return periods than are possible when using shorter lengths of observed or RCM data.
Exploitation Route These outcomes have been taken forward int he NERC KE grant (FRUITFUL: NE/H002420/1.
Sectors Environment

 
Description The result and outcomes have been taken forward in the NERC KE grant (FRUITFUL: NE/H002420/1.
Sector Environment
Impact Types Economic