Making better use of local data in flood frequency estimation

Dixon, Harry ORCID: https://orcid.org/0000-0002-1643-1046; Faulkner, Duncan; Fry, Matt ORCID: https://orcid.org/0000-0003-1142-4039; Kral, Filip; Lamb, Rob; Macklin, Mark; Prosdocimi, Ilaria; Reed, Duncan; Rogers, Peter; Sefton, Catherine ORCID: https://orcid.org/0000-0002-8157-0368; Stewart, Lisa ORCID: https://orcid.org/0000-0003-4246-6645; Vesuviano, Gianni ORCID: https://orcid.org/0000-0003-2157-8875. 2017 Making better use of local data in flood frequency estimation. Bristol, Environment Agency, 213pp. (Report SC130009/R, CEH Project no. C05366)

Before downloading, please read NORA policies.

Preview

Text N516408CR.pdf - Published Version Download (10MB) | Preview

Abstract/Summary

Flood frequency estimates are an essential part of flood risk management. Methods described in the Flood Estimation Handbook (FEH) published in 1999, and many subsequent updates, are the industry standard for flood frequency estimation in the UK. Even carefully calculated flood frequency estimates are associated with many sources of uncertainty. These hydrological uncertainties often constitute the most uncertain component in any flood study. Uncertainty, where it is recognised, can lead to difficulty in having confidence in the outputs of studies such as flood map outlines, designs of flood defences or other structures, decisions on new development or information needed for investment planning or insurance. It can also lead to a loss of public credibility. As a result, there is considerable benefit to be gained from any reduction in the uncertainty of flood frequency estimation. The raw material for flood estimation is high quality long-term records of river flow and rainfall. The FEH methods are based primarily on these types of data. There are many supplementary sources of information that can help to refine flood frequency and potentially reduce uncertainty. Examples include long-term flood history, river level records, temporary flow gauges, photographs of flood impacts, information obtained from field visits, measurements of channel width and evidence of flood deposits seen in the landscape (palaeoflood data). These and similar types of information are defined as local data. The FEH Local research project aimed to: quantify the uncertainty of design floods estimated from FEH methods; and, develop procedures and guidance for incorporating local data into flood estimation to reduce such uncertainties. This report describes the review of scientific developments and good practice, and the development of new procedures carried out during the FEH Local project. A companion output from the project is a document, ‘Using Local Data to Reduce Uncertainty in Flood Frequency Estimation’, giving guidance to practitioners on how to estimate uncertainty in flood frequency and how to find and incorporate local data. Following a wide-ranging review of the availability and use of local data, this report focuses on the evaluation and development of procedures for incorporating 2 main types of local data: historical (and palaeoflood) information; and, channel dimensions. A statistical simulation study examines methods for flood frequency analysis using historical data and tests the sensitivity of the results to uncertainty of aspects such as the length of the historical period or the possibility of missing some events. The study recommends a maximum likelihood technique for combining historical and gauged flood data. The technique is able to incorporate either historical floods for which discharges can be estimated or floods for which all that is known is that the discharge exceeded a given threshold. It can also be used with palaeoflood data and this report includes a critical review of the use of palaeoflood data in UK practice. A procedure is presented for estimating the median annual flood using a combination of catchment descriptors and bankfull channel width. This is an extension of a technique presented in the FEH using the same dataset of channel widths. The practitioner guidance explains how to implement these procedures and provides case studies showing how historical data, palaeofloods and several other types of local data can be incorporated in flood studies. The report presents a proposal for a new system to improve access to local data to be integrated with the National River Flow Archive. The development of this system is feasible, given clearly defined limits on the data types to be included and secure funding, with a national remit, both for the establishment and for the long-term operation and maintenance of the system. Also covered in this report is a separate aspect of the FEH Local project, a pilot study to develop high-resolution catchment descriptors and explore the potential for new catchment descriptors to replace or augment some of those currently used in FEH methods. Catchment boundaries and descriptors are derived from a 10m digital terrain model, considerably more detailed than the 50m terrain data used in the FEH. Existing FEH catchment descriptors are evaluated for all catchments down to a minimum size of 0.2km2 for a pilot area in north Cumbria. Several new descriptors are proposed, although it is not yet clear that they would necessarily lead to improved flood estimates. There is now a challenge for the flood management sector to put into practice the findings of FEH Local, so that it becomes common practice to seek and exploit local data (rather than, as currently, best practice if it is done at all). This will lead to several benefits: better estimates of design flows; reduced uncertainty; project results that are more robust to challenge; less need to seek reviews and improvement of hydrology studies; and, enhanced public credibility. Ultimately, it can be expected that an outcome will be improved protection of people and property. There are several ways to help meet this challenge, including dissemination of the practitioner guidance and encouragement or requirement of its implementation via appropriate wording of project scoping documents by the Environment Agency and other regulators. Appropriate use of local data should also be required in sector specific guidance or specifications. More rigorous reviews of flood estimates will help to challenge poor practice. There is a need for a change in culture that gets hydrologists out from their computer models more often and into the field.

Item Type:	Publication - Report
UKCEH and CEH Sections/Science Areas:	Rees (from October 2014) Reynard
ISBN:	9781849113878
Funders/Sponsors:	Environment Agency
Additional Information. Not used in RCUK Gateway to Research.:	Freely available via Official URL link.
NORA Subject Terms:	Hydrology
Date made live:	01 Mar 2017 13:05 +0 (UTC)
URI:	https://nora.nerc.ac.uk/id/eprint/516408

Actions (login required)

View Item

Document Downloads

More statistics for this item...