The FEH Methods
Originally published in 1999, and subject to continuous development, the Flood Estimation Handbook (FEH) comprises the standard methods for estimating flood frequency. There are two core handbook methods:
- the statistical method for estimating flood frequency curves implemented through WINFAP; and
- the event-based rainfall runoff method for estimating hydrographs for floods of a specified frequency and hence flood frequency curves implemented through ReFH 2.
The Revitalised Flood Hydrograph rainfall-runoff method version 2 (ReFH2) is the current FEH rainfall-runoff model and is the subject of this guidance. Guidance on choice of method for flood frequency estimation is given in Volume 1 of the FEH, with additional guidance published by government agencies with statutory responsibility for the management of flood risk.
Uncertainty in the FEH methods is summarised in the UKCEH short guide for FEH users (Griffin et al., 2021).
The original ReFH1 method was published in 2005 by Kjeldsen et al as a replacement for the original FEH rainfall-runoff method (Houghton-Carr, 1999), the FSR/FEH rainfall-runoff method. The methods are the subject of continuous improvement and the most up-to-date implementation of the methods is though the ReFH2 software. ReFH1 was developed for use in estimates only up to 150 years using the (now legacy) FEH99 rainfall DDF (depth duration frequency) model and was found to perform poorly for high permeability catchments. Despite the known issues, there was no public funding available to further develop the methods and tools at that time. The fully published ReFH1 method has been continuously improved by WHS to address gaps in the method. ReFH2 was first released in 2015, with the demonstrably improved ‘water balance’ ReFH2.3 model released in 2019. The latest calibration of the ReFH2.3 model was released in 2023 to utilise the FEH22 design rainfall.
The most common application of the ReFH2 software is the Design Application. Here, an estimate of a rainfall depth over a specified duration and frequency is used to estimate the flood hydrograph corresponding to that duration and frequency. In this application the software is used in conjunction with a Depth-Duration-Frequency (DDF) design rainfall model and a corresponding set of design initial conditions. This design application can be applied to river catchments to inform fluvial flood risk or at the scale of a site to inform pluvial flood risk and drainage design.
A second application of ReFH2 is the Observed Event Application. In this application the software is used to estimate the flood hydrograph, and hence peak flow, corresponding to an observed rainfall event. This estimate of peak flow can then be used with a flood frequency curve derived from the design application to estimate the frequency (rarity) of the flood corresponding to that observed event. The initial catchment conditions prior to the start of the observed rainfall event are estimated using a daily soil moisture accounting procedure instead of using the design initial conditions. This application can also be used in conjunction with a semi-distributed hydrological/hydraulic model to estimate inundation depths during specific events.
To use ReFH2 you will need to have access to the FEH catchment descriptors describing the climate, drainage characteristics and soils for your catchment or parcel of land together with the DDF rainfall model data for your application. These are available for purchase and download from the UK Centre for Ecology and Hydrology’s FEH Web Service. The FEH Web Service is an online GIS based data management and retrieval tool that enables the user to select the point or catchment of interest and export the relevant ReFH (xml) input file.
The ReFH2 model structure
This technical guide presents the basic principles of the model structure. A full description of the rural model component is provided in Chapter 2 of the FEH supplementary report No1 (Kjeldsen, 2007) which describes the original ReFH1 method (noting that the other chapters have been largely superseded). The development of the original ReFH1 urban model structure is described by Kjeldsen, et al (2013).