Design Storm Profiles
The design storm profiles used within ReFH that translate the rainfall depth to the hyetograph that forms the rainfall input to the model are the FSR 75% winter and 50% summer profiles. These are symmetrical and single peaked and the shape does not vary with storm duration or location. The original FSR rainfall profiles were based on a graphical procedure: Keers and Wescott, 1977. ReFH2 uses the equation that approximates these design profiles developed as part of the implementation within Micro-FSR (Institute of Hydrology, 1991).
The proportional depth of rain, y, falling in the temporal proportion, x, of the total duration, centred on the peak is given as:
where z = x^b and a and b are profile specific constants listed in Table 5.
Table 5. Parameters for derivation of design profiles
Note that this formula gives unrealistically large values for the 50% summer profile when a large number of time steps are used (see FEH Volume 2 (Faulkner, 1999) and Kjeldsen et al, 2005). In these cases the application of the model has been amended within the ReFH2.3 model to limit the peak intensity rainfall for both the Winter and Summer Profiles.
Faulkner (1999) (FEH Volume 2) presents a critical review of the simplicity of these storm profiles. The calibration of the ReFH design package is predicated on these profiles and thus the calibration compensates for the simplicity in terms of peak flow estimation. The event volume is independent of the profile whereas dynamic hydraulic modelling applications will potentially be sensitive to these profiles.
Research during the Small Catchments Project Phase 2 (see Report 6) compared the ReFH2 hydrograph shapes with the shapes derived using the empirical median hydrograph (EMH) method, outlined by Archer et al., 2000 across a set of 20 generally small catchments. The latter method was selected as a readily available empirical hydrograph method and was used as a comparator to ReFH2. The outcome was that the two methods agree well in 'essentially rural' and impermeable catchments. The limitations of both methods are exposed in difficult hydrological cases such as highly urbanised or groundwater catchments. In these catchments great care is required to obtain the best modelling outcome recognising the limitations of all methods.
This outcome gives weight to the argument that the ReFH2 hydrographs and the use of the recommended duration and resulting design hyetographs are appropriate for the majority of catchment types. However, as the duration extends to days, the likelihood that large events conform to a single unimodal event will be low. More typically in these cases the event will be a sequence of events. Although easy to conceptualise, representing this as a single, most appropriate design event storm profile is more problematic.
Reservoir safety is one application where the critical duration of large reservoired systems is likely to be several days reflecting the build-up of water levels over a series of storms. The latest version of the ICE Floods and Reservoir Safety: An Engineering Guide provides guidance on the development of bespoke storm profiles based upon historical observations for this case.
Within ReFH2 such rainfall hyetographs developed using bespoke storm profiles can be introduced using the observed event application within the software, see Simulating observed events using ReFH2.