Overview

The most common requirements for modelling observed events are:

• estimation of return period assessment of notable flood events for ungauged catchments; and
• evaluation of catchment scale hydrological/hydraulic models against evidence of flood inundation levels for observed events.

In the first of these applications there are unlikely to be suitable at site observed records to calibrate the model against (otherwise event severity could be assessed directly from the record). In the latter application, the sub-catchments draining to the reaches modelled explicitly using the hydraulic model may be a mix of gauged and ungauged catchments.

The observed modelling functionality within ReFH2 uses the ReFH2.3 water balance formation of the design package and assumes the model parameters are based on the design catchment descriptor estimates, with user modification as required. Instead of using the design rainfall hyetograph, the ReFH2 observed rainfall option for requires the user to import the catchment average hyetograph for the observed event. There are no restrictions on the timestep used for the hyetograph other than that it does not change within the event and is sub-daily. The $ARF$ and $SCF$ are not applicable within the observed event case as the input hyetograph is the catchment average rainfall hyetograph as derived from observations at appropriate rain gauges.

As for the design case, estimates of $C_{ini}$ and $BF_0$ are required. However, the requirement in this application is that these reflect the antecedent soil moisture content and baseflow in the river at the start of the observed event being modelled. In the absence of measured river flow the default $BF_0$ is estimated from the estimate of $C_{ini}$ using the relevant seasonal $BF_0$ equation presented under Initial Conditions Design Estimates.