Hydraulic Modelling of River Systems

Modelling river systems, using computers, is a powerful tool in river engineering. The model can be subjected to various storm events, and the behaviour studied. Modifications can be tested on the model before they are constructed. A flood defence scheme can be tested in extreme storm events which may not be seen in a lifetime without any risk.

Where possible, the model is verified from actual rainfall data (and data from flow gauging stations where available), and records of actual flooding events. Data is often sparse, however, and theoretical calculations of runoff into the rivers are limited in accuracy (see Hydrology).

Modelling a pipe distribution network is reasonably straightforward. Only water supplied to the system can contribute, and inflow and outflow can be predicted. Modelling a piped drainage system has the added complication of uncertainty of inflow and the possibility of flooding at manholes.

With river modelling, the hydraulic conduit (the river), changes capacity with water level. The flow may leave the banks, either to be stored during the flood event or finding it's own way downstream. The behaviour of hydraulic controls and constraints may change during a storm event as the river levels change. As a result, the output of river modelling is less certain than other forms of modelling, and requires much more data in the form of flood channel survey to set up the model.

There are two main types of modelling package. The simplest in concept is the 'Steady State' model, where the software will calculate the flows through the river system as defined. It cannot take account of any storage in the system. An example of this approach is HecRAS, initially developed by the US Bureau of Reclamation for major rivers, which respond so slowly to storm events that they can be treated as a series of steady state conditions.

In real storm events, the flow is constantly changing as floodwater is stored out of the river channel then returned as the level subsides.Fully dynamic models have been developed, which model the changes in storage and flows throughout the storm event. These are typified by ISIS, from Wallingford Software, and Mike11 from the Danish Hydraulics Institute. Although these are powerful tools, they require a great deal of input data and considerable care to yield reliable results.

Hydraulic modelling for rivers has only been generally available for practicing engineers in the last 10 years or so. It is increasingly common for modelling to be used to understand the behaviour of rivers and predict the consequences of change. These techniques were used to understand the 1998 Easter floods in Northampton, which led to revised guidance for developers in the Nene Valley by the Environment Agency.