Report No FR/D0023

LINKAGE OF FLOOD FREQUENCY CURVE
WITH MAXIMUM FLOOD ESTIMATE

FR/D0023

SEPT 1995

SUMMARY

EXECUTIVE SUMMARY

Some structures such as reservoir dams are designed to withstand very rare flood events. These events may be estimated in terms of their probability of being exceeded or in terms of a 'probable maximum' flood. A flood with a probability of being exceeded in any one year of 1 in 10 000 is also described as having a 10 000 year 'return period'. A graph displaying the variation of flood magnitude with probability is called a flood frequency curve; it does not normally extend up to the level of the probable maximum, although this may be shown on the same graph as a horizontal line limiting upwards extrapolation of the curve.

This report describes a systematic way of producing a composite flood frequency curve defined up to the level of the probable maximum flood. The method is generally applicable but is developed with the UK and the Flood Studies Report (FSR) principally in mind.

The FSR includes a number of complementary techniques for deriving a flood frequency curve and makes recommendations as to when each should be favoured. Methods of combining curves are less well treated and this report addresses that particular problem first.

Although the question of associating an annual exceedance probability with the probable maximum flood (PMF) has, despite the apparent contradiction in terms, often been debated, there has been no standard technique for plotting the PMF at a particular point on the same plot as the flood frequency curve.

This report adapts a technique first developed in Australia which enables such a point to be defined and goes on to describe a systematic interpolation between the defined (up to 1000 year return period in normal UK work) flood frequency curve and the probable maximum value.

The proposed method is independent of any particular distribution which might lie behind the development of the frequency curve, of any assumptions made, or of whether or not estimates are based on data or on regional relationships. It is a purely mechanical device which has the benefits of consistency and objectivity.

A final step is to describe a method for producing a complete hydrograph beneath a peak which lies between two other peaks of defined hydrographs. The two known hydrographs can be of any shape and size.

The methods are described fully in the text and are available also as computer programs on diskette.

Copies of the report are available from FWR, price £20.00, less 20% to FWR Members.