An appraisal of the regional flood frequency procedure in the UK Flood Studies Report

The algorithm for estimating the regional flood frequency hazard contained in the 1975 Natural Environment Research Council Flood Studies Report (FSR) can occasionally lead to upper quantile estimates that appear unrealistic when compared with engineering judgement. Tests with the FSR algorithm were made for several sets of observed flood sequences and a great variety of synthetic data in a Monte Carlo simulation study. Similar tests were conducted with many other regional and at-site flood frequency estimation procedures including a regional generalized extreme value distribution (GEV) procedure and a regional Wakeby distribution (WAK) procedure, both of which used biased probability weighted moments (PWM) in their formulation. For the Monte Carlo simulations, for which the true quantiles to be estimated were known, it was found that the FSR algorithm yielded quantile estimates that were always more variable, often by a factor of as much as 4 or 5, than those obtained by either the GEV/PWM or WAK/PWM algorithms. Further, the FSR algorithm was found to yield quantile estimates that were nearly always more biased than those obtainable by either the GEV/PWM or the WAK/PWM procedures. The above results were quite general, and were particularly noticeable when the simulations used data bases that were of similar size to those customarily encountered in hydrological practice. The available data base for the UK should now be much larger than it was at the time of the preparation of the Flood Studies Report, and it is suggested that an immediate effort is needed to produce an augmented, complete, consistent UK flood data base for the current period of record. Further, it is believed that analysis of this new data base using the GEV/PWM algorithm may well lead to more accurate and more consistent estimation of the extremal flood hazard for the UK.