Criteria for comparison of single event models

Abstract

Methods of assessing the goodness-of-fit of hydrological simulation models are discussed. The methods generally compare the goodness-of-fit of a simulated hydrograph with an observed one and invariably have a bias towards one or another aspect of the hydrograph. It appears that no single parameter is sufficient to assess adequately the overall measure of fit between a computed and an observed hydrograph, particularly in view of the many objectives behind hydrological modelling. Previous research has concentrated on assessing the goodness-of-fit of long duration hydrographs resulting from continuous simulation. This paper discusses a number of criteria for hydrograph comparison in single event modelling studies and applies them to a specific stormwater modelling exercise. It is shown that particular criteria highlight particular aspects. © 1986 Taylor & Francis Group, LLC.