The relative delay times of catchment model storages

The original aim of this research was to determine the best means of calculating the relative delay time parameter of the Laurenson runoff routing method of flood estimation. Early testing conducted as a part of achieving this end showed that the model could be quite insensitive to the means used to calculate relative delay. Model sensitivity to the means of calculating this parameter had apparently not yet been verified, and if it was insensitive, the best means of calculating the parameter was not required. Hence it was decided to test the thesis that the model was insensitive to the means of calculating the relative delay time parameter. The thesis was tested by applying the model to a range of flood sizes on two catchments using five different means of calculating relative delay time. The differences in model output produced by these delay hypotheses were examined in relation to the average model error. In a number of cases the effects on model performance of these different delay hypotheses were quite small, but in other cases they were quite large. Hence the conclusion had to be drawn that the model is sensitive to the means of calculating relative delay time. The fact that the means of calculating relative delay. time Has unimportant in a conditional sense left the possibility that the model could be used under these conditions, if they could be defined. However, in the long term it meant that the best means of calculating relative delay time was required. During the course of the investigation a number of problems in relation to model use were encountered. The attempts made to solve these problems should provide useful information on model application in the future.