Reconciliation between Flood Frequency Analysis and Rainfall-runoff Model Results

This study investigates the effectiveness of flood frequency analysis as a tool for verification of hydrological models. The study methodology is based on a case study at seven catchments on the Gold Coast, Australia. The study contributes to the body of knowledge by highlighting a number of pitfalls that hinder achieving reconciliation between flood frequency and hydrological modelling results. These pitfalls appear to be mainly associated with uncertain data. Rainfall maps that are used as input in hydrological modelling can be an important source of discrepancy, if they are not up-to-date and adequately filtered with respect to sub-duration inconsistencies. Inaccuracy of rating curves at measurement stations, inappropriate initial and continuous loss parameters; inaccurate Areal Reduction Factor and disregarding non-linear response of catchments are other contributing factors. A mismatch between the length of data set that is used for the generation of rainfall maps and the length of the data set that is used for flood frequency analysis can also be a reason for discrepancy between model outputs and flood frequency analysis results.