摘要: |
A comprehensive investigation was conducted to find the conditions under which a flood frequency analysis procedure performs better than alternative procedures. Five distributions and six estimation methods were evaluated. Three evaluation criteria were used in this investigation: the descriptive test, the predictive test, and the robustness test. By using the test results, a procedure-selection table was developed to help hydrologists or hydraulic designers select an appropriate flood-estimation model. The table is also useful for those researchers who need to select a flood frequency analysis procedure to evaluate performances of selected flood-estimation models. In addition, the study showed that under certain conditions the flood estimation procedure recommended since 1967 by the Interagency Advisory Committee on Water Data may not yield acceptable quantile estimates and, therefore, cannot serve as a standard procedure for calibrating or evaluating selected flood-estimation models. |