Estimation of Peak Discharges Using Flood Frequency Analysis and Hydrological Modeling System

Abstract

This study presents the methodology to estimate peak discharges using flood frequency analysis (FFA) and hydrological modeling system (HMS). FFA method was performed using two distribution functions, i.e., Gumbel’s extreme value distribution (Gumbel’s) and lognormal (LN) distribution, while HMS was carried out using the Hydrologic Engineering Center’s Hydrologic Modeling System (HEC-HMS) software. Long and reliable daily streamflow and rainfall data (i.e., 50 years) recorded at Sungai Rinching, Selangor, were used. Goodness-of-fit test indicates that the chi-square (χ²) values for Gumbel’s and lognormal distributions are 1.952 and 7.976, respectively, which are significantly less than the ${\upchi}_{\alpha ,v}^{2}$ value of 9.488. However, Gumbel’s is chosen to represent the distribution of annual maximum daily streamflow because the χ² test value was much lesser as compared to lognormal distribution. The calibration and validation processes were carried out to test the suitability of HEC-HMS software to simulate the streamflow of Sungai Rinching. The relative percentage difference (RPD) estimated for peak discharges from these processes was between 3 and 6 %, and maximum lag times were only 1 h. These results indicate that the model performance is very good. The peak discharges for several average recurrence intervals (ARIs), i.e., 2, 5, 10, 20, 50, and 100 years estimated using FFA and HMS, were also compared. Results show that the peak discharges estimated using HMS for all ARIs (except 2 years) are comparable to the values given by FFA, and the given percentage differences are between 2.7 and 12.2 %. This finding indicates that HEC-HMS may be used as a tool in simulating the daily streamflow and estimate the annual daily maximum streamflow of several ARIs at Sungai Rinching.