Combination of L-Moments Method and Hydrological Model for Design Flood Hydrograph Determination

Abstract

Major water resources projects such as design of reservoirs, flood control facilities (Dams, Levees, etc.), and flood plain delineation/mapping require projection of peak flows with different return periods (recurrence intervals, e.g., Q50, Q100, Q500). It is a common practice to use design flood (design hydrograph) when there is no data, or the available data are limited both in quantity and quality. In this paper, the L- Moment theory is used to estimate maximum precipitation for different recurrence intervals for the Rude Zard watershed in Khuzestan Province, Iran. The L- Moment is very suitable for frequency analysis of extreme (outlier data) values. In L- Moments analysis, linear combination of order statistics are used for outlier data since it is almost unbiased for small sample sizes. For this project, the Depth-Duration-Frequency relations were determined using the HEC-HMS model and simulated flood hydrographs for different return periods were developed. Based on regional analysis of data, rain gage stations were divided into two areas each fitting a specific frequency distribution. The Generalized Normal Distribution (GND) was found to be more appropriate for area I, and the Generalized Pareto Distribution (GPD) for area II. The results of model simulations show that the simulated hydrographs except for the two-year return period have the same base time. Also, the time to peak for all return periods except for two-year, seem to about the same.