Estimation of Design Flood using Fractal Theory and HEC-HMS Model (Case Study: Khorramabad River Basin)

Estimating the design flood of the basin for the design of hydraulic structures, stabilization of river banks, watersheds and flood zoning projects are the most important in hydraulic and hydrological issues and projects. The flood used to design structures and influenced by hydrological events is called design flood which depends on structure safety, cost, life expectancy, and possible damage. Intensity- duration- frequency (IDF) curves of rainfall is a hydrological tool for the estimation of the design flood and design of hydraulic structures. These curves are constructed for a region from rainfall data which are recorded at various continuations. Usually, in some countries such as Iran which has a large extent, there is not enough rain-gage station; or the length of the statistical period is low, so it is impossible to calculate the IDF curves. But since it is not usually possible to access daily rainfall data, the fractal theory can be used to estimate the precipitation data in different consistency and the IDF curve with a very good accuracy. Korramabad river basin, one of sub-basin of Karkhe basin, often has been exposed to destructive floods and damages caused by it. In this research, the intensity-duration-frequency curves of the catchment area are estimated using fractal theory at first, and then the design precipitations are obtained in different return periods. In the next step we calibrated HEC-HMS rainfall- runoff model and finally the design floods are estimated in different return periods. The HEC-HMS model is an extension of the HEC-1 model under Windows, with all its capabilities. Hydrographs calculated by this model are used directly or in combination with other software for various purposes such as water supply, urban drainage, flood and flow forecasting, land use change, flood control studies and exploitation of reservoir systems. In this research, SCS curve number method and recession method are used to calculate the losses and base flow. Also for estimation of runoff, SCS curve number, Snyder unit hydrograph and Clark unit hydrograph are used in three methods and after comparing the results of the three methods in the calibration stage of the model, the Clarke unit hydrograph method is identified as the best method for estimating runoff. Also the Maskingham method has also been used for flood routing. The data needed for this study include rain gage data, hydrometric data, physiographic data of the basin, and also amounts of CN or curve number of the basin. Rain gage and hydrometric station used in this research are Chamanjir station. In this research, due to the importance of peak discharge for designing hydraulic structures, the optimization of the parameters has been done using the peak-weighted RMS Error criterion. In the calibration step, Comparison between hydrographs shows that there is a good agreement between computational and observational hydrographs, in such a way that the difference between the simulated and the actual peak discharge are 0.6, 0.2 percent for the selected floods. After calibration of the HEC-HMS rainfall runoff model for the studied area, this model is used to estimate the design flood. In the process of conversion precipitation to runoff, it is necessary to determine the pattern for temporal distribution of rainfall at the stations and in the area. To do so, the non-dimensional rainfall data are plotted for some storms with different time durations. For making data for each storm non-dimensional, the accumulated depth of precipitation to the desired time step was divided by the total depth of storm’s rainfall. The same procedure was carried out to make the time axis non-dimensional. By analyzing the precipitation data of the recorder station at the basin, it was found that in the majority of the precipitations, 20% of rainfall occurs in the first quarter, 20% in the second quarter, 40% in the third quarter and 20% in the fourth quarter. The results of this research show that:Daily precipitation data have a fractal characteristic in the ranging from 1 to 8 days, and during this time interval, rainfall data can be converted from a continuity to another continuity.
Due to the lack of recorded rainfall statistics in different continuations, using fractal method can be a useful way to prepare IDF curves in this basin and these curves are obtained based on the daily rainfall data available.
There is high efficiency of fractal model and HEC-HMS model in this catchment and the Gambel probabilistic distribution is appropriated for the maximum daily rainfall data of this basin.