Calibration and Evaluation of HEC-HMS Hydrological Model Parameters in Simulation of Single Rainfall-Runoff Events (Case Study: Tajan Watershed)

Flood is one of the main natural disasters that has affected different regions of the country in recent years and caused a lot of damage. Scientists have always invented different methods to know the magnitude and time of flood and to estimate the height of runoff. Rainfall-runoff models are among the useful and powerful tools in achieving flood characteristics. The purpose of this research is to calibrate and validate the HEC-HMS model as well as to simulate floods as a single event in Tajan basin.

In this study, firstly the variables of precipitation and temperature were entered into the model, then the model was calibrated based on 3 single events and its parameters were calculated. Also, the parameters of curve number (CN) and impermeable surfaces were analyzed in Geographic Information System (GIS) environment and finally based on another single event the model was validated and its accuracy was estimated. The hydrographs generated from the scenarios of the adjusted model parameters were compared with the observed hydrographs of Kordkheil station.

Calibration results showed that the simulated data has a high correlation with the observed data (R>0.9). Also, the validation results using the correlation coefficient test and Student's t test showed that the correlation coefficient of the observation and simulation discharge is equal to 0.9, which shows an acceptable significance. Sensitivity analysis of the model was performed to determine the exact value of a sensitive parameter in exchange for changes of increase and decrease of parameters by 20%. The results of the sensitivity analysis of the model showed that the parameter of the curve number and the amount of initial losses are highly sensitive, so that in the range of -20 to +20% change of the parameter of the curve number and the initial losses, respectively, the sensitivity of the model varied from -2.33 to 7.15 and from 0.79 to -0.73. On the other hand, the model simulation results showed that the best objective function for estimating flood peak flow is Percent Error Peak, which the model predicted with a difference of 0.8% peak flow at Kordkheil station. Also, the objective function of the Percent Error Volume as the best estimator for estimated the flood volume with a difference of 0.

In general, the results of the study showed that the use of the hydrological model system using the geographic information system, rainfall–runoff modelling, flood simulation and predict the peak flow and flood volume, was very efficient and the obtained results are citable. Therefore, the tested models can be used in watershed management.