Sensitivity analysis of the spatially distributed hydrological WetSpa model parameters in estimation of daily flow in Khorramabad river

In this study, efficiency of WetSpa model to simulate daily runoff, sensitivity analysis parameters and their impacts on various components of the hydrological cycle was investigated in Khorramabad river in Lorestan Province using the PSET method. Watershed digital maps of elevation, soil and land use, time series of precipitaion, evapotranspiration and temperature are the main input data of the hydrological WetSpa model. The Observed discharge data at Doab Veisian station for the period of 2004-2010 (1383-1389 in Iranian solar calendar) was used for the model calibration and validation. Based on the results, the model is able to simulate the flow rate during the statistical period, so that simulated and observed data have NS coefficient around 83.76 and 66.44 for calibration and validation respectively (In confidence level of 95%). In analyzing the sensitivity of the model parameters, it was observed that the surface runoff factor for the near-zero rainfall intensity is the most sensitive, and the degree-day rainfall factor has the least sensitivity. The gmax parameter with the least sensitivity produces the greatest uncertainty in the simulation, and the accurate estimation of this parameter is almost impossible. The parameters of snow melt like T0 and Ksnow in the simulation of spring and autumn floods are not affected by rainfall, and sensitivity of these parameters is related to snow storms that occur in the winter. Krun and Pmax parameters have high sensitivity in the simulation of flood hydrographs and these parameters have a negative correlation with each other, affecting the flood discharge hydrograph, and have the greatest effect on the runoff simulation uncertainty band. Therefore, it is important to determine the appropriate range of the changes in these parameters in the simulation of flood hydrograph.
In this study, efficiency of WetSpa model to simulate daily runoff, sensitivity analysis parameters and their impacts on various components of the hydrological cycle was investigated in Khorramabad river in Lorestan Province using the PSET method. Watershed digital maps of elevation, soil and land use, time series of precipitaion, evapotranspiration and temperature are the main input data of the hydrological WetSpa model. The Observed discharge data at Doab Veisian station for the period of 2004-2010 (1383-1389 in Iranian solar calendar) was used for the model calibration and validation. Based on the results, the model is able to simulate the flow rate during the statistical period, so that simulated and observed data have NS coefficient around 83.76 and 66.44 for calibration and validation respectively (In confidence level of 95%). In analyzing the sensitivity of the model parameters, it was observed that the surface runoff factor for the near-zero rainfall intensity is the most sensitive, and the degree-day rainfall factor has the least sensitivity. The gmax parameter with the least sensitivity produces the greatest uncertainty in the simulation, and the accurate estimation of this parameter is almost impossible. The parameters of snow melt like T0 and Ksnow in the simulation of spring and autumn floods are not affected by rainfall, and sensitivity of these parameters is related to snow storms that occur in the winter. Krun and Pmax parameters have high sensitivity in the simulation of flood hydrographs and these parameters have a negative correlation with each other, affecting the flood discharge hydrograph, and have the greatest effect on the runoff simulation uncertainty band. Therefore, it is important to determine the appropriate range of the changes in these parameters in the simulation of flood hydrograph.
In this study, efficiency of WetSpa model to simulate daily runoff, sensitivity analysis parameters and their impacts on various components of the hydrological cycle was investigated in Khorramabad river in Lorestan Province using the PSET method. Watershed digital maps of elevation, soil and land use, time series of precipitaion, evapotranspiration and temperature are the main input data of the hydrological WetSpa model. The Observed discharge data at Doab Veisian station for the period of 2004-2010 (1383-1389 in Iranian solar calendar) was used for the model calibration and validation. Based on the results, the model is able to simulate the flow rate during the statistical period, so that simulated and observed data have NS coefficient around 83.76 and 66.44 for calibration and validation respectively (In confidence level of 95%). In analyzing the sensitivity of the model parameters, it was observed that the surface runoff factor for the near-zero rainfall intensity is the most sensitive, and the degree-day rainfall factor has the least sensitivity. The gmax parameter with the least sensitivity produces the greatest uncertainty in the simulation, and the accurate estimation of this parameter is almost impossible. The parameters of snow melt like T0 and Ksnow in the simulation of spring and autumn floods are not affected by rainfall, and sensitivity of these parameters is related to snow storms that occur in the winter. Krun and Pmax parameters have high sensitivity in the simulation of flood hydrographs and these parameters have a negative correlation with each other, affecting the flood discharge hydrograph, and have the greatest effect on the runoff simulation uncertainty band. Therefore, it is important to determine the appropriate range of the changes in these parameters in the simulation of flood hydrograph.