Effect of Digital Elevation Models Spatial Resolutions on Nash and Rosso Instantaneous unit Hydrograph Performance (case study: Kasilian Watershed)

Introduction and Goal
The use of digital elevation model (DEM) enables the extraction of physiographic characteristics used in Nash and Rosso instantaneous unit hydrograph (IUH) models. Therefore, the spatial resolution of the DEM is effective on the performance of Nash and Rosso IUH models. Therefore, the current research aims to evaluate the impact of DEMs including ALOS PALSAR, ASTER, SRTM and GTOPO with a spatial resolution of 12.5, 30, 90 and 1000 meters respectively and the DEM obtained from the topographic map (TOPO) of the Iran Geological Organization with a scale of 1:25000 and a spatial resolution of 10 meters was done in Nash and Rosso models in Kasilian watershed.
Material and Methods
Kasilian Watershed, with an area of about 66.43 km², is located in the north latitude of 35°58'30" to of 36°07'15"and east longitude of 53°08'44"to of 53°15'42". The watershed has a humid climate based on Domarten climatic classification and average annual rainfall of 783.4 mm. Physiographic characteristics including Horton ratios were calculated for each of the DEMs. Finally, IUH dimensions were estimated based on Nash and Rosso models for 64 rainfall-runoff events and five different DEMs.
Results and Discussion
The results showed that by reducing the spatial resolution of DEM, the evolution of the drainage network is lost, so that the maximum rank of the streamflow in TOPO and PALSAR ALOS DEMs is equal to six, and in ASTER its number is reduced to five. Only in SRTM and GTOPO, the number of streamflow ranks was significantly reduced to three. The parameters n and k are effective components in Nash and Rosso models, and in the Nash model, the amount of n was not significantly different in different DEMs. But the value of k has increased from 1.35 to 1.64 with the decrease of DEM spatial resolution. In the estimation of parameter n by Rosso's method, a constant trend is not seen in DEMs, so that in TOPO, ALOS PALSAR and ASTER, it was about six. While in SRTM, three and in GTOPO, four were estimated and the k parameter was different for each event and its value was different in different DEMs. The Nash method, the lowest and highest average relative error (RE) are related to the TOPO and GTOPO DEMs with values of 10.72% and 11.01% in the estimated runoff volume, respectively. The Rosso method, the average value of RE in the three TOPO, ALOSPALSAR and ASTER DEMs is almost similar in the estimated runoff volume. While the amount of RE in SRTM and GTOPO is higher than other DEMs. In fact, the results are close to each other due to the similar ability to extract the physiographic characteristics of the basin from these three models. Of course, it should be stated that with a small difference, the lowest amount of RE in estimating the runoff volume is related to the TOPO DEMs.
Conclusion
ASTER DEMs with lower spatial resolution compared to TOPO and PALSAR ALOS has provided an acceptable stream network. However, SRTM and GTOPO have not provided a suitable stream network. In general, with the increase in the spatial resolution of the DEM, drainage density developed more dense and real, but main stream is somewhat developed in DEMs with lower spatial resolution. Different methods have been presented to estimate the values of n and k, and in this research, in order to investigate the effect of different DEMs in the Nash method, among the different methods of estimating its parameters, the experimental method has been used. In the experimental method, the physiographic characteristics of the watershed, such as the slope, the length of the main river and the area of the basin, play the main role in estimating the peak discharge. In relation to the peak flow estimation, the Nash model performed better and in estimating the runoff volume, the Rosso model performed better using the TOPO DEMs. In fact, it can be stated that the estimation of the peak discharge in the Nash method based on the experimental method has provided acceptable results in this watershed. according to the necessity of estimating the peak discharge in determining the dimensions of the IUH in the watershed to accurately simulate and control future floods, using the results of the present study can be of great help.