Effect of the Spatial Resolution of Digital Elevation Models on the Performance of Nash, Rosso, and Clark Instantaneous Unit Hydrograph Models

Digital elevation models (DEMs) can be used to extract the physiographic characteristics adopted in Nash, Rosso, and Clark instantaneous unit hydrograph (IUH) models. Therefore, the spatial resolution of the DEM can affect the performance of these IUH models. The objective of this study was to evaluate the effect of DEMs specifically, ALOS PALSAR, ASTER, SRTM, and GTOPO (spatial resolutions of 12.5, 30, 90, and 1000 m, respectively), and the DEM obtained from the topographic map provided by the Iran National Cartographic Centre (scale of 1:25000 and spatial resolution of 10 m) on the performance of Nash, Rosso, and Clark IUH models for Amameh Watershed, Iran. To this end, Horton ratios were calculated from the abovementioned DEMs. The IUH dimensions were estimated based on Nash, Rosso, and Clark models for 34 rainfall–runoff events across the five DEMs. The results showed that as the spatial resolution of the DEM decreased, the evolution of the drainage network is lost. In general, parameters n and k are influential components in the Nash and Rosso models. The n values were similar across different DEMs in the Nash model, whereas the value of k increased from 0.92 to 1.43 h with the decrease in the DEM spatial resolution. The parameter n was constant for all DEMs, with a value of approximately 4, except for the GTOPO DEM (= 6). In contrast, the parameter k varied across different events and DEMs. The root mean square error of the estimated peak discharge decreased with the increase in the spatial resolution of the DEMs for the Nash IUH model. The Nash IUH model with TOPO DEM outperformed the Rosso and Clark models and other DEMs in terms of the IUH estimation accuracy. Acceptable results were also obtained using the ALOS PALSAR and ASTER DEMs in the Rosso model. However, the SRTM and GTOPO DEMs produced unsatisfactory results for all three models. Therefore, the application of these two DEMs in the Nash, Rosso, and Clark models is not recommended.