Physical-based hydrological modelling : Reliable method or excussive escape route? Case study in Sutami reservoir, Indonesia

Abstract

While numerous hydrological modeling are derived based on empirical studies, development of physical-based modeling are no less massive. Catchment characteristics are included in many models, although direct quantitative correlation between those features and model parameters are still found scarce. Soil Conservation Service (SCS) method for estimating hydrological loss and hydrograph transformation have been many applied, with the platform of hydrological soil type and land cover use. Using daily simulation in Sutami Reservoir, Indonesia, it is found that the usage of physical-based model does not guarantee excellent results. The study is conducted by performing number of scenarios of hydrological analysis methods combination; starting from Deficit & Constant (D&C), Clark and SCS method. From the simulation in December 2007, the loss calculation from D&C method delivers Nash-Sutcliffe and RMSE value of respectively 0.66 and 1.69, while SCS method has 0.51 and 2.56. In addition, hydrograph transformation of Clark method delivers closer approximation to the actual hydrograph, while the SCS approach indicates visually over-responsive runoff initiation. Moreover, Clark has identical NS and RMSE result of 0.51 and 2.56 with SCS. Lastly, simulation using D&C for loss and Clark for transform yields even better result with NS of 0.68 and RMSE of 1.68. In summary, specifically in this study area, physical-based modeling might produce decent simulations. However, the improvement for hydrological data collection coupled with parametric hydrological model is obviously necessary.