Puji Harsanto, Bambang Agus Kironoto, Bambang Triatmodjo


Hydrological models are classified as lumped and distributed. Lumped models ignore the spatial variability of precipitation, and other related processes. Even though lumped model are unable to account for internal variation of hydrological processes, they have the advantage of simplicity. Distributed hydrological model on the other hand account for spatial variation of hydrological processes and parameters. This type of model has the potential to give more accurate results but computationally more complex. The spatially distributed input and analysis required by spatially distributed model can be met by incorporating a system that can manage data on a grid basis. An approach to handle this problem is using geographic information system (GIS). The overall objective of this study was to comparing of distributed and composite model.

The SCS curve number method also known as the hydrologic soil cover complex method, is widely used procedure for runoff estimation. This method includes several important properties of the watershed namely soil’s permeability, landuse and antecedent soil water conditions which are taken into consideration. Daily runoff calculations were generated using the SCS curve number method, its based on the retention parameter, S, initial abstractions, Ia (surface storage, interception, and infiltration prior to runoff), and daily rainfall, Rday. Ratio of initial abstraction (Ia) to retention parameter (S) called λ is changes from time to time. Because of its, the hydrology analysis to estimating direct runoff need calibrate for this parameter. Goodness of fit analysis is used to comparing of both, distributed model and composite model.

The average of relatif error, correlation factor, and coefficient of determination, R2 for distributed model respectively are 25.70 %, 0.71 and 0.53, from composite model are 30.15 %, 0.66 and 0.44. The result from research is obtained that the distributed model is more accurate than composite model. The average initial abstraction ratio from distrubted model is 0.35 and composite model is 0.04.


Direct runoff, Distributed model, Lumped model, Initial abstraction ratio

Full Text: