Keywords
effective rainfall, unit hydrograph, transfer functions, calibration
Start Date
1-7-2010 12:00 AM
Abstract
This paper investigates three techniques for estimating the parameter val- ues for the unit hydrograph: full model optimisation (the CWI and CMD versions of IHACRES were used), simple rainfall scaling approaches for estimating e ective rainfall and an inverse ltering approach. The approaches are tested using synthetic stream ow data generated by the catchment moisture de cit (CMD) version of the IHACRES model, based on observed rainfall data for Cotter catchment (near Canberra, Australia) and the Singkarak catchment (Sumatra, Indonesia). The impact of rainfall and stream ow errors are considered by reducing the raingauge density and introducting errors in the rating curve. The inverse ltering apprach performed well for the synthetic stream ow derived from the rainfall data for the Cotter catchment. For the Singkarak catchment data, the di erences between the approaches was small, due possibly to either the higher frequency of rainfall in that area, or to the lower correlation between the rainfall gauge locations.
Robust estimation of the total unit hydrograph
This paper investigates three techniques for estimating the parameter val- ues for the unit hydrograph: full model optimisation (the CWI and CMD versions of IHACRES were used), simple rainfall scaling approaches for estimating e ective rainfall and an inverse ltering approach. The approaches are tested using synthetic stream ow data generated by the catchment moisture de cit (CMD) version of the IHACRES model, based on observed rainfall data for Cotter catchment (near Canberra, Australia) and the Singkarak catchment (Sumatra, Indonesia). The impact of rainfall and stream ow errors are considered by reducing the raingauge density and introducting errors in the rating curve. The inverse ltering apprach performed well for the synthetic stream ow derived from the rainfall data for the Cotter catchment. For the Singkarak catchment data, the di erences between the approaches was small, due possibly to either the higher frequency of rainfall in that area, or to the lower correlation between the rainfall gauge locations.
