Large Sample Basin Experiments for Hydrological Model Parameterization: Results of the Model Parameter Experiment–MOPEX. IAHS Publ. 307, 2006, 159–168.
Modelling ungauged basins with the Sacramento
model
Terri Hogue1, Koray Yilmaz2, Thorsten Wagener3 & Hoshin Gupta2
1 Dept of Civil and Environmental Engineering, University of California, Los Angeles, California, USA
thogue@seas.ucla.edu
2 Dept of Hydrology and Water Resources, University of Arizona, Tucson, Arizona, USA
3 Dept of Civil and Environmental Engineering, Pennsylvania State University, College Station, Pennsylvania, USA
Abstract This paper evaluates two contrasting approaches to parameter estimation for ungauged basins using the US National Weather Service’s SACramento Soil Moisture Accounting (SAC-SMA) model. An automatic calibration scheme (Multi-Step Automatic Calibration Scheme, MACS) provides deterministic parameter estimates using a three-step, multiple objective approach. The MACS estimates are then transferred to similar or “sister” watersheds for basins in the French MOPEX data set. Physically-based parameter estimates are also developed for the same basins based on the a priori approach of Koren et al. (2000). In general, the two methods, the transfer and the a priori approaches, show similar overall performance. Parameter estimates appear more consistent between basins using the a priori approach, but statistically the regionalized MACS parameters and the a priori parameters show very similar model performance for the three basins investigated in this study. Model simulated hydrographs are also very similar between the two methods, with both methods tending to underpredict most events (peak and volume) but matching the shape and pattern of flow well. However, both methods have worse performance than a calibrated model for the same basin, indicating the possibility for further refinement and adjustment of the techniques presented here.
Key words parameter
estimation; rainfall–runoff modelling; regionalization; Sacramento model