Large Sample Basin Experiments for Hydrological Model Parameterization: Results of the Model Parameter Experiment–MOPEX. IAHS Publ. 307, 2006, 196–208
Performance comparison of a complex physics-based land
surface model and a conceptual, lumped-parameter hydrological model at the
basin-scale
Thian Yew Gan1, Yeugeniy Gusev2, Stephen J. Burges3 & Olga Nasonova2
1 Department of Civil and Environmental Engineering, University of Alberta, Edmonton, Alberta T6G 2G7, Canada
2 Institute of Water Problem, Russian Academy of Sciences, Russia
3 Department of Civil and Environmental Engineering, University of Washington, Seattle, Washington 98195-2700, USA
Abstract The
Soil–Water–Atmosphere-Plants (SWAP) model of Gusev & Nasonova (2003) is a
one-dimensional, land-surface model describing heat and water exchanges between
the land surface and the atmosphere in a physics-based, analytical manner. The
Sacramento model (SAC-SMA) is a deterministic, lumped-parameter, conceptual,
rainfall–runoff model that does not explicitly consider spatial variability of
terrain features, land-use, soil heterogeneities, and energy fluxes. SWAP is
compared to SAC-SMA and both are tested using data from 12 MOPEX river basins
located in the middle to southern eastern USA. Two model simulation comparisons
are made: one using broad classifications of soil and vegetation to derive
parameters (base case), and the second using calibrated parameters derived from
measured data from the 12 MOPEX sites during the 20-year period (1960–1979). Only six of the SWAP parameters were automatically calibrated using a
stochastic or Monte-Carlo technique. SAC-SMA was calibrated using manual
and automatic means (Duan et al., 1992). In terms of the goodness-of-fit
between simulated and observed hydrographs, for the base case SAC-SMA was
slightly better than SWAP. Both yielded results were unsuitable for any
scientific inference. For the calibrated case SAC-SMA out-performed SWAP when
comparing simulated hydrographs, but the simulations are still unsatisfactory.
Analysis of the different behaviour of the models is presented.
Key words
a priori parameter estimation;
hydrological model; land surface model; MOPEX basins; parameter calibration