We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Multi-objective calibration of a distributed hydrological model (WetSpa) using a genetic algorithm.
- Authors
Shafii, M.; De Smedt, F.
- Abstract
A multi-objective genetic algorithm, NSGA-II, is applied to calibrate a distributed hydrological model (WetSpa) for predicting river discharge. The evaluation criteria considered are the model bias (mass balance), the model efficiency (Nash-Sutcliffe efficiency), and a logarithmic transformed model efficiency (to emphasize low-flow values). The concept of Pareto dominance is used to solve the multi-objective optimization problem and derive Pareto-optimal parameter sets. In order to analyze the applicability of the approach, a comparison is made with another calibration routine using the parameter estimator PEST to minimize the model efficiency. The two approaches are evaluated by applying the WetSpa model to the Hornad River (Slovakia) for which observations of daily precipitation, temperature, potential evapotranspiration, and discharge are available for a 10 year period (1991-2000). The first 5 years of the data series are used for model calibration, while the second 5 years for model validation. The results revealed that the quality of the solutions obtained with NSGA-II is comparable or even better to what can be obtained with PEST, considering the same assumptions. Hence, NSGA-II is capable of locating Pareto optimal solutions in the parameter search space and the results obtained prove the excellent performance of the multi-objective model calibration methodology.
- Subjects
HERNAD River (Slovakia &; Hungary); SLOVAKIA; HUNGARY; GENETIC algorithms; CHEMICAL models; PARETO analysis; EVAPOTRANSPIRATION; PARETO principle; CALIBRATION; METEOROLOGICAL precipitation
- Publication
Hydrology & Earth System Sciences Discussions, 2009, Vol 6, Issue 1, p243
- ISSN
1812-2108
- Publication type
Article
- DOI
10.5194/hessd-6-243-2009