loading page

The fallacy in the use of the “best-fit” solution in hydrologic modeling
  • Karim C. Abbaspour
Karim C. Abbaspour
Swiss Federal Institute of Environmental Science and Technology (EAWAG)

Corresponding Author:abbaspour@eawag.ch

Author Profile


The use of the parameters associated with the “best-fit” criterion to represent a calibrated hydrological model is inadequate. Furthermore, assessing the goodness of model calibration or validation based on performance criteria, such as NSE, R2, or PBIAS, is misleading because they only compare two signals, i.e., measurement and the best-fit simulation (i.e., simulation with the best objective function value). The reason is that the calibrated model’s best objective function value is usually not significantly different from the next best value or the value after that. This non-uniqueness of the objective function causes a problem because the best solution’s parameters are always significantly different from the next best parameters. Therefore, only using the best simulation parameters as the calibrated model’s sole parameters to interpret the watershed processes or perform further model analyses could lead to erroneous results. Furthermore, most watersheds are increasingly changing due to human activities. The lack of pristine watersheds makes the task of watershed-scale calibration increasingly challenging. Subjective thresholds of acceptable performance criteria suggested by some researchers to rate the goodness of calibration are based on the comparison of the two signals, and in most cases, the thresholds are not achievable. Hence, to obtain a satisfactory fit, researchers and practitioners are forced to massage and manipulate the input or simulated data, compromising the science behind their work. This article discusses the fallacy in using the “best-fit” solution in hydrologic modeling. It introduces a two-factor statistics to assess the goodness of calibration/validation while taking model output uncertainty into account.