<i>HESS Opinions</i> "Crash tests for a standardized evaluation of hydrological models"

Abstract

As all hydrological models are intrinsically limited hypotheses on the behaviour of catchments, models-which attempt to represent real-world behaviour-will always remain imperfect. To make progress on the long road towards improved models, we need demanding tests, i.e. true crash tests. Efficient testing requires large and varied data sets to develop and assess hydrological models, to ensure their generality, to diagnose their failures, and ultimately, help improving them.

The Court of Miracles of Hydrology: can failure stories contribute to hydrological science?

“…More demanding and severe tests should be designed to evaluate the reliability of models with the aim of improving them (Andréassian et al, 2009). We follow Mayo (2010) when she argues that: the growth of knowledge has not to do with replacing or confirming or probabilifying or "rationally accepting" large-scale theories, but with testing specific hypotheses in such a way that there is a good chance of learning something -whatever theory it winds up as part of.…”

“…In this respect, the post-evaluation of his own work made by Refsgaard (2007), or the critical evaluation of the TOPMODEL developments made by Beven (1997), should serve as examples for hydrologists. Comparative analysis and the use of reference data sets (though not very easy to implement in practice as new types of data emerge) could help in measuring more objectively the actual improvements made in modelling over several decades, as advocated by Andréassian et al (2009).…”

Exploration of sub-annual calibration schemes of hydrological models

“…This calibrating scheme is widely applied (Sorooshian, 1991;Gan and Biftu, 1996;Gupta et al, 1998;Gupta et al, 2009). The validation is a standard practice in hydrological modelling (Andréassian et al, 2009) to test the model with the data outside of calibration period to evaluate the model performance.…”

Watersheds are not static: Implications of climate variability and hydrologic dynamics in modeling

“…streamflow) (McMillan et al 2010), iii) the calibration method and performance measurements adopted (objective functions) (Efstratiadis and Koutsoyiannins 2010), and/or iv) the model structure (Andréassian et al 2009), contributing to model and parameter uncertainty. Therefore, adequate hydrological modeling is necessary to understand and simulate the dominant processes and dynamics that control the hydric balance in a basin, reduce model uncertainty, improve the output and prediction confidence degree, and predict in a realistic manner the future behavior of a basin under changing conditions (Merz et al 2011).…”