Keywords
rating curve, rainfall-runoff modelling, uncertainty, extrapolation
Start Date
1-7-2008 12:00 AM
Abstract
The inputs of rainfall-runoff (RR) models (precipitation, temperature) and the data used forthe calibration of their parameters (typically streamflow data) are known to be uncertain,due to measurement and spatial extrapolation errors. In the literature, the impact of inputvariables uncertainty [Oudin et al., 2006] and model structure [Perrin et al., 2001] on RRmodel simulations have been largely studied. However, there are very few studies dealingwith the impact of the uncertainty of streamflow data used for calibration on the outputs ofRR models. Those impacts can be potentially considerable since streamflows are used tocalibrate the RR models. The problems related to the construction and extrapolation ofrating curves are especially crucial and should deserve more attention.
Uncertainty and evolution of rating curves: a key issue for the reliability of rainfall-runoff models?
The inputs of rainfall-runoff (RR) models (precipitation, temperature) and the data used forthe calibration of their parameters (typically streamflow data) are known to be uncertain,due to measurement and spatial extrapolation errors. In the literature, the impact of inputvariables uncertainty [Oudin et al., 2006] and model structure [Perrin et al., 2001] on RRmodel simulations have been largely studied. However, there are very few studies dealingwith the impact of the uncertainty of streamflow data used for calibration on the outputs ofRR models. Those impacts can be potentially considerable since streamflows are used tocalibrate the RR models. The problems related to the construction and extrapolation ofrating curves are especially crucial and should deserve more attention.
