Calibrating discharge, bed friction, and datum bias in hydraulic models using water level and surface current observations

摘要

We show that observations of water surface elevation, surface velocity, and bathymetry can be used to calibrate multiple reach-scale parameters such as the discharge, the overall bed friction, and a water surface datum bias in a hydraulic model (FaSTMECH) for the depth-averaged flow in the braided reach of the Kootenai River near Bonners Ferry, ID. Remotely sensed surface velocity data acquired during a 2010 field experiment are converted to depth-averaged velocity using a uniform ratio of 0.79 computed from concurrent in situ observation of vertical profiles of velocity. The in situ measured discharge of the Kootenai River was approximately 220 m~3/s. For given discharge forcing and downstream boundary condition provided by the measured water surface elevation, the FaSTMECH model was used to predict the horizontal variation of the depth-averaged velocity and the water surface elevation over the braided reach. The calibration considered here consists of varying the discharge, the bed friction coefficient and the water surface datum offset to find the combination that minimizes error metrics based on the streamwise profiles of the measured water level and maximum depth-averaged velocity. Composite error metrics that combine water level and velocity measurements are found to be more sensitive with respect to variation of the datum bias and the bed friction. We find that a composite error based on the total head is minimized for a discharge of 220 m~3/s, friction coefficient of 0.005, and a datum offset of 40 cm.