CALIBRATION OF BRANCHED ESTUARY MODELS

Abstract

Unsteady flow models of one-dimensional channels or of channel networks are commonly calibrated using a combination of stage (water surface elevation) and discharge observations. In tidally influenced areas these observations may extend in time over several tidal cycles. Where there is an important mean flow discharge such as in a river harbor, observations must be repeated for a number of representative mean flow conditions. Stage observations are commonly collected using recorders coupled to water surface floats mounted at various points on the channel banks. The initial costs of purchasing the recorders, installing them, and in the leveling necessary to determine datum are the primary expenses involved in collecting stage observations. For these observations the cost per observation decreases as the number increases. In tidally influenced channels, accurate one-dimensional instantaneous discharge measurements are difficult if not impossible to obtain because of the rapidly changing flow velocity. In many locations, discharge measurements cannot be made throughout a tidal cycle because of safety restrictions. Finally, with respect to stage information, the cost of collecting discharge information can be extremely high. In addition, the cost per observation does not decrease with the number of observations. With respect to the boundary conditions used to drive them, there are two main classes of unsteady flow models, the stage-stage models and the discharge-stage models. The former are driven by inputs of stage at all important external boundary points. The primary purpose of the stage-stage model is to determine the time history, or an average value of downstream discharge. Discharge-stage models are driven upstream by steady or time variable observed or hypothetical discharges and downstream by observed or hypothetical stages. This type of model is used primarily in conjunction with solute transport models in predicting the location and concentration of dissolved substances, as in dissolved oxygen modeling.