Quantifying uncertainty in estuarine and coastal ocean circulation modeling

A methodology is developed to describe the effect of errors (or uncertainty) in the specification of certain drivers (bathymetry, river inflow, and wind speeds) on the circulation computed by a three-dimensional estuarine and coastal hydrodynamic circulation model. The methodology is based on first order variance analysis. Two analytical examples are used to illustrate the method and to provide a context for interpreting real world settings. An application of the method to a model of the N.Y./N.J. Harbor Estuary shows that current predictions are considerably sensitive to the accurate specification of bathymetry, and are usually more sensitive than water level predictions to errors in bathymetry. Estuarine properties with a strong seasonal (or spatial) component such as temperature or salinity did exhibit a sensitivity to driver accuracy that shifted from one season (or regime) to another. Bathymetry appears to control the circulation of the N.Y/N.J. Estuary more than the dynamic forcing of the winds and the Hudson River inflow, perhaps due to the fact that the estuary is primarily tidally driven.