The integration of a high-resolution baroclinic hydrodynamic model with multibeam echosounder data reduction: Examining depth uncertainty and model skill assessment

Multibeam echosounders require knowledge of the speed of sound in water to accurately convert and ray trace a measured two-way travel time (TWTT) into a depth estimate. In cases where the sound speed field is unknown or rapidly changing, due to variations in temperature and salinity, errors will become apparent in the depth calculation, and the resulting data quality will deteriorate.

The use of a high-resolution baroclinic hydrodynamic numerical model as a source of sound speed information for a multibeam echosounder is investigated in a highly stratified tidal estuary located in Saint John, New Brunswick, Canada. The model was run for four time periods representing the limits in variation of river flow. The model simulation periods were accompanied by high frequency CTD casts along the axis of the estuary from a Moving Vessel Profiler (MVP). The comparison of the CTD casts to the model output fields of temperature and salinity are used to determine the resulting error in the multibeam depth estimate. The depth error statistics are used as a quantifiable method of evaluating the model skill throughout the domain based on operational end user requirements.

The methodology of evaluating both the errors inherent in applying model sound speed fields to multibeam depth estimates and evaluating model skill based on operational error statistics is explored. The objective of this study is not only to provide improved multibeam error detection, quantification and minimization, but also to provide error statistic feedback to model development. This approach is now being extended to the coastal Northern Gulf of Mexico for various multibeam data collection scenarios.