Coupling 3D Ocean Baroclinicity into 2D Depth-Integrated Coastal Ocean Models

Two-dimensional (2D) depth-integrated coastal circulation models are commonly used to simulate shallow water physics such as storm surge, tides, and tsunamis. Although they require certain conditions to be ideally applicable (such as a well-mixed vertical column nearshore often induced during storm surge events) the fact that they are able to provide a high degree of horizontal resolution and numerical stability, usually outweighs the importance of three-dimensional (3D) baroclinicity on shallow water problems. In particular, coastal flooding related to coastal sea levels are dominated by 2D barotropic processes due to atmospheric and astronomical forcings (including the effects of wind waves which 2D circulation models are often coupled to). One issue related to the issue of storm tide and tsunami forecasting, is that long-term and seasonal variations driven by large-scale ocean baroclinicity will affect the coastal sea level in relation to the land, hence the risks of coastal flooding. Further, more localized impacts from coastal upwelling events, changes in ocean current transport rates, and changes to freshwater outflows are also important to the coastal sea level. 2D depth-integrated models cannot directly account for such effects.

To help to overcome the limitation of 2D depth-integrated models described above this study presents a process coupling paradigm to incorporate 3D ocean baroclinicity into a 2D depth-integrated model (ADCIRC, see http://adcirc.org/) from freely available data-assimilated global ocean model outputs (GOFS 3.1, see https://hycom.org/dataserver/gofs-3pt1/analysis). ADCIRC, commonly used in surge guidance systems, employs unstructured triangular meshes to provide high horizontal resolution to coastal regions over a large area without excessive computational expense. We find that through the prescription of appropriate lateral boundary conditions and the inclusion of baroclinic pressure gradient and momentum dispersion terms computed from GOFS 3.1 outputs, it is possible to capture large-scale long-term and seasonal changes in coastal sea levels in the 2D ADCIRC model. Additionally, some local baroclinic effects such as drawdown from cold water upwelling due to vertical mixing of the water column by a tropical cyclone are also captured.

This presentation will explore the effects of domain size and lateral boundary conditions; appropriate methods to correctly include momentum dispersion and other mixing effects; computational expense; and the overall model fidelity and its limitations. The setting for the model application is the Atlantic and Gulf Coasts of the United States. While in general preliminary results are very positive – the model skill in simulating coastal sea levels around Puerto Rico during the 2017 hurricane season were found to be increased at all tide gauges when 3D ocean baroclinicity was incorporated – it is anticipated that the model may be inadequate in areas where the global 3D ocean model is insufficiently resolved, such as in estuarine systems and enclosed bays. Thus, an important part of this study is to assess where this coupled baroclinic 2D depth-integrated modeling system appears to be largely sufficient and useful, and where further progress is required (i.e., facilitating model coupling into local high-resolution 3D baroclinic insets).