The estuarine parameterization problem: sensitivity of ROMS to the specifics of line source buoyancy input in the Coastal Gulf of Alaska

Albert J. Hermann and Elizabeth L. Dobbins

Joint Institute for the Study of the Atmosphere and Ocean
University of Washington, Seattle

For several years we have been using ROMS to simulate the buoyancy- and wind-driven Alaskan Coastal Current (ACC) in the Coastal Gulf of Alaska (CGOA), and its effect on biological productivity on the shelf. The buoyancy that drives the ACC is provided by innumerable streams and rivers which feed into deep, narrow fjords that, in many cases, are not resolved by the scale of our grid. We treat these inputs as a “line source” of buoyancy, and attempt to parameterize the unresolved processes within the fjord/estuaries by adjusting the vertical profile and salinity of the input freshwater. This has been complicated by a lack of knowledge regarding the physics which govern the buoyant plume of a line source (with the exception of Williams, 2003) relative to river point sources, and by the paucity of data collected within the fjords and the ACC. Recent data collected in the CGOA allow us to use transport within the ACC, and stratification across the shelf near Seward, AK, to validate the results of several buoyancy input schemes. The various possibilities for coastal freshwater input include the following, where Q represents the flux of freshwater into the system. Note that in each case we do not attempt to resolve the tidal flux across the model coast into the unresolved estuaries, but instead look to approximate the subtidal result.

1) Input of zero salinity water at rate Q to the surface level of the model. In this case, the estuarine conditions are essentially pushed onto the shelf, aliasing the stability of the water column there, to the detriment of the biological calculations;
2) Similar to scheme 1, but the outflow Q is distributed vertically, to approximate vertical mixing within the unresolved fjords/estuaries;
3) Input of brackish water at a rate greater than Q in the surface level. This represents continuous estuarine mixing of fresh and ocean waters (estuarine amplification) landward of the model’s coastline, but ignores the corresponding inflow of ocean water at depth;
4) Input of brackish water at rate Q in the topmost layer. This represents “perfect” estuarine mixing of fresh and ocean waters generated by each flood tide, followed by discharge of the resulting brackish product on the ebb.

At present, scheme 4 appears to give the most satisfactory result, and the closest match to observations. We discuss the merits and deficiencies of each approach, and other possible schemes for the parameterization of estuaries in the CGOA.