Abstract's details

Internal tides and waves in a high-resolution ocean general circulation model with data assimilation

Brian Arbic (University of Michigan, United States)


Maarten Buijsman (University of Southern Mississippi, USA); Hans Ngodock (Naval Research Laboratory, United States); James Richman (Florida State University, USA); Jay Shriver (Naval Research Laboratory, USA); Innocent Souopgui (University of New Orleans, USA); Arin Nelson (University of Michigan, United States)

Event: 2020 Ocean Surface Topography Science Team Meeting (virtual)

Session: Salient results from the 2017-2020 OSTST PIs

Presentation type: Type Forum only

Contribution: PDF file


A group of US scientists proposes to examine the frequency-wavenumber spectra and predictability of internal tides and the internal gravity wave (IGW) continuum in global high-resolution simulations of the HYbrid Coordinate Ocean Model (HYCOM) with simultaneous atmospheric and tidal forcing, the Navy Coupled Ocean Data Assimilation (NCODA) scheme acting on the mesoscale eddies, and an Augmented State Ensemble Kalman Filter (ASEnKF) acting on the tides. The simulations therefore simultaneously contain barotropic tides, IGWs including internal tides, the general circulation, and mesoscale eddies. The eddies render internal tides and other internal waves non-stationary, hence reducing their predictability. The NCODA scheme helps to place mesoscale eddies in the correct locations, and also moves the model stratification towards observations, while the ASEnKF increases the accuracy of the modeled barotropic tides. This work has implications for both present-day nadir altimetry (e.g., the Jason-series) and for planned future generation wide-swath altimetry (e.g., the SWOT mission). We have shown in a 2012 paper that the high-wavenumber end of the sea surface height wavenumber spectrum is dominated by high-frequency motions in regions of strong internal tides. Over the last year we have recognized that the high-frequency motions include the IGW continuum as well as internal tides. New results included in this proposal demonstrate that the continuum contributes more to sea surface height (SSH) variance than internal tides, at least in our models.
For the Ocean Surface Topography Science Team,
1- We will continue to improve the accuracy of the barotropic tides, and hence the baroclinic tides, in HYCOM. We are motivated by our earlier results showing that regions of large resonant coastal tides have a substantial back-effect on open-ocean tides. This in turn motivates us to (a) improve the bathymetric datasets in regions of large coastal tides, which will likely improve coastal tidal accuracy and feed back onto the open-ocean tides, (b) implement two-way nesting within HYCOM in order to improve the resolution of coastal areas at relatively low computational cost, and (c) utilize new ASEnKF perturbation fields with horizontal length scales that are more closely aligned with the spatially varying scales of ocean tides.

2- We will examine the frequency-wavenumber spectra and the predictability of internal tides and the IGW continuum in our simulations. The relative impacts of high-frequency internal tides and IGWs, and lower-frequency motions, on spectra will be examined. The modeled internal tides will be compared with empirical internal tide models derived from present-generation nadir altimeters. The accuracy of our modeled IGW continuum will be more difficult to check; we will examine whether it reduces the variance measured in along-track present-generation nadir altimeters.

3- We will also diagnose the spectra in 41-layer 1/50th degree forward (non-assimilative) global HYCOM simulations with atmospheric and tidal forcing, and in global 1/12th, 1/24th, and 1/48th degree non-assimilative simulations of the MIT general circulation model (MITgcm). Increasing the model horizontal resolution yields modeled internal wave fields that compare better with observations. These additional forward simulations, as well as comparisons with available in-situ data, will be used to further gauge the accuracy and limitations of our models.
Brian Arbic
University of Michigan
United States