Mesoscale to submesoscale ocean state estimation by 2-way nested 4-dimensional variational data assimilation utilizing multi-mission nadir altimetry with supporting high resolution satellite and in situ observations

Author:

John Wilkin - (Rutgers University, New Brunswick)

Co-Investigator(s):

Andrew Moore (University Of California, Santa Cruz)
Hernan Arango (Rutgers University, New Brunswick)

Abstract:

Using high resolution modeling and advanced data assimilation, the project is developing regional full-water-column ocean state estimates for the analysis of the role of submesoscale variability in coastal ocean processes. The data being assimilated include high-resolution along-track retracked and SAR-mode multi-mission altimetry and other high-resolution satellite and in situ observations.

The region of focus is the Mid-Atlantic Bight of the northwest Atlantic shelf, an exemplar of wide continental shelves adjacent to deep ocean regions with appreciable mesoscale variability. The project utilizes an established configuration of the Regional Ocean Modeling System (ROMS) comprising a hierarchy of nested refinement grids from mesoscale (7-km grid) through intermediate resolution (2.2 km) to submesoscale resolving (700-m) that assimilates all the data from a comprehensive observing network comprising the constellation of nadir altimeter satellites, multiple infrared temperature satellites, surface current measuring HF-radar (CODAR), and temperature, salinity and velocity from diverse in situ platforms (moorings, gliders, Argo profiling floats, drifters, vessels, and animal tags). Data from the NSF Ocean Observatories Initiative (OOI) Pioneer Coastal Array complement the regionwide network with seven moorings (at ~15 km separation) that profile temperature, salinity and velocity, augmented by of a fleet of autonomous gliders. The Pioneer Array thus captures scales comparable to or exceeding the resolution of high data rate retracked coastal altimetry, SAR-mode altimeters, and that anticipated from SWOT.

The proposal is motivated, in part, by recent success (shown in the adjacent figure) reproducing event-wise correspondence in the evolution of submesoscale variability in the nested modeling system using 4-dimensional variational (4D-Var) data assimilation at each successive level of grid refinement. Observations at very high resolution become increasingly influential in the assimilation as the grids transition from merely mesoscale resolving to capturing high Rossby number geostrophically unbalanced submesoscale turbulent circulation.

A systematic suite of numerical experiments is being undertaken to develop best practices for assimilating high resolution SAR-mode and retracked coastal altimeter data in coastal submesoscale resolving models. The data-constrained simulations will be used to examine connections between the ocean mesoscale and submesoscale, including how the former preconditions or drives the latter, and the subsequent feedback of submesoscale variability on shelf-wide processes such as net shelf-sea/deep-ocean exchange at the shelf-break and vertical fluxes into the coastal ocean mixed layer and euphotic zone.

Supported by NOAA