|
Ocean Model links: ROMS 4DVAR assimilation analysis/forecast
|
The Regional Ocean Modeling System (ROMS)
SW06 model covers the Mid-Atlantic Bight from the center of Long Island
southward to south of the mouth of the
The inner and mid-shelf domain coincides with the domain used for forecasting and analysis for the LaTTE program.
The prototype system is a 5-km horizontal, 30-level ROMS model with Incremental Strong Constraint 4DVAR assimilation of all glider observations (RU COOL), shipboard CTDs and XBTs of opportunity on the transit legs, scanfish profiles from RV Endeavor (Gawarkiewicz), underway thermosalinograph data from the first two RV Knorr legs, daily composite SST (RU COOL) and gridded altimeter SSH anomalies (AVISO).
Initial conditions were the Linder and Gawarkiewicz NJ shelf climatology adjusted by simple relaxation to the kinematic constraints of the model domain.
Meteorological forcing is NCEP/NAM 12-km 3-hourly forecast data.
Hudson River discharge is from daily average observations.
Tide boundary conditions are from the Oregon State OTPS harmonic analysis.
We assimilated data over 2-day intervals iterating the initial conditions to minimize the model-data misfit over each cycle. Each 2-day cycle begins with first guess initial conditions being the conclusion of the preceding interval.
Climatology was a poor estimator of the initial state in 2006 given the extreme precipitation and Hudson River discharge in mid-July. Assimilation results show the model quickly adjusts to in situ observations in the central SW06 region, influenced strongly by the low salinities observed by the gliders. However, in the absence of other data in the far field, the adjustment is unsatisfyingly local. Efforts to redress this bias will be pursued during future reanalysis.
As the model simulation proceeded, the salinity-corrected region advanced slowly northward and eastward consistent with the adjoint model propagating the model-data misfit information upstream. The barotropic transport of the shelf/slope front is not well constrained by the assimilation of predominantly temperature and salinity data from gliders, ships and satellites, arguing in favor of the value of complementary observing systems (CODAR, ship ADCP, moored current meters) to constrain ocean velocities.
Open boundary conditions
are simple gradient conditions because the Orlanksi-type radiation conditions
typically preferred in this type of application are not well-posed in the
adjoint model formulation. Air-sea heat and momentum
fluxes are calculated by the bulk formulae of Fairall et al. (1996,2002) using
the model sea surface temperature and sea level air temperature, pressure,
relative humidity, and 10-meter winds.
The inflow of the
