High-Resolution 3DVAR for Constraining Submesoscale Dynamics
Event: 2017 Ocean Surface Topography Science Team Meeting
Session: Application development for Operations
Presentation type: Type Oral
The launch of the Surface Water Ocean Topography (SWOT) satellite in 2021 will bring about new high-resolution surface observations that will resolve submesoscale variability like never before. Current high performance computing platforms allow the generation of submesoscale phenomena within regional ocean numerical models. However, due to non-deterministic processes the submesoscale features rapidly diverge from reality. The SWOT global observations are the first that will be capable of constraining the submesoscale for skillful ocean forecasts. Developed in anticipation of real-time SWOT data, an Observation System Simulation Experiment (OSSE) evaluates potential forecast improvements enabled by future SWOT data. A non-assimilative numerical simulation using a primitive equation 1 km resolution model for the entirety of 2016 provides a ‘truth’ for our OSSE case members. To generate a divergent simulation over the same time period, the initial condition was perturbed while using the same boundary conditions and surface forcing as was used in the simulated truth. As expected, mesoscale and submesoscale features between the two solutions diverge. The simulated truth was then sampled at real observation locations throughout 2016, and the observations were provided to the Navy Coupled Ocean Data Assimilation (NCODA) 3DVAR. The first simulation (the Control run) used only current observation systems to establish a baseline estimate of convergence by mainly constraining the mesoscale ocean features. Test SWOT data, created from the simulated truth using the Jet Propulsion Laboratory’s (JPL) SWOT simulator, is used for two additional experiments (one containing all estimated errors and the other with none). The addition of simulated SWOT data both increases the rate and magnitude of convergence when compared with the Control run. The difference between the experiments using SWOT data with full errors and no errors provides a range of expected performance the satellite data will provide when operating.