Variational Data Assimilation for the Global Ocean - Data Assimilation for Atmospheric, Oceanic and Hydrologic Applications - page 325

Fig. 13.11 Same as Fig. 13.10 , except for salinity observations

high ( 100

db) after the adjustment time period due to the assimilation of MODAS
synthetic profiles at high latitudes. MODAS synthetics were not thinned based on
stratification (Sect. 13.4.3 ) in these model runs. Layer pressure RMS errors are
reduced more than 50 % when weakly stratified MODAS synthetics are rejected
(not shown).
Figure 13.13 shows a verification result from the real-time run for sea surface
height in the Kuroshio region on 12 January 2012. The assimilation of SSH
anomalies is crucial to accurately map the circulation in these highly chaotic regions
dominated by flow instabilities. The white (black) line overlain is an independent
analysis of available infrared observations of the north edge of the current system
performed at the Naval Oceanographic Office. The frontal analysis clearly indicates
that the forecast system is able to accurately map the mesoscale features in the
western boundary current.
Tab le 13.2 gives run times for the 3DVAR conjugant gradient solver and post-
multiplication steps. The run times are listed for a typical day (28 January 2012) in
six of the global HYCOM analysis subdomains. A total of 2.2 million observations
were assimilated into the HYCOM grid that contained more than 520 million grid