Data Assimilation On Quasi Underwater Glider Observation

At present, ocean data assimilation is mainly used in fixed observation data obtainedby the observation platform，such as: traditional marine observation, satellite remotesensing data, the mooring observation data, etc. The mobile platform technologystarted relatively late and at present has not been widely used, so the amount ofobservation data is very limited. This paper aims to explore the data assimilation ofnew mobile platform: underwater glider observation. However, we does not haveunderwater glider observation data, so ARGO observations and small amounts oftowed CTD data are substituted in this article which have the similar trajectory..The OVALS assimilation system designed by the LASG and nudging method wereused to assimilate the Argo grid data with1°1°resolution and Argo profile data intoPOM numerical model, respectively. The3DVar assimilation of Argo grid data in2006indicated that the assimilation improved the simulated temperature results byreducing about1.0°C root mean square error (RMSE) in either horizontal or verticaldistribution except for SST. Furthermore the performance of assimilation increasedwith time, which was better in March and April than that in January and February. Incomparison with independent OFES data, the assimilation improved the temperaturepattern to a certain extent, because there was a big difference between OFES data andArgo grid data used by assimilation. The nudging method was applied to assimilatethe Argo profile data into POM numerical model in the summer of2006. The processof nudging was very rapid, which would complete in about half day when therelaxation coefficient was0.5. It was also found that Argo profiles could rapidlyimpact the temperature pattern even remotely, and the remote impact could remainseveral days after the completion of assimilation.