Global Ocean Data Assimilation and Prediction System2–ReAnalysis (GODAPS2-RA) Project : Preliminary Results

Abstract KMA has launched a project to produce global ocean reanalysis dataset from 1993 to 2020 by using KMA operational system– Global Ocean Data Assimilation and Prediction System2-ReAnalysis (GODAPS2-RA). We test-run GODAPS2-RA for recent five years (2016-2020) before the project formally initiated. The reason for prioritizing this period is that the year of Global Seasonal Forecasting System (GloSea6) hindcast is only available until 2016, highlighting the immediate necessity to extend to more recent years. In this study, we analyzed this preliminary experiment and examined whether GODAPS2-RA is properly implemented. It is found that the differences of GODAPS2-RA against the other validation dataset are within acceptable range. Compared with KMA operation system, GODAPS2-RA shows the reduction of observation innovation for most variables, signifying an improvement. However, the number of sub-surface temperature and salinity profile observations assimilated in the GODAPS2-RA is less than that in the operational run, consequently it turns out that further examination on the collection and processing procedure of profile data is necessary. In addition, the experimental GloSea6 hindcast using initial conditions produced by GODAPS2-RA is further tested. The results reveal that there is little difference between this experiment and existing operational GloSea6 hindcast, which indicating that GODAPS2-RA can be applied to the operational GloSea6 hindcast as oceanic initial conditions. By supplementing and preprocessing observations more appropriately and by further investigating the preliminary results, we will improve GODAPS2-RA to make it perform better, and finally official production will begin for releasing to publics.