Inter comparison of Tropical Indian Ocean features in different ocean reanalysis products

Abstract

This study makes an inter comparison of ocean state of the Tropical Indian Ocean (TIO) in different ocean reanalyses such as global ocean data assimilation system (GODAS), ensemble coupled data assimilation (ECDA), ocean reanalysis system 4 (ORAS4) and simple ocean data assimilation (SODA) with reference to the in-situ buoy observations, satellite observed sea surface temperature (SST), EN4 analysis and ocean surface current analysis real time (OSCAR). Analysis of mean state of SST and sea surface salinity (SSS) reveals that ORAS4 is better comparable with satellite observations as well as EN4 analysis, and is followed by SODA, ECDA and GODAS. The surface circulation in ORAS4 is closer to OSCAR compared to the other reanalyses. However mixed layer depth (MLD) is better simulated by SODA, followed by ECDA, ORAS4 and GODAS. Seasonal evolution of error indicates that the highest deviation in SST and MLD over the TIO exists during spring and summer in GODAS. Statistical analysis with concurrent data of EN4 for the period of 1980–2010 supports that the difference and standard deviation (variability strength) ratio for SSS and MLD is mostly greater than one. In general the strength of variability is overestimated by all the reanalyses. Further comparison with in-situ buoy observations supports that MLD errors over the equatorial Indian Ocean (EIO) and the Bay of Bengal are higher than with EN4 analysis. Overall ORAS4 displays higher correlation and lower error among all reanalyses with respect to both EN4 analysis and buoy observations. Major issues in the reanalyses are the underestimation of upper ocean stability in the TIO, underestimation of surface current in the EIO, overestimation of vertical shear of current and improper variability in different oceanic variables. To improve the skill of reanalyses over the TIO, salinity vertical structure and upper ocean circulation need to be better represented in reanalyses.