Circulation analysis in the northwest Indian Ocean using ARGO floats and surface drifter observations, and SODA reanalysis output

This study incorporates observations from Array of Real-time Geostrophic Oceanography (ARGO) floats and surface drifters to identify seasonal circulation patterns at the surface, 1000 m, 1500 m, and 2000 m in the northwest Indian Ocean, and quantify velocities associated with them. A skill comparison of the Simple Ocean Data Assimilation (SODA) reanalysis output was also performed to contribute to the understanding of the circulation dynamics in this region. Subsurface currents were quantified and validated using the ARGO float data. Surface currents were identified using surface drifter data and compared to the subsurface observations to enhance our previous understanding of surface circulations. Quantified Southwest Monsoon surface currents include the Somali Current (v(max) = 179.5 cm/s), the East Arabian Current (v(max) = 52.3 cm/s), and the Southwest Monsoon Current (v(max) = 51.2 cm/s). Northeastward flow along the Somali coast is also observed at 1000m (v(max) = 26.1 cm/s) and 1500 m (v(max) = 12.7 cm/s). Currents associated with the Great Whirl are observed at the surface (v(max) = 161.4 cm/s) and at 1000 m (v(max) = 16.2 cm/s). In contrast to previous studies, both ARGO and surface drifter data show the Great Whirl can form as early as the boreal Spring intermonsoon, lasting until the boreal Fall intermonsoon. The Arabian Sea exhibits eastward/southeastward flow at the surface, 1000 m, 1500 m, and 2000 m. Quantified Northeast Monsoon surface currents include the Somali Current (v(max) = 97.3 cm/s), Northeast Monsoon Current (v(max) =30.0 cm/s), and the North Equatorial Current (v(max) = 28.5 cm/s). Southwestward flow along the Somali coast extends as deep as 1500 m. Point-by-point vector and scalar correlations of SODA output to ARGO and surface drifter data showed that surface SODA output and surface drifter data generally produced a strong correlation attributed to surface currents strongly controlled by the monsoons, while subsurface correlations of SODA output and ARGO were mostly insignificant due to variability associated with intermonsoonal transitions. SODA output produced overall smaller velocities than both observational datasets. Assimilating ARGO velocities into the SODA reanalysis could improve subsurface velocity assimilation, especially during the boreal fall and spring when ARGO observations suggest that flow is highly variable. (C) 2017 Elsevier B.V. All rights reserved.