Impact of Assimilation of Moored Velocity Data on Low‐Frequency Current Estimation in Northwestern Tropical Pacific

Abstract

AbstractThis study investigates the impact of assimilation of Northwestern Tropical Pacific Ocean (NWTPO) moored currents observation system on low‐frequency (over a period of more than 90 days) current estimation. A low‐frequency scale‐selective ensemble optimal interpolation (EnOI) scheme is designed to assimilate the horizontally sparse moored currents into an eddy‐permitting ocean model. In this assimilation scheme, low‐frequency current signals are selected by removing intraseasonal signals in both observation and ensemble data, and a horizontal spatial filter is implemented to match the model forecast to the low‐frequency ensemble and observation. The assimilation of currents efficiently resolves simulated low‐frequency current displacements, especially for the Equatorial Undercurrent (EUC), the North Equatorial Countercurrent (NECC), and the North Equatorial Current (NEC). More realistic intensities of the NWTPO low‐frequency currents are also reproduced. With respect to the moored velocity profiles, the root‐mean‐square error (RMSE) of the simulated low‐frequency zonal velocity above a 500‐m depth is reduced from 10.9 to 5.8 cm/s using this low‐frequency EnOI scheme. The low‐frequency variation of NWTPO currents is reproduced by correcting the structures of different baroclinic modes. The increments of zonal velocities mainly correspond to the first baroclinic mode for the NECC and NEC, while for the EUC, they correspond to the leading three and higher‐order baroclinic modes. The improvements of the simulated stratification and pressure gradient remark the role of the assimilation of moored currents in adjusting available potential energy.