Evaluating a Hybrid Ensemble Data Assimilative Coupled Physical-Biogeochemical Ecosystem Model of the Red Sea

A hybrid ensemble data assimilation (DA) system is implemented for a coupled physical–biogeochemical ecosystem model of the Red Sea using MITgcm and NBLING at 4 km resolution, marking the first application of its kind in the region. The methodology combines a temporally varying ensemble from the Ensemble Adjustment Kalman Filter with a quasi-static monthly ensemble, implemented through the DA Research Testbed. Physical (satellite sea surface temperature, altimetry, in situ temperature and salinity) and biogeochemical (satellite chlorophyll) observations are assimilated, accounting for uncertainties in atmospheric forcing. Two configurations are evaluated: weakly coupled DA (weakly coupled data assimilation [WCDA]), which updates physical and biogeochemical states independently, and strongly coupled DA (strongly coupled data assimilation [SCDA]), which updates both using all observations. Sensitivity experiments assess the influence of assimilated observations on biogeochemical states, validated against independent temperature, salinity, sea surface height, chlorophyll, and oxygen data. Results demonstrate the benefits of joint assimilation in the Red Sea but also highlight challenges with SCDA. While SCDA improves the biogeochemical state relative to the free run, WCDA yields more robust physical estimates and better chlorophyll forecasts, particularly in subsurface layers. Physical assimilation through WCDA enhances biogeochemical fields throughout the water column, often exceeding 0.2 mg m⁻³ and the ensemble spread. Surface chlorophyll assimilation further improves WCDA surface predictions, though subsurface impacts are mixed. These findings emphasize both the value of WCDA and the need for further development to fully realize SCDA's potential for coupled physical–biogeochemical DA.