Data assimilation in a regional high-resolution ocean model by using Ensemble Adjustment Kalman Filter and its application during 2020 cold spell event over Asia-Pacific region

Cold spell events can bring strong wind and low-temperature freezing and induce a significant rise of the water level, which have negative impacts on the economy. Based on a regional high-resolution ocean model, the satellite observations, including sea surface temperature (SST) and absolute dynamic topography (ADT), and the Argo data are assimilated to improve the modeling results over the Asia-Pacific area during a strong cold spell event of 2020 by using Ensemble Adjustment Kalman Filter (EAKF). To reduce the calculation cost, the ensemble for data assimilation is obtained by a dynamic sampling method which is based on the model forecasting biases and the EAKF is implemented by using an efficient parallelization scheme. The root mean square (RMS) error of SST decreased from 0.72 degrees C to 0.07 degrees C after assimilation, which was reduced by 90.28%. Besides, the RMS error of ADT was decreased from 0.28 m to 0.12 m, which was reduced by 57.14%. Responses of the ocean during the 2020 cold spell event were better reproduced using optimal EAKF configurations. As the speed of the wind increased and the air temperature dropped, the heat transported from the ocean to the atmosphere in the ocean model increased. Furthermore, the temporal evolutions of the ocean state were captured by data assimilation, which was the decrease of temperature and the increase of salinity and density. The mixed layers in the Kuroshio Extension were thicker due to excessive surface cooling. The seawater first increased and then overflowed in the Bohai Bay and Laizhou Bay. In conclusion, the EAKF data assimilation in this regional high-resolution ocean model has successfully reduced the model biases and the physical processes in reality can be well produced.

Automated summary

This study utilizes a regional high-resolution ocean model and assimilates satellite observations and Argo data to improve modeling results during the 2020 cold spell event. The assimilation of data effectively reduces model biases and captures the ocean's response to the cold spell event.