This is the authors' abstract. We don't add key points for this paper.
The 4-dimensional multigrid analysis (4D-MGA) data assimilation system was developed for the 1/12° global Mass Conservation Ocean Model (MaCOM). 4D-MGA is a variational data assimilation method that accounts for error evolution across concurrent spatial and temporal scales. Dynamic height integration was incorporated into the cost function, thereby imposing a physical constraint for assimilating sea level anomaly observations. To evaluate the 4D-MGA system, sequential data assimilation was performed from 2024 April 5 to May 9. The analysis fields show that 4D-MGA effectively reduces large-scale model biases and improves the spatial representation of mesoscale eddies and surface currents. Both the root mean square error (RMSE) and bias were substantially reduced, and the sea surface height and current fields show substantially improved agreement with satellite observations. Forecasts initialized from the 4D-MGA analysis fields demonstrate consistent and moderate improvements in accuracy. The 1- and 7-d forecasts exhibit lower RMSE and bias than the free-run simulation and maintain stable performance. An intercomparison based on the GODAE Intercomparison and Validation Task Team metrics was conducted. The results place MaCOM with 4D-MGA above the average of the systems compared. MaCOM demonstrates relatively low RMSE and greater stability in salinity, sea surface temperature, and sea level anomaly. Although temperature forecasts rank below average during the first 3 d, they improve to a moderate level over the final 3 d. However, given the short duration of the current experiment, longer-term evaluation is required to fully assess the system’s sustained assimilation performance.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Wu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: