Reconstruction of marine three-dimensional thermohaline field based on the variational method

Abstract To enhance the accuracy of three-dimensional temperature and salinity data in the ocean, this paper adopts a three-dimensional variational (3D-Var) approach. This method integrates real-time observational data to reconstruct the three-dimensional temperature and salinity data that are consistent with the temporal dimension of the observational data, thereby obtaining more accurate temperature and salinity data. The paper first conducts quality control and preprocessing on the observational data, then uses the three-dimensional variational algorithm. By minimizing the differences between the observational data and the background field, it optimizes the estimation of the temperature and salinity fields. To verify whether the reconstructed temperature and salinity data are more accurate than the data before reconstruction, a reconstruction experiment was conducted in a small area of the Northwest Pacific for different years. By comparing the temperature and salinity data before and after the experiment, it was found that the data after reconstruction were more accurate.