Research On Data Assimilation And Energy-Preserving Method Based On Ocean Numerical Model

As the most widely distributed water on the earth,the ocean contains large amount energy and resources,and its heat and energy transportation can make a great contribution to the global climate change.Meanwhile,the disasters caused by the ocean would also threaten the survival of human beings,so the study focus on the environment and dynamical characteristic of the ocean is of great significance.Currently,the main approach we could explore the Three-Dimensional(3D)ocean structure is by the numerical simulation.The kernel of oceanic numerical model is to solve the fluid dynamics equations,so it can be applied to estimate the ocean state of any space-time range,but the discretization is mainly conducted by the finite difference method,causing the insufficient accuracy with the in-situ observations.Therefore,the main research of this article is focus on the optimization strategy of data assimilation and the construction of energy-preserving method,to benefit the ocean numerical simulation with higher accuracy.The data assimilation method could be considered as a posterior correction of the numerical simulation results,and it is realized by minimizing the error between the numerical solution and the in-situ obeservations.Moreover,constructing a better numerical method with structure-preserving property can make an essential contribution to the accuracy of the ocean modeling system.Firstly,we use the optimized objective analysis method to extend the along-track SSH(Sea Surface Height)observations to the blank region effectively.Moreover,an evaluation of the influence of the extending progress on the numerical simulation is conducted based on ROMS and Four-Dimensional Variational assimilation method.The altimeter missions regularly sample sea level along their ground tracks and the coverage is limited for the demand of continuous ocean process research,hence the enhancement of the model result is localized when it is assimilated.Though the gridded product obtained by the data merging could reconstruct the continuous structure of SSH,some noise is introduced at the same time unavoidably.By the comparison with in-situ observations,the impact of the extension progress on the model assimilation is evaluated.The results indicate that,the complete structure of ocean phenomenon is reconstructed by the benefit of the extension progress,and reconstructed information could enhance the eddy detection of the assimilation results.However,the authenticity of the observations is reduced in the extension,which may introduce a negative impact on the assimilation.Furthermore,we attempt to transfer the extension progress of SSH to the underwater and build an ocean Three-Dimensional(3D)temperature reconstruction model based on the available observations,therefore a 3D temperature field reconstruction algorithm is derived.Considering the vertical stratification of ocean temperature,the Argo temperature profiles are firstly fitted to generate a series of temperature functions of depth,accomplished with the vertical temperature gradient.By applying inverse distance weighting interpolation in the horizontal direction,a gridded 3D temperature gradient field could be obtained.Finally,combined with the processed SST,the 3D temperature field reconstruction is realized below the surface.To confirm the effectiveness of the algorithm,an experiment in the Pacific Ocean south of Japan is conducted,for which a 3D temperature field is generated.Compared with other similar gridded products,the reconstructed 3D temperature field derived by the proposed algorithm achieves satisfactory accuracy,with a higher spatial resolution,resulting in the capture of smaller-scale characteristics.Additionally,benefit from the proposed 3D temperature reconstruction algorithm,we extend the Argo profile to present more information.To validate the impact of the reconstructed information on the model assimilation,an experiment including four cases is conducted based on ROMS and 4DVAR.The comparison with the EN4 dataset shows that,the assimilation of reconstructed temperature profiles does enhance the accuracy.The net enhancement of reconstructed temperature profiles is comparable with Argo T-S observations.Hence,the positive impact of the developed algorithm on data assimilation is validated.Finally,we start with the derivation of the surface gravity wave equation,and then propose an energy-preserving numerical method based on the theory of Finite Volume Method(FVM).In each iteration,the theoretical energy of the system is computed by solving the derived energy evolution equation,accomplished with the estimation of the equation variable.By the comparison of the derived theoretical energy and numerical energy,we make an adaptation to the numerical solution by introducing a freedom in the numerical reconstruction.The experiment shows that the proposed energy-preserving algorithm has superiority on long-time simulation,with better convergence than the Godunov method.