Research Summaries

Back Computational Methods and Nonlinear Filters for Data Assimilation

Fiscal Year 2020
Division Graduate School of Engineering & Applied Science
Department Applied Mathematics
Investigator(s) Kang, Wei
Sponsor Naval Research Laboratory (Navy)
Summary A high resolution model for weather prediction has tens of millions of state variables. The extremely high dimension results in an intractable error covariance because of the required high computational cost and input/output (I/O) loads, as well as a large memory size needed in the process of the matrix. The goal of the proposed research is to explore new algorithms to strike a balance between the scalability and the accuracy of Kalman filters by taking the advantage of approximate sparsity of error covariances. The underlying philosophy of our research approach is fundamentally from existing ones. In addition to the statistical properties of error covariance, we explore the quantitative characteristics and the sparsity pattern of error covariance deduced from the model, including both the dynamic model and the observation model. The goal is to take the advantage of these quantitative characteristics for the purpose of improving data assimilation accuracy.
Keywords
Publications Publications, theses (not shown) and data repositories will be added to the portal record when information is available in FAIRS and brought back to the portal
Data Publications, theses (not shown) and data repositories will be added to the portal record when information is available in FAIRS and brought back to the portal