Data Assimilation Seminar

Ms. Jemima Tabeart (Jan. 28, 2019, 14:15-)

Affiliation U of Reading
Title Using reconditioning to study the impact of correlated observation errors in the Met Office 1D-Var system
Abstract

Reconditioning methods can be used to improve the condition number of correlated matrices. However, the impact of using reconditioned covariance matrices within a data assimilation framework is not well understood. We present the results of a detailed case study using the UK Met Office 1D-Var system where different levels of reconditioning are used. We consider changes to convergence, estimated variables and quality control associated with different choices of error covariance matrix. We find that convergence improves with the introduction of correlated observation error covariance matrices, and that the majority of changes to retrieved variables are small.

PDF Link to PDF file

Back to Seminar Series

Page Top