An adjoint-based adaptive ensemble Kalman filter

TitleAn adjoint-based adaptive ensemble Kalman filter
Publication TypeJournal Article
Year of Publication2013
AuthorsSong H.J, Hoteit I., Cornuelle BD, Luo X.D, Subramanian AC
JournalMonthly Weather Review
Volume141
Pagination3343-3359
Date Published2013/10
ISBN Number0027-0644
Accession NumberWOS:000324836800006
Abstract

A new hybrid ensemble Kalman filter/four-dimensional variational data  assimilation (EnKF/4D-VAR) approach is introduced to mitigate background  covariance limitations in the EnKF. The work is based on the adaptive  EnKF (AEnKF) method, which bears a strong resemblance to the hybrid  EnKF/three-dimensional variational data assimilation (3D-VAR) method. In  the AEnKF, the representativeness of the EnKF ensemble is regularly  enhanced with new members generated after back projection of the EnKF  analysis residuals to state space using a 3D-VAR [or optimal  interpolation (OI)] scheme with a preselected background covariance  matrix. The idea here is to reformulate the transformation of the  residuals as a 4D-VAR problem, constraining the new member with model  dynamics and the previous observations. This should provide more  information for the estimation of the new member and reduce dependence  of the AEnKF on the assumed stationary background covariance matrix.  This is done by integrating the analysis residuals backward in time with  the adjoint model. Numerical experiments are performed with the  Lorenz-96 model under different scenarios to test the new approach and  to evaluate its performance with respect to the EnKF and the hybrid  EnKF/3D-VAR. The new method leads to the least root-mean-square  estimation errors as long as the linear assumption guaranteeing the  stability of the adjoint model holds. It is also found to be less  sensitive to choices of the assimilation system inputs and parameters.

 

Short TitleMon. Weather Rev.
Integrated Research Themes: 
Student Publication: 
No
sharknado