Herschel Mitchell;
P.L. Houtekamer;
Sylvain Heilliette
ECCC
The 8th EnKF Data Assimilation Workshop
Assimilation of AMSU-A Radiances in an EnKF Column Environment
Talk:
H_Mitchell_presentation.pdf
A column EnKF, based on the Canadian global EnKF and using the RTTOV radiative transfer (RT) model, has been developed to investigate issues relating to the EnKF assimilation of AMSU-A radiances. Experiments have been performed with large and small ensembles, with and without localization. Two different descriptions of background temperature error have been considered: (A) using analytical vertical modes and hypothetical spectra and (B) using the vertical modes and spectrum of a covariance matrix obtained from the global EnKF after two weeks of cycling.
Rather different results were obtained with the two background-error descriptions: the EnKF performed well in experiments with background-error description A, but was virtually unable to reduce the background error (even when using a large ensemble) with background-error description B. To gain an understanding of these results, it seemed that it would be useful to view the different backgrounds that had been considered through the prism of the RT model. Indeed, a comparison of the trace of the matrix [H P H^{T}], where H is the RT model and P is the background-error covariance matrix, for background-error descriptions A and B did explain the difference in the results obtained with the two background-error descriptions.
The implications of these results for the operational assimilation of AMSU-A radiances will be discussed.
A column EnKF, based on the Canadian global EnKF and using the RTTOV radiative transfer (RT) model, has been developed to investigate issues relating to the EnKF assimilation of AMSU-A radiances. Experiments have been performed with large and small ensembles, with and without localization. Two different descriptions of background temperature error have been considered: (A) using analytical vertical modes and hypothetical spectra and (B) using the vertical modes and spectrum of a covariance matrix obtained from the global EnKF after two weeks of cycling.
Rather different results were obtained with the two background-error descriptions: the EnKF performed well in experiments with background-error description A, but was virtually unable to reduce the background error (even when using a large ensemble) with background-error description B. To gain an understanding of these results, it seemed that it would be useful to view the different backgrounds that had been considered through the prism of the RT model. Indeed, a comparison of the trace of the matrix [H P H^{T}], where H is the RT model and P is the background-error covariance matrix, for background-error descriptions A and B did explain the difference in the results obtained with the two background-error descriptions.
The implications of these results for the operational assimilation of AMSU-A radiances will be discussed.
Thanks to McGill's Atmospheric and Oceanic Sciences department for hosting this web-site.