Begin Main Content

Presenter: Dacian N. Daescu

Seminar Title: note on adjoint-based observation impact measures

A parametric approach is introduced as a tool to analyze and develop adjoint-based observation impact measures. High-order approximations of the variation in model functional output due to assimilation of data are derived and distinction is made between Taylor series methods and numerical quadrature methods. The equations of the sensitivity with respect to the observation/background error covariance in 3DVAR/4DVAR are derived and their close relationship to the observation/background sensitivity is established. Preliminary results are presented using the NASA/GMAO GEOS-5 DAS.

return to Seminars