Further development of data assimilation system for the Whole Atmosphere Model (WAM) at NCEP

Whole Atmosphere Model (WAM) represents an extension of NCEP's global forecasting model GFS with inclusion of more vertical levels (150 instead of 64, which corresponds to about 600 km) and parameterization of physics processes of critical interest for modeling of the middle and upper atmosphere. WAM has been developed at NOAA's Space Weather Prediction Center (SWPC), and preparations for its operational implementation at NCEP, from the modeling stand of view, are described in the accompanying presentation. Objective of this talk is to present similar preparation for implementation of WDAS, which is the data assimilation system for WAM. WDAS has been originally derived at SWPC in collaboration with NCEP by extensions of NCEP's grid point statistical interpolation (GSI) analysis system. The intended highlights of this talk will include testing of different strategies for initialization, including an implicit normal mode constraint modified to preserve tidal signal in the initial data while dealing with the volatile upper atmosphere and application of recursive digital filter, and further polishing of the background error covariance matrix for WDAS.