Whole atmosphere data assimilation and forecast experiments

We report data assimilation and forecast experiments with a whole atmosphere data assimilation and forecast system. The forecast model is an extension of the Global Forecast System (GFS) of the National Center for Environmental Prediction (NCEP), denoted as WAM, while the analysis subsystem uses the NCEP's Gridpoint Statistical Interpolation (GSI). We found that the conventional intermittent, 6-hourly cycling data assimilation strategy used in the WAM-GSI system does not produce the expected atmospheric state in the upper atmosphere (i.e., from about 60 km up): all the important tidal waves are severally damped. As an initial step of overcome this common problem for upper atmosphere data assimilation, the so-called Incremental Analysis Update (IAU) scheme is implemented in the WAM-GSI data assimilation cycles. The data assimilation and forecast experiments with the WAM-GSI-IAU system show significant improvements for the simulation of the 2009 sudden stratospheric warming (SSW) event. Specifically, tidal waves SW2 (semidiurnal, westward propagating, zonal wave number 2) and TW3 (terdiurnal, westward propagating, zonal wave number 3) are better captured in the WAM-GSI-IAU system. In both analysis and forecast, significant enhancement of both SW2 and TW3 in the ionospheric dynamo regions are shown during the 2009 SSW event with a few (~3) days delay.