Direct 4D-Var Assimilation of NCEP Stage IV Radar and Gauge Precipitation Data at ECMWF

AbstractDirect four-dimensional variational data assimilation (4D-Var) of NCEP stage IV radar and gauge precipitation observations over the eastern United States has been developed and tested in ECMWF’s global Integrated Forecasting System. This is the natural extension of earlier work using a two-step 1D+4D-Var approach. Major aspects of the implementation are described and discussed in this paper. In particular, it is found that assimilating 6-h precipitation accumulations instead of the original hourly data substantially improves the behavior of 4D-Var, especially as regards the validity of the tangent-linear assumption.The comparison of background and analysis precipitation departures demonstrates that most of the information contained in the new precipitation observations is properly assimilated. Experiments run over the periods April–May and September–October 2009 also show that local precipitation forecasts become significantly better for ranges up to 12 h, which indicates that a genuine precipitat...