Publications & Events

pubs home | list all entries

Implementation of a Digital Filter Initialization in the WRF Model and Its Application in the Rapid Refresh


Because of limitations of variational and ensemble data assimilation schemes, resulting analysis fields exhibit some noise from imbalance in subsequent model forecasts. Controlling finescale noise is desirable in the NOAA’s Rapid Refresh (RAP) assimilation/forecast system, which uses an hourly data assimilation cycle. Hence, a digital filter initialization (DFI) capability has been introduced into the Weather Research and Forecasting Model and applied operationally in the RAP, for which hourly intermittent assimilation makes DFI essential. A brief overview of the DFI approach, its implementation, and some of its advantages are discussed. Results from a 1-week impact test with and without DFI demonstrate that DFI is effective at reducing high-frequency noise in short-term operational forecasts as well as providing evidence of reduced errors in the 1-h mass and momentum fields. However, DFI is also shown to reduce the strength of parameterized deep moist convection during the first hour of the forecast.

View Citation