Presentation of Pierre Brousseau during the 2^nd AROME training course, Lisbon, 4-7 March 2008 : "A mesoscale data assimilation system for operations : the AROME RUC", PierreBrousseau, Y. Seity, L. Auger, L.Berre, T.Montmerle, E. Wattrelot, C. Faccani

• General ideas on data assimilation
• The AROME data assimilation system
• Background-error Statistics
• Assimilation experiments

• A prototype for the AROME data assimilation system is currently being tested at Météo-France.
  (* It is supplied by the same kind of observation as the ALADIN-FRANCE operational suite.
• The background-error statistics for AROME have been calculated by an ensemble methode using the same multivariate formulation as in ALADIN-France. Compared to the ALADIN-FRANCE ones, background-error standard deviations are increased and horizontal correlation length scales are much shorter: Analysis increments are strongerand narrower.
• This prototype shows its ability to improve analysis and forecasts, giving a more realistic depiction of initial conditions.
• The general benefit of the analysis appears during the first 12-hour forecast ranges, then lateral conditions mostly takeover the model solution.

• Works currently in progress on :
  • the use of observations at a higher spatial resolution (airep, IR radiances,…).
  • the assimilation of observations more representative of mesoscale like RADAR winds and reflectivities (see also the talk of Eric Wattrelot).
  • A surface assimilation for AROME (see also the talk of Rafiq Hamdi).
• Works are planned in orderto takea better advantageof high-frequencyobservations using:
  • 3D-FGAT (First Guess at Appropriate Time) assimilation scheme
  • Incremental Digital Filter Initialization allowing 1-h cycling
• AROME schedule :
  • April 2008 : AROME pre-operational suite managed by operational team (3-h RUC, RADAR winds,…).
  • October 2008 : AROME declared operational at Meteo-France (?)