A strategy for perturbing surface initial conditions in LAMEPS

The lack or inadequate representation of uncertainties in the surface initial conditions (ICs) affects the quality of ensemble forecast, in particular the near surface temperature and precipitation. In this paper, a strategy for perturbing surface ICs in limited area model ensemble prediction system, noncycling surface breeding (NCSB) is proposed. The strategy combines short‐range surface forecasts driven by perturbed atmospheric forcing and the breeding method for generating the perturbation to surface ICs. NCSB is implemented and tested in Aire Limitée Adaptation dynamique Développement InterNational‐limited area ensemble forecasting (ALADIN‐LAEF). Statistical verification demonstrates that the application of NCSB improves the ALADIN‐LAEF 2m temperature and precipitation forecast. Positive impacts are also obtained for temperature and specific humidity in the lower atmosphere. Copyright © 2010 Royal Meteorological Society