The definition of mesoscale selective forecast error covariances for a limited area variational analysis

摘要: ¶The paper deals with an alternative formulation of the so-called NMC (National Meteorological Center, now National Centers for Environmental Prediction) statistics to compute background error covariance matrix be used in a mesoscale variational analysis. While standard method uses differences forecasts valid same time, but starting from different analysis times, new required recomputation short-term forecast initial and lateral boundary data that come long-term run. In frame limited-area model, this approach forces variances at large scales decrease drastically, because those are controlled by (constant data) coupling. As result, cost function acts more scale selectively, emphasis on medium scales. The increments obtained 3D-VAR system show sharper concentrated formulation, both single observation full experiments. This work is part wider project building assimilation inside ALADIN model. complete should concentrate features it not reanalyse were already treated global model (ARPEGE). Some difficulties perspectives drawn concluding discussion.