P4 Multivariate 3DVar Analysis of Clouds

Descombes, Gael, Thomas auligne, Yann Michel, Francois Vandenberghe , National Center for Atmospheric Research

The specification of model background error statistics is a key component to data assimilation since it conditions the impact observations will have on the analysis. Classically, variational methods represent them for wind, temperature, pressure and humidity through a statistical model. Cloud data assimilation remains a challenge. In this study, the cloud hydrometeors are included as control variables and their background errors are represented in a multivariate approach. They are directly assimilated in a 3DVAR process (WRFDA). Sensible tests are leaded on a real case study over the CONUS domain to estimate the background errors formulation that performs the best using radiance data from different satellites.