Initialization of cloud water content in a data assimilation system

Abstract

Cloud water content (CWC) is not treated in most operational objective analyses and initialization schemes. When CWC is used as a prognostic variable in a forecast model, it is necessary to define this variable at the initial time. A commonly used method is to set the initial CWC to zero or use a forecast CWC field from the previous data-assimilation cycle (the first-guess field for the objective analysis) without any modification. The inconsistent treatment of CWC and other fields leads to an imbalance between the first-guess cloud water field and other analyzed fields (winds, temperature, humidity, and surface pressure). In this study, the diabatic digitalfiltering initialization scheme is used to alleviate this imbalance. It is shown that an intermittent data assimilation system with this initialization scheme can produce a better cloud evolution, a shorter spinup time, and a removal of the initial shock in precipitation.