Title: Control of lateral boundary conditions in WRF 4D-Var

Authors: Xin Zhang (NESL, National Center for Atmospheric Research, Boulder, CO)
Xiang-Yu Huang (NESL, National Center for Atmospheric Research, Boulder, CO)
Nils Gustafsson (Swedish Meteorological and Hydrological Institute, Sweden)

The regional WRF forecasting problem is a lateral boundary condition problem in addition to the initial condition problem. For regional WRF application, the lateral boundary conditions are generally provided by a global model. One may argue that the available boundary conditions from the host model have to be accepted (without change) for usage in the WRF. However, the observations close to the lateral boundaries might be assimilated in analysis stage and they might influence the initial conditions. In WRF model, the initial conditions are used as the lateral boundary conditions at the initial time, the subsequent forecast will also be influenced by these observations. With application of 4D-Var for the WRF forecasting, the lateral boundary conditions can be controlled in minimization during the period of the data assimilation window. Both ideal and realistic experiments with WRF 4D-Var show that this may be particularly important for assimilation of phenomena that are observed well inside the domain during the later part of the data assimilation window, while being propagated through the lateral boundaries during the early part of the data assimilation window. If we do not control the lateral boundary conditions, observed information related to these phenomena may be lost.

GMAO Head: Michele Rienecker Global Modeling and Assimilation Office NASA Goddard Space Flight Center

Curator: Nikki Privé Last Updated: May 27 2011