Hourly Data Assimilation With The Gridpoint Statistical Interpolation For The Rapid Refresh

Abstract

The Rapid Refresh (RR) model and analysis system, currently under development at the Earth System Research Laboratory (ESRL), will replace the Rapid Update Cycle (RUC) in 2009. The RUC has been operational at the National Centers for Environmental Prediction (NCEP) since 1994. It is a high-frequency (hourly) data assimilation and prediction system designed to provide meso-scale weather guidance for aviation, severe weather, and public forecasting needs. The RR will occupy the same niche as the RUC in the NCEP operational model suite, but will be based on one of the two dynamical cores of the Weather Research and Forecasting (WRF) model and utilize the Gridpoint Statistical Interpolation (GSI), under development at NCEP and other agencies, for the hourly data assimilation cycle. In addition, as seen in Fig. 1., the RR will run over a domain that is 2.6 times larger than the present RUC domain, including Alaska, Puerto Rico and the Caribbean. For further information about the general design of RR, see Benjamin et al. (2007). Experiments relating to the selection of a WRF dynamical core for the RR are described in Brown et al. (2007). This paper describes new developments related to the data assimilation aspects of RR. In the present RUC system, a three-dimensional variational (3-DVAR) scheme is used in RUC’s hourly update cycle for analysis of data from diverse sources. In the new RR, following a design decision in June 2005, the Gridpoint Statistical Interpolation (GSI) scheme will be used for data assimilation instead of the current RUC 3DVAR scheme.