Coupling Global Atmospheric Chemistry Transport Models to ECMWF Integrated Forecasts System for Forecast and Data Assimilation Within GEMS

Abstract

The paper presents the implementation of a coupled forecast and assimilation system developed within the subproject on Global Reactive Gases (GRG) of the GEMS-project (Global and regional Earth-system (Atmosphere) Monitoring using Satellite and in-situ data, FP6).

One of the main objectives of the GEMS project is to utilise ECMWFs 4D-VAR data assimilation system to assimilate satellite observations of atmospheric composition at the global scale. The GRG subproject focuses on the assimilation of the following gases: carbon monoxide (CO), ozone (O₃), nitrogen oxides (NO_x=NO+NO₂) formaldehyde (HCHO) and sulphur dioxide (SO₂). These gases play a key role in atmospheric chemistry and are observable from space.

ECMWF’s integrated forecast system (IFS) is able to simulate the transport of these tracers but does not contain the modules for the simulation of chemical conversion, emission and deposition. Instead of directly integrating (on-line coupling) the relevant modules into the IFS, a coupled approach was taken on which links the IFS to already established Atmospheric Chemistry Transport Models (CTMs). The coupled approach seemed to be a much smaller development effort, and it offers more flexibility in the choice of the modules for chemical conversion, emission and deposition by coupling different CTMs to the IFS.

The two-way coupled system consists of the IFS and a CTM. The three candidate CTMs for the coupled system are MOCAGE (Josse et al., Tellus 56B:339–356, 2004), MOZART (Horowitz et al., J Geophys Res, 2003), and TM5 (Krol et al., Atmos Chem Phys 5:417–432, 2005).The coupling software OASIS4 (Valcke and Redler, OASIS4 User Guide (OASIS4_0_2). PRISM Support Initiative Report, No. 4, 2006) has been implemented to facilitate the data exchange.

In the coupled system, IFS sends meteorological data at high temporal resolution to the CTMs. The CTMs provide concentration tendencies due to emissions and chemical conversion as well as initial tracer conditions to the IFS. The application of external tendencies is required in IFS because its 4DVAR data assimilation needs to account for tracer source and sink terms which are not simulated in the IFS model. Moreover, the tracer transport may benefit from the sophisticated vertical transport schemes of the IFS.

The coupled system has been applied in forecast mode for several months in 2003 in different configurations in terms of vertical transport and coupling synchronisation. Test assimilation runs of CO measured from space by the MOPITT instrument have been successfully carried out for several weeks. Experimental near-real time forecasts of the coupled system run since April 2007. However, this paper focuses on the design of the system.