A strong requirement for the next generation of atmospheric computer models is the ability to be able to perform 'Future' runs. Such models are becoming more and more important for making informed policy decisions as to the impact of the predicted growth in global Anthropogenic emissions on (e.g.) future air quality and the feedbacks associated with Climate Change. One type of such a model is commonly known as a 'chemistry-climate model' (or CCM). Here a large-scale chemistry-transport model (CTM) can be driven by meteorological output in different ways. When using a General Circulation Model (GCM) multi-decadal simulations can be made to investigate the possible consequences of a changing climate. Coupling a CTM to some type of operational forecasting system can be used to make integrated meteorological and chemical predictions on a daily basis, making use of an operational data assimilation system for both meteorological components and trace gases. At KNMI this strategy is currently being pursued in the framework of the EU-integrated research project GEMS, where the CTM TM5 has been coupled to the ECMWF forecasting system (IFS). This coupling is two way in that chemical fields for the gases O3, CO, NOx, HCHO and SO2 are exchanged into the IFS and the resulting meteorological output fields exchanged back into TM5. For the chemical fields this is done in terms of the 'chemical tendencies' (i.e.) whether chemical production or destruction terms exist for the selected species for certain locations as calculated by the CTM, which are subsequently exchanged once per hour with the IFS. As part of the development Vincent Huijnen, a post-doc in KS-CK, in collaboration with Johannes Fleming based at ECMWF, has performed sensitivity tests to investigate the differences introduced into these chemical tendencies when performing the transport by either the TM5 model or the IFS. The total tendency fluxes which are computed by TM5 have been found to agree with those calculated using the MOZART CTM model during an independent run for the same time period, which gives confidence that no erroneous tendencies are exchanged with the IFS. The figure below shows that, in general, for TM5 the consistency of the chemical fields obtained by each chosen transport method is good, apart from near the surface, where version 1 represents a CTM-constrained run in which both the tendencies and trace gas concentrations are exchanged with the IFS (one-way coupled, no IFS meteo. back into TM5) and version 2 represents a run where IFS applies vertical mixing to the trace gas distributions received from TM5 (two-way coupling with IFS meteo back into TM5). Version 2 is a crucial step towards obtaining an operational integrated assimilation system for trace gases. Here substantial differences are found which is most likely due to the different parameterizations used for convective transport between TM5 and IFS. Moreover, the use of the 'total tendency' approach (where the CTM performs both the chemical processing step and subsequent transport) is inherently unstable. Future experiments aimed at improving on this will focus on a better integration of TM5 with the IFS, in addition to a more detailed validation of the different contributions which determine the chemical tendencies (emission, deposition, etc) as compared with other participating CTM's (MOCAGE and MOZART). This work also plays a central role for the chemical component of the integrated Earth System model EC-Earth, which is being developed by a consortium of EU countries lead by The Netherlands.

See poster. (pdf)