• Introduction
• Brief description of the models
• Principle of the coupling
• Illustration of the effect of the coupling
• Other application of the ARPEGE/REPROBUS coupled model : sensitivity studies to chemical conditions
• MOCAGE : a new model
• Effect of tropospheric-stratospheric chemistry in climate studies
• Perpectives
• Acknowledgements

FIGURES

• Figure 1
• Figure 2
• Figure 3
• Figure 4
• Figure 5

Introduction

Classical Chemical-Transport Models (CTMs) usually use meteorological parameters (winds, temperature, …) from an external source (meteorological analyses, General Circulation Model (GCM) results) to compute chemistry and to transport chemical species, but they do not take into account the feedback processes of chemistry upon dynamics through radiative processes. This feedback can be taken into account by coupling the dynamical model to the chemical. We will present hereafter the results of such a coupling between the ARPEGE GCM and the REPROBUS CTM to illustrate its interest. As a new CTM -MOCAGE- is currently under development, with both stratospheric and tropospheric chemistries included whereas REPROBUS took only stratospheric chemistry into account, we will show the interest for climate studies to have a complete description of chemistry of the troposphere and stratosphere. Finally, we will give perspectives of such work.

1. Brief description of the models

The ARPEGE atmospheric GCM is a spectral global model that includes 41 layers from the surface up to 80 km of altitude. It has an Eulerian transport scheme and uses the Foucart-Morcrette radiation scheme. REPROBUS is a CTM that describes stratospheric chemistry with 55 chemical species and more than 100 chemical reactions. Heterogeneous chemistry upon Polar Stratospheric Clouds and aerosol particles is parameterised within the model. Transport of the chemical species are assured by a semi-Lagrangian scheme. More details about this CTM can be found in Lefèvre et al., J. Geophys. Res. [1994].

2. Principle of the coupling

Between times (t) and (t + 6h), the ARPEGE GCM computes winds and temperature that are sent to the REPROBUS CTM and used to compute chemistry and to advect the chemical tracers between time (t) and (t + 6h). This is the classical CTM approach (off-line mode) that just takes into account the effect of dynamics upon chemistry. The principle of the coupling is that the ozone three-dimensional field obtained at time (t + 6h) by the REPROBUS CTM is then sent back to the ARPEGE GCM. Thus, ARPEGE will compute new winds and temperature between times (t + 6h) and (t + 12h) initialized by the 3D ozone field from REPROBUS at time (t + 6h). This approach (on-line mode) allows feedback of chemistry upon dynamics throught radiative processes.

3. Illustration of the effect of the coupling

Two simulations were performed with the ARPEGE/REPROBUS model. Each started at January 1^st for 1995 conditions (Sea Surface Temperatures (SSTs), [CO₂], sources gases), but one uses a forced version of the models (off-line mode) while the other is a coupled simulation (on-line mode). The results (Figure 1) showed similar pre-conditioning and beginning date of the Antartic ozone hole (September 25) with the same minimum values and covered areas (October 13), but the coupled simulation gave a much longer lifetime for the stratospheric polar vortex with colder temperatures, delaying the vortex breakdown (November 20 and December 10) than the forced simulation.

Figure 1 : Southern hemisphere total ozone columns for the off-line forced model (left column) and the on-line coupled model (right column)

4. Other application of the ARPEGE/REPROBUS coupled model : sensitivity studies to chemical conditions

Another possible application of the coupled model is the sensitivity to chemical source gas concentrations. Thus, two simulations were performed with the coupled version of ARPEGE/REPROBUS (on-line mode) starting each at January 1^st and using 1995 conditions for SSTs and CO₂ concentrations, but one simulation used source gases concentrations of year 1982 (with 2.3 ppbv of total stratospheric chlorine corresponding to pre-ozone hole concentrations) while the second used concentrations of year 1995 (with 3.7 ppbv of total stratospheric chlorine, corresponding to ozone hole conditions). Results (Figure 2) showed no ozone hole for 1982 chemical conditions in contrast to the existence of an ozone hole for 1995 simulation, hence showing a good behaviour of the model in respect to the ozone hole theory.

A good agreement was found with the TOMS total ozone observations for values over Antarctica despite an underestimation of ozone at Southern mid-latitudes (related to the GCM dynamics).

Figure 2 : Southern hemisphere total ozone columns for the on-line coupled ARPEGE/REPROBUS model with different source gas concentration (above) and equivalent observations from the TOMS spectrometer (below).

5. MOCAGE : a new model

MOCAGE is a three-dimensional grid-point chemical-transport model with 3 domains and different horizontal resolutions (Figure 3). The model includes 47 vertical layers from the surface up to 5 hPa (Figure 4), meaning 1 million grid-points. MOCAGE is forced by ARPEGE analyses and uses a semi-Lagrangian transport scheme. Two convection schemes are available in the model [Tiedtke, Mon. Wea. Rev., 1989 ; Betchold et al., QJR Meteorol. Soc., 2000] and surface emissions for chemical species are from IGAC/GEIA and RIVM/EDGAR scaled by IPCC recommendations. This model can use several schemes to describe either stratospheric chemistry (REPROBUS, Lefèvre et al., J. Geophys. Res., 1994), or tropospheric chemistry (RACM, Stockwell et al., J. Geophys. Res., 1997 ; SAPRC99, California Air Resources Board ; Harvard-GISS, Wang et al., J. Geophys. Res., 1998 ; EMEP, Simpson et al., EMEP MSC-W note 2193, 1993 ; MELCHIOR, Beckmann et al ;, Atm. Env., 2000), or both tropospheric-stratospheric chemistry (meaning for instance, 115 chemical species and 380 chemical reactions for the RACM + REPROBUS scheme). Hence, MOCAGE is a flexible tool with a one-dimensional version for sensitivity studies and is also computationally efficient (1 day of simulation for 1 CPU hour of FUJITSU VPP-5000 for the full 3D version).

Future developments of the model will include a high-resolution regional domain (0.1° x 0.1°) for air quality studies and predictions, data assimilation and coupling MOCAGE to the ARPEGE GCM for source gases (N₂O, CH₄, CFCs) and ozone as it was already done for ozone the REPROBUS CTM as explained in parts 2, 3 and 4.

Figure 3 : Geographical domains taken into account in MOCAGE with associated horizontal resolution

Figure 4 : Vertical distribution (left column) in MOCAGE (47 sigma-pressure levels in the standard version). Intermediate layers are defined asP(i) = A(i) + B(i) x P_surfacewhere functions Aand Bare shown in the right column

6. Effect of tropospheric-stratospheric chemistry in climate studies

The Upper Troposphere / Lower Stratosphere (UTLS) is a region difficult to model but crucial as constraining the stratosphere and where occur exchanges of ozone and water vapour between the two layers. Thus, to show the importance of tropospheric chemistry upon the mean atmosphere, two experiments were performed with MOCAGE forced by ARPEGE analyses (off-line mode) : one uses only a stratospheric chemistry scheme (REPROBUS) and was used as reference while the second experiment includes both tropospheric and stratospheric chemistry (RACM + REPROBUS scheme). Each simulation was run over a few days in January and in July. Results upon ozone (Figure 5) showed an increase of ozone in the upper and middle stratosphere, wheres a decrease is found in the tropical UTLS in association with convection wich brings poorer ozone air from the lower troposphere to the UTLS region when tropospheric-stratospheric chemistry is taken into account.

Figure 5 : Snapshots of zonally averaged ozone distributions for January (left colum) and July (right colum) for ozone obtained with the tropospheric-stratospheric chemistry scheme (top), the stratospheric chemistry scheme (middle) and differences in percents (bottom) (yellow and red indicate ozone increase while green and blue indicate decrease)

7. Perpectives

When the MOCAGE CTM will be coupled to the ARPEGE GCM, then it will be used for climate studies and future atmosphere changes, with a consistent chemical simulation of both the troposphere and the stratosphere,including interactions between dynamics and chemistry. Also, the MOCAGE CTM will be used for case studies of the UTLS region and Stratospher-Troposphere Exchanges, with special emphasis upon comparison against the MOZAIC in-situ measurements and data assimilation.Several impact studies will be done to evaluate for instance, subsonic and European supersonic project aircraft, surface emissions abatments.

Acknowledgements

The authors are very grateful to the European Commission for funding this work through the EC-funded project MOZAIC-III (contract EVK2-1999-00141).