MOCAGE first results on the AMMA region

1
MOCAGE first results on the AMMA region

• Béatrice Josse1, David Kaczka1 and Brice Barret2
• AMMA Workshop, 2-3 July 2007

1 Météo France CNRM/GMGEC/CARMA 2 CNRS
Laboratoire dAérologie
2
Plan

• I . Which simulations?
• II . Some annual results
• III . Special look at August 2006
• IV. Next step LiNOx in MOCAGE

3
MOCAGE configuration for AMMA
I. Which simulations?
GLOBE 2x2
AMMA 0.5x0.5
Chemscheme RACMOBUS (strato-tropo)

• Subgrid parameterization
• - convection (Bechtold, 2001)
• scavenging (Mari et al, 2000 Liu et al, 2001)
• dry deposition (Wesely, 1989)
• eddy diffusion (Louis, 1979)
• Resolved-scale transport
• Semi-lagrangian (Williamson et Rasch,1989)

4
MOCAGE configuration for AMMA
I. Which simulations?

• Emissions as prescribed by IPCC on a 1 x1º
  grid
• averaged for 2 x 2 º global resolution, simply
  interpolated for 0.5 x 0.5 º
• Biomass burning emissions directly injected up
  to 6km.
• Other emissions at the surface
• Meteorological forcings ARPEGE Météo-France
  operational model
• Year 2006 simulated, 3-hour results.

5
Improvement of resolutionDoes it improve results?
I. Which simulations?
2x2º
0.5x0.5º
400hPa ozone mixing ratio, Aug 10th, 12hUTC

• Structures are the same
• Significative differences more poor-ozone air
  at 0.5º
• Convection is more efficient

6
Improvement of resolution Does it improve
results?
I. Which simulations?
ozone mixing ratio, Aug 10th, 12hUTC

• Convection more efficient, and realistic (better
  chemical tropopause )
• 0.5º grid resolution OK

7
Ozone tropospheric column
II. Annual results
Monthly means of O3 tropospheric columns

• Principal structures are well seen
• August very good
• February too high amplitude for MOCAGE
  concentrations

8
Low troposhere CO
II. Annual results
Monthly mean of CO mixing ratio (ppb) at 850 hPa
MOCAGE
MOCAGE
MOPITT
MOPITT
Wet season
Dry season

• Intensity and position of biomass fires well
  represented
• Atlantic export well seen
• Underestimation of concentrations above Sahara,
  Persian Gulf and
• southern Europe.

9
NOx
II. Annual results
Monthly means of NO2 tropospheric columns
Satellite OMI

• Intensity and position of biomass fires well
  represented
• LiNOx and wet ground-emissions are lacking

10
O3 and NO2 annual cycles at Nangatchori
II. Annual results

• Annual cycles are correct
• Over estimation of ozone during the dry season
• Under-estimation of NO2 Lack of
  ground-emissions?

11
Vertical sections around Cotonou
III. August 2006

• CO transport from the Southern Hemisphere
• fires to the Guinea Gulf
• Very high concentrations around 800hPa

12
Vertical sections around Cotonou
III. August 2006
III. August 2006

• Ozone production in the region of the fires, near
  CO production
• Very high low-levels concentrations, around
  800hPa again

13
Where do high CO concentrations come from?
III. August 2006

• Running MOCAGE in inverse mode
• Starting from a high CO concentration surface
  point
• Running back 10 days
• Back long-range transport
• Eddy diffusion, scavenging (self adjoint)
• No convection ( ajoint needed)
• Interpretation has to be careful this only
  gives the origin of the air mass

14
III. August 2006
Where do high CO concentrations come from?
Back trajectories at 870 hPa
15
O3 Observations at the surface
III. August 2006

• Diurnal cycle OK
• Ground concentrations OK

Mean diurnal cycle of O3 at Nangatchori
16
Altitude observations
III. August 2006
FALCON flight 16/08/06 6 km
All flights
ATR-42 07/08/06 lower altitudes
O3(MOCAGE)-O3(Obs)

• For all flights
• In the middle troposphere good representation
• Significative over-estimation in low levels.

17
Altitude observations
III. August 2006

• Gradients and ground concentrations OK
• Tropopopause well seen
• Maxima too low in altitude dipole between low
  and middle troposphere

CO measured by AMMA fleet
MOCAGE CO zonal (10W-10E) mean for August 2006
MOCAGE O3 zonal (10W-10E) mean for August 2006
O3 measured by AMMA fleet
18
Why this dipole?
III. August 2006

• 2 explanations
• Not enough exchange between PBL and free
  troposphere
• Altitude injection of emissions too weak

Percentage of emissions injected lower than 1km
19
Conclusions and plans

• Relatively good agreement with observations, but
  still work to do!
• Plans
• Go on with comparison to observations
• Work on the PBL and export in the free
  troposphere (tuning of Louisscheme, new one?)
• Data assimilation (MOPITT, OMI)
• Evaluate global and regional budgets of ozone,
  OH
• Participate in the model intercomparisons
• Include soil emissions and LiNOx

20
IV. LiNOx in MOCAGE
LiNOx parametrisation in the MOCAGE CTM

• Parametrisation based on the mass flux formalism
  of KFB convective scheme (Mari et al., 2006)
• Consistent with transport and scavenging of
  species
• No a priori vertical/horizontal NOx production
• No global yearly a priori NOx production

• Lightning frequencies parametrised with cloud
  top height (Hct)
• Lfcontinents Hct4.9
• Lfocéans Hct1.7
• (Price and Rind, 1992)
• Production of NOx per flash
• P(IC)P(CG)2.2x1026 molecs.
• (Ridley et al., 2005)

21
IV. LiNOx in MOCAGE
Convection and lightnings in the model
Lightnings
Convection
Mars 2005
MOCAGE ECMWF
Observations
OLR
LIS
Flashes/ km2/day
W/m2
Lightning frequencies too strong over the Amazon
bassin and too weak over the Congo bassin
Good representation of the location of deep
convection
22
IV. LiNOx in MOCAGE
Impact of LiNOx on the NOx ditributions
Maximum increase in the upper troposphere over
continental convective regions consistent with
lightning frequencies
23
IV. LiNOx in MOCAGE
Impact of LiNOx on the O3 ditributions
(O3 with LiNOx) / (O3 without LiNOx)

• Maximum O3 production from lightnings in the
  upper troposphere
• over convective regions (Bolivian and
  South-African highs)
• Recirculation over the oceans by the
• Subtropical Westerly Jets

24
IV. LiNOx in MOCAGE
Validation of the simulated O3 ditributions
TROCCINOX (Geophysica)
Obs. No LiNOx LiNOx
Colonnes troposphériques dO3 (Mars 2005)
25
SHADOZ
TROCCINOX
-25
SHADOZ (Radiosoundings at 8 tropical stations)

• Improved tropospheric columns in
• the southern hemispehere

• Improved ozone vmr in the middle
• to upper troposphere
• LiNOx ? correction of the
• negative bias in the simulated
• tropospheric O3

25
-25
25
(No Transcript)