African Aerosols : Emissions, Deposition and Impacts

1
African Aerosols Emissions, Deposition and
Impacts

• C. Liousse, C. Galy, M. Mallet, B. Guillaume, V.
  Pont, A. Poirson, A. Konaré, C. Junker, C.
  Granier, J.M. Grégoire, F. Solmon, H. Cachier, B.
  Guinot and R. Rosset.

2
Combustion Aerosol Emissions 1- African Fossil
fuel and Biofuel emissions 2- Biomass burning
emissions
African Aerosol Deposition
African Aerosol impacts
Permanent interactions between multiscale
modelling, experiments and satellite Models MNH-
C (local), RegCM3 (Regional) and TM4 (global)
with ORISAM-RAD and GAME models (aerosol
formation and aging, heterogeneous chemistry and
optical properties) Link with LAMP/LOA, ICTP
(UNESCO) and TM4-TM5 group, Measurements
IDAF/AMMA/SACCLAP and experiments in the LA
combustion chamber (Mass, Chemical speciation
of the aerosol,aerosol size/chemistry,physical
properties, in air and in rains) Strong
collaborations with African Researchers Ivory
Coast, Senegal, Benin, Cameroon, South Africa,
Niger, Mali, Congo 6 shared phD at this time
3
African Fossil fuel and Biofuel emissions from
1860-2030

• Africa data are extracted from our global model
  of emissions Junker and Liousse, ACPD 2006
• this global model For the first time to our
  knowledge a coherent inventory for gases and
  particles based on the same method and proxy data
  (fuel consumption, fuel usage..)
• Pollutants CO, CO2, NOx, NMVOC, SO2, BC, OCp,
  OCtot
• African Emissions are provided country by country

4
A bottom-up method (based on Junker and Liousse,
ACPD 2006)
United Nations Energy database Fuel consumption
data for 185 countries, 33 different fuels and
over 50 different usage/technology
categories Emissions are fuel-dependant, fuel
usage-dependant and technology-dependant Our
 lumping  Industrial/Domestic/Traffic Develo
ped/Semi developed/Developing gt Emission
factors for 3 countries classifications, 8
different fuels and 3 usage categories
Population density within each country
(population map) and emissions country/country
gt 1X1 spatial distribution of emissions
5
Year 2000 Africa
Fossil fuel and biofuel combustions
NMVOC
5.5 Mt(NMVOC)
0.64 Mt(BC)
BC
6
Regional differences on the predominant sources
over Africa
Fossil fuel BC
Biofuel BC
7
Trends of BC emissions in South Africa 1860-2030
8
To conclude this part

• We have constructed a flexible and coherent
  inventory for gases and particles with the same
  proxidata and assumptions for Africa

• African Biofuel and fossil fuel emissions
• On going in AMMA and SACCLAP EF
  characterization for unknown fuels (zems, trucks,
  south african coals) cf picture
• On going in our POLCA program phD E. Assamoi
• An improvement of African fuel consumption
  database
• By Replacing UNdata by African database (need to
  be created with Africaclean, APINA and others)
• ex diesel consumption in Ivory Coast 200
  when considering Africaclean database
• By considering the appropriate fuel consumption
  (zems, trucks..)

9
Urban emission characterization at Cotonou - May
2005
Measurements of emission factors ( zem ,
trucks) example for zem CO/CO2
0.42 EF(Black carbon) 0.79g/kgdm !!!
EF(Organic carbon) 9.1g/kgdm !!!
A fixed station at the most polluted place
(Aerosol size, chemical composition, optical
properties, CO/CO2. )
Guinot, Cachier, Liousse et al., in preparation
10
Example of EF results obtained for different
specific unknown fuels in our combustion
chamber (Guinot et al., in preparation)
11
Trends of BC emissions in some Western african
countries 1950-2003

UN fuel consumption data not adapted to Western
Africa
12
First tests show importance of regional focus on
emissions Examples for diesel consumptions for
some African countries
Important discrepancies between global inventory
and regional zoom for the traffic emission
inventory
The use of the African Clean data instead of that
of UN Stat could reduce the biais of AOD
13
Biomass burning emissions (savanna, forest and
agricultural fires)

• The most adapted method to derive AFRICAN BB
  emission gt a bottom up method based on satellite
  burnt area map (Liousse et al., 2004)
• Pollutants BC, OCp, OCtot, CO, CO2, NOx, NMVOC,
  SO2
• and all the species listed in Andreae and
  Merlet, 2004
• Emissions SB x GLCv x BEv x BDv x EFv
• SB area burned gt GBA 2000 product (0.5x0.5,
  monthly)
• GLCv quantity of vegetation v present in cell
  () gt GLC 2000 map
• BEv,BDv biomass density and burning efficiency
  by vegetation type
• EFv emission factor by vegetation type
• gt An important work based on Liousse et al.,
  2004, Michel et al., 2005 with inputs of P.
  Mayaux (Ispra) considering the GLC vegetation
  types

125000 km2
14
GLC map (Ispra) ( 0.5x0.5)
Our assumptions
15
07/2000
08/2000
01/2000
BC emissions in TgC in 2000
12/2000
448 kt(BC)

• Suitable comparison between UMD/GLC
• High difference if comparing with the old
  inventory

16
Interannual African BC emissions (TgC)
In white 1981-1991 AVHRR data and Matthiews
veget map (Liousse et al., 2004) In black
1900-2000 Mouillot et al. burnt areas and GLC
veget map (Liousse, Granier et al., in prep.) In
grey 1997-2003 coupling of GBA-ATSR products
(Granier et al., in prep)
17
African BC emissions by source types
Biomass burning
2.28 Mt(BC)
Biofuel combustions
Fossil fuel combustions
0.20 Mt(BC)
0.44 Mt(BC)
18
To conclude

• We have constructed a flexible and coherent
  inventory for gases and particles with the same
  proxidata and assumptions for Africa

• African Biomass burning emissions
• In AMMA and SACCLAP
• Algorithms are ready to welcome the burnt area
  (BA) for the years 2000- 2007 to provide BB
  emission inventories with a spatial resolution of
  25kmx25km for gases and particles
• African Biomass burning trends on going studies
  gt International
• Exercise called BBSO just launched

19
Sensitivity of BC concentrations to emission
inventory
BC surface concentrations
The model generally underestimate the BC
concentration. L96 inventory is better
performing than L06 since more appropriate for
the time period considered.
Wet deposited BC.
Triangles for measurements, squares for RegCM3
simulations with L96 inventory, diamonds for
RegCM3 simulations using L06 inventory.
Monthly RegCM3 simulated BC values averaged
during the period 1990-1992 at Lamto (6.22 N 
5.01 W ) vs experimental data ( Cachier et al)
20
Sensitivity of Aerosol Optical Depth to emission
inventory Comparison with MODIS Dec 2000

• Too large underestimation with the L96 inventory
• For the L06 inventory, spatial pattern is
  consistent as well as the location of Max AOD
  with MODIS
• The modelled AOD compare well with MODIS, despite
  a relatively underestimation within the Congo
  basin
• Low modelled AOD in West Africa indicates a
  possible under estimation of anthropogenic
  emission (fossil fuel biofuel) in the new
  emission inventory

21
Sensitivity of TOMS absorbing index to emission
inventory
Z gt 2500
BC with TM modelling/ TOMS index
22
DEPOSITION/AIR CHEMICAL COMPOSITION
IDAF Observation Network

IDAF African network of DEBITS II - Part of the
AMMA AC programme (tasks of IGAC II)
(African Monsson Multidisciplanry program
Atmospheric Chemistry)
The IDAF network is composed by 8 stations
representative of great african ecosystems  3 in
South Africa (coordination North West University,
ESKOM) 5 in West Central Africa (Cameroon,
Benin, Mali, Niger, Côte dIvoire) (label ORE
CNRS-coordination Laboratoire dAérologie
Toulouse).

• IDAF scientific objectives
• Chemical Composition of the Atmosphere
• Wet and dry atmospheric Deposition
• Aerosol radiative impact
• Impacts on natural ecosystems gt
• Our model outputs for Multidisciplinary Studies
• AMMA African Monsoon Multidisciplinary program-
  Atmospheric chemistry. Western AF.
• SACCLAP Air Pollution and Climate Change South
  AF.
• Next POLCA for Pollution in African capitals

Monthly Measurements Gas concentrations Weekly
Measurements Chemical composition of
aerosols/size Event/event Chemical composition
of precipitation
EDI/LA
23
IDAF stations in West Central Africa Dry savanna
Aérosols carbonés et minéraux
Hombori (Mali)
Hombori Wet Season
Dry season
Banizoumbou (Niger)
24
IDAF stations in West Central Africa
AMMA super site Benin
SO2, NO2, NH3, HNO3, O3
Aerosol sampling
Rain Collector
IDAF Station in Bénin (Djougou/Nangatchori)
25
Annual means of gas concentration (ppb) (DEBITS
passive samplers)
South Africa Dry savanna Dry savanna industrial
West Central Africa Dry Savanna Wet
Savanna Forest
South African sites Pienaar et al, 2006 (mean
1995-2005 for SO2, NO2,O3,NH3, for HNO3
2003-2005) West central african sites Ourabi,
pHD, mean 1998-2004.
26
NO2 surface concentration variation for the Niger
dry savanna site
Results from IDAF NO2 surface observations have
been coupled to NO2 data from the GOME satellite
biomass burning source in the dry season is
equivalent to soils NOx emission in the global
NOx budget over Africa. (Jaegle et al, JGR, 2004)
27
Particulate and Gaseous contribution to the
precipitation chemical composition
Lamto WET SAV
Banizoumbou DRY SAV
ZOETELE Cameroun
Skukuza DRY SAV
Louis Trichardt DRY SAV
Amersfoort DRY SAV Industial
Galy-Lacaux et al, 2006
28
Annual Mean Chemical Composition of AerosolsDry
savanna - Wet savanna - Forest
Terrigenous contribution soils particles CaCO3
calcite, CaSO4 gypsum, CaMg(CO3)2
dolomite Nitrogenous NO 3- heterogenous
chemistry neutralization of nitric acid by
terrigenous particles
29
BANIZOUMBOU
BC concentrations AMMA/IDAF (aethalometer)
DJOUGOU
BB season
30
ORISAM/TM4 focus on Africa
31
Aerosol concentrations at Lamto (ORISAM-TM4)
Mars-April-May
June-July-August
September-October-November
December-January-February
32
Aerosol concentrations at Djougou (ORISAM-TM4)
Mars-April-May
June-July-August
September-October-November
December-January-February
33
Rain concentrations at Lamto (JJA)
Observed IDAF network
ORISAM/TM4 preliminary results
34
J.P Lacaux et al, 2003,  IGAC synthesis, Global
change series, poster PS2P20
35
ABSORPTION
DJOUGOU (AMMA/IDAF)
SCATTERING
Aerosol direct radiative impact
36
Aerosol optical depth on Djougou
( coll. LOA)
ORISAM/TM4 results

• AOD440 ? 0.5 for major days (mean 0.90 ? 0.01

Aerosol direct radiative impact
37
Aerosol single scattering albedo on Djougou
ORISAM/TM4 results
Aerosol direct radiative impact
38
Aerosol vertical profiles over Djougou
used to improve aerosol vertical profiles in RTM
!
21 / 01 / 06 (14 UTC)
AODTOT_550 ? 0.75
(km)
AOD550 (1 4km) ? 0.22
ORISAM/TM4 results
AOD550 (0-1km) ? 0.49
J. Pelon (SA)
Aerosol direct radiative impact
39
Aerosol local direct radiative forcing on Djougou

• significant aerosol atmospheric forcing !
• 20 W.m-2 ? DFATM ? 70 W.m-2 !

what is the possible impact on the atmospheric
dynamic and on the cloud formation ?
Surface Forcing ATMospheric Forcing TOA
Forcing

• significant aerosol surface forcing !
• - 45 W.m-2 ? DFBOA ? - 100 W.m-2 ?17th Jan. due
  to SSA ? 0.85 AOD ? 1.5 !

what is the possible impact on the surface energy
budget, such as latent or sensible heat fluxes ?
40
Regional climatic Impact of dust particles with
RegCM3 model
PRECIP (NODUST DUST, mm.d-1)
JJA average , (1998 - 2002) period, 60 km SW
forcing only
Dust impacts on winds and precipitation
Wind (850 hpa)
DUST
NODUST
mB
Modélisations RegCM3 Solmon et al., 2007
41
Impact on HEALTH
In the frame on POLCA to improve the link
between Aerosol Exposition and Health impact in
African megacities (Dakar, Mali )
BC in BANGUI
BC in COTONOU
PM10 Cotonou 2005 12-133 µg/m3
(AMMA) Paris 2003-2004 24 µg/m3 (Guinot et
al., 2007) Normes 50 µg/m3 Pékin 2003-2004
135 µg/m3 (Guinot et al., 2007)