Post on 12-Jan-2016
Insight from the A-Train into Global Air Quality
Randall Martin, Dalhousie and Harvard-Smithsonian
Aaron van Donkelaar, Lok Lamsal, Akhila Padmanabhan, Dalhousie University
Chulkyu Lee, Dalhousie National Institute of Meteorological Research, Korea
with contributions from
Rob Levy, Ralph Kahn, Nick Krotkov, NASA
Mark Parrington, Dylan Jones, University of Toronto
OMI NO2 Team
A-Train Symposium, New Orleans
October 26, 2010
Two Applications of Satellite Observations for Air QualityTwo Applications of Satellite Observations for Air Quality
Key pollutants: PM2.5, O3, NO2
(AQHI)
Top-down Constraints on EmissionsTop-down Constraints on Emissions
(to improve AQ simulations)
Smog Alert
Estimating Surface ConcentrationsEstimating Surface Concentrations
(large regions w/o ground-based obs)
PM2.5 : fine aerosol
Column Observations of Aerosol and NOColumn Observations of Aerosol and NO2 2 Strongly Influenced Strongly Influenced
by Boundary Layer Concentrationsby Boundary Layer Concentrations
S(z) = shape factor C(z) = concentration Ω = columnNO2
Aerosol Extinction
O3
Martin, AE, 2008
0.30 0.36 0.43 0.52 0.62 2.2 4.7
Aerosol O3 NO2
0.75 9.6
Normalized GEOS-Chem Normalized GEOS-Chem Summer Mean Profiles Summer Mean Profiles over North Americaover North America
Strong Rayleigh Scattering
( )( )
C zS z
Weak Thermal Contrast
Vertical Profile Affects Boundary-Layer Information in Satellite ObsVertical Profile Affects Boundary-Layer Information in Satellite Obs
O3
Wavelength (μm)
Temporal Correlation of AOD vs In Situ PMTemporal Correlation of AOD vs In Situ PM2.52.5
Correlation over Aug-Oct 2010
Combined AOD Combined AOD from MODIS (and MISR) for 2004-2008from MODIS (and MISR) for 2004-2008Rejected Retrievals over Land Types with Monthly Error vs AERONET >0.1 or 20%Rejected Retrievals over Land Types with Monthly Error vs AERONET >0.1 or 20%
van Donkelaar et al., EHP, 2010
Spatial Correlation (r) of AOD vs in situ PM2.5 for North AmericaMODIS: r = 0.39MISR: r = 0.39Simple Average: r = 0.44 Combined: r = 0.61
Chemical Transport Model (GEOS-Chem) Chemical Transport Model (GEOS-Chem) Simulation of Aerosol Optical Depth Simulation of Aerosol Optical Depth
Aaron van Donkelaar
Ground-level “dry” PMGround-level “dry” PM2.52.5 = = ηη · AOD AOD
η affected by vertical structure, aerosol properties, relative humidity
Obtain η from aerosol-oxidant model (GEOS-Chem) sampled coincidently with satellite obs
GEOS-Chem Simulation of η for 2004-2008
van Donkelaar et al., EHP, 2010
τ(z)/τ(z=0)
Alti
tud
e [k
m]
Evaluate GEOS-Chem Evaluate GEOS-Chem Vertical Profile with Vertical Profile with
CALIPSO ObservationsCALIPSO Observations
• Coincidently sample model and CALIPSO extinction profiles
– Jun-Dec 2006
• Compare % within boundary layer
Model (GC)CALIPSO (CAL)
Optical depth above altitude zTotal column optical depth
van Donkelaar et al., EHP, 2010
Evaluation with measurements outside Canada/US
Global Climatology (2004-2008) of PMGlobal Climatology (2004-2008) of PM2.5 2.5 from MODIS (& MISR) AODfrom MODIS (& MISR) AOD
and GEOS-Chem AOD/PM and GEOS-Chem AOD/PM2.52.5 Relationship Relationship
Number sites Correlation Slope Offset (ug/m3)
Including Europe 297 0.75 0.89 0.52
Excluding Europe 107 0.76 0.96 -2.8
van Donkelaar et al., EHP, 2010
Evaluation for US/Canada
r=0.77 slope=1.07 n=1057
Better than in situ vs model (GEOS-Chem): r=0.52-0.62, slope = 0.63 – 0.71
US EPA standard
van Donkelaar et al., EHP, 2010
van Donkelaar et al., EHP, 2010
• 80% of global population exceeds WHO guideline of 10 μg/m3
• 35% of East Asia exposed to >50 μg/m3 in annual mean
• 0.61±0.20 years lost per 10 μg/m3 [Pope et al., 2009]
• Estimate decreased life expectancy due to PM2.5 exposure
PM2.5 Exposure [μg/m3]
Data Valuable to Assess Data Valuable to Assess Health Effects of PMHealth Effects of PM2.52.5
van Donkelaar et al., EHP, 2010
100
90
80
70
60
50
40
30
20
10
0
AQG IT-3 IT-2 IT-1
Pop
ulat
ion
[%]
5 10 15 25 35 50 100
WHO Guideline & Interim Targets
USA Today: Hundreds Dead from Heat, Smog, USA Today: Hundreds Dead from Heat, Smog, Wildfires in MoscowWildfires in Moscow
9 Aug 2010: “Deaths in Moscow have doubled to an average of 700 people a day as the Russian capital is engulfed by poisonous smog from wildfires and a sweltering heat wave, a top health official said Monday.”
MODIS/Aqua: 7 Aug 2010
PMPM2.52.5 Estimate from MODIS AOD and GEOS-Chem AOD/PM Estimate from MODIS AOD and GEOS-Chem AOD/PM2.52.5
Aaron van Donkelaar
Near Moscow
In Situ
MODIS-based
Evaluation Near Moscow
Regional Mean MODIS-based Estimate
Before Fires During Fires
Impact of TES Assimilation on Surface Ozone (Aug. 2006)
The model underestimates surface ozone in the west
[Parrington et al., GRL, 2009]
TES-based estimates of background O3 are 20-40 ppb
GEOS-Chem before assimilation GEOS-Chem after assimilation
Surface O3 difference (assim - no assim) AQS and NAPS surface O3 data
Background O3 at the surface before assim. Background O3 at the surface after assim.
Without assimilation the model underestimates background ozone by as much as 9 ppb (in western North America)
OMI Tropospheric NOOMI Tropospheric NO22 Column Proxy for Column Proxy for
Surface ConcentrationSurface Concentration
NO/NO2
with altitude
OMI Standard Product: October 2004 – September 2007 Inclusive
General Approach to Estimate Surface ConcentrationGeneral Approach to Estimate Surface Concentration
Daily OMI Tropospheric Column
S → Surface Concentration
Ω → Tropospheric column
In Situ
GEOS-Chem
Coincident Model (GEOS-Chem)
Profile
OM
MO S
S
Ground-Level NOGround-Level NO2 2 Inferred From OMI for 2005 Inferred From OMI for 2005
Temporal Correlation with In Situ Over 2005
Lamsal et al., JGR, 2008
Spatial Correlation of Mean vs In Situ for North America = 0.78
×In situ—— OMI
Application of Satellite Observations for Timely Updates Application of Satellite Observations for Timely Updates to NOto NOxx Emission Inventories Emission Inventories
Use GEOS-Chem to Calculate Local Sensitivity of Use GEOS-Chem to Calculate Local Sensitivity of Changes in Trace Gas Column to Changes in EmissionsChanges in Trace Gas Column to Changes in Emissions
E
Walker et al., ACP, 2010
Local sensitivity of column changes to emissions changes
Fractional Change
in Emissions
Fractional Change in
Trace Gas Column
Lamsal et al., GRL, in prep
Apply to regions where anthropogenic emissions dominate (>50%)
Forecast Inventory for 2009 Based on Bottom-up for 2006 Forecast Inventory for 2009 Based on Bottom-up for 2006 and OMI NOand OMI NO22 for 2006-2009 for 2006-2009
Lamsal et al., GRL, in prep
9% increase in global emissions
21% increase in Asian emissions
OMI SOOMI SO2 2 Column Retrievals Reflect Anthropogenic EmissionsColumn Retrievals Reflect Anthropogenic Emissions
Lee et al., JGR, 2009
OMI Improved SO2 Vertical Columns for 2006
Agreement with Aircraft Observations (INTEX-B): slope = 0.95, r=0.92
Use OMI SOUse OMI SO22 Columns to Map SO Columns to Map SO22 Emissions Emissions
Apply GEOS-Chem for the InversionApply GEOS-Chem for the Inversion
Combustion, Smelters, Volcanoes
Emission
SO2SO4
2-
~day
OH, cloud
Tropospheric SO2 column ~ ESO2 Over Land
Phytoplankton
DMSday
Deposition
Global Sulfur Emissions Over Land for 2006Global Sulfur Emissions Over Land for 2006
49.9 Tg S/yr
54.6 Tg S/yr
r = 0.77 vs bottom-up
SO2 Emissions (1011 molecules cm-2 s-1)
Lee et al., JGR, submitted
Top-Down (OMI)
Bottom-Up in GEOS-Chem (EDGAR2000, NEI99, EMEP2005, Streets2006) Scaled to 2006
Volcanic SO2 Columns (>10 DU) Excluded From Inversion
Top-Down Minus Bottom-up
ChallengesChallenges•Develop retrievals to increase boundary-layer information
•Continue to develop simulation to relate retrieved and desired quantity
•Develop comprehensive assimilation capability
(i.e. CALIPSO vertical profiles and OMI SO2 to inform AOD/PM2.5 relationship)
A-Train Has Provided Unprecedented Insight Into A-Train Has Provided Unprecedented Insight Into Global Air QualityGlobal Air Quality
Chemical Transport Model Plays a Critical Role in
Relating Retrieved and Desired Quantity
NASA, NSERC, Health Canada, Environment Canada
AcknowledgementsAcknowledgements