Transcript 2.5 van Donkelaar et al., EHP, 2010 van

Air Quality Applications of Satellite Remote Sensing
Randall Martin, Dalhousie and Harvard-Smithsonian
Aaron van Donkelaar, Akhila Padmanabhan, Dalhousie University
Lok Lamsal, Dalhousie U  NASA Goddard
with contributions from
Rob Levy, Ralph Kahn, NASA
1st Workshop on Satellite Observations for Air Quality Management
9 May 2011
Two Major Air Quality Applications of Satellite Observations of
Atmospheric Composition
Estimating Pollution Concentrations
(regions w/o ground-based obs)
(AQHI)
Smog Alert
Top-down Constraints on Emissions
(to improve AQ simulations)
Major Nadir-viewing Space-based Measurements of
Tropospheric Trace Gases and Aerosols (Not Exhaustive)
Solar Backscatter & Thermal Infrared
Sensor GOES GOME MOPITT
Imager
Platform GOES ERS-2
(launch) (varied) (1995)
MISR MODIS AIRS
Terra
(1999)
Equator
Crossing
n/a
10:30
Typical
Res (km)
4x4
320
x40
22x22
Global
Obs
n/a
3
3.5
Aerosol
X
X
Aqua
( 2002)
10:30
1:30
18x18 10x10
7
2
X
X
SCIA- TES OMI PARA- CALIOP GOME IASI
MACHY
SOL
-2
Envisat
(2002)
Aura
(2004)
10:00
1:45
14
x14
60x30
1
6
PARA- Calipso
SOL
(2006)
(2004)
1:30
8x5 >24 18x16
x13
n/a
MetOp
(2006)
1:30
9:30
40x40
80x40
12
x12
0.5
1
1
n/a
1
X
X
X
X
X
NO2
X
X
X
X
HCHO
X
X
X
X
CO
XX
Ozone
X
SO2
X
NH3
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Column Observations of Aerosol and NO2 Strongly Influenced
by Boundary Layer Concentrations Weak Thermal
Strong Rayleigh
Scattering
O3 HCHO
SO2
0.30
Aerosol
CO
CO
Contrast
O3
2.2
4.7
9.6
NO2
0.36
0.52
0.62
0.75
Wavelength (μm)
0.43
Vertical Profile Affects Boundary-Layer Information in Satellite Obs
Normalized GEOS-Chem
Annual Mean Profiles over
North America
Aerosol
Extinction
O3
CO
HCHO
SO2
NO2
C ( z)
S ( z) 

S(z) = shape factor
C(z) = concentration
Ω = column
Martin, AE, 2008
Temporal Correlation of AOD vs In Situ PM2.5
Correlation over Aug-Oct 2010
Aerosol Optical Depth (AOD) from MODIS and MISR over 2001-2006
MODIS
MODIS
r = 0.40
vs. in-situ PM2.5
1-2 days for global coverage (w/o
clouds)
AOD retrievals at 10 km x 10 km
Requires assumptions about surface
reflectivity
MISR
6-9 days for global coverage (w/o
clouds)
AOD retrievals at 18 km x 18 km
MISR
r = 0.54
vs. in-situ PM2.5
0
0.1
0.2
AOD [unitless]
0.3
Simultaneous retrieval of surface
reflectance and aerosol optical
properties
van Donkelaar et al., EHP, 2010
July
Agreement With AERONET Varies with Surface Type
MODIS
MISR
9 surface types, defined by monthly mean surface albedo ratios,
evaluation against AERONET AOD
van Donkelaar et al., EHP, 2010
Combined AOD from MODIS and MISR
Rejected Retrievals for Land Types with Monthly Error vs AERONET >0.1 or 20%
0.25
Combined
MODIS/MISR
r = 0.63 (vs. in-situ PM2.5)
0.2
0.15
0.1
0.05
MODIS
r = 0.40
MISR
r = 0.54
(vs. in-situ PM2.5)
(vs. in-situ PM2.5)
0
van Donkelaar et al., EHP, 2010
AOD [unitless]
0.3
Chemical Transport Model (GEOS-Chem) Simulation
of Aerosol Optical Depth
Aaron van Donkelaar
Ground-level “Dry” PM2.5 = η · 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 2001-2006
van Donkelaar et al., EHP, 2010
Model (GC)
CALIPSO (CAL)
•
•
Coincidently sample model
and CALIPSO extinction
profiles
– Jun-Dec 2006
Compare % within boundary
layer
Altitude [km]
Evaluate GEOS-Chem
Vertical Profile with
CALIPSO Observations
Optical depth above altitude z
Total column optical depth
τa(z)/τa(z=0)
Significant Agreement with Coincident In situ Measurements
0.40
MISR AOD
0.54
Combined AOD
0.63
Combined PM2.5
0.77
Satellite-Derived [μg/m3]
MODIS AOD
Satellite
Derived
In-situ
In-situ PM2.5 [μg/m3]
van Donkelaar et al., EHP, 2010
Annual Mean PM2.5 [μg/m3] (2001-2006)
r
Global Climatology (2001-2006) of PM2.5
Evaluation with measurements outside Canada/US
Number sites
Correlation
Bias (ug/m3)
Slope
Including Europe
244
0.83
0.86
1.15
Excluding Europe
84
0.83
0.91
-2.5
Better than in situ vs model (GEOS-Chem): r=0.52-0.62, slope = 0.63 – 0.71
van Donkelaar et al., EHP, 2010
van Donkelaar et al., EHP, 2010
Long-term Exposure to
Outdoor Ambient PM2.5
•
80% of global population
exceeds WHO guideline of
10 μg/m3
90
35% of East Asia exposed to
>50 μg/m3 in annual mean
70
Global mortality from PM2.5
2-8 million deaths/year (Evans
et al., EHP, submitted)
Used in WHO Global Burden
of Disease assessment
Significant association of
PM2.5 and health at low PM2.5
levels (Crouse et al., EHP, in
prep)
IT-2 IT-1
80
60
50
40
30
20
10
0
van Donkelaar et al., EHP, 2010
AQG IT-3
100
Population [%]
•
WHO Guideline & Interim Targets
5
10
15
25 35
PM2.5 Exposure
50
[μg/m3]
100
USA Today: Hundreds Dead from Heat, Smog,
Wildfires 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
Relaxed Cloud Screening Needed for Extreme Events
van Donkelaar et al., submitted
Application of Satellite-based Estimates to Moscow
Smoke Event
During Fires
Before Fires
MODIS-based
In Situ from PM10
In Situ PM2.5
van Donkelaar et al., submitted
General Approach to Estimate Surface NO2 Concentration
NO2 Column
Model Profile
In Situ
GEOS-Chem
 SM 
SO  O 


 M
S → Surface Concentration
Ω → Tropospheric column
Ground-Level NO2 Inferred From OMI for 2005
Works in Near-Real-Time!
Values Estimated Using Monthly NO2 Profiles for Different Year (2006)
Temporal Correlation with In Situ Over 2005
×In situ
—— OMI
Insignificant change in results if profiles are daily coincident
values from 2005
Lok Lamsal
Ground-Level NO2 Inferred From OMI for 2005
Spatial Correlation vs In Situ for North America = 0.78
Lok Lamsal
Bottom-Up Emission Inventories Take Years to Compile
Bottom-up Anthropogenic NOx Emission Inventory from
Land Sources for 2006
Based on EDGAR (2000), CAC (2005), NEI2005, BRAVO (1999), EMEP (2006),
Zhang (2006), scaled to 2006
Changes in Tropospheric NO2 Column Reflect
Changes in NOx Emissions
Trend in Tropospheric NO2 Column over 1996-2002 from GOME
1996 - 2002
Richter et al., 2005
Application of Satellite Observations for Timely Updates
to NOx Emission Inventories
Use GEOS-Chem to Calculate Local Sensitivity of Changes in Trace
Gas Column to Changes in Emissions
Fractional Change
in Emissions
E  
Fractional Change in
Trace Gas Column
Local sensitivity of column changes
to emissions changes
Insensitive to changes
in anthropogenic
CO and VOCs
Walker et al., ACP, 2010
Lamsal et al., GRL, 2011
Evaluate Hindcast Inventory Versus Bottom-up
Hindcast for 2003 Based on Bottom-up for 2006 and Monthly
NO2 for 2003-2006
Bottom-up
Hindcast
Lamsal et al., GRL, 2011
Forecast Inventory for 2009 Based on Bottom-up for 2006
and Monthly OMI NO2 for 2006-2009
Temporary Dataset Until Bottom-Up Inventory Available
9% increase in
global emissions
19% increase in
Asian emissions
6% decrease in
North American
emissions
Lamsal et al., GRL, 2011
Emerging Applications of Satellite Remote Sensing
of Atmospheric Composition
Chemical Transport Model Plays a Valuable Role in
Relating Retrieved and Desired Quantity
• Ground-level Estimates of PM2.5 & NO2
• Simple Method for Timely Updates to NOx Emission Inventories
Challenge
• Continue to develop retrieval capability
•Evaluate and improve simulation to relate retrieved and desired quantity
(includes AOD/PM2.5, NO2 / NOx emissions)
Acknowledgements:
NSERC, Environment Canada, Health Canada, NASA