Transcript PPT - Atmospheric Chemistry Modeling Group

US aerosols:
observation from space, interactions with climate
Daniel J. Jacob
with Easan E. Drury, Loretta J. Mickley,
Eric M. Leibensperger, Amos Tai
and funding from NASA, EPRI, EPA
IMPORTANCE OF ATMOSPHERIC AEROSOLS
Public health
Chemistry
Visibility
Cloud formation
Ocean fertilization
Climate forcing
number
AEROSOL CHARACTERISTICS
Typical size distribution
(Seinfeld and Pandis, 1998)
area
volume
Chemical composition of PM2.5
(NARSTO, 2004)
sulfate (coal combustion)
nitrate (fossil fuel combustion)
ammonium (agriculture)
black carbon (combustion)
organic carbon (combustion,
vegetation)
soil
other
PM2.5 (EPA std.)
SATELLITE OBSERVATIONS OF TROPOSPHERIC COMPOSITION:
a revolution over the past decade
The NASA “A-Train”
Integrated observing system
Satellites
Models
Surface sites
Principal tropospheric species measured from space:
• Ozone , NO2, formaldehyde, BrO, glyoxal
• CO, CO2, methane
• Aerosols, SO2
aircraft, ships,
sondes, lidars
HOW TO OBSERVE AEROSOLS FROM SPACE?
Solar occultation
(SAGE, POAM…)
Active system
(CALIPSO…)
Solar back-scatter
(MODIS, MISR…)
laser
pulse
EARTH
Surface
Pros: high S/N, vertical profiling
Cons: sparse sampling,
cloud interference,
low horizontal resolution
Pro: vertical profiling
Con: sparse sampling,
low S/N
Surface
Pro: horiz. resolution
Con: daytime only,
no vertical resolution
Aerosol observation from space by solar backscatter
Relatively easy to do qualitatively for thick plumes over ocean…
California fire plumes
Pollution off U.S. east coast
Dust off West Africa
…but difficult quantitatively! Fundamental quantity is aerosol optical depth (AOD)
Il ()
aerosol
scattering,
absorption
Il (0)=Il ()exp[-AOD]
Measured top-of-atmosphere reflectance
= f (AOD, aerosol properties,
surface reflectance, air scattering,
gas absorption, Sun-satellite geometry)
Aerosol optical depths (AODs) measured from space
Jan 2001 – Oct 2002 operational data
MODIS (c004)
return time 2x/day;
nadir view
known positive bias
over land
550 nm
AODs
MISR
9-day return time;
multi-angle view
better but much sparser
van Donkelaar et al. [2006]
MODIS AEROSOL RETRIEVAL OVER LAND
TOA
reflectance
0.47 mm
0.65 mm
2.13 mm
SURFACE
Operational retrieval:
 Use top-of-atmosphere (TOA) reflectance
at 2.13 mm (transparent atmosphere) to derive
surface reflectance
 Assume fixed 0.47/2.13 and 0.65/2.13 surface
reflectance ratios to derive atmospheric
reflectances at 0.47 and 0.65 mm by
subtraction
 Assume generic aerosol optical properties
to convert atmospheric reflectance to AOD
Our improved retrieval:
 Derive local values of 0.47/2.13 and 0.65/2.13
surface reflectance ratios from statistics of
low-aerosol scenes
 Use local aerosol column information from
the GEOS-Chem chemical transport model to
convert atmospheric reflectance to AOD
Drury et al. (JGR in press)
GEOS-Chem CHEMICAL TRANSPORT MODEL (geos-chem.org)
• Global model of atmospheric
composition driven by NASA/GEOS
assimilated meteorological data with
0.5ox0.625o (~50 km) resolution
• Simulates coupled oxidant-aerosol
chemistry for
• sulfate-nitrate ammonium
• organic aerosol
• black carbon aerosol
• dust (4 size classes)
• sea salt (2 size classes)
on 2o x2.5o grid
• Size distributions and optical properties
for different aerosol types are specified
TESTING THE MODIS AEROSOL RETRIEVAL
USING ICARTT AIRCRAFT DATA OVER US (Jul-Aug 2004)
fit AODs
synthetic TOA
reflectance = f(AOD,…)
MODIS satellite instrument:
TOA reflectance
NASA, NOAA, DOE aircraft:
speciated mass concentrations,
microphysical & optical properties
GEOS-Chem
model
evaluate
NASA
DC-8
EPA AQS/IMPROVE surface networks: mass concentrations
NASA AERONET surface network: AODs
MODIS local
surface
reflectance
and ratio
EASTERN U.S.
Drury et al. [JGR in press]
ORGANIC AEROSOL IN ICARTT
Water-soluble organic carbon (WSOC) measured on NOAA P-3
IMPROVE measurements of organic carbon
Fu et al.
(AE, 2009)
Two mechanisms for formation of secondary organic aerosol (SOA):
• Standard reversible SOA (Pankow/Seinfeld):
K (T , aerosol )
oxidation

 SOA
VOC 
 secondary organic gas (SOG) 

• Dicarbonyl SOA (Liggio/Fu):
oxidation
cloud uptake
VOC 

glyoxal,
methylglyoxal

SOA
multi-step
oxidation, oligomerization
AEROSOL OPTICAL PROPERTIES IN ICARTT
Single-scattering albedo = fraction of
aerosol extinction due to scattering
standard model
assumption (GADs)
improved fit
(this work)
AERONET
Drury et al., JGR in press
AEOSOL OPTICAL DEPTHS (0.47 mm), JUL-AUG 2004
bias=+2%
r = 0.84
c005 is current MODIS operational data;
statistics are relative to AERONET data
in circles
bias=-21%
r = 0.82
bias=-15%
r = 0.87
Beyond improving on the operational
products, our MODIS retrieval enables
quantitative comparison to model
results (consistent aerosol optical
properties); indicates model
underestimate in Southeast US, likely
due to organic aerosol
Drury et al., JGR in press
Can we use AODs measured from space as proxy for PM2.5?
 PM 2.5 
Infer PM2.5 from AOD by PM 2.5 = AOD*
 AOD 

 GEOS-Chem
EPA AQS surface network data
MODIS PM2.5 (this work)
MODIS captures general observed patterns in PM2.5
but is 50% higher than observed. Could reflect
• Clear-sky bias
• Time-of-day bias
• Model error in vertical aerosol distribution
Drury et al., JGR in press
RADIATIVE FORCING OF CLIMATE BY AEROSOLS
Anthropogenic aerosols may
have offset more than half of
global greenhouse warming
from 1750 to present
IPCC (2007)
But this aerosol radiative forcing is
very inhomogeneous: what are the
regional climate consequences?
Aerosol direct radiative forcing
Leibensperger et al., in prep.
CLIMATE IMPLICATIONS OF US AIR QUALITY POLICY
Radiative forcing of US
anthropogenic aerosols is small
globally but important regionally
Leibensperger et al., in prep.
US sulfur emissions are decreasing
rapidly: what are the regional climate
impacts?
today
CALCULATING THE CLIMATE RESPONSE
FROM SHUTTING DOWN U.S. AEROSOL
Use NASA/GISS general circulation model (GCM)
GISS GCM
Consider two scenarios:
Control: aerosol optical depths fixed at 1990s levels.
Sensitivity: U.S. aerosol optical depths set to zero
(radiative forcing of about +2 W m-2 over US)
Conduct ensemble of 3 simulations for each scenario.
Mickley et al. (AE, submitted)
Removing aerosols over US
causes 0.5-1o C annual mean warming in the East.
Mickley et al.
[AE, submitted]
Temperature (oC)
2010-2050 warming due to
greenhouse gases
Additional warming due to
zeroing of aerosols over the US.
No-US-aerosols case
Control, with US aerosols
Additional effects include increased summer heatwaves (1-2 o C warming)
and increased precipitation in the East
EFFECT OF CLIMATE CHANGE ON SURFACE AIR QUALITY
Expected effect
of 21st century
climate change
Ozone
PM (aerosol)
Stagnation
?
Temperature
?
?
?
?
Mixing depth
Precipitation
=
=
Cloud cover
Relative humidity
=
Jacob and Winner, AE 2009
EFFECT OF FUTURE CLIMATE CHANGE ON US AIR QUALITY
Models show consistent increase of ozone, mainly driven by temperature
Results from six coupled GCM-CTM simulations
2000-2050 change of 8-h
daily max ozone in summer,
MDA8
ppb
keeping anthropogenic emissions constant
5
4
3
2
1
0
-1
-2
-3
-4
-5
-6
-7
-8
-9
Weaver et al.
[BAMS, 2010]
Northeast
NE
Midwest
California
MW
CA
Harvard.A1B
CMU.A2
PGR.B1
NERL.A1B
Texas
TX
WSU.A2
Southeast
SE
PGR.A1Fi
…but model results for aerosols show no such consistency, including in sign.
How can we progress?
OBSERVED AEROSOL CORRELATION
WITH METEOROLOGICAL VARIABLES
Multilinear regression model fit to
1998-2008 deseasonalized EPA/AQS data
for PM2.5 (total and speciated)
9
yi =  0,i    k ,i xk ,i  interactio n terms
k =1
mostly precipitation
Tai et al.
[AE, submitted]
mostly temperature and stagnation
R2 fit
PM2.5 CORRELATION WITH METEOROLOGICAL VARIABLES
Tai et al. [AE , submitted]
TEMPERATURE COEFFICIENTS FOR SPECIATED PM2.5
Tai et al. [AE , submitted]
IMPORTANCE OF MID-LATITUDES CYCLONES
IN AIR POLLUTION METEOROLOGY
Cold fronts from mid-latitude cyclones are the principal ventilation process
for U.S. Midwest/Northeast, western Europe, China
Clean air sweeps
behind cold front
GCMs show decrease + N shift of cyclones from 21st-century climate change;
already seen in 1950-2000 climatological data
CORRELATIONS AND TRENDS
OF POLLUTION EPISODES AND CYCLONES IN NORTHEAST U.S.
# pollution episode days (O3>80 ppb) and # cyclones tracking across SE Canada
in summer 1980-2006 observations
# cyclones
# episodes
• Strong correlation; cyclone frequency is predictor of pollution episode frequency
• 1980-2006 decrease in cyclone frequency would imply a corresponding
degradation of air quality if emissions had remained constant
• Expected # of > 80 ppb days in Northeast dropped from 30 in 1980 to 10 in 2006,
but would have dropped to zero by 2001 in absence of cyclone trend!
Leibensperger et al. [ACP2008]
EFFECT OF INCREASED STAGNATION ON PM2.5
Difference in PM2.5 between stagnant days (wind < 8 m s-1, 500 hPa wind <13 m s-1,
no precipitation) and non-stagnant days, 1998-2008 data
PM2.5 is expected to be highly sensitive to Increasing stagnation in future climate
Tai et al. [AE , submitted]