U.S. aerosols: U.S. aerosols: observation from space, effects on climate observation from space, effects on climate

Daniel J. Jacob

and funding from NASA, EPRI

with Easan E. Drury, Tzung-May FuLoretta J. Mickley, and Eric M. Leibensperger

ATMOSPHERIC AEROSOLS:ATMOSPHERIC AEROSOLS:ensembles of condensed-phase particles suspended in airensembles of condensed-phase particles suspended in air

Typical aerosol size distribution

number

area

volume

Aerosols are the visible part of the atmosphere:

Pollution off U.S. east coast Dust off West AfricaCalifornia fire plumes

WHY CARE ABOUT ATMOSPHERIC AEROSOLS?WHY CARE ABOUT ATMOSPHERIC AEROSOLS?

Public health

Visibility

Ocean fertilization

Chemistry

Climate forcing

Cloud formation

MAJOR AEROSOL COMPONENTS IN U.S. MAJOR AEROSOL COMPONENTS IN U.S.

Dry mass concentrations

Sulfate: from atmospheric oxidation of SO2 emitted by combustion (mainly coal)Nitrate: from atmospheric oxidation of NOx emitted by combustionAmmonium: from NH3 emitted by agricultureCarbon: elemental carbon from combustion and organic carbon from combustion and vegetationCrustal: suspended mineral dust

Urban air concentrations of particulate matter <2.5 m diameter (PM2.5)

GLOBAL AEROSOL OBSERVATION FROM SPACEGLOBAL AEROSOL OBSERVATION FROM SPACEAerosol optical depths (AODs) at 0.55 m from MODIS and MISR sensors

MODIS

return time 2x/day

MISR

9-day return time

Why are the AODs so different?

van Donkelaar et al. [2006]

Jan 01 – Oct 02

MODIS RETRIEVAL MODIS RETRIEVAL OF AEROSOL OPTICAL DEPTHS (AODs) OVER LANDOF AEROSOL OPTICAL DEPTHS (AODs) OVER LAND

SURFACE

AEROSOL

0.47 m0.65 m2.13 m

• Interpretation of this top-of-atmosphere (TOA) reflectance in terms of AOD requires assumptions on surface reflectance, aerosol optical properties

• Use TOA reflectance at 2.13 m (transparent atmosphere) to derive surface reflectance

• Assume 0.47/2.13 and 0.65/2.13 surface reflectance ratios to obtain atmospheric reflectances at 0.47 and 0.65 m by subtraction

• Assume aerosol optical properties to convert atmospheric reflectance to AOD

• MISR does along-track multi-angle viewing of same aerosol column – better constraints but sparser data

MODIS measures backscatter solar reflectance in several wavelength channels

CONSTRAINING AND TESTING AEROSOL OBSERVATIONS CONSTRAINING AND TESTING AEROSOL OBSERVATIONS FROM SPACE DURING ICARTT CAMPAIGN (Jul-Aug 2004)FROM SPACE DURING ICARTT CAMPAIGN (Jul-Aug 2004)

EASTERN U.S.

IMPROVE surface network: speciated mass concentrations at background sitesAERONET surface network: aerosol optical depths

NASA, NOAA, DOE aircraft: speciated mass concentrations,microphysical & optical properties

MODIS satellite instrument:aerosol optical depths

NASADC-8

IMPROVING THE SURFACE REFLECTANCE CORRECTION IMPROVING THE SURFACE REFLECTANCE CORRECTION FOR MODIS AEROSOL RETRIEVALSFOR MODIS AEROSOL RETRIEVALS

Measuredtop-of-atmosphere (TOA)reflectances(ICARTT period)

2.13 m 0.65 m

Measured 0.65 vs. 2.13TOA reflectances: take lower envelope for given location to derive surface reflectance ratio

Derive aerosol reflectanceat 0.65 m(same procedure for 0.47 m)

Drury et al. [JGR 2008]

Fresno, CAICARTT period

0.65/2.13 surface reflectance ratio

CONVERTING TOA AEROSOL REFLECTANCES TO AODsCONVERTING TOA AEROSOL REFLECTANCES TO AODs

• Use GEOS-Chem model driven by NASA/GEOS assimilated meteorological data with 2ox2.5o resolution

• Model simulates mass concentrations of different aerosol types

• Size distributions and optical properties for different aerosol types are assumed (test with ICARTT data)

• Key advantage of approach is to allow quantitative test of model with the satellite aerosol reflectance data

Standard MODIS algorithm assumes generic aerosol optical properties

Better way is to use local info for given scene from a global 3-D aerosol model

PREVIOUS MODEL EVALUATION: sulfate-nitrate-ammoniumPREVIOUS MODEL EVALUATION: sulfate-nitrate-ammoniumAnnual mean concentrations at IMPROVE sites (2001) – CASTNET for NH4

+

• Sulfate is 100% in aerosol;• Ammonia NH3(g) neutralizes sulfate to form (NH4)2SO4;• Excess NH3(g) if present can combine with HNO3(g) to form NH4NO3 as function of T, RH

Park et al. [AE 2006]

r = 0.96 bias = +10% r = 0.60 bias = +30% r =0.94 bias = +10%

PREVIOUS MODEL EVALUATION: carbonaceous aerosolPREVIOUS MODEL EVALUATION: carbonaceous aerosol

• Primary sources: fossil fuel, biofuel, wildfires• Also large growing-season biogenic source of secondary organic aerosol (SOA)

Elemental carbon (EC) Organic carbon (OC)

volatile organiccompounds (VOCs)

oxidation, multi-stepSOA

Park et al. [AE 2006]

Annual mean concentrations at IMPROVE sites (2001)

r = 0.75 bias = -15% r = 0.70 bias = +20%

PREVIOUS MODEL EVALUATION: mineral dustPREVIOUS MODEL EVALUATION: mineral dust

GEOS-Chem

Local

Asian dust

Saharandust

Fairlie et al. [AE 2007]

Annual mean concentrations at IMPROVE sites (2001)

AEROSOL VERTICAL PROFILES IN ICARTTAEROSOL VERTICAL PROFILES IN ICARTT

NASA DC-8

IMPROVE (<2.5 m)

bulk filter (Dibb, UNH)

PILS (Weber, GIT)

• Sulfate model overestimate: excessive cloud processing?• Unresolved disagreement in ammonium and dust observations

Easan Dury, in prep.

ORGANIC AEROSOL IN ICARTTORGANIC AEROSOL IN ICARTT

PILS water-soluble organic carbon (WSOC) on NOAA P-3 IMPROVE measurements of organic carbon

• Standard reversible SOA (Pankow/Seinfeld):( , )oxidationVOC secondary organic gas (SOG) SOA

K T aerosol • Dicarbonyl SOA (Liggio/Fu):

oxidation cloud uptakemulti-step oxidation, oligomerization

VOC glyoxal, methylglyoxal SOA

Fu et al.(AE, in press)

MEAN AEROSOL VERTICAL PROFILES IN ICARTTMEAN AEROSOL VERTICAL PROFILES IN ICARTT

Obs erved

0 3 6 9 12 15 18

1

2

3

4

5

6

7

8

9

10

11

12

Hei

ght (

km)

Aeros ol Mas s (μg m-3 S T P )

B lack C arbon

Dus t

Ammonium

Nitrate

S ulfate

Organic C arbon

Modeled

0 3 6 9 12 15 18

1

2

3

4

5

6

7

8

9

10

11

12

Hei

ght (

km)

Aeros ol Mas s (μg m-3 S T P )

• Bulk of mass is in boundary layer below 3 km: sulfate, organic (dust?)• Dust, organic dominate above 3 km

Easan Drury, in prep.

AEROSOL OPTICAL PROPERTIES IN ICARTTAEROSOL OPTICAL PROPERTIES IN ICARTT

Single-scattering albedo is fraction ofaerosol extinction due to scattering

AERONET

standard modelAssumption (GADs)

improved fit(this work)

Easan Drury, in prep.

MEAN AEROSOL OPTICAL DEPTHS DURING ICARTTMEAN AEROSOL OPTICAL DEPTHS DURING ICARTT

Model results compared to observations from AERONET network (circles)

Model w/ GADs size distributions Model w/improved size distributions

Easan Drury, in prep.

r = 0.89 bias = -21% r = 0.89 bias = -7%

Main improvement was to reduce the geometric standard deviation in the log-normal size distributions for sulfate and OC from 2.0 to 1.6

IMPROVED MODIS IMPROVED MODIS RETRIEVALRETRIEVAL

OF AEROSOL OPTICAL OF AEROSOL OPTICAL DEPTH DEPTH

This work

standard MODISproduct (c005)

Easan Drury, in prep.

Circles are AERONET data

standard MODISproduct (c004)

c005 is the latest operationalMODIS AOD product (2006)

r=0.84bias =+2%

r=0.84bias =-20%

MAPPING SURFACE PMMAPPING SURFACE PM2.5 2.5 FROM IMPROVED MODIS AODsFROM IMPROVED MODIS AODs

• Shows model organic aerosol underestimate in Southeast (w/out dicarbonyl SOA) • Questions sulfate problem in Northeast

Easan Drury, in prep.

2.52.5

GEOS-Chem surface PMPM = (MODIS AOD)

GEOS-Chem AOD

GLOBAL RADIATIVE FORCING OF CLIMATE BY AEROSOLSGLOBAL RADIATIVE FORCING OF CLIMATE BY AEROSOLS

IPCC [2007]

Large historical offset of greenhouse warming by anthropogenic aerosols

Unlike, CO2, radiative forcing from aerosols is strongly regional and likely to decrease in future: what are the implications for future climate change?

Historical and projected U.S. trendof SO2 emissions

CLIMATE RESPONSE TO SHUTTING DOWN U.S. AEROSOL CLIMATE RESPONSE TO SHUTTING DOWN U.S. AEROSOL

Mickley et al. (in prep.)

CHANGE IN ANNUAL MEAN TEMPERATURECHANGE IN ANNUAL MEAN TEMPERATURE

Mickley et al. (in prep.)

THIS REGIONAL CLIMATE RESPONSE FROM U.S. AEROSOL THIS REGIONAL CLIMATE RESPONSE FROM U.S. AEROSOL VANISHES AFTER A FEW DECADESVANISHES AFTER A FEW DECADES

• Explains why previous studies (focusing on 2050 or 2100 endpoints) have found no regional climate response to aerosol emissions• May reflect regional feedbacks important in present atmosphere but already realized in future enhanced-greenhouse atmosphere

Mickley et al. (in prep.)