Emerging issues in air quality

Daniel J. Jacob

with Lin Zhang, Raluca Ellis, Fabien Paulot, Eloise Marais, Qiaoqiao Wang, Kevin Wecht, Alex Turner, Helen Amos, Hannah Horowitz

and Anne Perring, Joshua Schwartz, David Fahey (NOAA), Cui Ge, Jun Wang (U. Nebraska), David Streets (Argonne)

Support from BP, NASA, NSF

4th-highest annual maximum of daily 8-h average ozone,2008-2010

Current standard: 75 ppbProposed standard: 60-70 ppb

Long-term surface ozone trendTrend in 95th percentile, June-August 1990-2010

Cooper et al. [2012]

• Decrease in eastern US driven by NOx emission controls;• Increase or flat in Intermountain West

4th-highest annual maximum for daily 8-h average ozone,2008-2010

Intermountain West: The next ozone frontier!

High elevation, arid terrain → high ozone background

Current standard: 75 ppbProposed standard: 60-70 ppb

High background and stratospheric intrusionsin Intermountain West

• Most surface ozone in Intermountain West originates from outside N America

• Highest ozone events (>80 ppb) due to stratospheric filaments, cannot be reproduced in models (stretched-flow numerical diffusion)

• Have to be careful with definition of stratospheric influence

2007

Zhang et al., in prep.

S

T

O3 producedin stratosphere

Two definitions ofstratospheric influence on tropospheric ozone:

O3 transportedfrom stratosphere

tropopause

Ozone at Gothic, Colorado (2,900m)

ObservedGEOS-Chem N Am background

Strat (transported)

Strat (produced)

Wildfire plumes alone do not drive high-ozone events

• No evidence of high-ozone events from fire plumes unless mixed with pollution

• Models generate excessive ozone because they don’t account for emissions of highly reactive VOCs that lock up NOx as PAN in the plume – a difficult problem!

• Fires can still contribute to background ozone through PAN decomposition

Glacier NP, 2007: model has high O3 in fire plumes, observations show none

Organic aerosol (OC) correlates with western US fires but not ozone

OC (IMPROVE)Ozone (CASTNet)

2006-2008

Zhang et al., in prep.

Observed GEOS-Chem wildfires

N deposition at US national parks: critical load exceedances

Ellis et al. [2013]

More deposition is expected to originate from ammonia in future

Critical loads are 3-5 kg N ha-1 a-1

depending on ecosystem

2006 2050

NOx

NH3

Present and future (RCP)US emissions

Future exceedances driven byammonia emissions

Ellis et al. [submitted]

2006

2050RCP2.6

Optimizing NH3 emissions by adjoint inversion of 2005-2008 NH4

+ wet deposition flux dataNADP data (circles) and GEOS-Chem model after adjoint inversion

April:fertilizer

July:livestock

kgN ha-1 month-1

Error correlation between NH4+ wet deposition flux (F) and precipitation (P)

obtained by GEOS-Chem simulations with GEOS-4 vs. GEOS-5 meteorology:

PGEOS5/PGEOS4

FG

EO

S5/

FG

EO

S4 0.6 power dependence

Paulot et al. [submitted]

Optimized ammonia emissions …and new MASAGE bottom-up ammonia emission inventory

US EU E Asia x 0.5

crops

livestock

other anthronatural

2.7 2.9 8.4

2.8 3.1 8.4 (China)

Paulot et al. [submitted]

N deposition from agriculture is a global pollution problem

Annual mean agricultural ammonia emissions from MASAGE (2005-2008)

63% are from countries outside the US, European Union, and China

Paulot et al. [submitted]

Contribution of US food export to PM (NH4NO3) air pollution

MASAGE NH3 emissions from food export

wheat

PM due to food export (GEOS-Chem)

annual

Economic implications by state ($ per capita):

Paulot et al., in prep

beef

corn

Next frontier for air pollution: NigeriaOMI formaldehyde2005-2009

MISR SCIA

aerosol (AOD) NO2 HCHO glyoxal methane

• Population: 270 million (+2.6% a-1)• GDP: $273 billion (+7% a-1) – oil!• Most natural gas is flared• >80% of domestic energy from biofuel, waste

LagosPortHarcourt

An unusual mix of very high VOCs, low NOx –What will happen as infrastructure develops?

Marais et al., in prep.

gasflaring!

1015 molecules cm-2

Multimodel intercomparison and comparison to

observations

Multimodel intercomparisons and comparisons to observations

Koch et al. [2009], Schwarz et al. [2010]

ARCTAS (Arctic spring)

BC, ng kg-1 BC, ng kg-1

TC4 (Costa Rica, summer)

ObservedModels

• Models differ by order of magnitude between themselves and with observations

• Large overestimates of observations over oceans,

upper troposphere

• Discrepancy must be driven by model errors in scavenging

Large model errors for black carbon (BC) aerosol in remote airP

ress

ure

, h

Pa

obsmodels60-80N

obsmodels20S-20N

Pre

ssu

re,

hP

a

HIPPO over Pacific (Jan)

BC, ng kg-1 BC, ng kg-1

Ensemble of AeroCom models

Global BC simulation in GEOS-ChemSource (2009): 4.9 Tg a-1 fuel + 1.6 Tg a-1 open fires Lifetime: 4.2 days

NMB= -27%

NMB= -12%

NMB= 6.6%

Observations (circles) and model (background)

surfacenetworks

AERONETBC AAOD

NMB= -32%

Aircraft profiles in continental/outflow regionsHIPPO(US)

Arctic(ARCTAS)

Asian outflow(A-FORCR)

US(HIPPO)

observedmodel

Successful simulation in source regions and outflow

Wan

g e

t al

., i

n p

rep

HIPPO BC curtains across the Central Pacific, 2009-2011

• Minima in deep tropics• Model doesn’t capture

low tail, is also too high at N mid-latitudes; median bias is factor of 2, mean column bias is +48%

Wang et al., in prep

Observations by Perring et al. (in prep.)

Observed Model PDF

PD

F, (

mg

m-3 S

TP

)-1

BC top-of-atmosphere direct radiative forcing (DRF)

EmissionTg C a-1

Global load(mg m-2)[% above 5 km]

BC AAODx100

Forcing efficiency(W m-2/AAOD)

Direct radiative forcing (W m-2)

This work 6.5 0.15 [8.7%] 0.17 88 0.15 (0.13-0.26)

AeroCom [2006]

7.8 ±0.4 0.28 ± 0.08[21±11%]

0.22±0.10 168 ± 53 0.34 ± 0.07

Chung et al. [2012]

0.77 84 0.65

Bond et al. [2013]

17 0.55 0.60 147 0.88

• In our work, BC above 5 km contributes 30% of global DRF and BC over the oceans contributes 24%; these contributions would be higher in other models with less efficient scavenging.

• We find that BC radiative forcing is much less than previously estimated; need to better understand BC in free/remote troposphere!

Wang et al., in prep

Constraints on US methane emissions from SCIAMACHY data

1700 1800ppb

SCIA CH4 column mixing ratio, Jul-Aug 2004

• Adjoint inversion with EDGAR v4.2 (anthropogenic), Kaplan (wetlands) as priors• Focus on INTEX-A mission period to validate SCIAMACHY data and inversion

0.5 1.51.0

Adjoint inversion scaling factors

Wecht et al. [in prep]Livestock Oil & Gas Waste Coal Mining Other

0

5

10

15Total US anthropogenic emissions (Tg a-1) EDGAR v4.2 26.6

EPA 28.3

This work 32.7

Methane from GOSAT: preliminary adjoint inversion

GOSAT data for CalNex period (May-Jul 10) Scaling factors to EDGAR inventory

• GOSAT data show consistency with SCIAMACHY for constraining livestock and wetland sources, but also discrepancies

• The data are sparse; now applying a Gaussian Mixture Model to optimally reduce the state vector for the inversion

Turner et al. [in progress]

Can we monitor from space the evolving source from oil & gas?

Biogeochemical cycle of mercury

Hg(0) Hg(II)

particulate

Hg

burial

SEDIMENTS

uplift

volcanoeserosion

oxidation

Hg(0) Hg(II)reduction biological

uptake

ANTHROPOGENIC PERTURBATION:fuel combustion

mining

ATMOSPHERE

OCEAN/SOIL

VOLATILE WATER-SOLUBLE

(months)lifetime~6 months

History of global anthropogenic Hg emissions

Large past (legacy) contribution from N. American and European emissions; Asian dominance is a recent phenomenon

Streets et al. , 2011

Global source contributions to Hg in present-day surface ocean

• Human activity has increased 7x the Hg content of the surface ocean

• Half of this human influence is from

pre-1950 emissions

• N America, Europe and Asia share similar responsibilities for anthropogenic Hg in present-day surface ocean

Amos et al., in press

EuropeAsia

N America

S America

former USSR

ROW

pre-1850natural

emissions

from biogeochemical box model constrained with GEOS-Chem fluxes

Disposal of Hg in commercial products:a missing component of the Hg biogeochemical cycle?

Global production of commercial Hg peaked in 1970

Horowitz et al., in prep

• Commercial Hg enters environment upon use or disposal; much larger source than inadvertent emission

• Could explain observed atmospheric decrease of Hg(0) over past two decades

Environmental release from commercial products dwarfs current emission estimates

TEMPO geostationary UV/Vis satellite instrumentselected in November 2012 for 2018/2019 launch

PI: Kelly Chance, Harvard-Smithsonian

• Monitoring of tropospheric ozone (2 levels), aerosols, NO2, SO2, formaldehyde, glyoxal with 1-hour temporal resolution, 4-km spatial resoution

• To be part of a geostationary constellation with other sensors observing Europe and East Asia

TEMPO Sentinel-4 GEMS

Next frontier in satellite observationsof atmospheric composition!