NATURAL AND TRANSBOUNDARY POLLUTION INFLUENCES ON REGIONAL VISIBILITY STATISTICS IN THE UNITED STATES

1
NATURAL AND TRANSBOUNDARY POLLUTION INFLUENCES ON
REGIONAL VISIBILITY STATISTICS IN THE UNITED
STATES
Rokjin Park
with support from EPRI NASA Dalhousie
University May 19 2006
2
NATIONAL PARKS AND OTHER NATURAL AREAS IN THE
U.S. SUFFER SIGNIFICANT VISIBILITY DEGRADATION
FROM ANTHROPOGENIC AEROSOLS.
7.6 µgm-3
12.0 µgm-3
Glacier National Park
21.7 µgm-3
65.3 µgm-3
3
ATMOSPHERIC PARTICULATE MATTER (AEROSOLS)
Aerosol dispersed condensed matter suspended in
a gas Size range 0.001 mm (molecular cluster) to
100 mm (small raindrop)
Lifetime 4 6 days
Soil dust Sea salt
SO2 NOx NH3 VOCs

• Most important components of the atmospheric
  aerosol
• Sulfate- nitrate-ammonium
• Organic carbon (OC) elemental carbon (EC)
• Soil dust
• Sea salt

4
VISIBILITY METRIC

• VISUAL RANGE (km) - THE GREATEST DISTANCE AT
  WHICH AN OBSERVER
• CAN SEE A BLACK OBJECT VIEWED AGAINST
  BACKGROUND HORIZON
• - A quantitative measurement is subject to
  other conditions (Sun angle
• light condition) than aerosol
  concentrations.
• EXTINCTION (bext Mm-1) - THE AMOUNT OF LIGHT
  LOST AS IT TRAVELS OVER
• A MILLION METERS
• - Most useful for relating visibility directly
  to aerosol concentrations.
• - bext 3f(RH)(NH4)2SO4 NH4NO3 4OMC
  10EC SOIL 0.6CM 10
• DECIVIEWS (dv) THE LOGARITHM OF THE EXTINCTION
  
• - dv 10ln(bext/10)
• - A change in one dv is perceived to be the
  same under different conditions
• (clear and cloudy days).

Pitchford and Malm 1994
5
U.S. EPA REGIONAL HAZE RULE

• Federal class I areas (including national parks
  other wilderness areas) to return to natural
  visibility conditions by 2064
• State Implementation Plans to be submitted by
  2007 for linear improvement in visibility over
  the 2004-2018 period

visibility (deciviews)
from EPA 2001
Because visibility is a logarithmic (sluggish)
function of PM concentration The 2004-2018 phase
I implementation requires 50 reduction in
emissions highly sensitive to specification of
2064 endpoint
6
U.S. EPA HAS PROPOSEDDEFAULT ESTIMATED NATURAL
PM CONCENTRATIONS FOR APPLICATION OF THE
REGIONAL HAZE RULE
OBJECTIVE 1
These defaults are based on measurements at clean
remote sites NAPAP report 1990. A better
quantification of natural aerosol concentrations
is crucial.
7
TRANSBOUNDARY TRANSPORT COMPLICATES THE
DEFINITION OF NATURAL VISIBILITY
OBJECTIVE 2
Glen Canyon AZ
Dust storms provide visible evidence of
intercontinental transport of aerosols and
anthropogenic pollution is transported together
with the dust
satellite data
satellite data
Heald et al. 2005
8
GEOS-Chem GLOBAL 3-D MODELOF ATMOSPHERIC
TRANSPORT AND CHEMISTRY
http//www-as.harvard.edu/chemistry/trop/geos

• Developed by Harvard Atmospheric Chemistry
  Modeling Group used by 17 research groups in N.
  America and Europe 100 publications.

• driven by GEOS assimilated meteorological
  observations from NASA Global Modeling and
  Assimilation Office (GMAO) native resolution
  1ox1o
• applied to simulations of ozone aerosols (PM)
  CO2 methane mercury hydrogen
• Horizontal resolution 1ox1o to 4ox5o
  (user-selected) 48 levels in vertical

• Previous global evaluation of aerosol
  simulations in the United States Europe and
  East Asia by Park et al. 2003 2004a 2004b
  global evaluation by Martin et al. 2004.
• Conduct 1ox1o nested model simulations over
  North America with boundary condition from 4ox5o
  global model simulation.

9
SULFATE-NITRATE-AMMONIUM AEROSOL
Aerosol thermodynamic calculations using RPMARES
or ISORROPIA
Solid
Aqueous phase
f(T RH C)
pH 4.5
H2SO4
SO42-
NH4NO3
(NH4)2SO4
HNO3
H2O2 (aq) O3 (aq)
Lightning
OH
OH
(N2O5)
NOx
NH3
SO2
OH NO3
DMS
Ocean
Volcanoes Fossil fuel Domesticated Fertilizers
Fossil Fuel Biomass
Animals
burning
10
ORGANIC CARBON AEROSOL
Condense on pre-existing aerosol
SECONDARY ORGANIC AEROSOL (SOA) SIMULATION Chung
and Seinfeld 2002 VOCi OXIDANTj aijP1ij
aijP2ij Parameters (as Ks) from smog
chamber studies
Reactive Organic Gases
Oxidation by OH O3 NO3
Biogenic VOCs (Monoterpenes)
Direct Emission
Isoprene as a SOA source Claeys et al. 2004
Matsunaga et al. 2005 Lim et al. 2005 Kroll
et al. 2005 Henze and Seinfeld. 2006 van
Donkelaar et al. in review
Aromatics
Fossil Fuel Biomass
Vegetation combustion burning
11
BLACK CARBON IN THE ATMOSPHERE
BC is operationally defined as the
light-absorbing fraction of carbonaceous aerosols.
CHEMICAL AGING
WET DEPOSITION
PRIMARY EMISSION
coating by sulfate or organics
oxidation
Hydrophobic
Hydrophilic
How much
How long ()
Most global models assume 1 day for chemical
conversion of hydrophobic to hydrophilic BC.
12
2001 GEOS-Chem 1ox1o NESTED SIMULATIONS

• Uses the coupled oxidant-aerosol version of
  GEOS-CHEM (version 7.02) with 1ox1o horizontal
  resolution over North America (140-40oW 10-60oN)
  and 4ox5o horizontal resolution for the rest of
  the world.
• Includes weekday and weekend NEI99 anthropogenic
  emissions for NOx CO NMHC and SO2 in the
  United States EC and OC primary emissions from
  Bond et al. 2004 and Park et al. JGR 2003
  respectively.
• Include sulfur emissions in Canada and Mexico
  from EMEP and BRAVO emission estimates
  respectively.
• Include a global ship SO2 emission Corbett et
  al. 1999 Alexander et al. 2005.
• Includes a climatological biomass burning
  emission inventory with emission factors from
  Andreae and Merlet 2001.
• Includes a mechanistic simulation of secondary
  organic aerosols Chung and Seinfeld JGR 2002
  coupled to oxidant chemistry
• Applies HNO3 and NH3 dry deposition to the mixed
  layer column.

• Four simulations are conducted for 16 months
  starting from September 1 2000
• baseline (emissions as described above)
• natural (zero anthropogenic emissions worldwide)
• background (zero anthropogenic emissions in the
  U.S.)
• transpacific (zero anthropogenic emissions in
  North America)

13
ANNUAL MEAN SULFATE (2001) GEOS-Chem (1ox1o)
vs. IMPROVE (135 sites)
Highest concentrations in industrial Midwest
(coal-fired power plants)
14
SULFATE AT IMPROVE CASTNET NADP (deposition)
SITESmodel (1ox1o) vs. observed for different
seasons
High correlation in sulfate concentrations for
different seasons (R2 0.83 - 0.92) Low
bias in summer and high bias in other seasons
(Slope 0.84 - 1.32)
15
ANNUAL MEAN AMMONIUM AND NITRATE (2001)
GEOS-Chem vs. CASTNET (79 sites)
(no ammonium data at IMPROVE sites)
Highest concentrations in upper Midwest
NO3-
NH4
The spatial distribution of ammonium and nitrate
reflect the dominant ammonium nitrate formation
in North America.
16
Ammonium Nitrate Nitrate
AMMONIUM AND NITRATE AT CASTNET AND IMPROVE
SITES model (1ox1o) vs. observed for different
seasons
High correlation for different seasons (R2
0.82-0.85) High bias for NH4 in fall error in
seasonal variation of livestock emissions
High bias for NO3- esp. in summer/fall results
from bias on SO42--2NH4
17
ANNUAL MEAN EC AND OC (2001)GEOS-Chem (1ox1o)
vs. IMPROVE (135 sites)
GEOS-Chem
IMPROVE
µg m-3

• High OC in southeast U.S. vegetation
• High EC/OC in west fires

18
EC
OC
EC AND OC AT IMPROVE SITES model vs. observed
for different seasons
Low bias for EC indicates that Bond et al.
2004 EC emission could be low in the U.S.
No significant bias in OC with Park et al. 2003
emission but large scatter mostly from SOA
simulation dependent on preexisting primary
organic aerosols Chung and Seinfeld 2002
19
VISIBILITY DEGRADATION STATISTICS IN THE U.S.
(2001) IMPROVE vs. GEOS-CHEM (1ox1o)
R2 0.88
R2 0.63
Visibility extinction (deciviews dv
10ln(bext/10) ) from sulfate nitrate EC and
OMC.
20
CUMULATIVE DISTRIBUTION OF VISIBIILTY DEGRADATION
IN THE U.S. (2001) IMPROVE (black) vs. GEOS-CHEM
(red)
Deciviews
Model reproduces daily visibility degradation
successfully at 53 out of 87 sites in the west
and 24 out of 44 sites in the east.
Mexican sulfur emission in BRAVO inventory is
lower by 30 than Mexican NEI or global emission
inventory.
21
GEOS-CHEM SIMULATION OF TRACE-P OBSERVATIONS
Scavenging from Asian outflow is 80-90 efficient
for sulfate and BC 100 for nitrate
TRACE-P (Mar-Apr 2001) flight tracks for DC-8
P3-B aircraft
P3B DATA over NW Pacific (30 45oN 120 140oE)
Black carbon (BC)
Model underestimates BC observations by factor of
2 insufficient emissions Bond et al. 2004 or
excessive scavenging
22
EXPORT EFFICIENCY
BC emissions over East Asia are highly uncertain
Carmichael et al 2003.
Koike et al. 2003 Parrish et al 2004
X combustion-derived species RX emission
ratio (X/CO) enhancements relative to
background
NORMALIZED EXPORT EFFICIENCY
INDEPENDENT OF EMISSION RATIO R
We use the TRACE-P P-3B data north of 30oN for
which China provided a common source region.
23
OBSERVED EXPORT EFFICIENCYBC vs SOX
(SO2(g)SO42-) and HNO3T (HNO3(g)NO3-)
Export efficiency
Normalized export efficiency
BC AEROSOLS ARE SIGNIFICANTLY SOLUBLE BUT NOT AS
MUCH AS SULFATE OR NITRATE.
Park et al. 2005
24
BC NORMALIZED EXPORT EFFICIENCY IN ASIAN OUTFLOW
(GEOS-Chem vs TRACE-P )
Simulation with 11 days for BC scavenging
provides the best fit to the TRACE-P observations.
Park et al. 2005
25
IMPLICATION FOR CLIMATE
BC BURDEN ARCTIC DEPOSITION FLUX
BC lifetime is 5.8 1.8 days 50 longer than
that of sulfate global burden is 0.11 0.03 Tg
using Bond et al. 2004 inventory and resulting
decrease in Arctic snow albedo 3.2 2.5 with
1 1 days from the TRACE-P constraints Park
et al. 2005
26
CONTIGUOUS U.S. MAP (1ox1o)background
simulations with shutting off U.S anthropogenic
emissions
27
SIMULATED NATURAL AND BACKGROUND ANNUAL MEAN
AEROSOL CONCENTRATIONS IN THE UNITED STATES
28
SIMULATED NATURAL AND BACKGROUND ANNUAL MEAN
AEROSOL CONCENTRATIONS (CONT.)
29
AEROSOL CONCENTRATIONS IN THE U.S.contributions
from natural sources and transboundary pollution
Annual regional means averaged at IMPROVE sites
from GEOS-Chem standard and sensitivity
simulations

• Transboundary pollution influences from Canada
  Mexico are higher than those in Park et al.
  2003 2004 resulting in factor of 4 higher
  background concentration of ammonium sulfate than
  EPA default value.

30
END POINT VISIBILITY DEGRADATION FOR WORST 20
DAYS IN THE UNITED STATES

• The EPA default endpoint visibility shows a
  simple separation between the western and the
  eastern United States for which we find little
  basis.
• Our natural visibility endpoint has a
  considerable spatial variation and in general
  lower than the EPA default in the east.
• Background endpoint visibility is higher than
  natural visibility and is more spatially variable
  due to transboundary pollution influences.

31
IMPLICATIONS FOR EMISSION REDUCTIONS IN PHASE 1
(2004-2018) IMPLEMENTATION OF REGIONAL HAZE RULE
Illustrative calculation for NW IMPROVE sites
based on 20 worst visibility days statistics
assuming linear relationship between emissions
and PM concentrations and assuming constant
anthropogenic sources from foreign countries
between now and 2064
Desired trend in visibility
Phase 1
VISIBILITY DEGRADATION ON 20 WORST
VISIBILITY DAYS AT NORTHWESTERN IMPROVE SITES.
53
Required decrease of U.S. anthropogenic
emissions
28
32
PROJECTED SOx EMISSIONS IN ASIA
One projection suggests that emissions of SOx
will more than double in China between
1995-2020 Streets Waldhoff 2000
courtesy David Streets
Increasing SOx emissions from Asia will degrade
North American air quality and present a further
barrier to attainment of domestic air quality
regulations in the United States (eg. EPA Haze
Rule)