Transcript Tropospheric ozone response to methane emission controls: Implications for climate and global air quality Arlene M.

Tropospheric ozone response to
methane emission controls:
Implications for climate and global
air quality
Arlene M. Fiore
([email protected])
Acknowledgments:
Larry Horowitz, Chip Levy (NOAA/GFDL)
Jason West, Vaishali Naik (Princeton University)
Ellen Baum, Joe Chaisson (Clean Air Task Force)
Frank Dentener, Kees Cuvelier (JRC, Italy)
The TF HTAP Modeling Team
Funding from Luce Foundation via Clean Air Task Force
IGERT Joint Program Colloquium, Columbia University
April 5, 2007
The U.S. ozone smog problem is spatially widespread,
affecting >100 million people
Nonattainment Areas (2001-2003 data)
4th highest daily max
8-hr mean O3 > 84 ppbv
U.S. EPA, 2006
Radiative forcing of climate (1750 to present):
Important contributions from methane and ozone
IPCC, 2007
Air quality-Climate Linkage:
CH4, O3 are greenhouse gases
CH4 contributes to background O3 in surface air
Stratospheric O3
Stratosphere
~12 km
hn
O3
NO2
NO
OH
HO2
Free Troposphere
Hemispheric Pollution
Direct Intercontinental Transport
Boundary layer
(0-3 km)
VOC, CH4, CO
NOx
VOC
air pollution (smog)
O3
air pollution (smog)
CONTINENT 1
OCEAN
NOx
VOC
CONTINENT 2
O3
A.M. Fiore
IPCC [2001] scenarios project future growth
Projections of future CH4 emissions
(Tg CH4) to 2100
Change in 10-model mean
surface O3
2100 SRES A2 - 2000
Attributed mainly to increases
in methane and NOx
[Prather et al., 2003]
Rising background O3 has implications
for attaining air quality standards
Recent observational analyses suggest that
surface O3 background is rising
[e.g. Lin et al., 2000; Jaffe et al., 2003, 2005;
Vingarzan, 2004; EMEP/CCC-Report 1/2005 ]
Pre-industrial
background
20
Europe
seasonal
40
Current background
WHO/Europe
8-hr average
60
U.S. 8-hr
average
80
100
O3 (ppbv)
Future
background?
A.M. Fiore
Radiative Forcing* (W m-2)
Double dividend of Methane Controls:
Decreased greenhouse warming and improved air quality
Number of U.S. summer gridsquare days with O3 > 80 ppbv
50% 50% 50% 2030 2030
1995 50% 50% 50% 2030 2030
anth. anth. anth. A1 B1
(base) anth. anth.anth. A1 B1
VOC CH4 NOx
VOC CH4 NOx
GEOS-Chem Model (4°x5°)
IPCC
scenario
Anthrop. NOx emissions
(2030 vs. present)
Global
U.S.
Methane emissions
(2030 vs. present)
A1
+80%
-20%
+30%
B1
-5%
-50%
+12%
CH4 links air quality
& climate via
background O3
Fiore et al., GRL, 2002
Impacts of O3 precursor reductions on
U.S. summer afternoon surface O3 frequency distributions
GEOS-Chem Model Simulations (4°x5°)
NOx controls
strongly decrease
the highest O3
(regional pollution
episodes)
CH4 controls
affect the entire
O3 distribution
similarly
(background)
Results add linearly when both methane and NOx are reduced
Fiore et al., 2002; West & Fiore, ES&T, 2005
Methane trends and linkages with
chemistry, climate, and ozone pollution
1) Climate and air quality benefits from CH4 controls
 Characterize the ozone response to CH4 control
 Incorporate methane controls into a future emission scenario
2) Recent methane trends (1990 to 2004)
 Are emission inventories consistent with observed CH4 trends?
 Role of changing sources vs. sinks?
Research Tool:
MOZART-2 Global Chemical
Transport Model [Horowitz et al., 2003]
NCEP, 1.9°x1.9°, 28 vertical levels
 Fully represent methane-OH relationship
 Test directly with observations
3D model structure
More than half of global methane emissions
are influenced by human activities
~300 Tg CH4 yr-1 Anthropogenic [EDGAR 3.2 Fast-Track 2000; Olivier et al., 2005]
~200 Tg CH4 yr-1 Biogenic sources [Wang et al., 2004]
>25% uncertainty in total emissions
Clathrates?
Melting permafrost?
PLANTS?
BIOMASS BURNING
+ BIOFUEL
ANIMALS
30
WETLANDS
90
180
60-240 Keppler et al., 2006
85 Sanderson et al., 2006
10-60 Kirschbaum et al., 2006
0-46 Ferretti et al., 2006
GLOBAL METHANE
SOURCES
(Tg CH4 yr-1)
TERMITES RICE 40
20
LANDFILLS +
WASTEWATER
50
GAS + OIL
60
COAL
30
A.M. Fiore
Characterizing the methane-ozone relationship
with idealized model simulations
Reduce global anthropogenic
CH emissions by 30%
Surface Methane
Tropospheric O34burden
0
0
-50
 Surface
Methane
Abundance
(ppb)
-2
-100
-4
-150
-200
-6
-250
-8
-300
Tropospheric
O3 Burden
(Tg)
-10
-350
-400
-12
1
6
11
16
21
26
31
Simulation Year
Model approaches a new steady-state after 30 years of simulation
Is the O3 response sensitive to the location of CH4 emission controls?
A.M. Fiore
Change in July surface O3 from 30% decrease in
anthropogenic CH4 emissions
Globally uniform emission reduction
Percentage
Difference:
Asia – uniform
Asia
Emission reduction in Asia
Enhanced effect in
source region
<10% other NH source regions
< 5% rest of NH
<1% most of SH
 Target cheapest controls worldwide
A.M. Fiore
Decrease in summertime U.S. surface ozone
from 30% reductions in anthrop. CH4 emissions
MAXIMUM DIFFERENCE
(Composite max daily afternoon mean ozone JJA)
NO ASIAN ANTH. CH4
 Largest decreases in NOx-saturated regions
A.M. Fiore
Tropospheric O3 responds approximately linearly to
anthropogenic CH4 emission changes across models
MOZART-2 [West et al., PNAS 2006; this work]
TM3 [Dentener et al., ACP, 2005]
GISS [Shindell et al., GRL, 2005]
X GEOS-CHEM [Fiore et al., GRL, 2002]
IPCC TAR [Prather et al., 2001]
Anthropogenic CH4 contributes ~50 Tg (~15%) to tropospheric O3 burden
~5 ppbv to global surface O3
A.M. Fiore
Multi-model study shows similar surface ozone decreases
over NH continents when global methane is reduced
ppbv
ppbv
ANNUAL
MEAN OZONE DECREASE FROM 20%
ANNUAL MEAN SURFACE OZONE DECREASE DUE TO
DECREASE
IN GLOBAL
METHANE
20% GLOBAL METHANE
REDUCTION
2
2
Full range of
12 individual
models
1.5
1.5
1
1
0.5
0.5
0
0
EUROPE
EU
EU
N. AMER.
NA
NA
S.SA
ASIA
Receptor region
SA
E.EA
ASIA
EA
 >1 ppbv O3 decrease over all NH receptor regions
 Consistent with prior studies
TF HTAP 2007 Interim report draft available at www.htap.org
A.M. Fiore
How much methane can be reduced?
Ozone reduction (ppb)
Cost-saving
reductions
0.7
North America
Rest of Annex I
Rest of World
(industrialized nations)
1.4
<$10 / ton
CO2 eq.
All identified
reductions
00
1.9
10% of anth.
emissions
20% of anth.
emissions
20
40
60
80
100
120
20
40
60
80
100 -1
120
Methane
Methane reduction
reductionpotential
potential(Mton
(MtonCH
CH4 4yryr)-1)
IEA [2003] for 5 industrial sectors
Comparison: Clean Air Interstate Rule (proposed NOx control)
reduces 0.86 ppb over the eastern US, at $0.88 billion yr-1
West & Fiore, ES&T, 2005
Will methane emissions increase in the future?
Anthropogenic CH4 emissions (Tg yr-1)
Dentener et al., ACP, 2005
A2
B2
MFR
Current
Legislation
(CLE)
Scenario
PHOTOCOMP for IPCC AR-4 used CLE, MFR, A2 scenarios for all O3 precursors
[Dentener et al., 2006ab; Stevenson et al., 2006; van Noije et al., 2006; Shindell et al., 2006]
Our approach: use CLE as a baseline scenario & apply methane controls
Emission Trajectories in Future Scenarios (2005 to 2030)
Anthropogenic CH4
Emissions (Tg yr-1)
Surface NOx Emissions
2030:2005 ratio
CLE Baseline
A
B
C
0.3
0.8
1.4
1.9
2.5
Control scenarios reduce 2030 CH4 emissions relative to CLE by:
A) -75 Tg (18%) – cost-effective now
B) -125 Tg (29%) – possible with current technologies
C) -180 Tg (42%) – requires new technologies
Additional 2030 simulation where CH4 = 700 ppbv (“zero-out anthrop. CH4”)
A.M. Fiore
Reducing tropospheric ozone via methane controls decreases
radiative forcing (2030-2005)
+0.16 Net Forcing
(W m-2)
+0.08
0.00 -0.08
-0.58
OZONE
METHANE
CLE
A
B
C CH4=700 ppb
Methane Control Scenario
 More aggressive CH4 control scenarios offset baseline CLE forcing
A.M. Fiore
Future air quality improvements from CH4 emission controls
USA Percentage of grid-square days > 70 ppb
Percentage of model grid-cell days where surface ozone > 70 ppbv
CLE2005
CLE2030
A2030
B2030
C2030
CH4_700
18
16
14
12
10
8
6
4
2
0
DJF
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
Controls on global CH4
reduce incidence
of O3 > 70 ppbv
USA
MAM
Europe
JJA
SON
Cost-effective
CLE2005
controls
prevent
CLE2030
increased
A2030
occurrence
of
B2030
O3 > 70
ppbv in 2030
C2030
relative
to 2005
CH4_700
Europe
DJF
MAM
JJA
SON
2030 European high-O3 events under CLE emissions scenario
show stronger sensitivity to CH4 than in USA
A.M. Fiore
Summary: Connecting climate and air quality via O3 & CH4
Ozone reduction (ppb)
North America
Rest of Annex I
Rest of World
Cost-saving
reductions
• Independent of reduction location (but depends on NOx)
Target cheapest controls worldwide
<$10 / ton
CO2 eq.
All identified
reductions
0
20
40
60
80
100
120
Methane reduction potential (Mton CH 4 yr-1)
Anthrop. CH4 emissions
CLE
A
B
C
Tg yr-1
CLIMATE AND AIR QUALITY BENEFITS FROM CH4 CONTROL
• Robust response over NH continents across models
~1 ppbv surface O3 for a 20% decrease in anthrop. CH4
• Complementary to NOx, NMVOC controls
• Decreases hemispheric background O3
 Opportunity for international air quality management
How well do we understand recent trends in atmospheric methane?
How will future changes in emissions interact with a changing climate?
A.M. Fiore
Observed trend in surface CH4 (ppb) 1990-2004
Global Mean CH4 (ppb)
Hypotheses for leveling off
discussed in the literature:
1790
1780
1. Approach to steady-state
1770
1760
NOAA GMD Network
1750
1740
1730
1720
2. Source Changes
Anthropogenic
Wetlands/plants
(Biomass burning)
3. (Transport)
1710
1990
1992
1994
1996
1998
2000
2002
2004
Data from 42 GMD stations with 8-yr minimum
record is area-weighted, after averaging in bands
60-90N, 30-60N, 0-30N, 0-30S, 30-90S
4. Sink (CH4+OH)
Humidity
Temperature
OH precursor emissions
overhead O3 columns
Can the model capture the observed trend (and be used for attribution)?
A.M. Fiore
BASE simulation
EDGAR 2.0 emissions held constant
1790
1780
Global Mean Surface Methane (ppb)
OBSERVED
MOZART-2
1770
1760
Mean Bias (ppb)
Bias and correlation vs. observed surface CH4: 1990-2004
Overestimates
interhemispheric
gradient
1750
1740
1730
1720
1990
1992
1994
1996
1998
2000
2002
2004
Overestimates 1990-1997
but matches trend
Captures flattening post1998 but underestimates
abundance
r2
1710
Correlates poorly
at high N latitudes
S
Latitude
N
Biomass Burning
Ruminants
Estimates for changing
methane sources in the 1990s
Rice
Biogenic
Wastewater
Inter-annually varying wetland emissions
Biomass Burning
Landfills
1990-1998 from Wang et al. [2004]
Ruminants
Energy
(Tg CH4 yr-1); different distribution
Rice
Biogenic adjusted to maintain
270
constant total source
Wastewater
260
Landfills
547
Energy
1995
500v3.2 2000 v3.2
250
Tg CH4 yr-1
240
Biogenic
230
Biomass Burning
Ruminants
220
Rice
210
Wastewater
Landfills
200
Energy
1990
1995
400
300
200
100
2000
2005
Apply climatological mean
(224 Tg yr-1) post-1998
0
1990 v2.0 1990 v3.2 1995 v3.2 2000 v3.2
BASE
ANTH
EDGAR anthropogenic inventory
ANTH + BIO
A.M. Fiore
Bias & Correlation vs. GMD CH4 observations: 1990-2004
Global Mean Surface Methane (ppb)
Mean Bias (ppb)
1790
1780
1770
1760
1750
OBS
BASE
ANTH
ANTH+BIO
1740
1730
1720
1710
1990
1995
2000
2005
ANTH+BIO simulation with timevarying EDGAR 3.2 + wetland
emissions improves:
 Global mean surface conc.
 Interhemispheric gradient
 Correlation at high N latitudes
Fiore et al., GRL, 2006
r2
S
Latitude
N
How does meteorology influence methane abundances?
Why does BASE run with constant emissions level off post-1998?
 Examine sink
CH4 Lifetime (t) against Tropospheric OH
t=
[CH 4 ]
k[OH][CH 4 ]
t
Temperature
(88% of CH4 loss
is below 500 hPa )
Humidity
Photolysis
Lightning NOx
t = 0.17 yr = 1.6%)
What drives the change in
methane lifetime in the model?
A.M. Fiore
Small increases in temperature and OH
shorten the methane lifetime against tropospheric OH
tOH
0.18
Deconstruct t (-0.17 years)
from 1991-1995 to 2000-2004
into individual contributions by
varying OH and temperature
separately
0.16
0.14
0.12
0.1
0.08
+
0.06
=
0.04
0.02
Lightning NOx
3.4
Global Lightning NOx (TgN yr-1)
0
T(+0.3K)
climT
OH(+1.2%)
climOH
BASE
BASE
3.2
LNOx (TgN/y)
3
2.8
An increase in lightning NOx drives the
OH increase in the model
2.6
2.4
2.2
But lightning NOx is highly parameterized
…how robust is this result?
2
1.8
1990
1992
1994
1996
1998
Year
2000
2002
2004
A.M. Fiore
Additional evidence for a global lightning NOx increase?
Estimate lightning NOx changes using options available in the
GFDL Atmospheric General Circulation Model:
• Convection schemes (RAS vs. Donner-deep)
• Meteorology (free-running vs. nudged to NCEP reanalysis)
Lightning
NOx NOx
change,
1991-95
to 2000-04
Lightning
change,
1991-95
to 2000-04
20% 20%
Lightning NOx % change (91-95 to 00-04)
RAS Donner
MOZART
18% 18%
More physically-based lightning
NOx scheme [Petersen et al., 2005]
16% 16%
Evidence from observations?
14% 14%
MOZART-2
MOZART-2
10% 10%
AM2 AM2
% change
12% 12%
8%
8%
6%
6%
4%
4%
2%
2%
0%
0%
AM2-D
AM2-D
NCEP(nudged)
free-running GCM
Lightning NOx increase robust;
magnitude depends on meteorology
c/o L.W. Horowitz
LIS/OTD
Flash counts
Magnetic field variations
in the lower ELF range
[e.g. Williams, 1992;
Füllekrug and FraserSmith, 1997; Price, 2000] Negev Desert
Station, Israel
Summary: Connecting climate and air quality via O3 & CH4
Ozone reduction (ppb)
North America
Rest of Annex I
Rest of World
Cost-saving
reductions
CLIMATE AND AIR QUALITY BENEFITS FROM CH4 CONTROL
• Independent of reduction location (but depends on NOx)
Target cheapest controls worldwide
<$10 / ton
CO2 eq.
All identified
reductions
0
20
40
60
80
100
120
Methane reduction potential (Mton CH 4 yr-1)
Tg yr-1
Anthrop. CH4 emissions
CLE
A
B
• Robust response over NH continents across models
~1 ppbv surface O3 for a 20% decrease in anthrop. CH4
• Complementary to NOx, NMVOC controls
• Decreases hemispheric background O3
 Opportunity for international air quality management
C
Methane (ppb)
1790
1780
1770
1760
OBS
BASE
ANTH
ANTH+BIO
1750
1740
1730
1720
1710
1990
0.18
1995
2000
METHANE TRENDS FROM 1990 TO 2004
• Simulation with time-varying emissions and meteorology
best captures observed CH4 distribution
2005
tOH 91-95 to 01-04
0.16
• Model trend driven by increasing T, OH
• Trends in global lightning activity?
0.14
0.12
0.1
0.08
0.06
+
0.04
0.02
0
T
climT
 Potential for climate feedbacks (on sources and sinks)
=
OH
climOH
BASE
BASE
A.M. Fiore