Transcript ppt - Harvard University

Diagnosing the sensitivity of
O3 air quality
to climate change
over the United States
Moeko Yoshitomi ([email protected]),
Daniel J. Jacob, Loretta J. Mickley,
Philippe Le Sager, Julia M. Sygiel, Shiliang Wu
Department of Earth & Planetary Sciences
Harvard University
presented at GCAP Phase 2 Science Team Meeting
Harvard University, MA, US
October 12th, 2007
Projecting Future O3 Air Quality in the US
Δ(2050s－present) in simulated summer daily 8-hr max O3 levels
[1] Harvard CTM
[Wu et al., 2007]
[2] CMU CTM
[Racherla and Adams, 2006]
O3 [ppb]
・ Many regions show substantial future O3 increases (1-5 ppb) due to
climate change
・ Spatial patterns of climate-induced increases in O3 is different
- where the highest O3 is predicted
- how much O3 increases at a maximum
Projecting Future O3 Air Quality in the US
Δ(2050s－present) in simulated summer daily 8-hr max O3 levels
[1] Harvard CTM
[Wu et al., 2007]
[2] CMU CTM
[Racherla and Adams, 2006]
Higher isoprene emissions in SE
⇒ O3 ↑ or ↓???
O3 [ppb]
isoprene
emission
[10e-8 g C/m2/s]
Key issue: modeling of the fate of isoprene emissions over the SE
[g/s]
Uncertain NOx-Isoprene-O3 Chemistry
There is a controversial competition on isoprene effects on O3
OH
NO
Isoprene
RO2
k298=1.0x10-10
NO2
O3 ↑
NOx recycled ??
NO
RONO2
(isoprene nitrate)
branch ratio??
Deposition
O3
Products
Isoprene
k298=1.3x10-17
O3 ↓
O3 ↓
Understanding the Relationship of Surface O3 and Temperature
Major Questions from Previous GCAP Research
1. Can we apply observed O3-temperature relationship to validate
chemical models used to investigate the sensitivity of surface O3 to
climate change?
2. Is the O3-temperature relationship sensitive to assumptions about
the fate of isoprene nitrate?
Observations
Hourly surface O3 concentrations from AIRS/EPA and daily maximum
temperatures from NCDC, for JJA 1980-1998, which is gridded at 4x5
Model Simulations
[1] GCAP run at 4x5 resolution = GEOS-Chem + calculated GISS GCM
meteorology (JJA 1999-2001)
[2] GEOS-4 run at 2x2.5 resolution = GEOS-Chem + assimilated
GEOS4/GMAO meteorology (JJA 1999-2001)
Regression Analysis on O3 and T
Data Sets
・ Daily maximum of 8-hr running average O3 concentrations (MDA8)
・ Daily maximum temperature
Approach
Reduced-major-axis (RMA) linear regression: to allow for errors in both
x and y coordinates
[For Question 1 & 2] Analyze the relationship in a gridbox-scale
[For Question 2] Analyze the relationship in a regional scale
(Northeast: 7 boxes, Southeast: 9 boxes, Midwest, Northwest,
Southwest, Central, Mid-Atlantic)
Observed Correlation between O3 and T
From RMA linear regression
Slope dO3/dT
-10
5
Correlation Coefficient
0
・ Sensitivity of O3 to
temperature varies
across the US
・ The NE shows stronger
O3/T correlation than the
SE
5
10
-1
-.5
0
.5
1
Observed Correlation between O3 and T
From RMA linear regression
Slope dO3/dT
-10
5
Correlation Coefficient
0
・ Sensitivity of O3 to
temperature varies
across the US
・ The NE shows stronger
O3/T correlation than the
SE
5
10
SE
-1
-.5
0
NE
.5
1
O3 and Temperature Observed by AIRS
150
150
100
100
50
50
MDA8 [ppb]
0
280
300
320
0
280
150
150
100
100
50
50
0
280
300
T max [K]
0
320 280
300
320
300
T max [K]
320
Temperature has an inter-annual variability
- 19-yr mean/1988 mean/1992 mean; NE: 300.53/302.14/298.79 [K],
SE: 305.92/306.10/305.26 [K]
MDA8 [ppb]
O3 and Temperature Simulated by GCAP
150
150
100
100
50
50
MDA8 [ppb]
0
280
300
320
0
280
150
150
100
100
50
50
0
280
300
320
0
280
The GCAP
results do not
vary too much
year-to-year
300
320
300
T max [K]
320
NE has stronger O3/T correlation than SE
- GCAP O3/T correlations are slightly stronger in both NE and SE than
observed O3/T correlations
GCAP Model Captures Correlation Trends
1980-1998 obs
-1
-.5
1999-2001 GCAP
0
.5
1
・ Sensitivity of O3 to temperature
varies across the US in both obs.
and models
・ Both the observations and
GCAP show stronger correlation in
1999-2001 GEOS4
the NE than the SE
・ GEOS4 does not show the difference between NE and SE that obs.
and GCAP model show
GCAP Captures the Variation in dO3/dT Slope
Slope of dO3/dT from RMA linear regression
1980-1998 obs
-10
5
1999-2001 GCAP
0
5
10
・ Cool Gulf of Mexico air with lowlevel O3 gives higher slopes in the
SE; nothing to do with local
chemistry
・ GEOS4 does not show the
difference between NE and SE
that GCAP shows
1999-2001 GEOS4
My Future Work
[Short-term work]
・ Explore reasons for the differences in O3-T correlation between obs.
and models
- meteorology issue (e.g., difference between GCAP and GEOS4)
- missing chemistry (e.g., isoprene sensitivity analysis)
・ Interpret whether the NE-SE contrast reflects transport or chemistry
[Long-term work]
・ Estimate the sensitivity of ozone pollution to future global change
with our GCAP model for each SRES scenario
・ Investigate effects of 2000-2050 changes in global anthropogenic
emissions and in general circulation to the intercontinental transport
of air pollutants to the US
GCAP Phase 2 Science Team meeting
at Maxwell Dworkin 223
GCAP Phase 2 Science Team meeting
at Maxwell Dworkin 223
GCAP Phase 2 Science Team meeting
at Maxwell Dworkin 223
GCAP Phase 2 Science Team meeting
at Maxwell Dworkin 223