Evaluating NOx emission Inventories for regulatory air quality

Report
Evaluating NOx Emission Inventories for
Regulatory Air Quality Modeling using
Satellite and Model Data
Greg Yarwood, Sue Kemball-Cook and Jeremiah Johnson
ENVIRON International Corporation
January 16, 2014
Template
Introduction
Episode Average Normalized Bias: June 2006
• A high bias for modeled
ozone in the southeast
may be affecting the
TCEQ’s SIP modeling
– Confounds ozone
transport assessments
– May result from biased
NOx emissions
• Purpose of this project:
Use satellite and CAMx
model data to assess
whether bias is present
in NOx emission
inventories in the TCEQ’s
SIP modeling
2
Mass Balance Method for Evaluating NOx Emissions
   
Ω
=
×   × ℎ 
Ω
• Comparison of satellite-retrieved and CAMx
modeled NO2 columns
– Ω are integrated tropospheric NO2 vertical column
densities (VCD)
– Method of Leue et al. (2001); Martin et al. (2003)
 Used by Boersma et al. (2008); Tang et al. (2013)
• NO2 columns from Ozone Monitoring
Instrument
3
Necessary Conditions for Mass Balance Method to
Provide Constraints on NOx Emissions
• Model must accurately simulate formation,
transport and fate of NO2 and its reservoir species
– Meteorology, chemistry, boundary conditions
– Largest uncertainty should be in the emission
inventory
• Satellite column NO2 retrieval must have error
smaller than the perturbation in the NO2 columns
caused by the uncertainty in the emission
inventory
4
NO2 Column Retrievals
• Used KNMI DOMINO
Stratosphere
Troposphere
Slant Path
Aura Satellite
Vertical
Stratospheric
NO2 Column
Vertical
Tropospheric
NO2 Column
v2.0 and NASA SP2
retrievals
– Estimate of uncertainty
introduced by the
retrieval in the topdown emission
estimates
• Retrievals begin with
the same OMI slant
columns, differ in:
– Method for
stratospheric column
determination
– AMF calculation
5
CAMx Model
• TCEQ June 2006 modeling
platform
• 36/12/4 km nested grids
• CAMx v5.41
– CB6r1 chemical
mechanism
• Evaluated model against
surface obs (CAMS,
CASTNet, SEARCH, AQS),
and INTEX-A aircraft data
• Calculated modeled VCD
for June episode
6
Upgrades Required to Simulate Tropospheric NO2
VCD with TCEQ Modeling Platform
• Improved simulation of NO2 sources/sinks in UT
– Lightning NOx emission inventory
– TCEQ aircraft emission inventory based on detailed flight
track data
– Day-specific wildfire emissions (FINN)
– CAMx simulation of
 Transport of NOy and ozone from stratosphere to troposphere
 Vertical transport of chemical species via convection
– Revised chemical mechanism
• CAMx showed sufficient agreement with retrieved
columns and INTEX-A data for project to proceed
– Does not demonstrate that the CAMx NO2 columns are
correct or that they agree with OMI for the right reasons
7
June 2006 Episode Average Retrieved NO2 VCD
• Overall patterns of high and low VCD are similar, but there are
differences between the retrievals
– Southeastern U.S.
– Atlanta, New York maxima
– Offshore of Carolinas
• When used together with a single set of CAMx columns, retrievals will
give different top-down emissions estimates
8
Episode Average NO2 VCD Comparison
• Differences in OMI/CAMx over East Texas, southeastern U.S. and ocean
• I-35 <0 in DOMINO/CAMx, not SP2/CAMx
9
No Smoothing: Top Down - Bottom Up NOx EI
Using CAMx with Two Retrievals
DOMINO
SP2
• Results differ over East Texas, broad areas of southeast
• How to use this information to improve TCEQ NOx EI?
10
Alternative Method: State-Level Comparison
CAMx VCD – DOMINO VCD
Northeast
Texas Southeast
and
adjacent
Ohio River
Valley
North
West
• Simple column comparison
by state
• No smoothing of EI required
11
State-Level Comparison using Two Retrievals
CAMx VCD – DOMINO VCD
• Results differ
Northeast
Texas
and
adjacent
Southeast
Ohio River
Valley
North
West
CAMx VCD - SP2 VCD
Texas
and
adjacent
Southeast
Ohio River
Valley
Northeast
North
West
near Texas,
in southeast,
better
agreement
further north
• Retrievals /
model give
contradictory
results on
TCEQ NOx
inventory in
southeast
12
Summary
• Top-down emissions estimates derived with
current generation of regional air quality models
and retrievals not recommended for Texas
– Uncertainty in modeling of NO2 and reservoir species
and in retrievals
• Use of satellite data is not straightforward
– Analyzing multiple retrievals was very useful in this
project
• Satellite data used together with aircraft flight
data are powerful tools for evaluating and
improving air quality models
13
Acknowledgements
• We acknowledge the free use of tropospheric NO2
column data from the OMI sensor from
www.temis.nl and the use of NASA SP2 retrieval.
• We wish to thank the University of WisconsinMadison for the use and development of the
Wisconsin Horizontal Interpolation Program for
Satellites (WHIPS). WHIPS was developed by Jacob
Oberman, Erica Scotty, Keith Maki and Tracey
Holloway, with funding from the NASA Air Quality
Applied Science Team (AQAST) and the Wisconsin
Space Grant Consortium Undergraduate Award.
14
End
15

similar documents