A Study of Photochemical Processes of Houston Air Quality with EPA

1

• A Study of Photochemical Processes of Houston Air
  Quality with EPAs CMAQ

Daewon W. Byun
Contribution Dr. G. Lubertino, Dr. S.-B. Kim,
Dr. N.-K. Moon, Dr. K. Lazarova Ms. B Czader, Ms.
F.-Y. Cheng
University of Houston Air Quality Modeling and
Monitoring
2
Houston Air Quality Characteristics

• TexAQS 2000 campaign and historic O3 data
  analysis have identified massive and frequent
  spikes of ozone (THOEsTransient High Ozone
  Events)
• Dependent on the Land/Sea Breeze Circulations
• Appear to be associated with large releases of
  reactive unsaturated hydrocarbons
• Models with standard emissions often miss the
  observed peaks large uncertainties in VOC
  emissions

3
Texas Air Quality Study Aug./Sept. 2000
0830 1300 CDT
0830 1200 CDT
Transient High-Ozone Events (THOEs)
4
Texas Air Quality Study Aug./Sept. 2000
0830 1500 CDT
0830 1400 CDT
This picture may not be right
Transient High-Ozone Events (THOEs)
5
Transient High-Ozone Events (THOEs)
6
Large petrochemical facilities
50 of USA Capacity
7
Propylene Emissions
CH2CH2
Ethylene Emissions
CH3-CHCH2
8
Unsaturated Hydrocarbon Photochemistry

• Ethylene reactions

CH2CH2

• Propylene reactions

CH3-CHCH2
9
TexAQS NOAA Aircraft Measurement (Ryerson et al
2001)
10
Land/Sea Breeze Circulation
N
N
W
W
E
E
S
S
Hourly Resultant Wind Vectors Analysis on (a)
exceedance days (b) non-exceedance days (from
Pete Breitenbach, TNRCC)
11
Application of ARL/UH trajectory model
Backward Trajectories for 6 hours ending at C408,
C1, C607 and C608 (10m)
12
Science Issues To Study

• How well does MM5 driver run represent the
  reality?
• Do synoptic and land/sea breeze circulations well
  represented?
• Does PBL evolution well captured?
• How important is the chemical mechanism in
  simulating Houstons high ozone problem?
• Does CB-4 represent olefin chemistry well?
• How about SAPRC99?
• Do we need further enhancement of CB-4 or
  SAPRC99?
• How bad is the VOC emissions uncertainty?
• EPA NET96 only TCEQ 2000 data not for SMOKE/CMAQ
• Even the recent inventory may have 3-10 times
  less olefins

13
TAMU MM5 Simulations (Driver Run)
 
J. Nielsen-Gammon (TAMU)
14
 
Urban Area
Temporally modified soil moisture value () in
TAMU simulation (S2)
25-category (USGS) vegetation categories
15
TAMU MM5 Driver Simulations 4-km (Domain Average)
 
16
PBL Height Evaluation with Wind Profiler Data
La Marque (LM) Ellington (EL) Wharton
(WH) Liberty (LB) HS (Western Houston)
17
TexAQS Profiler data
August 27, 2000
August 26, 2000
Profiler
Profiler
MM5-S2
MM5-S2
18
TexAQS Profiler data
August 29, 2000
August 28, 2000
Profiler
Profiler
MM5-S2
MM5-S2
19
TexAQS Profiler data
August 31, 2000
August 30, 2000
Profiler
Profiler
MM5-S2
MM5-S2
20
Impact of Different Mechanism in CMAQ

• TEXAQS 2000 (August 22-September 2)
• EPA 96 NET
• Waiting for EPA NET 99 V2 (IDA Format)
• SMOKE Tool/SMOKE emissions processor
• TAMU/TNRCC MM5 driver base simulation
• IC 2-day initialization runs with emissions data
• BC CMAQ seasonal default at 36-km domain
• CMAQ simulations with CB-4 and SAPRC99 mechanisms

21
VOC emission categories for CB4 SAPRC
SAPRC99 mechanism
CB-4 mechanism
FORM Formaldehyde ALD2 High molecular weight
aldehydes C2O3 Peroxyacyl radical PAN
Peroxyacyl nitrate PAR Paraffin carbon
bond ROR Secondary organic oxy radical OLE
Olefine carbon ETH Ethene TOL
Toluene CRES Cresol and higher molecular
weight phenols TO2 Toluene-hydroxy radical
adduct CRO Methylphenoxy radical OPEN
High molecular weight aromatic oxidation ring
fragment XYL Xylene MGLY Methylglyoxal ISOP
Isoprene XO2 NO-to-NO2 operation XO2N
NO-to-nitrate operation
22
CB-4 and SAPRC Emissions 96 NET
Ethylene Emissions
23
CB-4 and SAPRC Emissions 96 NET
Olefin Emissions
CB-4
SAPRC
24
CB-4 and SAPRC Emissions
96 NET
CB-4
Aldehyde Emissions
SAPRC
25
Air Quality Modeling with US EPAs CMAQ
Community Multi-pollutant Multi-scale Air
Quality Modeling System
26
Air Monitoring Site
27

28
CAMS and EPA Site Number Table
29
CB4 vs. SAPRC 99 Aug 23 27 with NEI96
30
Trajectories (Aug 25, Aug 26, 2100 UTC)
31
Trajectories (Aug 27, Aug 28, 2100 UTC)
32
O3 simulations (CB-4 vs. SAPRC99) Process
Analysis (Aug 31, 1800 UTC)
33
Trajectories High Ozone Day (Aug 30/31, 2100
UTC)
Propylene
34
Aug 28 31 with NEI96
CB4
SAPRC
35
Aug 28 31 with NEI96
CB4
SAPRC
36
Aug 28 31 with NEI96
SAPRC
CB4
37
Aug 28 31 with NEI96
SAPRC
CB4
38
CMAQ, 96NET-SMOKE, CB-4 vs. SAPRC99
CB-4
SAPRC99
39
O3 simulations (CB-4 vs. SAPRC99) and Process
Analysis (Aug 31, 2100 UTC)
40
Max. O3 Difference between SAPRC99 and CB4
Mechanism Aug 23 Sept. 1, 2000
41
O3 (CB-4 vs. SAPRC99) at 12-km resolution
42
O3 (CB-4 vs. SAPRC99) at 4-km resolution
43
Impact of Emissions Uncertainty

• Use of TCEQs most recent CAMx-ready emissions
  data
• EPA 96 NET (current UH database) too old
• Projection possible, but less reliable
• Waiting for QAed EPA 99 NET IDA format data
• Current UH emissions processor is SMOKE
  Tool/SMOKE
• Has not fully learned TCEQs emissions processing
  steps with EPS2 yet.
• Compare with the NET96 CMAQ runs
• very large emissions sensitivity (from 96NEI to
  current TNRCC)
• Help evaluating robustness of meteorological data
• Evaluate with observation data
• Help performance comparison with CAMx simulations

44
Benefits of using CAMx-ready data

• TCEQ has invested many emissions processing
  components Major differences from EPA NET 96/99
• Mobile
• (on-road/off-road, commercial marine, aircraft,
  railroads, etc)
• Biogenics GLOBEIS
• Satellite derived radiation and surface
  temperature
• Landuse data Land cover categories
• Point sources
• More recent, better QAed data
• More time dependent information (upsets, etc)
• Speciated VOC (Ship Channel point sources)

45
Limitations of relying on CAMx-ready data

• Emissions
• Only for CB-4 mechanism available now
• Subject to data availability
• Impossible to make additional data linkage
  between emissions, met data, and/or model
  configurations
• Difficult to characterize CMAQ model performance
  in relation to each emissions component with
  mixing, transport, and chemical mechanism.
• Limited emissions control sensitivity runs
  possible

46
Emissions Data Processing
I/O converter
netCDF Area Data
CAMx Area Data
Species renaming Unit conversion Coordinate
conversion
Major stack
camxPt
Major emis
CAMx Point Data
MEPSE stack
Unit conversion Species renaming Coordinate
conversion
MEPSE emis
47
Emissions-Chemistry Interface Processor
netCDF Area Data
MCIP Output
Major stack
CMAQ Emissions Data
ECIP
Major emis
MEPSE stack
MEPSE emis
Unit conversion Merging Plume Rise Plume Dynamics
Model (if MEPSE used)
48
Sample Comparison of Emissions August 30, 2000
CAMx-emissions
EPA NET 96
49
Sample Comparison of NOx Emissions Aug 30
CAMx-emissions
EPA NEI 96
50
Sample Comparison of NOx Emissions Aug 30
CAMx-emissions
EPA NEI 96
51
Sample Comparison of CO Emissions Aug 31, 20UTC
CAMx CO emissions
CAMx-NET96 CO
52
Biogenic Emissions Need to use BGUSE, but
BCUSE BEIS2
GLOBEIS
53
EPA NEI 96
CAMx-emissions
54
CMAQ Simulation with CAMx-ready emissions
CAMx-emissions
EPA NEI 96
55
CMAQ, CAMx-ready, CB-4
56
CMAQ, CAMx-ready, CB-4
57
O3 (NEI96 vs. CAMX-ready) at 4-km resolution
58
Where to go from here? (1)

• General Garbage in - Garbage out problem
• There is so much uncertainty in the inventory
  that the numbers we get out of any model (CAMx or
  CMAQ) will be suspect.  Is multiplying the VOC's
  by 6 (or, making OLE the same as NOx) really
  accurate?
• Continual changes in TCEQ processing are creating
  re-work. 
• - Nielsen-Gammon s new Meteorology,
• - Use PBL height 30 less than MM5 simulations,
• - New KS1f
• - Toluene reaction rates in CB-4, etc.,

59
Where to go from here? (2)

• Data processing for Texas/Houston specific
  emissions is not possible with current system
• Need to develop software able to connect to
  TCEQ/Environ emissions inventory
• Suggested ISSC and TCEQ for the development of
• Texas Emissions Inventory Preparation System for
  SMOKE
• that will allow basic emissions processing
  capability
• Need new resources and commitment from TCEQ and
  Community

60

1. CMAQ is wave of the future EPAs CMAS effort
2. But, Texas is not ready to jump off from CAMx in
   the middle of developing the SIP revision. 
3. We don't want to duplicate TCEQs CAMx runs
   unless it helps science understanding. 
4. But, still should have capability to cross-check
   the control scenarios with CMAQ.
5. We had hoped to win a funding contract to compare
   CAMx and CMAQ science components in detail, but
   lost out to others due to 5