Title: Modeling Air Quality in Houston, Texas
1Modeling Air Quality in Houston, Texas
- Byeong-Uk Kim
- November 17, 2004
- Department of Environmental Sciences and
Engineering - University of North Carolina at Chapel Hill
2Content
- Background
- Uniqueness of Houston
- Air Quality Modeling for Houston
- Dissertation
3Background
- Is ozone bad or good?
- Ozone formation science
- Non-attainment area and State Implementation Plan
(SIP) - Difficulties in developing control strategy
- Non-linearity
- Ozone transport
- Photochemical Air Quality Modeling (PAQM) system
4Is ozone bad or good?
http//www.epa.gov/oar/oaqps/gooduphigh/
5Ozone formation science
Jeffries, 1993
6Ozone formation science
- Where do we get the precursors?
- NOx (NONO2)
- Automobiles
- Power plants
- VOCs (Volatile Organic Compounds)
- Automobiles
- Refineries
- Trees
- Other area sources
7Non-attainment and SIP
- Non-attainment area
- NAAQS for ozone daily maximum one hour average
concentration lt 0.12 ppm - fourth highest reading in 3 years gt 125 ppb is a
violation at that monitor - highest violation in 3 years is the design
value for control - State Implementation Plan (SIP)
- Control strategy development
- Attainment demonstration
NAAQS National Ambient Air Quality Standards
8Difficulties in developing control strategy
Non-linearity
Philadelphia
Chicago
Blanchard and Reynolds, 2002
9Difficulties in developing control strategy
Ozone transport
100km Travel in 6 hrs 100km between New York
City- Philadelphia
10Formulation of PAQM
Jeffries, 1993
11Operation of PAQM
12Operation of PAQM
13Operation of PAQM
TCEQ developed 9114 new emission point speciation
profiles
From Byun, UH
14Reported EI
- Normal Non-EGU VOC EI with Special EI additions
- hg_02km.tx_negu_si4a Total Point Source CB-IV HC
Emissions, 08/25/2000 - Base inventory of 240 T/D
- Nearly constant emissions
TCEQ Cantu, 2002
15SIP modeling
Base case explain current ozone with current
inventory
Future case predict ozone with growth and
existing regulations.
Future controlled case predict ozone with
growth, existing regulations, and newly proposed
regulations.
All three cases use the same episodic meteorology.
From Byun, UH
16Uniqueness of Houston
- Transient High Ozone Events (THOEs)
- Large variability of VOC emissions
- Different Modeling Approach Needed
- Complex meteorological conditions
17Transient High Ozone Event in Houston
Jeffries, 2004
18gt10,000 lbs/hr ethylene release at La Porte,
(6700 lbs between 1100 AM and 1125 AM)
3/27/2002
19Short term ozone enhancements of up to 100 ppb
20Large Variability of VOC Emissions
- Permitted emission components
- Nearly Constant
- Routinely Variable
- Allowable Episodic
- Off-permit event emission releases
- Start-up, shut-down, maintenance operations,
accidents, emergencies - Subject to RQ value
Webster, 2003
21Different Modeling Approach Needed
Jeffries, 2003
22Complex Meteorological Conditions
http//www.islandnet.com/see/weather/elements/sea
brz.htm
23Complex Meteorological Conditions
GOES Geostationary Operational Environmental
Satellites
24Complex Meteorological Conditions
August 31, 2000 Estimated Mixing Height by Kv
profile
1000 AM
0200 PM
0600 PM
Land/sea breeze does not happen clearly all the
time. The synoptic weather pattern plays a key
role.
25Air Quality Modeling for Houston
- Modeling Domain
- General Description
- Base case results
- Future case results
26Air Quality Modeling for Houston
27Air Quality Modeling for Houston
- UNC Beowulf servers
- More than 150 nodes, MPI support
- CAMx (Comprehensive Air quality Model with
eXtensions Ver. 4.x) - Serial/OpenMP
- Modeling period
- August 22, 2000 September 6, 2000
- Run statistics
- About 4 hours clock time for an episodic day on
an Intel 2.8GHz CPU node - About 2 GB size of outputs per episodic day
28Base case results
Imputed base
Not imputed (max 112 ppb) Event
29Base case results
30Future case results
Without Event Imputed future
With Event Imputed future
31Dissertation
- Does this model show or have all necessary
components to produce the phenomena that we can
expect from the current best perceptual/conceptual
model? - Can this model distinguish what precursor to
control for ozone reduction? - Does it estimate the control requirement
unambiguously? - What are the possible biases in the prediction
and its impact on the policy choice? - The answers to these questions may vary
spatially over the modeling domain and temporally
during the modeling period.
32Why do we care those questions?
- The current SIP framework requires the use of
PAQM for the control strategy developments. - Once developed, the control strategies will
influence our daily life. Some of them are also
very expensive to implement. - Highway speed limit change
- Restriction of the construction equipments
- Installation of NOx control device
- It may consume all important material for the
device that is available in the USA to make the
demanded amount of devices for a region. - What if these are not the primary reason for the
ozone violation in your area?
33Lots of analysis and tools
- Model performance analysis
- Potential reliability of the modeling results for
the policy use - Area of influence analysis
- Source-receptor relationships
- Uncertainty/Sensitivity analysis
- Identification of the important
process/parameter/input - Process analysis
- Models internal ozone concentration by process
contributions