48
Geographic Scale -The explicit physical scale of an air quality model affects more abstract levels of scale of data inputs Atmospheric Chemistry -The dynamic interaction of chemicals and meteorology Modeling Visibility in the National Parks -Modeling at the regional scale – Long range transport in addition to chemistry Using the Results, Dealing With Uncertainty -Modeling complex systems rarely yields definitive results Tyler Cruickshank Patrick Barickman 4/5/05

Geographic Scale

  • Upload
    sileas

  • View
    47

  • Download
    0

Embed Size (px)

DESCRIPTION

Geographic Scale -The explicit physical scale of an air quality model affects more abstract levels of scale of data inputs. Atmospheric Chemistry -The dynamic interaction of chemicals and meteorology. Modeling Visibility in the National Parks - PowerPoint PPT Presentation

Citation preview

Page 1: Geographic Scale

Geographic Scale -The explicit physical scale of an air quality model affects more abstract levels of scale of data inputsAtmospheric Chemistry -The dynamic interaction of chemicals and meteorology

 

Modeling Visibility in the National Parks -Modeling at the regional scale – Long range transport in addition to chemistry

Using the Results, Dealing With Uncertainty -Modeling complex systems rarely yields definitive results

Tyler CruickshankPatrick Barickman

4/5/05

Page 2: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry ModelingLogan, Utah Logan, Utah

Particulate Air PollutionParticulate Air Pollution

NHNH33

NONO22NONO

NHNH 33+HNO+HNO 33

NH N

H 44NONO 33 83%83%

-21-21°°

250m250m

SOSO22

NONO22+O+O

33 NO NO33

OH+NOOH+NO22 HNO HNO

33

EmissionsEmissions

MeteorologyMeteorology

ChemistryChemistry

A Box Model ApproachA Box Model Approach

Page 3: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

Page 4: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

Dirty pixDirty pix

Page 5: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

News conference pixNews conference pix

Source: M. Lipsett, California Office of Environmental Health Assessment Source: M. Lipsett, California Office of Environmental Health Assessment

Particulate Matter (PMParticulate Matter (PM2.52.5))

Page 6: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

This image shows a magnified view of aerosol particles collected in the industrial city This image shows a magnified view of aerosol particles collected in the industrial city of Port Talbot, England. Many of the particles measure roughly 2.5 microns across, of Port Talbot, England. Many of the particles measure roughly 2.5 microns across, small enough to easily enter and damage human lungs. (Micrograph adapted fromsmall enough to easily enter and damage human lungs. (Micrograph adapted from Sixth Annual UK Review Meeting on Outdoor and Indoor Air Pollution Research Sixth Annual UK Review Meeting on Outdoor and Indoor Air Pollution Research 15th–16th April 2002 (Web Report W12), Leicester, UK, MRC 15th–16th April 2002 (Web Report W12), Leicester, UK, MRC Institute for Environment and Health)

Page 7: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

Logan, Utah Particulate Matter (PM2.5)

0

20

40

60

80

100

120

140

160

1/1/2004 1/7/2004 1/13/2004 1/19/2004 1/25/2004 1/31/2004

Con

cent

ratio

n (u

g/m

3)

2 weeks of violations

Page 8: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

Bad news when your town is a USA today feature!Bad news when your town is a USA today feature!

Page 9: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

Even worse when your town is USA today graphic!Even worse when your town is USA today graphic!

Page 10: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

Me??Me??

Page 11: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

Yes, You!!Yes, You!!

Page 12: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

… … and us …and us …

Page 13: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

Page 14: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

Open ValleyOpen ValleyClosed ValleyClosed Valley

Page 15: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

Brief Chemistry BackgroundBrief Chemistry Background

PMPM2.52.5 typically consists of: typically consists of:

• • Ammonium Nitrate (NHAmmonium Nitrate (NH44NONO33))

• • Ammonium Sulfate (NHAmmonium Sulfate (NH44SOSO44))

• • Organic CarbonOrganic Carbon• • Elemental CarbonElemental Carbon• • Sodium, Potassium, Calcium, …Sodium, Potassium, Calcium, …

Logan PMLogan PM2.52.5 is ~90% Ammonium Nitrate is ~90% Ammonium Nitrate

• • Lots of chemistry going on.Lots of chemistry going on.

Page 16: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

Brief Chemistry BackgroundBrief Chemistry Background

Ammonium Nitrate (NHAmmonium Nitrate (NH44NONO33))

NHNH33 + HNO + HNO33 NHNH44NONO33

(Ammonia + Nitric Acid) Ammonium Nitrate Particle(Ammonia + Nitric Acid) Ammonium Nitrate Particle

AgricultureAgriculture

Chemical transformationChemical transformationof NOof NOxx from Cars/combustion from Cars/combustion

Depending on temperature and humidity Depending on temperature and humidity ammonium nitrate will exist as a:ammonium nitrate will exist as a:1)1) Solid particle (cold and dry)Solid particle (cold and dry)2)2) Aqueous droplet (high humidity)Aqueous droplet (high humidity)3)3) Dissociate into gas (warm)Dissociate into gas (warm)

Page 17: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

Brief Chemistry BackgroundBrief Chemistry Background

In a strong Cache Valley inversion:In a strong Cache Valley inversion:

1)1) Ammonium nitrate concentration rapidly increases when:Ammonium nitrate concentration rapidly increases when:

• • Temperature < 0Temperature < 0°° C (Dissociation) C (Dissociation)

• • Humidity < ~80% (Deliquescence point)Humidity < ~80% (Deliquescence point)

2)2) Ammonium nitrate concentration decreases when:Ammonium nitrate concentration decreases when:

• • Humidity > ~80% (Aqueous droplets larger than PMHumidity > ~80% (Aqueous droplets larger than PM2.52.5))

3)3) Indoor ammonium nitrate concentration rapidly decreases.Indoor ammonium nitrate concentration rapidly decreases.

• • Temperature is warm (> 0Temperature is warm (> 0°° C ) C )

Page 18: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry ModelingWhat Kind OF Model Should I Use?What Kind OF Model Should I Use?

Cache Valley:Cache Valley:• • Interested in secondary chemistry (90% NHInterested in secondary chemistry (90% NH44NONO33))

- NH- NH33 + HNO + HNO33 NHNH44NONO33

- Model must treat this and “lead-up” chemical reactions- Model must treat this and “lead-up” chemical reactions• • Cache Valley is small and contained. Cache Valley is small and contained. • • Transport by wind is limited. Weather data is limited.Transport by wind is limited. Weather data is limited.• • PMPM2.52.5 concentrations uniform throughout valley. concentrations uniform throughout valley.

What is my goal?:What is my goal?:1)1) Represent average airshed conditions.Represent average airshed conditions.

• • Valley concentrations are uniform.Valley concentrations are uniform. • • Don’t need to tie results to widely spaced monitors.Don’t need to tie results to widely spaced monitors.

2) Test NO2) Test NOxx, VOC, NH, VOC, NH33 control strategies control strategies

• • Basic control strategies are needed.Basic control strategies are needed.

Page 19: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

Two model options:Two model options:

1)1) Complex grid based model.Complex grid based model. • • Gridded emissions, meteorology, chemistry.Gridded emissions, meteorology, chemistry. • • Spatially and temporally oriented.Spatially and temporally oriented.

2) Simple box model.2) Simple box model. • • Single volume, no transport, “shake the box”.Single volume, no transport, “shake the box”. • • Provides average conditions.Provides average conditions.

What Kind OF Model Should I Use?What Kind OF Model Should I Use?

Results can be tied to the Logan monitor.Results can be tied to the Logan monitor.Valley concentrations seem mixed and uniform.Valley concentrations seem mixed and uniform.Valley PMValley PM2.52.5 problem tied to NO problem tied to NOxx, VOC, NH, VOC, NH33 dynamics. dynamics.

Page 20: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

Fixed XFixed X

Fixed

Y

Fixed

Y

Emissions InEmissions In

0.0

20.0

40.0

60.0

80.0

100.0

0 24 48 72 96

Hours

CO

nce

ntr

atio

n (

pp

b)

O3 NO2NO O3 NO NO2

February 14-17, 2004#74

Model PredictionModel Prediction

Chemistry PerformedChemistry Performed

NOxNOx

SO4SO4

H2OH2O

NH3NH3

NONO

VOCVOC

NH3NH3

SO4SO4NOxNOx

Box ModelBox Model

Variable Z

Variable Z

Page 21: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

EmissionsEmissions NO2NO2 NONO VOCVOC NH3NH3

PredictionsPredictions NO2NO2 NONO O3O3

Model VolumeModel Volume How big a box?How big a box?

MeteorologyMeteorology TemperatureTemperature RHRH PhotolysisPhotolysis

Are my emissions reasonable?Are my emissions reasonable?Are my emissions reasonable relative to the box size?Are my emissions reasonable relative to the box size?

Meteorology is measured and hence, fixed.Meteorology is measured and hence, fixed.How does the meteorology impact the chemistry?How does the meteorology impact the chemistry?

Represent the whole valley?Represent the whole valley?Just the populated portion?Just the populated portion?

Do my predictions look reasonable?Do my predictions look reasonable?No? Why not? Try again …No? Why not? Try again …

Page 22: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

Model Chemistry – How, exactly, does it work?Model Chemistry – How, exactly, does it work?

InputsInputs::Initial Concentrations: NO, NOInitial Concentrations: NO, NO22, CO, etc ..., CO, etc ...

Hourly Emissions: NO, NOHourly Emissions: NO, NO22, CO, etc …, CO, etc …

ChemistryChemistry::Carbon Bond-IV Chemical MechanismCarbon Bond-IV Chemical Mechanism

- Used for urban smog modeling- Used for urban smog modeling- About 80 different reactions ie.- About 80 different reactions ie.

• • NONO22 + Sunlight NO+O + Sunlight NO+O

•• OO33 + NO O + NO O33

•• NONO22 + O + O33 NO NO33

Page 23: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry ModelingResults !!!Results !!!

0.0

20.0

40.0

60.0

80.0

100.0

0 24 48 72 96

Hours

Co

nce

ntr

atio

n (

pp

b)

Obs O3 Obs NO2Obs NO Mod O3 Mod NO Mod NO2

February 14-17, 2004#72

Page 24: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

0

20

40

60

80

100

0 24 48 72 96

Hours

Co

nce

ntr

atio

n (

pp

b)

Obs NO

Mod NO

February 14-17, 2004#72

NO appears to be progressively depleted.NO appears to be progressively depleted.

Some Problems …Some Problems …

Page 25: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

0

20

40

60

80

100

0 24 48 72 96

Hours

Co

nce

ntr

atio

n (

pp

b)

Obs NO2

Mod NO2

February 14-17, 2004#72

NONO22 appears to be progressively depleted. appears to be progressively depleted.

Some Problems …Some Problems …

Page 26: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

0.0

20.0

40.0

60.0

80.0

100.0

0 24 48 72 96

Hours

Co

nce

ntr

atio

n (

pp

b)

Obs O3

Mod O3

February 14-17, 2004#72

OO33 progressively increases. progressively increases. OO33 depleted at night. depleted at night.

Some Problems …Some Problems …

Page 27: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry ModelingWhat Might Explain My Problems?What Might Explain My Problems?

#1 : NO#1 : NO22 appears to be progressively depleted. appears to be progressively depleted.

• • Is our “box” too big relative to emissions coming in?Is our “box” too big relative to emissions coming in?• • Is there too much photochemistry happening – NOIs there too much photochemistry happening – NO22 depletion? depletion?

#2 : O#2 : O33 progressively increases. progressively increases.

• • Is the “box” not big enough – too much VOC?Is the “box” not big enough – too much VOC?• • Is the photochemistry happening too quickly?Is the photochemistry happening too quickly?• • What else is increasing to allow OWhat else is increasing to allow O33 to build? to build?

#3 : O#3 : O33 depleted at night. depleted at night.

• • Is there a reaction happening too fast?Is there a reaction happening too fast?• • Does subtle meteorology explain the discrepancy?Does subtle meteorology explain the discrepancy?

Page 28: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

I must remember that:I must remember that:

• • The model cannot capture the subtleties.The model cannot capture the subtleties.• • I shouldn’t expect to match the observations exactly.I shouldn’t expect to match the observations exactly.• • The model results represent average airshed conditions.The model results represent average airshed conditions.

What Might Explain My Problems?What Might Explain My Problems?

Considering the above: Considering the above:

• • I want my ratios to be reasonable -I want my ratios to be reasonable - NONOxx:NH:NH33, NO:NO, NO:NO22

• • Trends match observations.Trends match observations.• • Does it make sense?Does it make sense?

Page 29: Geographic Scale

Atmospheric Chemistry ModelingAtmospheric Chemistry Modeling

• • Identify dynamics behind AQ problem.Identify dynamics behind AQ problem.

• • IIdentify scale of AQ problem.dentify scale of AQ problem.

• • Select model as appropriate to above and goals.Select model as appropriate to above and goals.

• • Understand uncertainty.Understand uncertainty.

• • AQ modeling is not a research project – it is applied modeling.AQ modeling is not a research project – it is applied modeling.

What’s the Skinny?What’s the Skinny?

Page 30: Geographic Scale

Atmospheric Chemistry ModelingWhat’s the Skinny?What’s the Skinny?

Identify dynamics behind AQ problem.Identify dynamics behind AQ problem.Identify scale of AQ problem.Identify scale of AQ problem.Select model as appropriate to above and goals.Select model as appropriate to above and goals.Understand uncertainty.Understand uncertainty.AQ modeling is not a research project – it is AppliedAQ modeling is not a research project – it is Applied.

Page 31: Geographic Scale

Geographic Scale -The explicit physical scale of an air quality model affects more abstract levels of scale of data inputsAtmospheric Chemistry -The dynamic interaction of chemicals and meteorology

 

Modeling Visibility in the National Parks -Modeling at the regional scale – Long range transport in addition to chemistry

Using the Results, Dealing With Uncertainty -Modeling complex systems rarely yields definitive results

Tyler CruickshankPatrick Barickman

4/5/05

Page 32: Geographic Scale

 

http://www.wrapair.org/

Page 33: Geographic Scale

 

http://coha.dri.edu/index.html

Page 34: Geographic Scale

 

Overall Goal of the Regional Haze Rule (CAA, 1990)

Page 35: Geographic Scale

 

Page 36: Geographic Scale

 

MM5 Modeling at the Continental Scale

Page 37: Geographic Scale

 

Emissions Inventory

Page 38: Geographic Scale

 

What is the model trying to evaluate?

Light Extinction (Bext): The attenuation of light due to scattering and absorption as it passes through a medium Benefit:  Light extinction can be directly related to gaseous and aerosol concentrations.Drawback:  Light extinction is non-linearly related to a person perception of changes in haze.

How is it done?The model converts concentrations of pollutants into the extinction coefficients These are based on known relationships between the type of particle and its effect on visibility

How is the evaluation made?

Page 39: Geographic Scale

 

http://vista.cira.colostate.edu/views/Observations in the National Parks and Wilderness Areas

Page 40: Geographic Scale

Geographic Scale -The explicit physical scale of an air quality model affects more abstract levels of scale of data inputsAtmospheric Chemistry -The dynamic interaction of chemicals and meteorology

 

Modeling Visibility in the National Parks -Modeling at the regional scale – Long range transport in addition to chemistry

Using the Results, Dealing With Uncertainty -Modeling complex systems rarely yields definitive results

Tyler CruickshankPatrick Barickman

4/5/05

Page 41: Geographic Scale

 

Page 42: Geographic Scale

 

Page 43: Geographic Scale

 

The Weight of Evidence Approach

• The Causes of Haze• Data for each state’s Regional Haze State Implementation Plan (RH SIP)• Sulfates and Nitrates • Uncertainty in the model results require supporting evidence

Two models used•Tagged Species Source Apportionment (TSSA)•Trajectory Regression Attribution Method

Page 44: Geographic Scale

 

Back Trajectory Analysis

Page 45: Geographic Scale

 

Utah Emissions Inventory•S02 + NH3 = SO4

•Nox + NH3 = NH3

Page 46: Geographic Scale

 

Weight of EvidenceThis is a descriptive approach that is narrative and qualitative rather than purely quantitative

Page 47: Geographic Scale

Tyler CruickshankPatrick Barickman

4/5/05

Summary

• Complex System – “One not describable by a single rule. Structure exists on many scales… not reducible to only one level of

description.” http://www.calresco.org/glossary.htm

• Usually not possible to give definitive answers – a conceptual model puts the results in context

Page 48: Geographic Scale

Atmospheric Chemistry Modeling