atmospheric chemistry module for tics · an atmospheric chemistry module for modeling toxic...

An Atmospheric Chemistry Module for Modeling Toxic Industrial Chemicals

(TICs) in SCIPUFF

Douglas S Burns, Veeradej Chynwat, Jeffrey J Piotrowski, Kia Tavares, and Floyd Wiseman

ENSCO, Inc.

Science and Technology for Chem-Bio Information Systems(S&T CBIS)

October 28, 2005

BAA TYN 03-001Atmospheric Chemistry Module for

Toxic Industrial Chemicals

Tyndall AFB / DTRA

Michael HenleyAFRL/MLQ Tyndall AFB, FL

Martin BagleyDTRA / TDOC Alexandria, VA

Outline

• Project Goals• Methodology

– Integration in SCIPUFF– Chemistry of 1-butene– Derivation of keff, Xeff– Parameter Space

• Results– Model output

• Decay of TICs (1-Butene, Methylpropene)• Product Formation

• Summary

Project Goals• Develop initial atmospheric chemistry capability

– Develop Atmospheric Chemistry Algorithm• Algorithm MUST run rapidly.• Develop generic algorithm so that a detailed chemical kinetics approach

is not required.• Algorithm must account for all (most) modeling scenarios (e.g., CC, T,

ambient conditions).• Algorithm must be robust enough to account for diurnal changes to

degradation rates.• Algorithm should account for the potential generation of intermediate

toxic compounds.– Develop Chemical data for the Chemistry Algorithm

• Review existing chemistry data for nine alkenes (and H2S)• Develop mechanisms used to generate chemistry algorithm.

• Couple Algorithm to SCIPUFF– Work with Dr. Sykes to create interface with SCIPUFF

• Launch Chemistry Module from HPAC

Methodology: Minor Modification to SCIPUFF

SCIPUFF Dispersion

−= ( )(min)(max) ),sin(max AAA kkk φ=

∫=.),,,],[,( etchumidityccTambfkk effA φ==

Method: Create Degrade Dynamic Link LibraryDetails in the Software Development Plan

API

SCIPUFF Function Library Plottool

Interface Functions

FileManager

PlotManager

ProjectManager

Dispersion

MeteorologyGenerator

SWIFT

MC-

SCIPU

FF

FileReaders

UtilityFunctions

PlotGenerator

EffectsManager

EffectM

odule

EffectM

odule

DEGRADE DLL

• Algorithm is transparent to the User

• Code alwayscalls chemistry

Methodology: Chemistry of 1-butene

Determine ID Rxn’s, EA, k(T)

(w/ OH, NO3, H2O, O3, etc.)

Implement inDetailed

Mechanism

Run PBM as

f(met parm’s)

Obtain cTIC(t) as

f(met parm’s)

Derive Empiricalkeff (met parm’s)

Populate SCIPUFF data tables w/

keff for TICs

Methodology: Chemistry of 1-buteneCH2=CHCH2CH3 + OH ⎯⎯→ 0.94 C2H5CHO + Other products

CH2=CHCH2CH3 + NO3 ⎯⎯→ 0.12 C2H5CHO + 0.6 C3H5O-CH-O-NO2+ Other products

CH2=CHCH2CH3 + O3 ⎯⎯→ 0.35 C2H5CHO + 0.41 OH + Other products

Rate = -(kOH[OH] + kNO3[NO3] + kO3[O3]) [1-butene]

Rate = -keff [1-butene]

Methodology: Chemistry of 1-butene

Determine ID Rxn’s, EA, k(T)

(w/ OH, NO3, H2O, O3, etc.)

Implement inDetailed

Mechanism

Run PBM as

f(met parm’s)

Obtain cTIC(t) as

f(met parm’s)

Derive Empiricalkeff (met parm’s)

Populate SCIPUFF data tables w/

keff for TICs

krxn as f(T)

5

10

15

20

25

260 270 280 290 300 310 320

Temperature [K]

k NO

3 and

kO

3 [pp

m-1

min

-1]

40000

45000

50000

55000

60000

k OH [p

pm-1

min

-1]

kO3 * 10^3kNO3kOH

Methodology: Detailed Mechanism

Determine ID Rxn’s, EA, k(T)

(w/ OH, NO3, H2O, O3, etc.)

Implement inDetailed

Mechanism

Run PBM as

f(met parm’s)

Obtain cTIC(t) as

f(met parm’s)

Derive Empiricalkeff (met parm’s)

Populate SCIPUFF data tables w/

keff for TICs

• Carbon Bond Mechanism– Mass consistent atmospheric chemistry

mechanism.– EPA Model (Adelman, 1999).– Used to model the ambient conditions.– [OH], [NO3], [O3], NOX, VOCs, etc..

• Append chemistry for TIC– Data provided for 9 alkenes and H2S

Methodology: Run Detailed Chemistry

Implement inDetailed

Mechanism

Run PBM as

f(met parm’s)

Obtain cTIC(t) as

f(met parm’s)

Derive Empiricalkeff (met parm’s)

Populate SCIPUFF data tables w/

keff for TICs

or

NOX VOC’s

O3 CO

Location (lat,long)

ZA = f(Lat, Lon, DOY, Time of day) H2O

Temperature

keff is a function of solar elevation, cloud cover, air quality, temperature, humidity, etc

Methodology: Parameter Space

Parameter Units SCIPUFFSolar Zenith Angle 0 – 90 Deg X

Location (lat, lon) 0 – 70 Deg X

Time of Day 1440 min X

Day of Year 3/21, 6/20, 12/20 X

Photochemistry (Cloud Cover) 0 – 8 Eighths X

Temperature 230 – 310 K X

Water Concentration 100 – 40000 PPM

Moisture Mixing ratio X

Air Quality [NOX], VOC, O3, …

Land Use Urban, ocean, forest, … X

Methodology: Surrogate for Air Quality• Land Use 1=Developed

2=Dry Cropland & pasture3=Irrigated Cropland5=Cropland/Grassland6=Cropland/Woodland7=Grassland8=Shrubland9=Shrubland/Grassland10=Savanna11=Deciduous Broadleaf12=Deciduous Needleleaf13=Evergreen Broadleaf

14=Evergreen Needleleaf15=Mixed Forest16=Water17=Herbaceous Wetland18=Wooded Wetland19=Barren20=Herbaceous Tundra21=Wooded Tundra22=Mixed Tundra23=Bare Tundra24=Snow or Ice25=Partly Developed

1001=Urban Superclass1002=Grassland Superclass1003=Forest Superclass1004=Desert Superclass1005=Water Superclass

Methodology: Surrogate for Air QualityNOx vs VOC (vary by Latitude)

(Mar, Jun, Dec, 2000, T = 280K, CC = 0, Lat 0-60)

1.00E-06

1.00E-05

1.00E-04

1.00E-03

1.00E-02

1.00E-01

1.00E+00

0.001 0.01 0.1 1

Propene-Eq (ppm-C)

NO

x (p

pm)

Water

Forest

Urban

Grassland

Desert

Default

Methodology: Refined Parameter Space (T, H2O)

• Surface Stations Nov 2003 – Sep 2004.• Global 0.5 km LU Data Set

Temperature (K) [H2O] (x103) ppm)Min12.47.054.55

30 274 310 3.81 34.940 265 304 1.54 28.550 257 299 1.02 19.860 245 294 0.400 14.270 231 291 0.113 11.6

LatitudeMin Max Max

0 288288288

310 37.110 310 37.420 310 37.1

1. Extracted data using 3 hr interval instead of 30 sec data. (both day and night)

2. Removed extreme data points (i.e., T<-60 °C or T<Dew point).

3. Matched weather station data with LU data before analysis (5 categories).

Methodology: Run Detailed Chemistry

[ ][ ] [ ][ ] [ ][ ] [ ] .....33 33−−−−−=⎟

⎠⎞

⎜⎝⎛

∂∂

−= iiOiNOiOHChemistry

ii ckcOkcNOkcOHk

tcr

Implement inDetailed

Mechanism

Run PBM as

f(met parm’s)

Obtain cTIC(t) as

f(met parm’s)

Derive Empiricalkeff (met parm’s)

Populate SCIPUFF data tables w/

keff for TICs

[ ]ieffchemistry

i ckdtdc

=⎟⎠⎞

⎜⎝⎛−

[ ]cdtdc

keff

−=

Methodology: Obtain CTIC as f(t)[ ]

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.10

0 360 720 1080 1440

20253035404550

Implement inDetailed

Mechanism

Run PBM as

f(met parm’s)

Obtain cTIC(t) as

f(met parm’s)

Derive Empiricalkeff (met parm’s)

Populate SCIPUFF data tables w/

keff for TICs

[ ]cdtdc

keff

−=

[Butene] as f(time)Latitude = 20 – 50º N

20º

50ºT = 290 K, Land Use = Urban

Time of Day (min from midnight)

[But

ene]

[PPM

]

Methodology: Obtain keff as f(met parms)

0.0000

0.0004

0.0008

0.0012

0.0016

0.0020

0 360 720 1080 1440

253035404550

Implement inDetailed

Mechanism

Run PBM as

f(met parm’s)

Obtain cTIC(t) as

f(met parm’s)

Derive Empiricalkeff (met parm’s)

Populate SCIPUFF data tables w/

keff for TICs

[ ]cdtdc

keff

−=

T = 291 K, Land Use = Urban

25º

50º

keff as f(time)

k eff

[min

-1]

Time of Day (min from midnight)

Methodology: Derive Empirical keff

• Generate keff for various combinations of meteorological parameters for each land use

• Transform data to center on all parameters• Perform statistical regression - correlation

– Review Equation– Review Statistical Parameters (e.g., r2)– Weigh fit vs number of parameters

• Derive an empirical keff = f(SE, T, lat, tod, CC, [H2O])• Compare the keff (empirical model) with the PBM derived keff.

Obtain cTIC(t) as

f(met parm’s)

Derive Empiricalkeff (met parm’s)

Populate SCIPUFF data tables w/

keff for TICs

Results: keff (polynomial) vs keff (PBM) for butene

Model: 7 Parametersr2 = 0.95

keff [min-1] from PBM

Land Use = Urbank e

ff[m

in-1

] fro

m P

olyn

omia

l

> 500,000 points

Results: keff (polynomial) vs keff (PBM) for butene

Land Use = Grassk e

ff[m

in-1

]

PBM

Polynomial

Lat 0°, Temp 300 K, Cloud Cover 0/8, [H2O] = 20000 ppm,

Time from Solar Noon [min]

1.50x10-3 min-1

1.43x10-3 min-1

Results: keff (polynomial) vs keff (PBM) for butene

Land Use = WaterLat 0°, Temp 300 K, Cloud Cover 0/8, [H2O] = 20000 ppm,

k eff

[min

-1]

PBM

Polynomial

Time from Solar Noon [min]

9.3x10-4 min-1

9.2x10-4 min-1

Methodology: Obtain Xeff

CH2=CHCH2CH3 + OH ⎯⎯→ 0.94 C2H5CHO + Other products

CH2=CHCH2CH3 + NO3 ⎯⎯→ 0.12 C2H5CHO + 0.6 C3H5O-CH-O-NO2+ Other products

CH2=CHCH2CH3 + O3 ⎯⎯→ 0.35 C2H5CHO + 0.41 OH + Other products

Rate = -(kOH[OH] + kNO3[NO3] + kO3[O3]) [1-butene]

Rate = -keff [1-butene]

Rate = +(0.94 kOH[OH] + 0.12 kNO3[NO3] + 0.35 kO3[O3])[butene]

Rate = +Xeff keff [1-butene]

Methodology: Obtain XeffT = 295 K, Land Use = Water

Stoichiometry for Propanal Formation

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 360 720 1080 1440

Time from Midnight [Min]

Stoi

chio

met

ric C

oeffi

cien

t [-]

OH

O3

NO3

Results: Nine Alkenes

• Priority I– 1-Butene

• Products (Propanal, Nitroxybutanone).– Ethene– Propene– Methylpropene– 1,3-Butadiene

• Priority II– Styrene

• Priority III– cis-2-Butene– trans-2-Butene– Isoprene

Why Chemistry is Important

in AT&D Modeling

Results: T&D Compared to T&D + Chemistry (Butene)

T&D Only T&D + Chemistry

SCIPUFF

Tracer 1-Butene

Results: Methylpropene8 hr continuous release starting at 8 am local time

No Chemistry

Chemistry

8 AM9 AM10 AM11 AM12 PM1 PM2 PM3 PM Local Time

Results: Calculated Plume is TIC Dependent

3-Methyl Propene

1,3-Butadiene

Ethene Propene

Results: TIC Decay and Product Formation

1-Butene Propanal

At 4 hrs and 8 hrs after release

2 hr continuous release starting at noon local time

Results: Test and Evaluate (Output)Comparison of Original SCIPUFF and “Degrade”

kmin / kmax

keff (met parms)

Diffusion

Summary & Future Work

• Developed Chemistry Model for 10 TICs– 9 Alkenes + H2S

• No slow down in SCIPUFF• Ability to model product formation• Future Work

– Site specific keff’s– keff’s for other TICs

• Complementary lab / theory development of fundamental kOH, kO3, kH2O, etc.

– Chamber Studies (Chemistry Validation)– Field Studies (Model Validation)

End of slides

Example: Crop Dusting Scenario – Bottom line

03 July 06:00 (Local) 13:00 Z 03 July 07:00 (Local) 14:00 Z 03 July 08:00 (Local) 15:00 Z 03 July 09:00 (Local) 16:00 Z 03 July 10:00 (Local) 17:00 Z 03 July 11:00 (Local) 18:00 Z 03 July 12:00 (Local) 19:00 Z 03 July 13:00 (Local) 20:00 Z 03 July 14:00 (Local) 21:00 Z 03 July 15:00 (Local) 22:00 Z 03 July 16:00 (Local) 23:00 Z 03 July 17:00 (Local) 00:00 Z 03 July 18:00 (Local) 01:00 Z 03 July 19:00 (Local) 02:00 Z 03 July 20:00 (Local) 03:00 Z 03 July 21:00 (Local) 04:00 Z 03 July 22:00 (Local) 05:00 Z

Chemistry vs T&D

Degrade Generation