Recent Developments in the Community Emissions Model CONCEPT

Recent Developments in the Community

Emissions Model CONCEPT

5th Annual CMAS Model User Conference

Tuesday October 17, 2006Mark Janssen

LADCO

Background

• CONsolidated Community Emissions Processing Tool - CONCEPT.

• PostgreSQL, Perl, PostGIS. Some fortran for M6, Nonroad, CMAQ

• Build with a community in mind: easy to read/modify and customize QA.

• Now operational and will be part of LADCO/MWRPO SIP.

Models in CONCEPTAnd what makes them different

• Industrial Point- advanced temporal

• Area- advanced temporal

• Nonroad- integrated NONROAD, day specific meteorology

• Biogenics – BEIS2/BEIS3/MEGAN

• NH3- Processed based model (NAS recommendation)

• Onroad- Complex temporal, Link Based

If you want a modeling system that needs extensive and flexible QA

and is foundation community model

Onroad Motor Vehicle• Link based onroad model with embeded version

of mobile6. • T3 – Conversion tool reads many travel demand

models makes them usable by CONCEPT• Handle very complex temporal(month, day,

hour, ) by road and vehicle type for VMT, Speed, Mix.

• Adjust speed based on volume after temporalization

• Flexible speed, temperature bins for accelerated M6 Processing

• Toxics being added by USEPA

LADCO 2002 Networks Data Processed State Network TDM # Links

Illinois CATS - Chicago Area EMME2 33,774

Illinois ILDOT Statewide Generated from Observed 303,297

Indiana MPO - Indianapolis TransCAD 7,599

Indiana NIRPC - Northwest Indiana EMME2 9,023

Indiana INDOT Statewide TransCAD 31,181

Michigan SEMCOG - Detroit Area TransCAD 14,985

Michigan MIDOT Statewide TransCAD 9,227

Minnesota MMC - Minneapolis St. Paul Area TP+ 20,898

Missouri EWGCOG – St. Louis Area CUBE - TP+/Voyager 40,394

Minnesota MNDOT Statewide Generated from Observed 4,402

Ohio Nine Urban Areas CUBE-TRANPLAN 3,723 to 29,796

Ohio OHDOT Statewide TransCAD 50,644

Wisconsin SEWRPC - Milwaukee Area TRANPLAN 17,054

Wisconsin WIDOT Statewide TP+/VIPER 143,327

Total Links 801,467

Example of Complex TemporalMichigan Urban Interstate Hourly Profiles, Sunday through Saturday

0

0.02

0.04

0.06

0.08

0.1

1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 13 17 21

Hour

Fra

cti

on

of

To

tal T

raff

ic

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

Example Vehicle Mix Temporal Profile 

Michigan Urban Interstate Hourly VMT Mix Fractions, Sunday through Saturday

0

0.1

0.2

0.3

0.4

0.5

0.6

1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 13 17 21

HDDV HDGV LDDT LDDV LDGT1 LDGT2 LDGV MC

LADCO NetworksAll of region 5

EGU Temporal based on CEM• Gold Standard in Emissions, Direct

measurement

• Use the Hourly CEM Data to derive more represenative EGU inventories.

• Automated tool other users can use not a one-off analysis. Create complex NIF3.0

• Need profiles because actual hourly not usable in future years.

EGU Temporal based on CEMContinued…

• Build hourly weekday, Saturday, Sunday for each month.

• Why Not use: Large Data, What is Typical, Matching CEM to the NEI.

• CEM Analysis tool will ship with CONCEPT v.66

• Use CONCEPT’s temporal processor that can use the complex temporal NIF allows.

Ratio of July weekday EGU NOx with and without CEM based temporal allocation

July 12, 2002 hourly NOx for Georgia EGUs

July 12, 2002 hourly NOx for Illinois EGUs

Significant weekday/weekend variation is apparent with CEM based temporal allocation

Species BIOME MEGANALD2 9,720 11,310CG3 0 52,995CO 0 16,524ETOH 0 3,880FORM 0 991ISOP 104,655 75,294MEOH 0 14,017NOX 3,101 351NR 1,584 2,221OLE 6,409 3,150PAR 46,372 17,346SQT 0 7,259TOL 0 208TRP 5,670 28,431

Biogenic Emissions: MEGAN

• July 3, 2002 episode day

• Total emissions (TPD) from: – EMS/BIOME (BEIS3)

and– CONCEPT/MEGAN

models

• Less Isoprene using MEGAN model

0

20000000

40000000

60000000

80000000

100000000

120000000

140000000

160000000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

BIOME MEGAN

Isoprene:Difference plot of BIOME(BEIS3) – MEGAN

Hourly isoprene emissions (right)

Species BIOME MEGANALD2 9,720 11,310CG3 1,846 52,995CO 0 16,524ETOH 0 3,880FORM 0 991ISOP 104,655 75,294MEOH 0 14,017NOX 3,101 351NR 1,584 2,221OLE 6,409 3,150PAR 46,372 17,346SQT 5,332 7,259TOL 0 208TRP 3,961 28,431

Common Hybrid Approach

CAMx SOA pre-cursor species: CG3, SQT, and TRP

The application of a BEIS post-processor to “estimate” SOA pre-cursor species are shown in the table on the right in bold

Much higher PM SOA pre-cursor emissions using MEGAN

Agricultural Ammonia

• Processed based NH3 model based on recommendations from National Academy of Science.

• Model runs every farm and requires complex input data like N content of feed, barn color.

What We are Trying to Model?

Actual Farm Data

Animal Allocation Processor

(AAP)

Default Data Tables

Combined Real and Synthetic Farms by Grid Cell

Farm Emissions Model(FEM)

Farm Emissions Model Output

EPA State Animal Populations

EPA County Animal Population Distributions

EPA State-Level MMT Distributions

County-Level Fertilizer Data

Fertilizer Data by Grid Cell

Customized MMT Data

Ammonia Emission Rate from Swine Housing (two days in February)

NH3-Next Steps

• Poultry Housing Model science not complete

• May years of validation work because there are limited field measurements.

• Model too slow to be operational( months to run 7 days nationally.

• Speed up the model. C, perl, fortran• PC single facility version so people doing

validation studies can test tool.

Performance Problems

• We have a model that works, but is slow• Link based networks, biogenics, nh3 model. • Projects underway to improve them. • Example is CAMX/CMAQ writer that used to

take 12 hours, now runs in 67 seconds. • Opinion: Better to have a transparent model that

is slow than a fast black box. Can fix speed problems where needed.

• If speed was a primary concern we would all be running CAMX not CMAQ.

Thanks to Community!

• Environ (Pollack, Haasbeek, Shepard, et al) Onroad Networks

• Alpine Geophysics(Wilkinson) MEGAN • NCAR(Alex Guenther) MEGAN• LADCO(Baker) - MEGAN • MI-DEQ(Edick) – CEM/EGU• UC-Davis(Zhang, et al) NH3 Model• UC- Riverside(Wang, et al) NH3 Model • Users who make sure the model works