much ado about emission modeling

Much ado about emission modeling …

Sigh no more, [colleagues], sigh no more,[Emissions] were deceivers ever,One foot in sea and one on shore,

To one thing constant never.*

*w/ apologies to Wm. Shakespeare, from Much Ado About Nothing.

T. PierceAtmospheric Modeling Division/NOAA

U.S. EPA/Office of Research and DevelopmentResearch Triangle Park, North Carolina

Much ado about emission modeling support and research …

• AMD’s role: To obtain the best available emission estimate data for regional air quality modeling; to improve these estimates by building emission models that account for meteorological conditions; and, to develop innovative ways of evaluating emissions.

• AMD staff: W. Benjey, T. Pierce, G. Pouliot (w/ contributions from J. Ching, D. Gillette, A. Gilliland, G. Gipson, P. Bhave, and J. Godowitch)

• Outside collaborators: EPA/OAQPS, EPA/ORD, RPOs, States/EIIP, CSC, USFS, IGAC/GEIA, Environment Canada, NCAR, and university researchers.

• Selected R&D areas: SMOKE, geographical data files, air quality forecasting, biogenic emissions, sea salt, fugitive dust, and NH3 inverse modeling

EPA’s National Emissions Inventory

PreliminaryNEI for Base Year 20XX

States & Other

Stakeholders

Improved NEI for Base Year 20XX

Local Activity

Levels & Variables

Factorand Model

Improvements

Defaults for Emissions

Related Variables

Databases for Source

Activity Levels

Existing Point Source

Data

Growth Factors for

Some Categories

Emission Factors and

Models

Refining & ImprovingInputs(Process Repeated Yearly - Emphasized every 3 Years)

Starting Point for NEI

Emissions Processor

(eg SMOKE)

Input to AQ Model

State / Local / Tribe Improvements

Speciationfactors

Speciationfactors

AMD’s role

Courtesy T. Pace (EPA/OAQPS)

Much ado about emission modeling …

Purpose:

• to report on emission focus areas w/in AMD• to suggest topics for future research• to encourage collaboration

Outline:

• Ammonia• Biogenic emissions• Fugitive dust• Air quality forecasting• Wildfires

Much ado about ammonia emissions …

• 1994 – Battye et al., Development and selection of ammonia factors (EPA work assignment, W. Benjey)

• 1999 – Pierce and Bender, Examining the temporal variability of ammonia and nitric oxide emissions from agricultural processes (www.epa.gov/asmdnerl/awma_ei99.pdf)

• 2000 – CMU NH3 emissions model (www.cmu.edu/ammonia)

• 2003 – Gilliland et al., Seasonal NH3 emission estimates for the eastern United States based on ammonium wet concentrations and an inverse modeling method, J. Geophys. Res., 108 (D15), 4477, doi:10.1029/2002JD003063.

• 2004 – National Emission Inventory - Ammonia Emissions from Animal Husbandry Operations, Draft Report (www.epa.gov/ttn/chief/net/2002inventory.html)

• 2004 – Battye and Barrows, Review of ammonia emission modeling techniques for natural landscapes and fertilized soils, Draft Report (avblfor review at www.epa.gov/ttn/chief)

http://www.cmu.edu/ammonia

Much ado about ammonia emissions …

1998 National Emission Inventory for the U.S.

Source NH3 emissions (x1000 tons)

Animal husbandry 3,520

Fertilizer 724

Industrial 691

Natural 0

Total 4,935

National Emission Inventory - Ammonia Emissions from Animal Husbandry Operations, Draft Report (2004)

33% reduction in NH3 emissions

Review of Ammonia Emission Modeling Techniques for Natural Landscapes and Fertilized Soils, Draft Report

Recommended Default Emission Factors for Natural Landscapes

Vegetation

Emissionfactors

(ng m-2 s-1)

Estimated emissions incontinentalUS (Gg/yr)

Forests 1.2 58

Grasslands 0.9 32

Shrublands 1.3 46

Deserts 0.3 <1

Total 137

=> ~5% of total U.S. inventory


Proposed temporal adjustment factors for NH3 emissions from natural landscapes

Proposed seasonal

allocation:Spring = 0.143

Summer = 0.714Autumn = 0.143Winter = 0.000


Proposed temporal adjustment factors for NH3 emissions from fertilized soils

Proposed seasonal

allocation:Use CMU’s county-

wide, monthly values

Much ado about biogenic emissions …

• Biogenic Emissions Inventory System (BEIS): – introduced by AMD in 1988 to estimate VOC emissions

from vegetation and NO emissions from soils

• BEIS3.09:– current default version used for regulatory modeling– 1-km vegetation database (BELD3) – emission factors for isoprene, monoterpenes, OVOCs,

and nitric oxide (NO) – environmental corrections for temperature and solar

radiation (isoprene only) – speciation factors for the CBIV, RADM2, and SAPRC99

mechanisms


• BEIS3.12:– current research version of CMAQ and in 2001 NEI– emission factors for 34 chemicals, including 14

monoterpenes and methanol– MBO, methanol, isoprene modulated by solar radiation – soil NO dependent on soil moisture, crop canopy

coverage, and fertilizer application– avbl as a module to SMOKE and can be downloaded

from www.epa.gov/asmdnerl/biogen.html

http://www.epa.gov/asmdnerl/biogen.html

Creating Gridded Vegetation Data for the Biogenic Emissions Landuse Database (BELD3)

USGS North AmericanLand Characteristics Data

(1 km, 21 broad classes, satellite-derived)1990 US Census

(urbanized boundaries)1992 US

Agricultural Census(county-level)

Assign urban/rural/water identifier (1-km grid)

Assign crop areas to 1-km grid cells in a county, using

USGS class and urban/rural IDs

Derive crop species distribution and total

crop area (county-level)

Assign forest areasto 1-km grid cells

Zhu & Evans (1994)US forest density data(1 km, satellite-derived)

Assign tree species distribution to

1-km grid forest areas

BELD3 – a 1-km hybrid dataset with crop types, tree species,

urban/rural IDs,and 19 USGS classes

Develop tree species distribution

within county

US Forest Inventory and Analysis (FIA) data

(county-level, 300,000 survey plots)

USGS_drycropUSGS_irrcropUSGS_cropgrassUSGS_cropwdlndUSGS_grasslandUSGS_shrublandUSGS_shrubgrassUSGS_savannaUSGS_decidforesUSGS_evbrdleafUSGS_coniferforUSGS_mxforestUSGS_waterUSGS_wetwoodsUSGS_sprsbarrenUSGS_woodtundrUSGS_mxtundraUSGS_snowiceUSGS_urbanAlfalfaBarleyCornCottonGrassHayMisc_cropOatsPasturePeanutsPotatoesRiceRyeSorghumSoybeansTobaccoWheatAcaciaAilanthusAlderApple

AshBasswoodBeechBirchBumelia_gumCajeputCalifor-laurelCascara-buckthorCastaneaCatalpaCedar_chamaecypCedar_thujaChestnut_buckeyChinaberryCypress_cupressCypress_taxodiuDogwoodDouglas_firEast_hophornbeaElderElmEucalyptusFir_balsamFir_CA_redFir_corkbarkFir_fraserFir_grandFir_nobleFir_Pacf_silverFir_SantaLuciaFir_Shasta_redFir_sppFir_subalpineFir_whiteGleditsia_locusHackberryHawthorn

HemlockHickoryHolly_AmericanHornbeamIncense_cedarJuniper KY_coffeetreeLarch Loblolly_bayMadroneMagnoliaMahoganyMaple_bigleafMaple_bigtoothMaple_blackMaple_boxelderMaple_FLMaple_mtnMaple_NorwayMaple_redMaple_RkyMtnMaple_silverMaple_sppMaple_stripedMaple_sugarMesquiteMisc-hardwoodsMixed_coniferMountain_ashMulberryNyssaOak_AZ_whiteOak_bearOak_black Oak_blackjackOak_blueOak_bluejackOak_burOak_CA_black

Oak_CA_liveOak_CA_whiteOak_canyon_liveOak_chestnutOak_chinkapinOak_delta_postOak_DurandOak_EmeryOak_EngelmannOak_evergreenOak_GambelOak_interio_livOak_laurelOak_live Oak_MexicanblueOak_Northrn_pinOak_Northrn_redOak_nuttallOak_OR_whiteOak_overcupOak_pinOak_postOak_scarletOak_scrubOak_shingleOak_Shumrd_redOak_silverleafOak_Southrn_redOak_sppOak_swamp_cnutOak_swamp_redOak_swamp_whiteOak_turkeyOak_waterOak_whiteOak_willowOsage-orangePaulowniaPawpaw

PersimmonPine_ApachePine_AustrianPine_AZPine_BishopPine_blackjackPine_brstlconePine_chihuahuaPine_CoulterPine_diggerPine_EwhitePine_foxtailPine_jackPine_JeffreyPine_knobconePine_limberPine_loblollyPine_lodgepolePine_longleafPine_MontereyPine_pinyonPine_pinyon_brdPine_pinyon_cmnPine_pitchPine_pondPine_ponderosaPine_redPine_sandPine_scotchPine_shortleafPine_slashPine_sprucePine_sugarPine_SwwhitePine_tablemtnPine_VAPine_WashoePine_whitebarkPine_Wwhite

Pine_yellowPopulusPrunusRedbayRobinia_locustSassafrasSequoiaServiceberrySilverbellSmoketreeSoapberry_westrSourwoodSparkleberrySpruce_blackSpruce_blueSpruce_BrewerSpruce_EnglemanSpruce_NorwaySpruce_redSpruce_SitkaSpruce_sppSpruce_whiteSweetgumSycamoreTallowtree-chinsTamarixTanoakTorreyaTung-oil-treeUnknown_treeWalnutWater-elmWillowYellow_poplarYellowwoodYucca_Mojave

BELD3 – 229 vegetation classes

tree species/genera (FIA)

Cro

p ty

pes

(USD

A)

USG

S cl

asse

s

BEIS3 - Improved spatial resolution

BEIS2/BELD2 BEIS3/BELD3

BEIS3 – Chemical species

34 chemical speciesisoprene ethene

methyl-butenol propene

a-pinene ethanol

b-pinene acetone

d3-carene hexanal

d-limonene hexenol

camphene hexenylacetate

myrcene formaldehyde

a-terpinene acetaldehyde

b-phellandrene butene

sabinene ethane

p-cymene formic acid

ocimene acetic acid

a-thujene butenone

terpinolene carbon monoxide

g-terpinene ORVOCs

methanol nitric oxide

0

50000

100000

150000

200000

250000

300000

350000

400000

450000

Jan-01 Feb-01 Mar-01 Apr-01 May-01 Jun-01 Jul-01 Aug-01 Sep-01 Oct-01 Nov-01 Dec-01

Date

VOC

(met

ric to

nsC

/day

)

B309 B312 Isoprene

Comparison of BEIS3.09 v BEIS3.12 for 2001 Domain total emissions

Animation of isoprene

Comparison of BEIS3.09 v BEIS3.12 for 2001 Domain total emissions

0

500

1000

1500

2000

2500

3000

3500

4000

Jan-01 Feb-01 Mar-01 Apr-01 May-01 Jun-01 Jul-01 Aug-01 Sep-01 Oct-01 Nov-01 Dec-01

Date

NO

tons

/day

B309 B312 Animation of NO

Comparison of BEIS3.09 v BEIS3.12 for 2001 Domain total emissions (103 metric tons)

Compound BEIS3.09 BEIS3.12 % change

Soil NO 467

50,320

22,141

+30%

Total VOC

609

48,365 -4%

Isoprene 22,141 0%

The latest modeling “crisis” – high biogenic organic aerosol concentrations in the northwestern U.S. …

Are the monoterpene emission factors too high?

Preliminary CMAQ modeling plots

BEIS3.12 -- Change in the monoterpene emission factor for spruce and fir from 2.6 ug/g-hr to 1.5 ug/g-hr

Example for Lassen County, California – Effect of changing the monoterpene emission factor for spruce and fir from 2.6 ug/g-hr to 1.5 ug/g-hr

Tree type Fraction B3.12 efac(ug/m2-hr)

B3.13 efac(ug/m2-hr)

Jeffrey pine 0.29 1820 1820

Ponderosa pine 0.28 1820 1820

White fir 0.22 3900 2250

Lodgepole pine 0.05 1820 1820

Incense cedar 0.04 132 132

Calif. red fir 0.04 3900 2250

Calif. blk oak 0.01 86 86

Juniper 0.01 371 371

Oreg. wht oak 0.01 86 86

Douglas fir 0.01 3900 2250

Total/Wtd avg 0.97 2265 1800

=> Changing efac for spruce & fir lowered monoterpene emissions by ~20%


• Proposed future priorities:

Examine climatic adjustments of monoterpene emissions and evaluate flux estimates w/ field data

Incorporate isoprene seasonality

Combine the BEIS3 and the WRF-Chem/MEGAN platforms

Much ado about fugitive dust emissions…

• Fugitive dust emissions tend to be overestimated in atmospheric transport models (Gillette, 2001, www.wrapair.org/forums/dejf/documents/).

• Dr. Shan He has developed an wind blown dust algorithm for CMAQ based on the work of Gillette. The algorithm uses threshold friction velocity parameterizations and incorporates gridded databases of soil type, surface soil moisture content, meteorology, and vegetation. It also estimates the interceptionof dust particles by vegetation. Unfortunately w/ Shan’s departure, this effort is “collecting dust”.

• G. Pouliot has been working on an improved algorithm for other fugitive dust sources.

http://www.wrapair.org/forums/dejf/documents/

Proposed Fugitive Dust Emission Model

CMAQMM5

MCIP

Aerosol

Fugitive Dust Emission

Windblown Dust Road Dust

Paved Road Dust

Unpaved Road Dust

)uu(ugkAF 2t

2*** −

ρ=

EF = 0.016(SL/2)0.65(W/3)1.5

EF = 2.6(SC/12)0.8(W/3)0.4(M/0.2)-0.3

thresholdfrictionvelocitysoil moisture

soil texture/landuse data

frictionvelocity

area of dust source

roughnesslength of soil

silt loading/silt contenton road surface

fleet average vehicle weight

moisture ofroad surface

gridded road coverageroad type data

griddedVMT data

Local scale dust fluxRegional scale vertical flux

KVV

mm

d

d

horizontal

vertical 1 +−==Φ

frictionvelocity

surfacewind

depositionvelocity

landusedata

county levelVMT data

androad coverage

spatialtemporalallocation

Courtesy of Dr. Shan He

Windblown Fugitive Dust Emissions Estimated with the “He” Algorithm

Fugitive Dust Emissions from Unpaved Roads

Current method Proposed method

Does not account for removal by vegetation

Incorporate transport fraction developed by Dr. Shan He

FHWA road mileage TIGER data to grid unpaved roads from county data

Uses monthly rainfall from a single station in a state

Simulate the moisture content of the road surface using griddedsolar radiation, dew point, wind speed and rainfall data.

Based on published AP-42 methodology and used in EPA’s NEI.

Status: Unpaved road data have been gridded and the emissions algorithm will be tested later in 2004.

Much ado about wildland fire emissions …

• Current 2001 inventory includes monthly estimates by state

• For 2002, the RPOs are producing an event-based, location-specific inventory

• With support from OAQPS and EIIP, an IAG is being established w/ USFS to integrate BlueSky into SMOKE:

will provide framework for producing model-ready emissions from wildfires

will include updated fuel characteristics

will need support for plume rise and vertical allocation

BlueSky is a working system developed by the USFS for forecasting smoke from prescribed burns and wildfires. It has many of the components needed for an emissions module to CMAQ/SMOKE. (www.blueskyrains.org)

Actual simulation from the BlueSky system

Much ado about wildland fire emissions …


Include RPO inventory as soon as practical for 2002 CMAQ simulations

Integrate BlueSky into SMOKE and evaluate CMAQ model performance

Much ado about AQ forecast emissions …

• Work performed by G. Pouliot

• Priorities for 2004:

Incorporate and streamline Mobile6 calculations

Account for significant reductions in electrical generating unit (EGU) NOx emissions

Emissions for air quality forecasting – using simple regression to account for hourly temperature effects in Mobile6

Comparison of VOC emissions from Mobile6 vs. NLSQ regression across the NE AQ forecast domain for

August 4, 2003

Emissions for air quality forecasting – change in NOxEGU emissions from 2001 to 2004

-22%

-10%

-21%

-43%

-45%

-20%

Source:www.eia.doe.gov/oiaf/aeo/supplement/

Note: according to the 1998 NEI, electric utilities contributed ~25% of NOxemissions

Much ado about AQ forecast emissions …


Adjust EGU emissions as a function of predicted temperature (based on analysis of existing CEM data)

Incorporate satellite-derived wildfire emissions

Perform sensitivity analysis of the effect of emission changes resulting from voluntary control measures

Much ado about emission modeling…

Then sigh not so, but let them [emissions] go,And be you blithe and bonny,

Converting all your sounds of woeInto hey nonny, nonny.*

*w/ apologies once again to Wm. Shakespeare, from Much Ado About Nothing.

Disclaimer: This work has not been reviewed by EPA and may not reflect official Agency policy.

much ado about emission modeling

Documents