Comparison of Ambient Measurements to Emissions Representations in Modeling Presented by: Lyle R. Chinkin and Stephen B. Reid Sonoma Technology, Inc. Petaluma, CA Presented to: The CCOS Technical Committee Sacramento, CA December 14, 2005 905044.01-2838
49
Embed
Comparison of Ambient Measurements to Emissions Representations in Modeling Presented by: Lyle R. Chinkin and Stephen B. Reid Sonoma Technology, Inc. Petaluma,
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Comparison of Ambient Measurements to Emissions Representations in Modeling
Presented by:Lyle R. Chinkin and Stephen B. Reid
Sonoma Technology, Inc.Petaluma, CA
Presented to:The CCOS Technical Committee
Sacramento, CADecember 14, 2005
905044.01-2838
2
Statement of Work – Phase 1
Task 1 – Kickoff meeting, workplan, and EI data and EI documentation gathering
Task 2 – Identify available air quality data
Task 3 – Review ARB chemical speciation profiles
Task 4 – Workshop to justify and discuss merits of Phase 2
3
Summary of Findings: In general, ARB speciation profiles for
key source categories are up-to-date A handful of speciation profiles were
identified that need updating Updates would likely result in a slight
lowering of the reactivity of the organic gas inventory
Task 3 – Review Speciation Profiles (1 of 18)
4
ARB TOG Speciation Profile Database 425 available organic gas profiles 252 profiles applied to the 2002 CCOS
placeholder EI Individual profiles prioritized by
summing TOG, ROG, and reactivity-weighted emissions associated with each profile
Task 3 – Review Speciation Profiles (2 of 18)
5
Reactivity-Weighted Emission Calculation
Rx = ∑(MIR)ywy
where:Rx = Weighted reactivity for profile x
(MIR)y = maximum incremental reactivity for species ywy = weight fraction of species y in profile x
Rank Profile # Profile Name Application Source Vintage
3 422 Hot soak emissions - CA light-duty vehicles
Running and hot soak evaportive emissions from light-duty gasoline vehicles
19 SHED tests conducted in 1999 and 2000
2000
8 906 Diurnal and resting evaporatives
Diurnal and resting evaporatives from on-road gasoline vehicles
Study of gasoline samples collected at Berkeley service stations
1996
9 419 Liquid gasoline - MTBE 11% - commercial
Vehicle refueling and petroleum storage and marketing operations
ARB study of gasoline blends containing MTBE and ethanol
1997
11
Task 3 – Review Speciation Profiles (9 of 18)
0
1
2
3
Berkeley 1995 Berkeley 1996/ARB 906
Berkeley 1999 Berkeley 2001 Sacramento2001
Nor
mal
ized
Rea
ctiv
ity
0
1
2
3
4
5
ARB 419 Berkeley1995
Berkeley1996
Berkeley1999
Berkeley2001
Sacramento2001
Nor
mal
ized
Rea
ctiv
ity
Weighted reactivity of liquid gasoline speciation profiles
Weighted reactivity of gasoline headspace speciation profiles
12
ARB diesel exhaust profile
Task 3 – Review Speciation Profiles (10 of 18)
Total ROG = 79.1 tons/day
Point1%On-road
Mobile23%
Area andOff-road Mobile
76%
Emissions associated with
Profile 818
Rank Profile # Profile Name Application Source Vintage
4 818 Diesel Farm Equipment
Exhaust emissions from on-road diesel vehicles and off-road diesel equipment
Cal Poly heavy-duty diesel equipment engine tests
1991
13
Reactivity of various diesel exhaust profiles
Task 3 – Review Speciation Profiles (11 of 18)
0
1
2
3
4
5
6
ARB 818 EPA 2520 DRI Schauer
No
rmal
ized
Rea
ctiv
ity
14
ARB animal waste decomposition profile
Task 3 – Review Speciation Profiles (12 of 18)
Rank Profile # Profile Name Application Source Vintage
6 203 Animal waste decomposition
Livestock husbandry operations
EPA's SPECIATE 3.2 database - based on 1978 study in SOCAB
1978
Species Name CAS CodeWeightPercent
MIR
Methane 74-82-8 70 0.0139
Ethane 74-84-0 20 0.31
Acetone 67-64-1 2 0.43
Isopropyl alcohol 67-63-0 2 0.71
Propyl acetate 109-60-4 2 0.86
Ethanol 64-17-5 2 1.69
Trimethyl amine 75-50-3 1 7.06
Ethyl amine 75-04-7 1 7.79
FROG for profile 203 = 8%
15
Task 3 – Review Speciation Profiles (13 of 18)
Process Type Process
Emissions (lbs/day)ROG
PercentTOG ROG
Milk Cow Bedding 1.3 0.5 38.5%
Flush Lane 10.5 1.4 13.3%
Feeding 5.7 5.4 94.7%
Turnout 500.5 2.1 0.4%
Dry Cow Bedding 0.0 0.0 0.0%
Flush Lane 0.1 0.1 100.0%
Feeding 0.5 0.4 80.0%
Turnout 0.7 0.7 100.0%
Solids Piles Fresh 3.1 0.0 0.0%
Aged 873.4 0.0 0.0%
Bedding Storage 0.5 0.3 60.0%
Lagoon Lagoon 164.1 1.1 0.7%
Milk Parlor Effluent Stream 0.2 0.2 100.0%
Total All Processes 1560.6 12.2 0.8%
Dairy organic gas emissions by process type (Schmidt, et al., 2005)
16
ARB jet exhaust profile
Task 3 – Review Speciation Profiles (14 of 18)
Comparison of jet exhaust profiles
Rank Profile # Profile Name Application Source Vintage
7 586 Composite jet exhaust
Military, commercial, and civil jet aircraft
Composite of 3 EPA profiles developed from engine tests
1984
0%
20%
40%
60%
80%
100%
Profile 586 Environment Canada
Wei
ght
% o
f T
OG
alkenes
alkanes
aromatics
carbonyls
17
ARB wildfire profile
Task 3 – Review Speciation Profiles (15 of 18)
Rank Profile # Profile Name Application Source Vintage
-- 307 Forest fires Unplanned fires on grasslands and forested lands
EPA's SPECIATE 3.2 database - based on literature search
1975
0
1
2
3
4
5
ARB 307 LADCO Forest LADCO Grassland
Nor
mal
ized
Rea
ctiv
ity
Reactivity of various wildfire
profiles
18
Findings & Recommendations Gasoline exhaust and evaporative profiles
appear to be appropriate for on-road vehicles in CA in 2000.
A 1997 lawnmower-based profile is more appropriate for off-road gasoline equipment than the current ARB profile (401).
A Schauer speciation profile is more appropriate for on-road diesel vehicles than ARB’s current farm equipment-based profile (818).
Further study of the reactivity of animal waste emissions is needed.
Task 3 – Review Speciation Profiles (16 of 18)
19
Findings & Recommendations (cont’d) Further study of the composition of organic
gas emissions from jet exhaust is needed. Wildfires can be a significant ROG source on
given days; a new California-specific profile should be developed to replace the current EPA profile used by ARB for this source category.
ARB industrial surface coating, medium-cure asphalt, and all-category composite profiles need to be updated (new industrial coating profiles identified).
Task 3 – Review Speciation Profiles (17 of 18)
20
Findings & Recommendations (cont’d) Application of the recommended profiles is
likely to result in a slight decrease in MIR-weighted TOG emissions for the CCOS domain (from 4,946 tpd to 4,922 tpd).
Task 3 – Review Speciation Profiles (18 of 18)
21
Task 2 – Identify Available Air Quality Data (1 of 13)
Site selection based on Data availability (distinct counts of
Task 2 – Identify Available Air Quality Data (9 of 13)
Site TierTotal Count 7/
1/20
00
7/2/
2000
7/3/
2000
7/4/
2000
7/5/
2000
7/6/
2000
7/7/
2000
7/8/
2000
7/9/
2000
7/10
/200
0
7/11
/200
0
7/12
/200
0
7/13
/200
0
7/14
/200
0
7/15
/200
0
7/16
/200
0
7/17
/200
0
7/18
/200
0
7/19
/200
0
7/20
/200
0
7/21
/200
0
7/22
/200
0
7/23
/200
0
7/24
/200
0
7/25
/200
0
7/26
/200
0
7/27
/200
0
7/28
/200
0
7/29
/200
0
7/30
/200
0
7/31
/200
0
FSF 1 29 X X X X X X X X XBGS 1 27 X X X X X X X X X XNAT 1 27 X X X X X X X X XCLO 1 22 X X X X X XSDP 1 21 X XVTE 2 32 X X X X X X X XFLN 2 27 X X X X X X X X XPLR 2 27 X X X X X X X X X XM29 3 26 X X X X X XARV 3 23 X X X X X X X X X XELK 3 11 X
Site TierTotal Count 8/
1/20
00
8/2/
2000
8/3/
2000
8/4/
2000
8/5/
2000
8/6/
2000
8/7/
2000
8/8/
2000
8/9/
2000
8/10
/200
0
8/11
/200
0
8/12
/200
0
8/13
/200
0
8/14
/200
0
8/15
/200
0
8/16
/200
0
8/17
/200
0
8/18
/200
0
8/19
/200
0
8/20
/200
0
8/21
/200
0
8/22
/200
0
8/23
/200
0
8/24
/200
0
8/25
/200
0
8/26
/200
0
8/27
/200
0
8/28
/200
0
8/29
/200
0
8/30
/200
0
8/31
/200
0
FSF 1 29 X X X X X X X X X X XBGS 1 27 X X X X X X X X X X XNAT 1 27 X X X X X X X X XCLO 1 22 X X X X X X X X XSDP 1 21 X X X X X X X X XVTE 2 32 X X X X X X X X X X XFLN 2 27 X X X X X X X XPLR 2 27 X X X X X X X X X XM29 3 26 X X X X X X X X X X XARV 3 23 X X X X X X X XELK 3 11 X X X
July VOC
Aug. VOC
30
Task 2 – Identify Available Air Quality Data (10 of 13)
Site TierTotal Count 9/
1/20
00
9/2/
2000
9/3/
2000
9/4/
2000
9/5/
2000
9/6/
2000
9/7/
2000
9/8/
2000
9/9/
2000
9/10
/200
0
9/11
/200
0
9/12
/200
0
9/13
/200
0
9/14
/200
0
9/15
/200
0
9/16
/200
0
9/17
/200
0
9/18
/200
0
9/19
/200
0
9/20
/200
0
9/21
/200
0
9/22
/200
0
9/23
/200
0
9/24
/200
0
9/25
/200
0
9/26
/200
0
9/27
/200
0
9/28
/200
0
9/29
/200
0
9/30
/200
0
FSF 1 29 X X X X X X X X XBGS 1 27 X X X X X XNAT 1 27 X X X X X X X X XCLO 1 22 X X X X X X XSDP 1 21 X X X X X X X X X XVTE 2 32 X X X X X X X XFLN 2 27 X X X X X X X X X XPLR 2 27 X X X X X X XM29 3 26 X X X X X X X X XARV 3 23 X X X X XELK 3 11 X X X X X X X
Sept.
VOC
31
Task 2 – Identify Available Air Quality Data (11 of 13)
Site Tier 1 2 3 4
BGS 1 A A A ACLO 1 A A A AFSF 1 A A A ANAT 1 A,M A A ASDP 1 A A A ASUN 1 M A,M P MFLN 2 A A A APLR 2 A A A AARV 3 A A A PELK 3 M A,M M MSJ4 3 A A A,M A,MM29 3 A A A ABTI 3 A N A ABODB 3 N A N NTSM 3 A A A A
Wind Quadrant
Dominant emission source types by wind quadrant
Legend:
A = AreaM = Mobile (on-road)N = Non-roadP = Point
32
Task 2 – Identify Available Air Quality Data (12 of 13)
Dominant emission source types by wind quadrant
33
Task 2 – Identify Available Air Quality Data (13 of 13)
Dominant emission source types by wind quadrant
34
Phase 1 Summary
12 monitoring sites identified with sufficient data to perform Phase 2 analyses that have a high probability of identifying specific biases/uncertainties in the emission inventory that will lead to improved air quality modeling results.
35
Phase 2 – Objective (1 of 2)
To gather corroborative evidence using different analysis techniques that will result in recommendations for specific, meaningful improvements to the CCOS emission inventory.
36
Phase 2 – Objective (2 of 2)
Sample questions (example outcomes in red):
• Do the methods used to characterize wildfires in the EI adequately represent the spatial and temporal dimensions of large fire events? (e.g., development of a new temporal profile for wildfires by pollutant species.)
• Do any discrepancies exist between ambient data and emissions data for those species with strong diurnal patterns? What are the likely sources of those differences? (e.g., recommendations for adjustments to the temporal distribution of biogenic emissions.)
37
Original Phase 2 Techniques
Integrate the results of previous research
Perform EI reconciliation with pollutant ratios (VOC/NOx, CO/NOx)
Perform EI reconciliation with speciated VOCs (ratios, TNMOC composition)
• Ratio comparisons (VOC/NOx and individual species)
• Fingerprint analyses• Wildfire analyses• Analysis of species that vary temporally• Source apportionment (e.g., CMB and
PMF) - as a corroborative tool
39
Key QuestionsDoes the current EI preparation methodology incorporate the latest results from available research (e.g. speciation profiles,temporal profiles, emission factors, etc.)?
Review Previous Findings
Literature review to identify previous, relevant work• For example
– SJV Emissions Reconciliation (STI)– DRI advanced data analysis study
40
Analysis Methods (1 of 5)
Ratio Comparisons Convert emission inventory (EI) from mass to
moles and compare VOC/NOx ratios in EI to ambient data ratios by hour and wind quadrant
Individual species ratios by hour and wind quadrant (e.g., acetylene/benzene, benzene/toluene, benzene/xylene)Key QuestionsHow do pollutant ratios derived from the EI compare with those from ambient data? How do these ratios vary by site/wind quadrant due to the influence of various emission sources?
41
Analysis Methods (2 of 5)
Fingerprint analyses Comparison of speciated emissions to
speciated VOCs in ambient air by hour and wind quadrant
Key QuestionsHow does the EI-predicted VOC species composition compare with the ambient data? Do any variations appear to be a result of differences in mass, speciation, or both?
42
Analysis Methods (3 of 5)
Wildfire analyses 2,000 tpd of TOG on July
31, 2000 EI (33% of total TOG)
Manter fire consumed 74,000 acres over an 18-day period
Flaming and smoldering emissions generatedKey QuestionsDo the methods used to characterize wildfires in the EI adequately represent the spatial and temporal dimensions of large fire events?
0
2
4
6
8
10
12
14
16
18
1 3 5 7 9 11 13 15 17 19 21 23
Hour of Day
% D
aily
Em
issi
on
s
Wildfire Diurnal Profile
43
Analysis Methods (4 of 5)
Analysis of species that vary temporally Identify and assess those species such as
isoprene that exhibit diurnal patterns (i.e., isoprene, evaporative VOCs)
Analyze morning and afternoon data for selected abundant species
Key QuestionsDo any discrepancies exist between ambient data and emissions data for those species with strong diurnal patterns? What are the likely sources of those differences?
44
Analysis Methods (5 of 5)
Source apportionment Chemical mass balance (CMB) or positive
matrix factorization (PMF) Use as a tool to corroborate findings from
previous analyses
Key QuestionsDoes the source mix produced by source apportionment tools match up with the mix calculated from the EI? How does this analysis corroborate the findings of other techniques?
45
Develop Recommendations
Synthesize Findings• Formulate overarching conclusions• Summarize the apparent strengths and
weaknesses of the EI including a discussion of possible biases in the EI
• Make recommendations for “corroborative adjustments” to the EI
46
Examples (1 of 4)
0.00
2.00
4.00
6.00
8.00
10.00
1993 1994 1995 1996 1997 1998 1999 2000 2001
Year
VO
C/N
Ox
Ambient Data Emission Inventory Data
Emission inventory- andambient-derived VOC/NOxratios at Los AngelesNorth Main during summermornings.
Results can be used to explore segments of the inventory that may be underestimated.
47
Examples (2 of 4)
Houston area EI reconciliationspike in ambient concentrations of n-butane when winds were from the southeast.
Results can be used to identify the speciation profiles that are need revision.
Clinton 2000 - Wind Quadrant 2
0%
5%
10%
15%
20%
25%
30%
35%
Eth
ane
Eth
ylen
e
Pro
pane
Pro
pyle
ne
Isob
utan
e
1,3-
buta
dien
e
n-B
utan
e
Ace
tyle
ne
C4
olef
ins
C5
para
ffins
isop
rene
C5
olef
ins
C6
para
ffins
met
hyl p
ente
nes
hexe
nes
C7
para
ffins
C8
para
ffins
tolu
ene
benz
ene
ethy
lben
zene
xyle
nes
styr
ene
n-no
nane
/C9
C3/
C4/
C5
alky
lben
zene
n-de
cane
/C10
n-un
deca
ne/C
11
TN
MO
C W
eig
ht%
Ambient - Avg
Ambient - Median
EI - Low LevelOnly
EI - With ElevatedSources
48
Examples (3 of 4)
Reconciliation of some speciated VOCs such as isoprene will require looking at concentrations during the daytime, rather than in the morning.
Isoprene concentrations are highest during the day, due to the higher emissions from biogenics. 0 8 16 24
HOUR
0
5
10
15
Isop
rene
(pp
bC)
3-hr duration isoprene concentrations at Clovis in 2001
49
Examples (4 of 4)
Source apportionment (PMF) used for multiple Houston sites to provide a breakdown of emissions from various sources.
The EI appears to be overestimating mobile source emissions at this site by a factor of 1.5 (15% in the EI vs. 6% in the PMF results).