Emissions from laboratory combustion of wildland fuels: emission factors and source profiles.

Abstract

Combustion of wildland fuels represents a major source of particulate matter (PM) and light-absorbing elemental carbon (EC) on a national and global scale, but the emission factors and source profiles have not been well characterized with respect to different fuels and combustion phases. These uncertainties limit the accuracy of current emission inventories, smoke forecasts, and source apportionments. This study investigates the evolution of gaseous and particulate emission and combustion efficiency by burning wildland fuels in a laboratory combustion facility. Emission factors for carbon dioxide (CO2), carbon monoxide (CO), total hydrocarbon (THC), nitrogen oxides (NO(x)), PM, light extinction and absorption cross sections, and spectral scattering cross sections specific to flaming and smoldering phases are reported. Emission factors are generally reproducible within +/- 20% during the flaming phase, which, despite its short duration, dominates the carbon emission (mostly in the form of CO2) and the production of light absorption and EC. Higher and more variable emission factors for CO, THC, and PM are found during the smoldering phase, especially for fuels containing substantial moisture. Organic carbon (OC) and EC mass account for a majority (i.e., > 60%) of PM mass; other important elements include potassium, chlorine, and sulfur. Thermal analysis separates the EC into subfractions based on analysis temperature demonstrating that high-temperature EC (EC2; at 700 degrees C) varies from 1% to 70% of PM among biomass burns, compared to 75% in kerosene soot. Despite this, the conversion factor between EC and light absorption emissions is rather consistent across fuels and burns, ranging from 7.8 to 9.6 m2/g EC. Findings from this study should be considered in the development of PM and EC emission inventories for visibility and radiative forcing assessments.

Full text

Page 1

Emissions from Laboratory
Combustion of Wildland Fuels:
Emission Factors and Source
Profiles
L . - W . A N T O N Y C H E N , *, †
H A N S M O O S M U ¨ L L E R ,†
W . P A T R I C K A R N O T T ,† , §
J U D I T H C . C H O W ,†J O H N G . W A T S O N ,†
R O N A L D A . S U S O T T ,‡
R O N A L D E . B A B B I T T ,‡C Y L E E . W O L D ,‡
E M I L Y N . L I N C O L N ,‡A N D W E I M I N H A O‡
Division of Atmospheric Sciences, Desert Research Institute,
Nevada System of Higher Education, Reno, Nevada, and Fire
Sciences Laboratory, Rocky Mountain Research Station, USDA
Forest Service, Missoula, Montana
Combustion of wildland fuels represents a major source
of particulate matter (PM) and light-absorbing elemental
carbon (EC) on a national and global scale, but the emission
factorsandsourceprofileshavenotbeenwellcharacterized
with respect to different fuels and combustion phases.
These uncertainties limit the accuracy of current emission
inventories, smoke forecasts, and source apportionments.
This study investigates the evolution of gaseous and
particulate emission and combustion efficiency by burning
wildland fuels in a laboratory combustion facility. Emission
factors for carbon dioxide (CO2), carbon monoxide (CO), total
hydrocarbon (THC), nitrogen oxides (NOx), PM, light
extinction and absorption cross sections, and spectral
scattering cross sections specific to flaming and smoldering
phases are reported. Emission factors are generally
reproducible within (20% during the flaming phase, which,
despite its short duration, dominates the carbon emission
(mostly in the form of CO2) and the production of light
absorption and EC. Higher and more variable emission
factors for CO, THC, and PM are found during the smoldering
phase, especially for fuels containing substantial moisture.
Organic carbon (OC) and EC mass account for a majority
(i.e., >60%) of PM mass; other important elements include
potassium,chlorine,andsulfur.Thermalanalysisseparates
the EC into subfractions based on analysis temperature
demonstrating that high-temperature EC (EC2; at 700 °C)
variesfrom1%to70%ofPMamongbiomassburns,compared
to 75% in kerosene soot. Despite this, the conversion
factor between EC and light absorption emissions is rather
consistent across fuels and burns, ranging from 7.8 to
9.6 m2/gEC. Findings from this study should be considered
in the development of PM and EC emission inventories
for visibility and radiative forcing assessments.
1. Introduction
Biomass burning in the U.S. accounts for more than one-
thirdofprimaryPM2.5(particleswithaerodynamicdiameter
<2.5µm) emissions and is a major source of light-absorbing
elemental carbon (EC), strongly affecting the atmospheric
visibility and radiation budget (1, 2). Today prescribed
burning is commonly used as a land management tool to
maintain forest health. The U.S. Environmental Protection
Agency(3)listsprescribedburningasthethirdlargestsource
of primary anthropogenic PM2.5, behind only utility fuel and
residential wood combustion. Major uncertainties in the
assessmentofprescribedburningimpactsonairqualityand
climatehavebeenattributedtoinadequateemissionfactors,
activity data, or both.
Innaturalandmanagedfires,aflamefrontpassesrapidly
through a fuel bed followed by sustained smoldering
combustion, so emissions from both combustion phases
coexistwiththeirproportionsquicklychanginginspaceand
time. While the most accurate emission factors would be
those measured directly in the field during real fires, these
by necessity represent a combination of vegetation and
combustion phases and are influenced by the underlying
soil moisture and other biomass (e.g., wood and grass duff).
Because of the variability in actual burning conditions,
extrapolatingthemeasuredemissionfactorsfromonestudy
to other fires is uncertain. Laboratory-controlled fires are
usefulforisolatingeffectsofindividualparameters(e.g.,fuel,
moisture, combustion phase) on emission factors, but they
do not fully anticipate or reproduce the complex real-world
fires(e.g.,4).Itis,however,possibletoestimategaseousand
PMemissionsusingfieldemissionfactorsadjustedtoreflect
laboratoryobservationswithmodernemissionmodelssuch
as the Fire Emission Production Simulator (FEPS) (5).
This laboratory experiment was designed to monitor
individual combustion events that proceed from ignition,
through flaming and smoldering combustions, and ending
with fire extinction. Measuring properties of burning emis-
sions such as light scattering (σscat), light absorption (σabs),
andparticlemassconcentrationwithsecondtimeresolution
helps differentiate between flaming and smoldering emis-
sions that are known to have distinct characteristics (4, 6).
The fuels tested were common wildland fuels mostly from
mid-latitude biomes. This paper presents gas and particle
emission factors and source profiles, and discusses their
dependence on fuel and combustion phase.
2. Experimental Section
The experiment was conducted at the United States Forest
Service (USFS) Fire Science Laboratory (FSL, Missoula, MT)
from 11/19/03 to 11/26/03 (7, 8). The combustion facilities
at the FSL include a continuously weighed fuel bed under
a 1.6 m diameter exhaust stack with a 3.6 m inverted-funnel
opening extending from 2 m above the fuel bed through the
ceiling (∼22 m high). Goode et al. (9) show well- mixed and
cooled plumes at the height of a sampling platform ∼17 m
abovethebed.Sampleairwasdrawndirectlyintoinstruments
without additional dilution. This minimized the change in
particle mass due to organics volatilization under different
dilutionratios(e.g.,10).Eightwildlandfuelswereexamined:
(1) ponderosa pine wood; (2) ponderosa pine needles; (3)
white pine needles; (4) sagebrush; (5) excelsior; (6) Dambo
grass; (7) Montana grass; and (8) tundra core (see detailed
description in Table S1, Supporting Information). The first
six fuels were “dry” from long-term indoor storage with a
fuelmoisturecontentof<10%ofthedrymassandrepresent
* Corresponding author e-mail: antony@dri.edu.
†Desert Research Institute.
‡Rocky Mountain Research Station.
§Current affiliation:Department of Physics, University of
Nevada, Reno.
Environ. Sci. Technol. 2007, 41, 4317-4325
10.1021/es062364i CCC: $37.00
Published on Web 05/10/2007
 2007 American Chemical Society VOL. 41, NO. 12, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY94317

Page 2

wildland fuels during drought episodes. Montana grass and
tundra core were relatively “wet” fuels (moisture content
>10%) from Montana and Alaska, respectively. In addition,
reference soot aerosol was generated by a kerosene lamp in
one of the experiments.
About250goffuelwasusedforeachcontrolledfire(except
tundra core which used 1250 g). Continuous gas and mass
measurements of the effluent included: (1) CO2by infrared
absorption (LI-840, Li-COR Environmental); (2) CO by
infrared absorption (model 48C, Thermo Electron Corpora-
tion); (3) reactive nitrogen species (NO, NO2, and NOx) by
chemiluminenscence (model 42C, Thermo Electron Corpo-
ration);(4)THC,includingCH4,byflameionizationdetection
(model 42C THC Analyzer, Thermo Electron Corporation);
and (5) PM mass by inertial microbalance (TEOM, Series
1105,Rupprecht&PatashnickCompony).Therewasnosize-
selective inlet upstream of the TEOM, though it is believed
that most of the PM mass is in the fine fraction (i.e., <2.5
µm).TheTEOMmasswasnormalizedtogravimetricanalysis
ofsimultaneoustime-integratedTeflon-filtersamples.Optical
measurements included: (1) σabs at 532 and 1047 nm by
photoacoustic detection (11); (2) σext(σscat+ σabs) at 532 nm
by cavity ring-down/cavity enhanced detection (CRD/CED)
(12); (3) σscatat 450, 550, and 700 nm by nephelometry (TSI
3563, TSI). The nephelometer reported total (σscat) and
hemispheric back-scattering (?σscat), with their accuracy
limited by the nephelometer truncation angle (i.e., 7°) (13).
Thenephelometerdataareusefulforcalculatingthespectral
Ångstro ¨m exponent (i.e., Rs) -ln σscat,λ1/σscat,λ2/ln[λ1/λ2]),
a semiquantitative indicator of particle size (8). The continu-
ous instruments operated with a time resolution between 1
and 10 s. Precisions of PM mass and optical measurements
have been discussed in Chen et al. (8).
Smoke PM was collected on Teflon-membrane and
quartz-fiber filters for the quantification of mass (by gravim-
etry),multi-elements(byX-rayfluorescence(14)),andorganic
carbon (OC), EC and TC (OC + EC) concentrations. These
samples represent time-integrated averages over individual
burns, and contributions from different combustion phases
werenotseparated.OCandECweredeterminedbyboththe
IMPROVE and STN protocol (15) using a DRI model 2001
carbon analyzer. Source profiles based on elements and
carbon fractions are presented and compared in this paper.
Emissionfactorsfororganicspeciesarepresentedelsewhere
(16).
3. Results and Discussions
3.1 Definitions. The fundamental definition of fuel-based
emissionfactoristhemassofacompoundreleasedpermass
offuelconsumed(17).Thisisrelatedtotheamountofcarbon
in the fuel:
where EFj) emission factor of species j; Mfuel) mass of the
fuelburned;Mj)massofthespeciesjemitted;Cash)carbon
mass in ash; Ci) carbon mass in every combustion product
i (CO2, CO, etc., including species j); and xc,fueland xc,ash)
carbon mass fraction in fuel and ash, respectively.
For the fuels used here, xc,fuelis in the range of 45-50%
(Table S1) except for kerosene fuel and tundra core. Since
ash typically accounts for <5% of the fuel mass (7) and
contains little carbon, the “ash term” in eq 1 can be ignored,
making this equation for biomass burning identical to
equations used for liquid fuel combustion (18).
To obtain time-resolved (dynamic) EFs, eq 1 is further
modified for use with continuous measurements as
where ∆[Mj] is the average of excess mass concentration ∆-
[Mj] (i.e., above background level) of species j measured in
the smoke for a sampling period. Mi includes MCO2, MCO,
MTHC, and MPM with xc,i ≡ Mi/Ci. THC and PM consist of
multiplespecies,butTHCwasalreadyinunitsofppmcarbon.
For xc,PM, a nominal value of 0.75 is used, according to the
carbon mass fraction of 0.5-1 in PM estimated from filter
samples (see Section 3.3). TotC represents the total (gas +
PM)carbonemission.Theinfluenceofxc,PMuncertaintieson
theEFcalculationisnegligible,i.e.,<(2%becausethecarbon
emission is predominantly (>95%) in the form of CO2and
CO.
Flaming and smoldering combustion are distinguished
by their different combustion efficiencies (CE), defined as
the fraction of carbon emitted in the form of CO2. Al-
ternatively, one may use modified combustion efficien-
cy (MCE) if only CO2 and CO are measured (MCE )
∆[CCO2]/(∆[CCO2] + ∆[CCO])). CE and MCE are usually close
to 1 during the flaming phase due to near stochiometric
combustion. When most volatiles have been expelled from
the fuel surface, flaming ceases, and smoldering begins.
Smoldering combustion is a lower-temperature oxidation
process (<850 K) in the char layer that yields more CO and
otherincompletelyoxidizedpyrolysisproducts.Leeetal.(4)
reports good correlations between CO2 and many hydro-
carbon species for smoldering emissions. Typical CE and
MCE during smoldering phase are 0.7-0.9 (6, 19, 20).
Figure 1a and b exemplify the evolution of CE and MCE
during a burn of white pine needles and sagebrush, respec-
tively. CE and MCE exceed 0.98 for the first 1.5 min after
ignition, and then rapidly decrease to <0.9 indicating the
transition from flaming to smoldering. Despite a relatively
short duration, the flaming phase dominates the carbon
emission (TotC) and has a relatively high emission factor for
σabs(note: σabsat 532 nm is caused by both PM and NO2).
As the smoldering combustion takes over, EFPMand EFσext
increase,especiallyfortheburningofwhitepineneedles.In
terms of TotC emission, smoldering combustion is less
intense but can persist for a long time. In Figure 1b, the
excessCO2/COconcentrationsremainedat2/0.2ppmatthe
endofsamplingwhenthePMleveldroppedbelowtheTEOM
detection limit. Smoke emissions during the smoldering
phaseexhibitamuchhighersinglescatteringalbedo(ωo,the
fraction of σextdue to σscat) than that from the flaming phase.
The other dry fuels show similar patterns (e.g., 8). It is
EFj)
Mj
Mfuel
Ci(
)
Mj
Cash+∑
∑
i
i
Ci
xc,fuel)
Mj
∑
i
Ci
Cash+∑
i
Ci)
xc,fuel)
Mj
∑
i
Ci(
xc,fuel-
Mash
Mfuel
xc,ash)
(1)
EFj)
∆[Mj]
∑
i
xc,i∆[Mi]
xc,fuel)
∆[Mj]
∑
i
∆[Ci]
xc,fuel)
∆[Mj]
[TotC]
xc,fuel(2)
CE )
xCO2∆[MCO2]
∑
i
xc,i∆[Mi]
)
∆[CCO2]
[TotC]
(3)
43189ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 12, 2007

Page 3

desirable to separate emission factors for flaming and
smolderingphasesforemissioninventoryandsmokeforecast
applications. A perfect split, however, is unlikely because
thereisatransitionperiod(0.9<CE<0.98;typically<1min
in this study) when flaming and smoldering combustion
coexist. The lowest CE usually occurs at the end of this
transition period. CE increases thereafter may be due to an
increasing oxygen level as observed by Hays et al. (21), and
this complicates the separation of flaming and smoldering
phases based solely on a threshold CE or MCE (e.g., MCE )
0.9 as noted by Reid et al. (6) and references therein). Since
COisthemostimportantmarkerforsmolderingcombustion,
in this study the split is made so that the difference in
∆[MCO]/[TotC] or average EFCO between the flaming and
smoldering phases is maximized. This occurs during the
transition (e.g., Figure 1a and b) and is consistent with the
visual observation of flaming and smoldering. The same
criterion is applied to all the burns except for tundra core
and kerosene combustion where only one phase could be
identified.
3.2 Emission Factors. Average fuel-based emission fac-
tors for the flaming and smoldering phases are shown in
Table1.Theconsistencywithin6-7replicateburnsforeach
fuelnotonlyreflectsthenaturalvariabilityofthecombustion
process but also corroborates the estimate for xc,PMand the
flaming/smoldering split point.
Most of the carbon emissions occur during the flaming
phase, particularly for ponderosa pine wood (98 (1%),
excelsior (97 ( 1%), and Dambo grass (97 ( 2%). The dry
fuelsexhibitCEg0.97andEFsreproducibilitytypicallywithin
( 20% for CO, THC, NO, and PM, averaged over the flaming
phase. However, large variations are observed among the
different fuels. Sagebrush burning yields the highest EFCO
and EFTHC, more than twice those of white pine needles.
EFPMvaries by a factor of 2.5. CE and MCE remain below 0.9
throughout the burning of wet tundra core (fuel moisture
113 ( 126%), and therefore no flaming phase is reported.
Montanagrass,awetfuelcontaining17.5%water(TableS1),
shows a flaming phase but its fraction varies substantially
between two replicate burns (CF60-90%). The smoke ωofor
this flaming phase is 0.95, much higher than those of dry
fuels (i.e., 0.32-0.73), and the absorption is largely caused
by NO2(Table 1). These could reflect a lower combustion
temperature for wet fuels, or an unclear separation between
FIGURE 1. Time series of combustion efficiency (CE, MCE), total carbon (TotC) concentration, and instant (10 s) emission factors of
extinction, absorption cross section (at 532 nm), and PM mass measured during combustion of (a) white pine needles and (b) sagebrush.
Emission factors are normalized to TotC (i.e., in unit of g/kgC). The dashed box defines the flaming phase (see text for details).
VOL. 41, NO. 12, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY94319

Page 4

TABLE 1. Emission Factors (Average and 1σ) of Burning for the Flaming Phase, Smoldering Phase, and Overall Combustiona
Flaming Phase
WPNEEDEFPPWOOD PPNEEDSAGEEXCEL DGRASS MTGRASSb
KEROc
# of burn
CF(%)
CE
MCE
CO2
CO
THCd
NO
NO2
PM
σext(532 nm)
σabs(532 nm)
σabs(532 nm) by NO2e
σabs(1047 nm)
σscat(450/550/700 nm)f
?σscat(450/550/700 nm)f
77666621
98% ( 1%
0.98 ( 0.00
0.99 ( 0.00
1763.1 ( 2.4
15.1 ( 1.2
0.4 ( 0.2
0.5 ( 0.0
0.3 ( 0.0
3.2 ( 0.6
29.4 ( 2.9
20.1 ( 2.3
0.1 ( 0.0
11.0 ( 1.3
7.8/5.6/3.6
1.5/1.1/0.9
88% ( 5%
0.97 ( 0.00
0.98 ( 0.00
1781.9 ( 8.7
19.9 ( 4.3
3.3 ( 0.4
2.9 ( 0.2
0.9 ( 0.3
4.0 ( 0.5
28.6 ( 3.4
17.0 ( 2.0
0.1 ( 0.0
9.3 ( 1.1
10.0/7.0/4.3
1.8/1.4/1.0
74% ( 1%
0.98 ( 0.00
0.99 ( 0.00
1758.5 ( 4.7
11.2 ( 1.9
2.5 ( 0.6
2.5 ( 0.2
0.6 ( 0.2
5.0 ( 0.4
30.8 ( 1.8
9.8 ( 1.4
0.1 ( 0.0
5.0 ( 0.8
19.1/15.0/10.6
2.9/2.2/1.8
66% ( 5%
0.96 ( 0.00
0.98 ( 0.00
1692.2 ( 7.3
22.6 ( 4.1
6.3 ( 1.0
1.6 ( 0.5
0.2 ( 0.1
5.4 ( 0.6
31.2 ( 4.2
21.4 ( 3.6
0.0 ( 0.0
10.5 ( 1.9
11.0/7.3/4.3
2.2/1.6/1.2
97% ( 1%
0.98 ( 0.00
0.99 ( 0.00
1721.8 ( 2.3
16.1 ( 0.8
1.3 ( 0.3
0.8 ( 0.1
0.1 ( 0.1
3.4 ( 0.4
23.0 ( 3.2
14.5 ( 1.3
0.0 ( 0.0
7.5 ( 0.7
11.1/7.8/4.8
1.9/1.4/1.0
97% ( 2%
0.98 ( 0.00
0.98 ( 0.00
1612.0 ( 5.0
16.8 ( 1.0
2.1 ( 0.4
1.7 ( 0.1
0.8 ( 0.1
2.1 ( 1.0
12.1 ( 6.8
3.2 ( 0.6
0.1 ( 0.0
1.5 ( 0.4
9.2/6.6/4.3
1.4/1.0/0.8
75% ( 15%
0.97-0.96
0.99-0.99
1553.8 ( 11.6
11.7 ( 1.8
10.4 ( 1.6
4.3 ( 1.1
4.8 ( 0.3
4.5 ( 1.6
15.5 ( 7.5
0.7 ( 0.1
0.8 ( 0.1
100%
0.96
1.00
3520.2
2.8
n.d.
0.4
2.6
51.6
578.7
328.8
0.4
178.4
g/kg fuel
g/kg fuel
g/kg fuel
g/kg fuel
g/kg fuel
g/kg fuel
m2/kg fuel
m2/kg fuel
m2/kg fuel
m2/kg fuel
m2/kg fuel
m2/kg fuel
16.0/13.0/9.9
2.3/1.8/1.5
131.2/95.4/62.7
24.2/18.9/15.2
Smoldering Phase
WPNEEDEFPPWOODPPNEED SAGEEXCELDGRASS MTGRASSb
TUNDRAc
# of burn
CS(%)
CE
MCE
CO2
CO
THCd
NO
NO2
PM
σext(532 nm)
σabs(532 nm)
σabs(532 nm) by NO2e
σabs(1047 nm)
σscat(450/550/700 nm)f
?σscat(450/550/700 nm)f
77666621
2% ( 1%
0.88 ( 0.02
0.89 ( 0.03
1573.8 ( 44.9
126.7 ( 30.0
1.7 ( 3.6
1.3 ( 0.7
1.1 ( 0.3
7.0 ( 11.5
6.7 ( 0.9
2.6 ( 1.5
0.2 ( 0.1
0.8 ( 0.8
5.2/3.1/1.8
1.0/0.7/0.4
12% ( 5%
0.87 ( 0.04
0.88 ( 0.04
1596.9 ( 73.3
134.9 ( 47.0
4.3 ( 1.7
4.9 ( 1.9
2.1 ( 1.1
4.6 ( 2.0
13.5 ( 6.9
4.6 ( 3.9
0.3 ( 0.2
0.5 ( 0.5
8.6/5.6/3.4
1.4/1.0/0.8
26% ( 1%
0.80 ( 0.02
0.85 ( 0.01
1445.0 ( 34.5
157.4 ( 7.2
14.5 ( 4.4
2.5 ( 0.5
0.9 ( 0.2
23.4 ( 6.8
114.4 ( 28.2
1.1 ( 0.2
0.2 ( 0.0
0.3 ( 0.1
100.9/85.0/66.3
12.4/10.2/9.2
34% ( 5%
0.86 ( 0.02
0.88 ( 0.02
1506.4 ( 35.1
131.5 ( 17.9
8.8 ( 2.0
6.6 ( 0.5
0.7 ( 0.2
8.2 ( 2.2
27.8 ( 13.6
9.8 ( 2.1
0.1 ( 0.0
4.5 ( 1.1
26.0/17.0/9.7
4.0/3.0/2.2
3% ( 1%
0.82 ( 0.04
0.86 ( 0.02
1436.0 ( 62.2
154.4 ( 14.5
9.9 ( 6.0
5.0 ( 2.6
2.2 ( 1.4
19.7 ( 11.7
76.6 ( 49.4
9.3 ( 4.0
0.4 ( 0.2
3.2 ( 1.7
73.4/54.8/36.8
9.8/7.4/6.2
3% ( 2%
0.88 ( 0.03
0.88 ( 0.02
1447.3 ( 41.3
121.7 ( 18.9
1.2 ( 4.6
0.4 ( 0.5
0.4 ( 0.3
3.0 ( 6.2
4.5 ( 6.1
0.3 ( 0.3
0.1 ( 0.0
0.0 ( 0.3
5.2/3.3/2.2
0.7/0.4/0.3
25% ( 15%
0.50-0.78
0.82-0.94
1031.7 ( 226.1
82.5 ( 31.8
125.3 ( 44.0
4.8 ( 0.5
10.2 ( 1.9
39.1 ( 20.1
194.7 ( 115.4
1.8 ( 0.7
1.7 ( 0.3
100%
0.76
0.87
2784.0
270.9
124.9
2.0
1.1
41.3
668.4
3.4
0.2
g/kg fuel
g/kg fuel
g/kg fuel
g/kg fuel
g/kg fuel
g/kg fuel
m2/kg fuel
m2/kg fuel
m2/kg fuel
m2/kg fuel
m2/kg fuel
m2/kg fuel
153.5/125.1/95.4
20.2/16.3/14.6
441.4/360.8/276.9
51.9/42.9/37.9
43209ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 12, 2007

Page 5

theflamingandsmolderingphases.Thekeroseneflameemits
littleCOandTHCbutanextremelyhighquantityofparticles
with an overall ωoof 0.43.
EFsforthesmolderingphasearemorevariableforseveral
reasons. In the combustion of ponderosa pine wood,
excelsior, and Dambo grass, only 2-3% of carbon emissions
came from the smoldering phase, resulting in very low gas
andparticleconcentrationswithlargerelativemeasurement
uncertainties. In addition, smoldering combustion is a slow
oxidation process for which emission factors may be more
sensitivetotheinhomogeneityinfuel,char,andcombustion
temperature.EFCOincreasedbynearlyanorderofmagnitude
when entering the smoldering phase for most of the burns.
Particleemissionsdidnotincreaseasmuch,buttheparticles
appear to be less light-absorbing and larger in size (7, 8).
Tundra core EF values for THC, PM, and σextare among the
largest of all the fuels tested. Smoldering CE and MCE are
between 0.8 and 0.9 for all dry fuels and lower values (CE <
0.8) are found for wet fuels.
Althoughsmolderingcombustionconsumesonlyaminor
fraction of fuel mass, it dominates CO, THC, and PM
emissions during the white pine needles and Montana grass
burns (Figure S1, Supporting Information). For sagebrush,
75%ofCOandnearlyhalfofTHCandPMareemittedduring
the smoldering phase. NO and NO2emissions are suggested
to occur mostly during flaming combustion (e.g., 22);
sagebrush burning represents the only exception observed
inthisstudy.Whileσabsisanothermainproductoftheflaming
phase, the smoldering phase can contribute substantial σext
and σscat. Fuels containing plant leaves, such as sagebrush,
grasses, and pine needles, often have a higher nitrogen
content(TableS1).OverallNOxemissionsareconsistentwith
the nitrogen content in the fuel.
Emission factors reviewed by Andreae and Merlet (17)
representing integrated field burning generally fall within
the range of EFs determined for the flaming and smoldering
phases for similar fuels in this study. For example, EFCOfor
Andreae and Merlet’s savanna grass burning is 65 ( 20 g/kg
fuel, compared with 16.8 ( 1.0 g/kg fuel (flaming) and 121.7
( 18.9 g/kg fuel (smoldering) from Dambo grass burning;
EFPM(fine)for Andreae and Merlet’s extratropical forest fire is
13 ( 7 g/kg fuel, compared with 5.0 ( 0.4 g/kg fuel (flaming)
and23.4(6.8g/kgfuel(smoldering)fromwhitepineneedle
burning. Lee et al. (4) report EFCOof ∼31 and ∼75 g/kg fuel
(assumingxc,fuel)0.5)forflamingandsmolderingcombustion
of a pine-dominated forest in the southeastern U.S., both of
which are between the flaming EFCO(11.2-19.9 g/kg fuel)
andsmolderingEFCO(126.7-157.4g/kgfuel)determinedfor
ponderosa pine wood/needles and white pine needles in
this study. Flaming and smoldering combustion is less
separated during the measurements of real-world fires.
Moreover, due to the influence of underlying soil and
biomass,thepartitionofflamingandsmolderingcontribution
(e.g., CF and CS in Table 1) may differ between real and
simulated fires. This partly explains why the overall (entire-
burn) EFs in Table 1, which are dominated by the flaming-
phase EFs, differ from those given by Lee et al. and Andreae
and Merlet. If CO2 and CO are measured, however, MCE
should provide an estimate for the flaming and smoldering
partition in any particular burn, and that could be used to
construct the total emission with phase-specific emission
factors.
3.3 PM Source Profiles. Mass fractions of species mea-
sured in PM are shown in Table 2. These are composite
chemicalprofilesresultingfromtheaverageofmassfractions
fromreplicatesamplesburningthesamefuel,astheapproach
described by Chow et al. (23). Carbon is the dominant
componentinPM,thoughtheTCcontentinPMvariesfrom
63.7% (sagebrush) to nearly 100% (ponderosa pine wood,
Dambo grass, and tundra core). The carbon content was
Overall
EF
PPWOOD
PPNEED
WPNEED
SAGE
EXCEL
DGRASS
MTGRASSb
# of burn
7
7
6
6
6
6
2
CE
0.98 ( 0.00
0.96 ( 0.00
0.93 ( 0.01
0.93 ( 0.01
0.97 ( 0.00
0.97 ( 0.01
0.91-0.89
MCE
0.99 ( 0.00
0.97 ( 0.00
0.96 ( 0.00
0.95 ( 0.01
0.98 ( 0.00
0.98 ( 0.00
0.98-0.97
CO2
g/kg fuel
1759.8 ( 2.1
1762.5 ( 5.5
1677.4 ( 9.1
1628.2 ( 14.7
1712.2 ( 4.6
1606.6 ( 10.2
1456.4 ( 15.0
CO
g/kg fuel
17.0 ( 0.5
32.0 ( 3.3
49.0 ( 2.0
60.0 ( 6.5
20.7 ( 1.5
20.1 ( 4.0
24.8 ( 1.5
THCd
g/kg fuel
0.4 ( 0.2
3.5 ( 0.5
5.6 ( 1.5
7.2 ( 1.2
1.6 ( 0.4
2.1 ( 0.5
32.6 ( 5.4
NO
g/kg fuel
0.5 ( 0.0
3.1 ( 0.2
2.5 ( 0.1
3.3 ( 0.2
0.9 ( 0.0
1.7 ( 0.1
4.5 ( 0.9
NO2
g/kg fuel
0.3 ( 0.0
1.0 ( 0.3
0.7 ( 0.2
0.3 ( 0.1
0.2 ( 0.1
0.8 ( 0.1
6.5 ( 1.1
PM
g/kg fuel
3.3 ( 0.8
4.0 ( 0.6
9.8 ( 1.8
6.4 ( 0.9
3.9 ( 0.6
2.2 ( 1.1
10.3 ( 0.9
σext(532 nm)
m2/kg fuel
29.0 ( 2.8
26.7 ( 2.5
52.3 ( 6.9
30.2 ( 5.1
24.9 ( 4.7
11.9 ( 6.6
43.7 ( 7.0
σabs(532 nm)
m2/kg fuel
19.8 ( 2.2
15.5 ( 1.2
7.6 ( 1.0
17.4 ( 2.3
14.3 ( 1.3
3.2 ( 0.7
0.9 ( 0.0
σabs(532 nm) by NO2e
m2/kg fuel
0.1 ( 0.0
0.2 ( 0.0
0.1 ( 0.0
0.1 ( 0.0
0.0 ( 0.0
0.1 ( 0.0
1.0 ( 0.2
σabs(1047 nm)
m2/kg fuel
10.8 ( 1.3
8.2 ( 0.7
3.8 ( 0.6
8.4 ( 1.3
7.4 ( 0.7
1.5 ( 0.4
σscat(450/550/700 nm)f
m2/kg fuel
7.7/5.5/3.5
9.9/6.9/4.2
40.2/33.1/25.0
16.3/10.7/6.3
13.4/9.5/6.0
9.1/6.5/4.2
38.5/30.9/23.4
?σscat(450/550/700 nm)f
m2/kg fuel
1.5/1.1/0.8
1.8/1.3/1.0
5.4/4.3/3.7
2.9/2.1/1.5
2.2/1.6/1.2
1.4/1.0/0.8
5.4/4.2/3.6
aCFand CSrepresent the percentage of carbon emission measured from the flaming phase and smoldering phase, respectively. PPWOOD: ponderosa pine wood; PPNEED: ponderosa pine needles; WPNEED:
white pine needles; SAGE: sagebrush; EXCEL: shredded aspen wood product; DGRASS: Dambo grass; MTGRASS: Montana grass; TUNDRA: tundra core; KERO: kerosene flame soot.bUncertainties for MTGRASS
are half the difference of two replicate burns.cEmission factors are presented in g/kgC (per kg carbon) in fuel since the carbon mass fraction in fuel was not determined.dEFTHCis reported as gCH4/kg fuel, i.e., an
average molecular weight of 16 per carbon atom.eσabsby NO2is calculated using an absorption efficiency of 0.306 Mm-1/ppb NO2at 532 nm.fNephelometer data are without any truncation correction.
VOL. 41, NO. 12, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY94321

Page 6

TABLE 2. Mass Percentage (Average and 1σ) of Elements and Carbon Fractions in PM Emitted from Burning of Wildland Fuels and Kerosene. Carbon Fractions Are Determined by the IMPROVE
and STN Thermal Protocols.
speciesponderosa pine wood ponderosa pine needleswhite pine needles sagebrush excelsior Dambo grassMontana grasstundra corekerosene soot
OCa(IMPROVE)
EC (IMPROVE)
TC (IMPROVE)
OC1 (IMPROVE)
OC2 (IMPROVE)
OC3 (IMPROVE)
OC4 (IMPROVE)
OP (IMPROVE)
EC1 (IMPROVE)
EC2 (IMPROVE)
EC3 (IMPROVE)
OC (STN)
EC (STN)
Al
Si
P
S
Cl
K
Ca
Ti
Mn
Fe
Ni
Cu
Zn
Se
Br
Pb
SUMb
OM/OCc(IMPROVE)
15.64 ( 7.52
80.41 ( 4.85
96.05 ( 8.88
1.72 ( 2.29
5.51 ( 3.58
6.60 ( 1.85
1.81 ( 1.63
0.00 ( 1.61
10.33 ( 1.97
70.01 ( 5.35
0.06 ( 1.61
41.65 ( 9.37
79.83 ( 6.50
0.287 ( 0.119
0.371 ( 0.056
0.063 ( 0.041
0.540 ( 0.136
0.635 ( 0.136
2.108 ( 0.034
0.503 ( 0.042
0.017 ( 0.024
0.013 ( 0.025
0.152 ( 0.053
0.050 ( 0.038
0.241 ( 0.162
0.214 ( 0.092
0.000 ( 0.005
0.000 ( 0.006
0.015 ( 0.021
101.26 ( 8.89
0.93 ( 0.55
31.42 ( 4.09
57.27 ( 3.87
88.70 ( 6.98
2.29 ( 2.28
7.85 ( 2.01
13.37 ( 2.46
7.01 ( 2.03
0.90 ( 1.90
55.92 ( 6.35
2.26 ( 1.90
0.00 ( 1.89
44.19 ( 10.08
44.64 ( 4.92
0.505 ( 0.162
0.638 ( 0.211
0.115 ( 0.056
1.292 ( 0.211
5.605 ( 2.502
7.816 ( 2.119
0.458 ( 0.095
0.002 ( 0.006
0.058 ( 0.027
0.132 ( 0.080
0.046 ( 0.036
0.209 ( 0.124
0.165 ( 0.033
0.000 ( 0.005
0.007 ( 0.010
0.000 ( 0.020
105.74 ( 7.72
1.07 ( 0.19
51.45 ( 6.22
17.48 ( 3.94
68.93 ( 4.79
8.47 ( 3.99
14.05 ( 1.40
19.32 ( 2.73
8.12 ( 1.71
1.50 ( 2.12
15.57 ( 3.65
3.10 ( 1.40
0.30 ( 0.63
63.98 ( 13.58
7.23 ( 1.27
0.297 ( 0.103
0.611 ( 0.174
0.013 ( 0.007
0.342 ( 0.029
0.701 ( 0.227
1.063 ( 0.169
0.352 ( 0.063
0.035 ( 0.021
0.029 ( 0.011
0.243 ( 0.024
0.025 ( 0.006
0.163 ( 0.008
0.083 ( 0.018
0.000 ( 0.002
0.002 ( 0.003
0.005 ( 0.008
72.89 ( 4.81
1.51 ( 0.20
28.01 ( 7.50
35.68 ( 3.45
63.69 ( 4.61
2.14 ( 0.93
6.88 ( 1.58
12.46 ( 5.88
5.96 ( 1.44
0.56 ( 0.92
35.14 ( 4.18
1.10 ( 0.92
0.00 ( 0.91
41.06 ( 8.87
22.39 ( 2.42
0.498 ( 0.077
0.270 ( 0.138
0.188 ( 0.075
2.929 ( 0.153
9.643 ( 0.447
23.709 ( 0.050
0.192 ( 0.102
0.002 ( 0.004
0.008 ( 0.016
0.090 ( 0.044
0.024 ( 0.012
0.117 ( 0.009
0.140 ( 0.014
0.010 ( 0.014
0.008 ( 0.004
0.008 ( 0.012
101.52 ( 4.64
1.78 ( 0.49
29.44 ( 8.00
51.21 ( 3.19
80.65 ( 10.45
1.52 ( 1.64
7.05 ( 2.12
13.82 ( 2.95
6.63 ( 1.43
0.43 ( 1.23
47.09 ( 5.17
4.37 ( 1.25
0.19 ( 1.20
47.55 ( 10.33
35.38 ( 3.46
0.221 ( 0.111
0.276 ( 0.052
0.094 ( 0.024
1.005 ( 0.127
0.331 ( 0.161
5.757 ( 0.099
1.112 ( 0.011
0.000 ( 0.005
0.019 ( 0.027
0.093 ( 0.042
0.022 ( 0.017
0.164 ( 0.051
0.202 ( 0.010
0.000 ( 0.005
0.011 ( 0.015
0.014 ( 0.017
89.97 ( 10.46
1.41 ( 0.40
62.6 ( 8.05
35.5 ( 13.04
98.21 ( 9.30
3.09 ( 1.83
15.11 ( 5.00
24.86 ( 3.73
13.91 ( 2.37
5.72 ( 2.47
16.66 ( 10.38
23.50 ( 6.67
1.08 ( 2.01
88.24 ( 19.00
6.79 ( 3.65
0.532 ( 0.379
0.698 ( 0.549
0.104 ( 0.066
1.248 ( 0.689
0.321 ( 0.080
2.900 ( 2.598
0.295 ( 0.180
0.000 ( 0.013
0.023 ( 0.058
0.234 ( 0.331
0.069 ( 0.052
0.259 ( 0.103
0.146 ( 0.070
0.000 ( 0.011
0.006 ( 0.013
0.006 ( 0.042
105.05 ( 9.71
0.88 ( 0.24
69.60 ( 14.07
5.62 ( 7.15
75.22 ( 21.22
0.04 ( 3.93
16.28 ( 4.35
33.04 ( 9.24
12.38 ( 4.35
7.86 ( 5.03
5.38 ( 3.96
7.50 ( 3.94
0.59 ( 3.94
76.35 ( 12.74
-0.03 ( 3.93
0.259 ( 0.233
0.424 ( 0.115
0.061 ( 0.067
0.183 ( 0.047
0.078 ( 0.042
0.486 ( 0.062
0.011 ( 0.019
0.000 ( 0.011
0.045 ( 0.050
0.165 ( 0.074
0.083 ( 0.087
0.548 ( 0.451
0.309 ( 0.245
0.000 ( 0.010
0.014 ( 0.020
0.000 ( 0.036
77.88 ( 21.23
1.32 ( 0.29
93.5 ( 18.67
2.64 ( 1.65
96.2 ( 11.60
11.74 ( 1.64
30.09 ( 5.13
31.92 ( 6.24
12.36 ( 3.06
7.47 ( 4.52
5.97 ( 1.83
3.81 ( 1.65
0.33 ( 1.65
98.2 ( 20.84
0.46 ( 1.63
0.148 ( 0.097
0.279 ( 0.047
0.020 ( 0.009
0.163 ( 0.003
0.142 ( 0.007
0.407 ( 0.001
0.218 ( 0.008
0.003 ( 0.005
0.029 ( 0.021
0.137 ( 0.031
0.017 ( 0.006
0.084 ( 0.005
0.044 ( 0.006
0.000 ( 0.004
0.000 ( 0.005
0.012 ( 0.015
97.92 ( 11.60
1.01 ( 0.20
39.75 ( 11.91
79.50 ( 12.05
119.25 ( 16.80
6.77 ( 8.91
16.04 ( 9.28
13.90 ( 9.29
3.03 ( 8.94
0.00 ( 8.91
3.21 ( 8.92
75.40 ( 9.97
0.89 ( 8.94
60.96 ( 15.67
65.95 ( 10.06
0.683 ( 0.529
1.065 ( 0.249
0.143 ( 0.049
1.113 ( 0.021
0.064 ( 0.037
0.842 ( 0.002
0.207 ( 0.043
0.000 ( 0.025
0.000 ( 0.113
0.366 ( 0.170
0.048 ( 0.030
0.620 ( 0.025
0.286 ( 0.033
0.000 ( 0.022
0.095 ( 0.025
0.000 ( 0.081
124.78 ( 16.81
0.27 ( 0.32
aOC and organic fractions may contain organic vapors adsorbed on the quartz-fiber filter.bSum of measured species includes TC (IMPROVE) and all elemental species.c[OM] ) [PM] - 4.125[S] - 2.2[Al] - 2.49[Si]
- 1.63[Ca] - 2.42 [Fe] - 1.94[Ti] - [EC] (IMPROVE).
43229ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 12, 2007

Page 7

determined from quartz-fiber filters in which adsorbed
organic vapors are measured as OC (24). This may have
inflated the particulate OC fraction and explains why TC
exceeds 100% PM mass in kerosene soot (Table 2, the PM
massisdeterminedgravimetricallyusingTeflon-membrane
filters that do not adsorb organic vapors). EC is not strongly
influenced by vapor adsorption. The actual carbon content
in PM is likely between 50% and 100%. A wide range of the
organic mass (OM)/OC ratio estimated from mass closure
(Table 2) reflects uncertainties from the sampling artifact,
OC/EC split, and potential unaccounted species.
Elementalpotassium(K),chlorine(Cl),andsulfur(S)make
up most of the remaining PM mass. The mass percentages
of K and Cl can be as high as 23.7% and 9.6%, respectively,
insagebrush,buttheyareparticularlylow(<1%)inthesmoke
of the two wet fuels (Montana grass and tundra core). In the
ponderosa pine wood smoke, the PM contains 2.1%, 0.64%,
0.54%, and 0.21% of K, Cl, S, and Zn, respectively, consistent
with other measurements for ponderosa pine slash burning
(25). The copper (Cu) content, however, is almost 2 orders
of magnitude higher (0.24% in this study versus 0.0038% in
Turnetal.(25)),possiblyduetocontaminationfromcopper
samplinglines.Thesumofspeciesappearstoexplainallthe
PM mass except for white pine needles and Montana grass,
wheretheOCfractionisrelativelyhigh.Oxygen(O),nitrogen,
and hydrogen associated with OC, S, and crustal elements
are not included in the summed mass.
EC abundances from the IMPROVE protocol are higher
than those from the STN protocol for all cases in Table 2.
This difference is mostly a result of different charring
corrections (i.e., reflectance versus transmittance), as dis-
cussedinChowetal.(15)andChenetal.(26).TheIMPROVE
protocol further segregates the PM carbon into eight
operationally defined fractions with OC1, OC2, OC3, and
OC4 evolved at 120, 250, 450, and 550 °C, respectively, in a
purehelium(He)atmosphereandEC1,EC2,andEC3evolved
at 550, 700, and 800 °C in a 2% O2/98% He atmosphere. OP
representsthepyrolyzedcarbondeterminedfromthechange
in filter reflectance (26). These OC and EC subfractions are
reported by the nationwide IMPROVE monitoring network
and have been found useful for PM source apportionment
(27-29).
As shown in Figure 2, the wood smoke is relatively rich
inEC,whileneedles,grass,andtundracorecombustionyield
higher OC fractions. Most of the EC should originate from
the flaming combustion. The EC subfractions vary dramati-
cally between different fuels. Ponderosa pine wood burning
resemblesthekeroseneflame,emittingECthatisdominated
(88-95%)bythehigh-temperatureEC2(i.e.,700°C).Dambo
grass also produces a higher EC2 fraction (∼66% of EC), but
EC1ismoreabundantinotherburns(Figure2).EC2isoften
used as a marker for diesel soot (e.g., 28, 30). Due to lower
combustion temperatures, gasoline vehicle and wood com-
bustion are not expected to produce abundant EC2. This
experiment shows otherwise. For OC subfractions, the
majorityishigh-temperatureOC3andOC4inallthesamples
tested(exceptforkerosenesoot),incontrasttogasolineand
diesel exhausts where OC1 and OC2 are more abundant (30,
31). OC3 and OC4 likely represent polar and/or high-
molecular-weight organic compounds that are less volatile.
The amount of OP also tends to increase with the OC3 and
OC4 content.
3.4 Relations of EC with σabsand CO. The mass specific
absorptionefficiencyofkerosenesootwasdeterminedtobe
FIGURE 2. Mass percentage of thermally resolved carbon fractions in PM. The numbers indicate the mass percentage of EC. OP is the
difference between EC1 + EC2 + EC3 and EC (i.e., OP is the gray area above the red bar).
FIGURE 3. Relation between light absorption (at 532 nm and
correctedforNO2lightabsorption)emissionfactorsandECemission
factors measured by the IMPROVE and STN carbon analysis
protocols. The scatter plot includes all fuels in this study except
kerosene. Unweighted linear regression is used to calculate the
slope and intercept.
VOL. 41, NO. 12, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY94323

Page 8

6.3 m2/gPM at 532 nm (8). Thermal analysis found its EC
mass fraction at 79% and 66% of PM by the IMPROVE and
STN protocols, respectively. This translates to an EC mass
absorption efficiency of 8.0 m2/g (IMPROVE) and 9.6 m2/g
(STN). The conversion factor between EFECand EFσabs(532
nm)emittedfrombiomasscombustioninthisstudyisinthe
same range, i.e., 9.6 ( 0.8 m2/g by the IMPROVE protocol to
7.8(0.9m2/gbytheSTNprotocolbasedonregressionslopes
(Figure 3). These values might have been influenced by a
largeregressioninterceptforbothIMPROVEandSTN,which
isconsistentwithbiasesinopticalcharringcorrectionwhen
EC concentrations are low (e.g., 26). We can also calculate
the conversion factor for individual burns using the
EFσabs/EFEC ratio. The highest two EC emission factors
(IMPROVE EC) are 2.6 ( 0.8 g/kg fuel for ponderosa pine
wood burning and 2.3 ( 0.5 g/kg fuel for ponderosa pine
needleburning.AlthoughthethermalpropertiesofECdiffer
(i.e., abundant EC2 in ponderosa pine wood versus EC1 in
ponderosa pine needles), the conversion factors are similar
(7.5 ( 0.9 m2/g EC versus 6.6 ( 0.5 m2/g EC at 532 nm).
COhasbeenusedtoestimatetheECemissionsonregional
andglobalscales(e.g.,32,33)becausebothCOandECresult
fromincompletecombustionsandCOemissioninventories
are relatively well developed. This study suggests that EC
and CO are predominately emitted from flaming and
smolderingcombustion,respectively.Thiscallsintoquestion
the use of CO emissions as surrogate for EC emissions from
biomass burning since their relationship depends on the
contribution from each combustion phase. Information on
fuel and combustion conditions within the geographical
domain of interest will be needed to refine the current
approach. The IMPROVE EC/CO ratios determined in this
study are as follows: ponderosa pine wood (0.15 g/g),
ponderosapineneedles(0.072g/g),whitepineneedles(0.035
g/g),Dambograss(0.039g/g),sagebrush(0.038g/g),excelsior
(0.097g/g),Montanagrass(0.023g/g),andtundracore(0.004
g/g). These values are higher than the ∆[MEC]/∆[MCO] ratio
measuredinambientairdominatedbyon-roadmotorvehicle
emissionsinNorthAmerica(0.003-0.007g/g(32))butcloser
to measurements in South Asia that represent a mixture of
biomass and fossil fuel combustions (0.013-0.027 g/g (33)).
Acknowledgments
This work was supported in part by the Joint Venture
Agreement 03-JV-11222049-102 between the USFS, Rocky
Mountain Research Station, Research Work Unit Number
4404andDRI.AdditionalsupportwasprovidedbyCalifornia
AirResourceBoardGrant04-307,USEPAGrantRD-83108601-
0, and Joint Fire Science Project through NPS Task
J8R07060005. Development of the DRI photoacoustic and
cavityringdown/cavityenhanceddetectioninstrumentswas
supported under NSF grants ATM-0340423 and ATM-
9871192. We acknowledge S. Baker, V. Kovalev, S. Leininger,
J. Newton, L. Rinehart, and C. Ryan for their efforts at the
USFS Fire Science Laboratory.
Supporting Information Available
Specifications of wildland fuels used for the 2003 pilot study
and fractional contributions of flaming phase and smolder-
ing phase to the gaseous and particulate emissions. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Literature Cited
(1) Watson, J. G. Visibility: Science and regulation. J. Air Waste
Manage. Assoc. 2002, 52, 628-713.
(2) Kasischke, E. S.; Penner, J. E. Improving global estimates of
atmospheric emissions from biomass burning. J. Geophys. Res.
2004, 109, D14S01, doi: 10.1029/2004JD004972.
(3) USEPA. Air Quality Criteria for Particulate Matter; EPA/600/
P-99/002aF; EPA/600/P-99/002bF; U.S. EPA: Washington, DC,
2004.
(4) Lee, S.; Baumann, K.; Schauer, J. J.; Sheesley, R. J.; Naeher, L.
P.; Meinardi, S.; Blake, D. R.; Edgerton, E. S.; Russell, A. G.;
Clements, M. Gaseous and particulate emissions from pre-
scribed burning in Georgia. Environ. Sci. Technol. 2005, 39,
9049-9056.
(5) Anderson, G.; Sandberg, D.; Norheim, R. Fire Emission Produc-
tion Simulator (FEPS) User’s Guide; U.S. Forest Service: Wash-
ington, DC, 2004.
(6) Reid, J. S.; Koppmann, R.; Eck, T. F.; Eleuterio, D. P. A review
of biomass burning emissions part II: intensive physical
properties of biomass burning particles. Atmos. Chem. Phys.
2005, 5, 799-825.
(7) Chakrabarty, R. K.; Moosmu ¨ller, H.; Garro, M. A.; Arnott, W. P.;
Walker, J.; Susott, R. A.; Babbitt, R. E.; Wold, C. E.; Lincoln, E.
N.; Hao, W. M. Emissions from the laboratory combustion of
wildland fuels: Particle morphology and size. J. Geophys. Res.
2006, 111, D07204, doi: 10.1029/2005JD006659.
(8) Chen, L.-W. A.; Moosmu ¨ller, H.; Arnott, W. P.; Chow, J. C.;
Watson, J. G.; Susott, R. A.; Babbitt, R. E.; Wold, C.; Lincoln, E.;
Hao, W. M. Particle emissions from laboratory combustion of
wildlandfuels: Insituopticalandmassmeasurements.Geophys.
Res. Lett. 2006, 33, 1-4, doi: 10.1029/2005GL024838.
(9) Goode, J. G.; Yokelson, R. J.; Susott, R. A.; Ward, D. E. Trace gas
emissions from laboratory biomass fires measured by Fourier
transforminfraredspectroscopy: Firesingrassandsurfacefuels.
J. Geophys. Res. 1999, 104, 21237-21245.
(10) Lipsky, E. M.; Robinson, A. L. Effects of dilution on fine particle
massandpartitioningofsemivolatileorganicsindieselexhaust
and wood smoke. Environ. Sci. Technol. 2006, 40, 155-162.
(11) Arnott, W. P.; Moosmu ¨ller, H.; Walker, J. W. Nitrogen dioxide
and kerosene-flame soot calibration of photoacoustic instru-
ments for measurement of light absorption by aerosols. Rev.
Sci. Instrum. 2000, 71, 4545-4552.
(12) Moosmu ¨ller, H.; Varma, R.; Arnott, W. P. Cavity ring-down and
cavity-enhanced detection techniques for the measurement of
aerosol extinction. Aerosol Sci. Technol. 2005, 39, 30-39.
(13) Anderson, T. L.; Covert, D. S.; Marshall, S. F.; Laucks, M. L.;
Charlson, R. J.; Waggoner, A. P.; Ogren, J. A.; Caldow, R.; Holm,
R. L.; Quant, F. R.; Sem, G. J.; Wiedensohler, A.; Ahlquist, N. C.;
Bates, T. S. Performance characteristics of a high-sensitivity,
three-wavelength, total scatter/backscatter nephelometer. J.
Atmos. Ocean. Technol. 1996, 13, 967-986.
(14) Watson, J. G.; Chow, J. C.; .Frazier, C. A. X-ray fluorescence
analysisofambientairsamples.InElementalAnalysisofAirborne
Particles,Vol.1;Landsberger,S.,Creatchman,M.,Eds.;Gordon
and Breach Science: Amsterdam, 1999; Ch. 2.
(15) Chow, J. C.; Watson, J. G.; Chen, L.-W. A.; Arnott, W. P.;
Moosmu ¨ller, H.; Fung, K. K. Equivalence of elemental carbon
by Thermal/Optical Reflectance and Transmittance with dif-
ferent temperature protocols. Environ. Sci. Technol. 2004, 38,
4414-4422.
(16) Mazzoleni, L. R.; Zielinska, B.; Moosmu ¨ller, H. Emissions of
levoglucosan, methoxy phenols, and organic acids from pre-
scribedburns,wildlandfuels,andresidentialwoodcombustion.
Environ. Sci. Technol. 2007, 41 (7), 2115-2122.
(17) Andreae, M. O.; Merlet, P. Emission of trace gases and aerosols
frombiomassburning.GlobalBiogeochem.Cycles2001,15,955-
966.
(18) Moosmu ¨ller, H.; Mazzoleni, C.; Barber, P. W.; Kuhns, H. D.;
Keislar,R.E.;Watson,J.G.On-roadmeasurementofautomotive
particle emissions by ultraviolet lidar and transmissometer:
Instrument. Environ. Sci. Technol. 2003, 37, 4971-4978.
(19) Yokelson, R. J.; Susott, R.; Ward, D. E.; Reardon, J.; Griffith, D.
W. T. Emissions from smoldering combustion of biomass
measured by open- path Fourier transform infrared spectros-
copy. J. Geophys. Res. 1997, 102, 18865-18877.
(20) Hao, W.-M.; Ward, D. E. Methane production from global
biomass burning. J. Geophys. Res. 1993, 98, 20657-20661.
(21) Hays, M. D.; Fine, P. M.; Geron, C. D.; Kleeman, M. J.; Gullett,
B. K. Open burning of agricultural biomass: Physical and
chemicalpropertiesofparticle-phaseemissions.Atmos.Environ.
2005, 39, 6747-6764.
(22) Ferek, R. J.; Reid, J. S.; Hobbs, P. V.; Blake, D. R.; Liousse, C.
Emissionfactorsofhydrocarbons,halocarbons,tracegases,and
particles from biomass burning in Brazil. J. Geophys. Res. 1998,
103, 32107-32118.
(23) Chow, J. C.; Watson, J. G.; Ashbaugh, L. L.; Magliano, K. L.
Similarities and differences in PM10chemical source profiles
for geological dust from the San Joaquin Valley, California.
43249ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 12, 2007

Page 9

Atmos.Environ.2003,37,1317-1340,doi: 10.1016/S1352-2310-
(02)01021-X.
(24) Turpin, B. J.; Huntzicker, J. J.; Hering, S. V. Investigation of
organic aerosol sampling artifacts in the Los Angeles Basin.
Atmos.Environ.1994,28,3061-3071,Research&Development.
(25) Turn,S.Q.;Jenkins,B.M.;Chow,J.C.;Pritchett,L.C.;Campbell,
D. E.; Cahill, T. A.; Whalen, S. A. Elemental characterization of
particulatematteremittedfrombiomassburning: Windtunnel
derived source profiles for herbaceous and wood fuels. J.
Geophys. Res. 1997, 102, 3683-3699.
(26) Chen,L.-W.A.;Chow,J.C.;Watson,J.G.;Moosmu ¨ller,H.;Arnott,
W. P. Modeling reflectance and transmittance of quartz-fiber
filter samples containing elemental carbon particles: Implica-
tions for thermal/optical analysis. J. Aerosol Sci. 2004, 35, 765-
780, doi: 10.1016/j.jaerosci.2003.12.005.
(27) Cao, J. J.; Wu, F.; Chow, J. C.; Lee, S. C.; Li, Y.; Chen, S. W.; An,
Z. S.; Fung, K. K.; Watson, J. G.; Zhu, C. S.; Liu, S. X.
Characterization and source apportionment of atmospheric
organic and elemental carbon during fall and winter of 2003 in
Xi’an, China. Atmos. Chem. Phys. 2005, 5, 3127-3137, 1680-
7324/acp/2005-5-3127.
(28) Kim, E.; Hopke, P. K. Source apportionment of fine particles at
Washington, DC, utilizing temperature-resolved carbon frac-
tions. J. Air Waste Manage. Assoc. 2004, 54, 773-785.
(29) Chen, L.-W. A.; Chow, J. C.; Watson, J. G.; Magliano, K.
Quantifying PM2.5 source contributions for the San Joaquin
Valleywithmultivariatereceptormodels.Environ.Sci.Technol.
2007, 41 (8), 2818-2826.
(30) Watson, J. G.; Chow, J. C.; Lowenthal, D. H.; Pritchett, L. C.;
Frazier, C. A.; Neuroth, G. R.; Robbins, R. Differences in the
carbon composition of source profiles for diesel- and
gasoline-powered vehicles. Atmos. Environ. 1994, 28, 2493-
2505.
(31) Chow, J. C.; Watson, J. G.; Kuhns, H. D.; Etyemezian, V.;
Lowenthal, D. H.; Crow, D. J.; Kohl, S. D.; Engelbrecht, J. P.;
Green, M. C. Source profiles for industrial, mobile, and area
sources in the Big Bend Regional Aerosol Visibility and
Observational (BRAVO) Study. Chemosphere 2004, 54, 185-
208.
(32) Chen, L.-W. A.; Doddridge, B. G.; Dickerson, R. R.; Chow, J. C.;
Mueller, P. K.; Quinn, J.; Butler, W. A. Seasonal variations in
elemental carbon aerosol, carbon monoxide, and sulfur diox-
ide: Implicationsforsources.Geophys.Res.Lett.2001,28,1711-
1714.
(33) Dickerson, R. R.; Andreae, M. O.; Campos, T.; Mayol-Bracero,
O. L.; Neusuess, C.; Streets, D. G. Analysis of black carbon and
carbonmonoxideobservedovertheIndianOcean: Implications
for emissions and photochemistry. J. Geophys. Res. 2002, 107,
8017, doi: 10.1029/2001JD000501.
Received for review October 4, 2006. Revised manuscript
received February 23, 2007. Accepted March 22, 2007.
ES062364I
VOL. 41, NO. 12, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY94325

Page 10

Table S1. Specifications of wildland fuels used for the 2003 pilot study at the Fire Science Laboratory, Missoula, MT.
Fuel Name Descriptions Origin C
(%)a,b
N
(%)a,b
Ash
(%)a
FM
(%)a,c,d
Mass
(g/per
burn)e
Ponderosa Pine Wood

Ponderosa Pine (Pinus ponderosa) is
commonly seen in temperate/boreal
forests and one of America's abundant
tree species, covering approximately 27
million acres of land. The wood was
prepared in the form of dried sticks.
US 48 0.07 0.28 6.0 ± 0.4 250
Ponderosa Pine Needles

See above. The needles were removed
from the trees and dried for combustion.
US 50 0.6 3.8 7.3 ± 0.8 250
White Pine Needles

White Pine (Pinus strobes) has soft
blue-green needles borne in groups of
five. It is the state tree of Maine and
Missouri. The needles were removed
from the trees and dried for combustion.
US 49 0.5 4.3 8.2 ± 0.7 250

Page 11

Table S1. (Continued)
Sagebrush

Sagebrush (Artemisia tridentate) is a
perennial shrub that grows in many arid
regions of the western U.S. The dried
fuel contains both leaves and branches.
US 48 0.92 3.5 8.3 ± 1.9 250
Excelsior

Excelsior is a dried wood product made
from Aspen poplar trees, which are found
throughout North America.
US 48 0.07 0.64 5.9 ± 0.5 250
Dambo Grass

Dambo grasses commonly occur in the
savannas of southern Africa, particularly
in Zambia and Zimbabwe. It was dried
and shipped to U.S.
Zambia 45 0.34 4.3 6.3 ± 0.5 250

Page 12

Table S1. (Continued)
Montana Grass

Grasses that were harvested around the
buildings of Fire Science Laboratory
(Missoula, MT) a few hours before
burning.
US/
Montana
44 0.17 4.2 17.5 250
Tundra Core

Tundra cores were acquired from Alaska.
The cores consisted of live moss at the
top and dead organic mass below it.
Carbon mass fraction was not measured.
Fuel moisture was determined only for
the live part, which was also the only part
that was burned.
US/
Alaska
N/Af
N/A N/A 113 ±
126
1250
a All percentages are relative to dry fuel mass.
b Carbon and nitrogen mass fraction was measured by a CHN analyzer (Perkin Elmer, Waltham, MA).
c Fuel moisture was determined by measuring the mass loss after holding the sample at 90ºC overnight.
d Fuel Moisture (FM) ranges are given as standard deviation, no errors are given for Montana grass as only two measurements (FM =
18.3 and 16.8%) were made. The large variability for the Tundra core is in part due to large inhomogeneity of the material.
e Fuel mass is the nominal fuel mass used for each burn.
f May be between 15 – 48% for the surface organic layer (1,2).

Page 13

Figure S1. Fractional contributions of flaming phase and smoldering phase to the gaseous and particulate emissions for (a) ponderosa
pine wood (b) ponderosa pine needles (c) white pine needles, (d) sagebrush (e) Excelsior (f) Dambo grass, and (g) Montana grass
burns, determined from the product of phase-specific fuel consumption (CF and CS) and emission factors (Table 1). Combustion of
tundra cores (not shown) shows only smoldering phase while combustion of kerosene shows only flaming phase.
(1.0)
(0.8)
(0.6)
(0.4)
(0.2)
0.0
0.2
0.4
0.6
0.8
1.0
COTHCNONO2
NO2
PMBext
σext
Bap
σabs
Fractional Contribution
Flaming Phase
Smoldering Phase
Ponderosa Pine Wood
(a)
(1.0)
(0.8)
(0.6)
(0.4)
(0.2)
0.0
0.2
0.4
0.6
0.8
1.0
COTHCNONO2
NO2
PMBext
σext
Bap
σabs
Fractional Contribution
Flaming Phase
Smoldering Phase
Ponderosa Pine Needles

(b)
(1.0)
(0.8)
(0.6)
(0.4)
(0.2)
0.0
0.2
0.4
0.6
0.8
1.0
COTHCNONO2
NO2
PMBext
σext
Bap
σabs
Fractional Contribution
Flaming Phase
Smoldering Phase
White Pine Needles
(c)
(1.0)
(0.8)
(0.6)
(0.4)
(0.2)
0.0
0.2
0.4
0.6
0.8
1.0
COTHCNONO2
NO2
PMBextBap
Fractional Contribution
Flaming Phase
Smoldering Phase
Sagebrush
σext
σabs

(d)

Page 14

(1.0)
(0.8)
(0.6)
(0.4)
(0.2)
0.0
0.2
0.4
0.6
0.8
1.0
COTHCNONO2
NO2
PMBext
σext
Bap
σabs
Fractional Contribution
Flaming Phase
Smoldering Phase
Excelsior
(e)
(1.0)
(0.8)
(0.6)
(0.4)
(0.2)
0.0
0.2
0.4
0.6
0.8
1.0
COTHC NONO2
NO2
PMBext
σext
Bap
σabs
Fractional Contribution
Flaming Phase
Smoldering Phase
Dambo Grass

(f)
(1.0)
(0.8)
(0.6)
(0.4)
(0.2)
0.0
0.2
0.4
0.6
0.8
CO THCNO NO2
NO2
PMBextBap
Fractional Contribution
Flaming Phase
Smoldering Phase
Montana Grass
σext σabs
(g)

Page 15

Literature Cited

1. Kasischke, E. S.; Hyer, E.; Novelli, P.; Bruhwiler, L.; French, N. H. F.; Sukhinin, A. I.; Hewson, J. H.; Stocks, B. J. Influences
of boreal fire emissions on northern hemisphere atmospheric carbon and carbon monoxide. Global Biogeochem.Cycles 2005, 19,
GB1012, doi:10.1029/GB002300.
2. Harden, J. W.; Neff, J. C.; Sandberg, D. V.; Turetsky, M. R.; Ottmar, R.; Gleixner, G.; Fries, T. L.; Manies, K. L. Chemistry of
burning the forest floor during the FROSTFIRE experimental burn, interior Alaska, 1999. Global Biogeochem.Cycles 2004, 18,
GB3014, doi:10.1029/2003GB002194.