Subscriber access provided by WEBSTER UNIV
Energy and the Environment
Emission measurements from traditional biomass cookstoves in South Asia and Tibet Cheryl L. Weyant, Pengfei Chen, Ashma Vaidya, Chaoliu Li, Qianggong ZHANG, Ryan Thompson, Justin Ellis, Yanju Chen, Shichang Kang, Ganesh Ram Shrestha, Mahesh Yagnaraman, Joseph Arineitwe, Rufus Edwards, and Tami C. Bond Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.8b05199 • Publication Date (Web): 25 Feb 2019 Downloaded from http://pubs.acs.org on February 25, 2019
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 30
Environmental Science & Technology
Emission measurements from traditional biomass cookstoves in South Asia and Tibet Cheryl L. Weyant,† Pengfei Chen,‡,¶ Ashma Vaidya,§ Chaoliu Li,‡ Qianggong Zhang,‡ Ryan Thompson,†,k Justin Ellis,†,⊥ Yanju Chen,†,# Shichang Kang,¶,‡,@ Ganesh Ram Shrestha,§ Mahesh Yagnaraman,4 Joseph Arineitwe,∇ Rufus Edwards,†† and Tami C. Bond∗,† †Environmental Engineering, University of Illinois Urbana-Champaign, IL ‡Institute of Tibetan Plateau Research, Chinese Academy of Sciences, Center for Excellence in Tibetan Plateau Earth Sciences, Beijing 100101, China 1
¶State Key Laboratory of Cryospheric Science, Northwest Institute of Eco-environment and Resources, Chinese Academy of Sciences, Lanzhou 730000, China §Center for Rural Technology - Nepal kMountain Air Engineering, Cottage Grove, Oregon ⊥National Oceanic and Atmospheric Administration (NOAA) #California Air Resources Board, Sacramento, California @University of Chinese Academy of Sciences, Beijing 100039, China 4First Energy Pvt. Ltd. Pune, Maharashtra, India ∇Center for Integrated Research and Community Development (CIRCODU), Kampala, Uganda ††Department of Epidemiology, School of Medicine, University of California Irvine E-mail:
[email protected] 1
ACS Paragon Plus Environment
Environmental Science & Technology
2
Abstract
3
Traditional biomass stoves are a major global contributor to emissions that impact
4
climate change and health. This paper reports emission factors of particulate matter
5
(PM2.5 ), carbon monoxide (CO), organic carbon (OC), black carbon (EC), optical ab-
6
sorption, and scattering from 46 South Asian, 48 Tibetan, and 4 Ugandan stoves. These
7
measurements plus a literature review provide insight into the robustness of emission
8
factors used in emission inventories. Tibetan dung stoves produced high average PM2.5
9
emission factors (23 and 43 gkg−1 for chimney and open stoves) with low average EC
10
(0.3 and 0.7 gkg−1 , respectively). Comparatively, PM2.5 from South Asian stoves (7
11
gkg−1 ) was in the range of previous measurements and near values used in inventories.
12
EC emission factors varied between stoves and fuels (p < 0.001), without corresponding
13
differences in absorption; stoves that produced little EC, produced enough brown car-
14
bon to have about the same absorption as stoves with high EC emissions. In Tibetan
15
dung stoves, for example, OC contributed over 20% of the absorption. Overall, EC
16
emission factors were not correlated with PM2.5 and were constrained to low values,
17
relative to PM2.5 , over a wide range of combustion conditions. The average measured
18
EC emission factor (1 gkg−1 ), was near current inventory estimates.
TOC Art
2
ACS Paragon Plus Environment
Page 2 of 30
Page 3 of 30
Environmental Science & Technology
19
Introduction
20
In populations without sufficient access to gas or electricity, household energy demand is
21
often met with biomass fuels that are burned for heating, cooking, lighting, and other social
22
and cultural purposes. About three billion people use solid fuels for household energy and
23
the emissions negatively impact health, air quality, and climate on a global scale.
24
Household biomass combustion contributes a large, but uncertain, fraction of total par-
25
ticulate matter (PM), black carbon (BC), and organic carbon (OC) emissions. Zhang et al.
26
(2009) estimated that about half of the total anthropogenic PM, 60% of BC and 85% of OC
27
emissions are from household solid fuel combustion in China (in 2006), 1 while Ohara et al.
28
(2007) 2 attributed 81% of the BC and 96% of OC emissions to household solid fuel combus-
29
tion in Asia (in 2000). Globally, BC from household solid fuel combustion contributes an
30
estimated 57% of total anthropogenic emissions (in 2010), 32% is from biomass (2100 Tg). 3
31
These emission estimates are uncertain, in part, because they rely on few measurements that
32
do not cover the range of solid fuel types, stove designs, and operation observed in real-world
33
cooking. 3
34
Previous measurements of emissions from traditional biomass stoves may poorly represent
35
real-world cooking emissions. Approaches to measuring emissions have included (1) open
36
burning of cooking fuels without a stove, 4,5 (2) laboratory stove testing with controlled tasks
37
and fuel feeding, usually using a water boiling test (WBT), 6–10 (3) field tests in homes with
38
uncontrolled fuel type and feeding rate but a controlled task (field WBT), 11–13 and (4) field
39
tests in homes with uncontrolled fuel, feeding rate, and task (uncontrolled test). 9,14–22 Fuel
40
type, feeding rates, and cooking tasks influence emissions; laboratory studies and field tests
41
with the WBT have both been shown to have lower emissions and less variability compared
42
to uncontrolled field tests. 15,23 Field studies with uncontrolled operation have typically had
43
small sample sizes for each stove-fuel combination and high variability. 9,14–16 Given the
44
global variability in fuels, stoves, cooking practices, and observed emissions, these studies
45
may not reflect emissions outside their region of measurement. Tests in other regions would 3
ACS Paragon Plus Environment
Environmental Science & Technology
46
provide more confidence in global applicability of household biomass cooking emission factors.
47
In addition, few measurements are available from the Himalayas and Tibet, where biomass
48
burning emissions have an important influence on glacial melting. 24
49
This study reports emission measurements from 94 biomass combustion events in village
50
homes in Tibet, Nepal, India, and Uganda. Fuels include wood, dung, agricultural waste,
51
and combinations of these fuels. Emission factors and real-time variability of stove-fuel
52
combinations were computed and compared with other biomass stoves in this and previous
53
studies.
54
Methods
55
Emissions were measured from household cooking events from commonly used biomass stoves
56
in households in Tibet, India, and Nepal. These measurements were part of a campaign to
57
characterize BC emissions that deposit and accelerate glacial melting in the Himalayas. 25,26
58
Tests in Uganda were measurements of opportunity, not published previously. Twelve previ-
59
ously published tests in Honduras, from a study using the same methods, are also reported
60
here for comparison. 14,23 All tests were of traditional stoves in each region.
61
Stoves and fuels
62
Table 1 provides summary results for each stove group and Table S1 summarizes the stove
63
groups characterized.
64
Traditional clay stoves called “chulhas” are widespread in South Asia. They are typically
65
homemade, clay structures that are built into the floor of a kitchen with side ports for fuel
66
feeding. There were regional variations in chulha construction and fuels, though all chulhas
67
used a mix of biomass fuels. Six chulhas in Maharashtra, India were primarily fueled with
68
wood and dung. Dung in this region was usually mixed with rice husks and packed around
69
a stick, so the long pieces could be easily fed into the stove. Four traditional chulhas and
4
ACS Paragon Plus Environment
Page 4 of 30
Page 5 of 30
Environmental Science & Technology
70
seven improved chulhas (ICS) were measured in Nepal Midhills. The ICSs were similar to
71
the traditional stoves in outward appearance, but had an internal baffle that was intended
72
to improve the combustion. Dung in this region was made into patties that were broken and
73
fed into the stove. Fuel consumption for chulhas in India and Nepal Midhills are described
74
in Johnson et. al (2013). 27 In the Nepal Terai, 29 chulhas were measured with dung, wood,
75
and agricultural waste fuels. Multiple fuels were almost always used, sometimes in series but
76
often concurrently. The fuel was categorized as a particular fuel when it made up more than
77
60% of the fuel mass, and otherwise was categorized as mixed. These categories were used
78
to explore whether emission differences could be attributed to fuel type.
79
In Tibetan homes, two stove types were found in mobile tent homes: open combustion
80
stoves without chimneys (termed “Open dung”) and metal stoves with chimneys (termed
81
“Chimney dung”). Measurements were taken in the Nam Co region of the Tibetan Plateau
82
(4730 m above sea level ). Yak dung is the primary fuel at high elevations where woody plants
83
cannot survive. Similar stoves are described in a study reporting indoor air concentrations by
84
Xiao et. al (2015). 28 At lower elevations in the Tibetan Plateau, metal stoves with chimneys
85
(termed “Chimney wood”) burned wood in permanent houses.
86
Biomass fuel stoves can be used for cooking and heating. Stoves in Uganda, India,
87
and Nepal Midhills were measured in warm climates or in summer months and were used
88
primarily for cooking. Chulhas in the Nepal Terai were measured in winter and partially
89
provided heating energy. In the winter, cooks discontinued fuel feeding when the meal was
90
complete, but the heat from the chulha replaced other forms of household heating, such as
91
straw burning. Stoves in Tibet were used for both cooking and heating; fuel was fed for
92
heating even when food was not being prepared. Cooking and heating activities were not
93
monitored during individual tests nor distinguished in data processing.
5
ACS Paragon Plus Environment
Environmental Science & Technology
94
Sampling methods
95
All measurements were uncontrolled cooking tests where household cooks performed normal
96
daily cooking activities using the stoves and fuels of their choice.
97
Emissions were sampled using partial plume capture, as described previously. 14,29 Briefly,
98
emissions were drawn from the plume at 1-3 ft above the cooking vessel using multiple inlet
99
sampling probes described in Roden et al. (2006). 14 The height was determined by selecting
100
a location as near to the stove as possible, while allowing the cook clear and free access to
101
cook. When a chimney was present, a single inlet sample was drawn from the chimney.
102
The sampling system has been described in Weyant et. al (2014) and is summarized
103
here. 29 Particles were measured downstream of a 2.5 µm cut cyclone (URG 2000-30ED).
104
PM2.5 mass was collected on 47 mm PTFE (polytetrafluoroethylene) fiber filters (Fluoro-
105
pore Membrane Filters, FALP04700, Millipore) and measured gravimetrically using a mi-
106
crobalance (Cahn C-31, Thermo Electron Corp) in a temperature and humidity controlled
107
environment (20-25 deg C and 45-50% RH). EC and OC were collected on 47 mm quartz
108
fiber filters (TISSUQUARTZ 2500QAT-UP, Pall). An additional quartz filter behind the
109
Teflon filter collected adsorbed gas-phase carbon and the measured OC mass was subtracted
110
from the primary quartz filter OC mass to correct for gas-phase adsorption. 30 Filters were
111
sealed and stored in ice-packed insulated containers after collection and maintained at -4 ◦ C
112
prior to analysis to prevent loss of volatile organic carbon. 31
113
Particle optical properties were measured in real-time (1 Hz). Particle scattering was
114
measured with a narrow-angle red-wavelength light sensor (635 nm, Aprovecho Research
115
Laboratory). 75% percent of open dung and chimney dung stove tests in Tibet had an
116
overranging scattering signal and reports of scattering were omitted for these tests. Absorp-
117
tion was measured using a 3-wavelength (467, 530, and 660 nm) Particle Soot Absorption
118
Photometer (PSAP, Radiance Research). The filter spot size correction was applied as in
119
Bond et al. (1999). 32 Downstream of particle measurements and filtration, CO and CO2
120
concentrations were measured in real-time (1 Hz) using an electrochemical sensor (SS1128, 6
ACS Paragon Plus Environment
Page 6 of 30
Page 7 of 30
121
Environmental Science & Technology
Senko) and a non-dispersive infrared sensor (Telaire T6615, GE), respectively.
122
Flows were measured in-line at six sampling points in the system using a thin film flow
123
calibrator (mini-BUCK calibrator APB-805000) before and after each measurement. In addi-
124
tion, flows through each filter were measured in real-time with mass flow meters (Honeywell
125
AWM3300V). The system was checked for leaks by confirming that the system could sus-
126
tain negative pressure to the entire closed system. The type, mass, and moisture content
127
(conductance meter) of the fuel was recorded for some events and is included in the SI
128
dataset.
129
Data analysis methods
130
Emission factors in grams of pollutant per kilogram fuel (gkg−1 ) were calculated using the
131
carbon-balance method, which relies on conservation of carbon mass before and after com-
132
bustion. The method is described in the SI section 3 and in previous literature. 14,33 A carbon
133
mass fraction of 0.5 was used for wood and agricultural waste and 0.3 was used for dung
134
fuel, 34 with uncertainties of 2% and 10%, respectively. Emission factors are reported for
135
CO (EFCO ), PM2.5 (EFPM ), EC (EFEC ), and OC (EFPM ). Optical emission factors were
136
calculated similarly for scattering (EFscattering ) and absorption (EFabsorption ). The average
137
emission factor uncertainty was ±22% for wood burning stoves and ±24% for dung stoves.
138
The uncertainty was driven by calibration drift during measurement campaign.
139
Several common metrics were computed to characterize the emissions. (1) The Absorp-
140
tion ˚ Angstr¨om Exponent (AAE) quantifies the wavelength dependence of absorption. An
141
AAE of 1 indicates particles with a constant refractive index over a range of wavelengths 35
142
and is associated with black carbon, and higher values indicate brown or yellow material. (2)
143
The mass scattering cross-section (MSC) is the scattering observed for a given aerosol mass,
144
and is useful to translate real-time scattering measurements to approximate mass measure-
145
ments. (3) The mass absorption cross-section (MACEC ) is an indication of the absorptivity
146
of EC. Metric calculations are described in the SI section 3. 7
ACS Paragon Plus Environment
Environmental Science & Technology
147
All quartz filters were analyzed using a Sunset Laboratory OC/EC analyzer equipped
148
for analysis using the thermal optical transmittance (TOT) method 36 with the temperature
149
procedure in the SI section 4. A reflectance detector was added to the Sunset analyzer mid-
150
project and 19 of 29 tests from the Nepal Terai and all tests in Tibet were also analyzed using
151
thermal-optical reflectance (TOR). Neither TOT nor TOR analysis have been conclusively
152
determined to be more accurate, but Chow et al. (2004) suggested that reflectance is more
153
consistent across different temperature profiles and agrees better with EC determined by
154
filter absorption. 37 Reflectance derived EC was used when available due to previous research
155
and the observation that EC concentrations were better correlated with measured absorption
156
(discussed further in Results and SI section 7).
157
For real-time analyses, each minute of data was averaged and metrics and emission factors
158
were calculated. These are termed “one-minute” averages, while the average of an entire
159
cooking event is termed an “event average”. Statistical tests of comparison of means were
160
done with the Welch test for unequal variances and unequal sample sizes and the Games-
161
Howell post-hoc comparison test. 38,39 Significance was reported at the 0.05 level. Dispersion
162
of the data was reported as the mean ± standard deviation.
163
Results and discussion
164
Average CO and PM2.5 emission factors
165
Table 1 shows emission factors and particle optical metrics for each stove group in the
166
measurement dataset and select metrics are presented in Figure 1. Previous studies are also
167
summarized in Figure 1, tabulated in Table S6, and discussed in a later section.
8
ACS Paragon Plus Environment
Page 8 of 30
Page 9 of 30
Environmental Science & Technology
Table 1: Average and standard deviation of emission factors and emission metrics separated by stove group. Stove group
N
EFCO gkg−1
EFPM gkg−1
EFEC gkg−1
EFOC gkg−1
AAE -
MSC m2 g
MACEC m2 g
South Asian chulha India chulha Nepal Hills chulha Nepal Terai chulha Nepal ICS chulha Average chulha
6 4 29 7 46
69.0 (21.9) 56.2 (37.4) 117.7 (35.9) 70.3 (15.2) 98.8 (40.2)
11.9 (7.6) 4.3 (0.8) 6.7 (2.6) 5.8 (2.6) 7.0 (4.0)
0.9 (0.3) 1.1 (0.3) 1.7 (0.8) 0.8 (0.6) 1.4 (0.8)
5.6 (5.2) 3.4 (2.1) 2.3 (1.8) 3.3 (1.7) 3.0 (2.7)
2.3 (0.6) 1.8 (0.1) 1.6 (0.2) 1.7 (0.2) 1.7 (0.3)
1.8 (0.6) 2.1 (0.9) 1.7 (0.6) 3.0 (1.9) 1.9 (1.1)
21.6 (4.7) 25.7 (20.7) 10.9 (2.3) 25.4 (7.8) 14.7 (8.6)
Tibetan stoves Open dung Chimney dung Chimney wood Average Tibetan
14 14 20 48
130.4 (43.1) 107.4 (29.9) 121.4 (31.4) 120.0 (35.2)
42.7 (26.0) 23.2 (12.6) 12.7 (10.2) 24.5 (20.8)
0.7 (0.8)c 0.3 (0.1)c 0.9 (0.8)c 0.7 (0.7)c
31.6 (19.8) 16.2 (10.2) 8.5 (7.3) 17.5 (15.9)
3.8 (1.0) 3.2 (0.5) 2.3 (0.6) 3.0 (0.9)
6.2 (2.2)b 2.0 (2.4)b 4.1 (2.2) 4.4 (2.3)b
41.3 (13.8) 69.2 (27.5) 27.3 (6.1) 43.8 (24.8)
Other traditional Uganda 3-stone fire Hondurasa
4 12
67.2 (14.5) 116 (55)
14.6 (11.8) 8.5 (1.6)
0.8 (0.6) 1.5 (0.3)
4.2 (3.0) 4.0 (0.9)
1.6 (0.3) 1.5 (0.3)
1.4 (0.9) 2.2 (0.6)
29.2 (9.9) 17.0 (6.3)
a Results
were previously published in Roden et al. (2006), and are included here for comparison. 14 with over-ranging scattering sensor were omitted (discussed below). c Thermal-optical reflectance was used for the charring correction instead of transmittance. b Tests
168
CO emission factors varied significantly by stove group. They ranged from an average
169
of 56±37 gkg−1 for chulhas measured in the Nepal Midhills to 130±43 gkg−1 for Tibetan
170
open dung stoves. The average for all South Asian chulhas (99±40 gkg−1 ) was lower than
171
for traditional stoves measured in Honduras and Tibet (116±55 gkg−1 and 120±35 gkg−1 , p
172
= 0.001, respectively). 14
173
PM2.5 and CO emission factors followed a similar pattern; stove groups with a lower CO
174
emission factor also had a lower PM2.5 emission factor. The PM2.5 emission factor for the
175
South Asian chulhas, (7.0±4.0 gkg−1 ) was lower than for Tibetan chimney dung (23±13
176
gkg−1 , p = 0.003) and for open dung stoves (43±26 gkg−1 , p = 0.003). The CO and PM2.5
177
emission factors for event averages across all regions were weakly correlated (r = 0.3, p
0.05). The
182
highest values were from Tibetan open dung stoves (6.2 m2 g−1 ). Most average MSC values
183
were in the range of previous estimates; Roden et. al (2006) reported an average of 2.2 m2 g−1
184
(530 nm) from field tests in Honduras 14 and values of 2.5 – 4.2 m2 g−1 (550 nm), have been
185
reported from biomass burning. 40–42 In the Tibetan open and chimney dung stoves, tests
186
were omitted when overranging occurred in the scattering sensor. When partial scattering
187
data were included (overranging seconds treated as missing and no events omitted), the MSC
188
tended to be slightly higher; 6.6 m2 g−1 for open dung stoves and 3.5 m2 g−1 for chimney dung
189
stoves. Yet, these averages are likely to be lower bounds because high scattering events were
190
omitted while particle mass was collected.
191
Scattering and PM2.5 emission factors for event averages were correlated across all stove
192
groups (r = 0.77, p < 0.001), yet tended to have lower levels of correlation within groups (SI
193
section 6). Low correlation coefficients (r = 0.0 – 0.6) occurred in all groups except for chulhas
194
(r = 0.77). A combination of high variability and relatively few events contribute to these
195
low correlation coefficients. Low correlation coefficients highlight the inherent uncertainties 10
ACS Paragon Plus Environment
Page 11 of 30
Environmental Science & Technology
196
when using scattering as a proxy for particle mass as is done with many low cost continuous
197
particle monitors.
198
Elemental and organic carbon
199
In contrast to CO and PM2.5 emission factors, EC emission factors were highest for South
200
Asian chulhas (1.4 gkg−1 ) and lowest in Tibetan chimney dung stoves (0.3 gkg−1 , p < 0.001).
201
The chimney dung stoves also produced significantly lower EC emission factors compared
202
to the chimney wood stoves (p = 0.009). The EC emission factor for the Tibetan open
203
dung stoves was between the other Tibetan stove groups, without significant differences (p
204
> 0.05), yet was significantly lower than for chulhas (p = 0.03).
205
The EC emission factors were notably small in Tibetan stoves, considering the large
206
PM2.5 emission factors. Figure 2 shows the emission factors of EC, OC, and PM2.5 for all
207
events, ordered by the EC emission factor magnitude. EC ranged from near zero to 4 gkg−1
208
while OC and PM2.5 ranged from about 4 - 100 gkg−1 and had little relationship with the
209
EC emission factors (r = -0.03, p > 0.05). Thus, estimations of EC as a fraction of PM2.5
210
would produce spurious results. Only in the Tibet open dung stoves is a relationship between
211
EC and OC (or PM2.5 ) apparent, and this is largely driven by a single outlier. It appears
212
that processes that produce high OC (or PM2.5 ) have little bearing on the amount of EC
213
formed, and that EC may be constrained to a fairly low emission factor over a wide range
214
of combustion conditions.
11
ACS Paragon Plus Environment
100
PM emission factor
80
OC emission factor EC emission factor
Page 12 of 30
60 40 20
4 3 2 1 0
3 stone fire
South Asian Chulha
Tibet chimney wood
Tibet chimney dung
Tibet open dung
−1
0
EC (gkg )
−1
Emission factor (gkg )
Environmental Science & Technology
Events Figure 2: Particle emission factors for individual events. The events are sorted by the EC emission factor magnitude within each stove group. The top panel shows the emission factors of PM2.5 as a line segment, OC as a light colored bar, and EC as a dark colored bar. EC is low in comparison with OC and PM2.5 , and the scale is enhanced in the bottom panel to more clearly show the range of EC emission factors.
215
Figures S4 - S7 show traces from the thermal-optical analysis used to determine OC and
216
EC. Samples from Tibet underwent large amounts of OC charring during the analysis which
217
confounded the accurate detection of EC. The estimated amount of EC varied depending on
218
use of the transmittance or reflectance signal (TOT or TOR) as exemplified in Figure 3 and
219
discussed in the SI section 7.
12
ACS Paragon Plus Environment
Environmental Science & Technology
Tibet: Chimney dung
−1
Absorption emission factor (m kg )
Page 13 of 30
2
40 30 Transmittance > reflectance Transmittance < reflectance Transmittance EC Reflectance EC 2 −1 MAC transmittance 113.6 m g 2 −1 MAC reflectance 69.2 m g
20 10 0 0
.2 .4 −1 EC emission factor (gkg )
.6
Figure 3: Scatter plot showing the relationship between absorption and EC emission factors determined by TOT and TOR methods in Tibetan chimney dung stoves. For each event, two EC emission factors are shown; determined by optical transmittance (black circle) or reflectance (gray square). The dotted lines are average MACEC derived from each method.
220
Tests in the Nepal Terai had similar OC and EC emission factors when either TOR
221
or TOT methods were used (p>0.5). However, stove groups in Tibet had large differences
222
depending on the optical analysis method (p 0.05). This is striking because it is in contrast to the order-of-magnitude differences in
243
the PM2.5 emission factors between groups (Figure 4 compared to Figure 1: center). This
244
contrast is also apparent in the similar distributions of absorption emission factors within
245
events compared to the range of scattering emission factors (Figure 5: left and center,
246
discussed in next section).
247
Most absorption is expected to be due to EC, yet stove groups with statistically lower
248
EC emission factors did not have lower absorption. This is expressed by the range in the
249
amount of absorption per EC (MACEC ) between groups. The MACEC ranged from 10.9 –
250
69.2 m2 g−1 and was highest in stoves with low EC emission factors (Tibetan dung stoves)
251
and lower when EC emissions factors were high (chulhas).
252
The MACEC values observed were higher than typically found for other sources of black
253
carbon. A MACEC value recommended for pure BC in models is 7.5±1.2 m2 g−1 at 550 nm. 44
254
South Asian chulhas had values that were a factor of two higher and MACEC of Tibetan
255
stoves was a factor of five higher than this value. Some absorption is likely attributable to
256
absorbing OC (brown carbon), 45 so the ratio of all particle absorption over EC mass would
257
overestimate the true value of MACEC . In cases where the EC:OC ratio is small, the relative
258
impact of OC absorption on MACEC can be high.
259
The impact of absorption by OC can be approximated by applying linear regression using
14
ACS Paragon Plus Environment
Page 14 of 30
Page 15 of 30
260
Environmental Science & Technology
Equation 1.
EFabsorption = β1 EFEC + β2 EFOC + β3 (EFEC × EFOC )
(1)
261
β1 is an estimate of the absorption per mass of EC, but unlike MACEC , Equation 1 allows
262
apportioning some absorption to OC with the coefficient β2 and OC and EC interactions
263
2 −1 2 −2 with β3 . Over all events, β1 was 12.8 m2 g−1 EC , β2 was 0.7 m gOC , and β3 was -0.17 m gEC×OC
264
(R2 = 0.9, p < 0.001 for all coefficients). EC does explain most of the variability in the
265
2 absorption: β1 = 15.5 m2 g−1 EC , R = 0.7, p < 0.001), in the case where β2 and β3 terms are
266
omited. However, estimates of the absorption emission factors using purely EC regression
267
are prone to errors when the EC:OC ratio is either high or low; EFabsorption is underestimated
268
by 40% on average when the EC:OC ratio is less than 0.2 and is overestimated by about
269
40% when the EC:OC ratio is greater than 0.8. When Equation 1 is used instead, the error
270
(20%) is not impacted by different EC:OC ratios (Figure S11). The regression is explored
271
further in Figure S10 and Table S7.
272
When Equation 1 was applied to emission factors of EC and OC reported in the lit-
273
erature (Table S6), calculated absorption emission factors were in the same range as the
274
stove groups in this study. This suggests that the combination of EC and OC emitted from
275
traditional stoves tends to result in a fairly stable absorption emission factor (confidence
276
interval: ±15%, compared to ±25% for EC). The average value for the measurements re-
277
ported here was 20.6±11.5 m2 kg−1 . When regression-derived literature emission factors were
278
included, the average absorption emission factor was 14.84±6.6 m2 kg−1 . This suggests the
279
average absorption emission factors found here are fairly similar across most traditional stove
280
combustion conditions.
15
ACS Paragon Plus Environment
Environmental Science & Technology
Page 16 of 30
Tibet: Open dung Tibet: Chimney dung Tibet: Chimney wood S. Asia: Clay chulha Uganda: 3 stone fire Literature: field Literature: lab 0
20
40
60 2
−1
Absorption emission factor (m kg )
Figure 4: Absorption emission factors of fresh particle emissions are similar across all stove groups. For the literature emission factors, the two red points are from Roden et al. (2006) and Grieshop et al. (2017) 14,17 and all others are determined using measured OC and EC values (Table S6) and a regression equation derived from this study.
281
In some stove groups, most absorption was attributed to OC, not EC. Using the Equation
282
1, for Tibetan open dung stoves, 70% of the absorption was attributed to OC. In comparison,
283
OC contributed to about 45% of the absorption in Tibetan chimney dung stoves, 20% in
284
Tibetan chimney wood stoves, and 10% in chulhas (Figure S9).
285
Values of AAE also indicate that cookstoves produce a significant amount of absorbing
286
organic brown carbon. AAE values ranged from 1.7 for chulhas to 3.8 for Tibet open dung,
287
where values higher than 1 suggest that absorbing OC is present. When AAE was used to
288
attribute absorption to OC, 46 using a delineating value of AAE of 1, 22% of the absorption
289
(530nm) was attributed to OC for Tibetan open dung stoves, 21% for Tibetan chimney dung
290
stoves, 12% for Tibetan chimney wood stoves, and 1% for chulhas. Although the attribution
291
of absorption to OC is uncertain, it remains significant for Tibetan dung stoves. Given the
292
proximity of Tibetan dung stoves to glaciers, ice melting in this region is likely accelerated
293
by absorbing OC in addition to EC.
294
Measurements of AAE in this study are in the same range as Roden et al. (2006) (1-5) 14
295
and higher than those observed by Grieshop et al. (2017) (1.2-1.3), 17 suggesting variable
296
brown carbon emissions in biomass cookstoves.
16
ACS Paragon Plus Environment
Page 17 of 30
Environmental Science & Technology
297
Variability within events
298
One-minute average metrics were used to assess variability within events. Probability den-
299
sity functions (PDFs) in Figure 5, show the range and density of one-minute metric values
300
throughout all events in each stove group. The one-minute average scattering and absorption
301
emission factors and AAE all have wider ranges than observed in event averages. Particle
302
scattering emission factors ranged over five orders of magnitude, absorption ranged over
303
three orders of magnitude, and AAE ranged from 0 to 8. The convergence of the absorption
304
emission factor between stoves, discussed above, is even more apparent it the distribution of
305
real-time data; the EC emission factor from Tibetan open dung stoves was half that from
306
South Asian chulhas, yet the distributions in absorption were nearly identical. Those same
307
stove groups produced distributions of particle scattering that were the most divergent. In
308
Tibetan stoves the scattering emission factor for the chimney wood stove shows a bimodal
309
distribution and this feature is due both to biomodal events and differences between events
310
as shown in Figure S13 . .8
(a)
1.5
(b)
.6 Density
Density
1 .4
.5 .2 0 −1 10
10
0
10
1
10
2
10 2
3
10
4
0 −1 10
−1
0
10
1
10 2
Scattering emission factor (m kg ) S. Asia: Clay chulha Tibet: Chimney wood
10
2
10
3
−1
Absorption emission factor (m kg )
Tibet: Chimney dung Tibet: Open dung
S. Asia: Clay chulha Tibet: Chimney wood
Tibet: Chimney dung Tibet: Open dung
Figure 5: Probability density plots of one-minute averages of emission parameters for all tests within each group. (a) Scattering emission factor. (b) Absorption emission factor.
17
ACS Paragon Plus Environment
Environmental Science & Technology
Page 18 of 30
(b)
(a) .8
Density
.6 0
2
4
6
8
0
2
4
6
8
0
2
4
6
8
.4 .2 0
0
2
4
6
8
Absorption Angstrom exponent S. Asia: Clay chulha Tibet: Chimney wood
Tibet: Chimney dung Tibet: Open dung
Figure 6: (a) Probability density functions of one-minute averages for the absorption Angstrom exponent for four stove groups. (b) Three figures show the overall PDF for each stove group in bold and the PDFs of each individual event measured within that stove group.
311
The variability within stove groups is apparent in the range of results in Figure 1, but
312
there are also variations in the shape of PDFs for events within stove groups. This is
313
shown in Figure 6 and Figures S12- S14. These figures indicate that individual events with
314
similar averages can have widely varying emission characteristics, such as wide or narrow
315
distributions or bimodality. Variation in emission distributions
316
that features other than stove, fuel, and region influence the combustion and emissions.
317
Subtle elements of stove construction and fuels likely affect overall emission factor averages,
318
but the variable features of the distributions within groups are likely caused by user cooking
319
practices. Probability distributions of real-time data may reflect variations in stove operation
320
better than event averages, and the use of distributions coupled with observations may aid
321
our ability to explain emission variability.
322
Effect of fuel choice on emission characteristics
323
The effect of fuel choice on emissions was not part of the study design. However, for two
324
stove groups, wood stoves in the Nepal Terai and chimney stoves in Tibet, stove design and 18
ACS Paragon Plus Environment
between events suggests
Page 19 of 30
Environmental Science & Technology
325
cooking practice were relatively constant while fuel choice varied. Emissions within these
326
groups were examined to evaluate the influence of fuel choice.
327
In the Nepal Terai, CO and PM2.5 emission factors were lowest in wood stoves and
328
higher when dung or agricultural waste was predominant, but these differences were not
329
significant (p > 0.05). The EC emission factor for fuel mixtures that were predominantly
330
agricultural waste (2.2±0.9 gkg−1 ) was higher than other fuel groups (1.0±0.6 gkg−1 , p =
331
0.02). Similarly, EC/(EC+OC) was 0.5±0.1 for tests with agricultural waste fuel and an
332
average of 0.3±0.1 for other fuels in the region. However, the difference was not significant
333
(p > 0.05). Dry, low density, agricultural waste fuels may be more likely to undergo flaming
334
(i.e. EC producing) combustion compared higher density, higher moisture retaining fuels
335
such as wood and dung. Guofeng et al. (2012) similarly found higher EC emission factors
336
for shubs (70% higher) and agricultural waste (120% higher) compared to woody fuels, 47
337
but others have found the opposite. 19,48
338
The chimney stoves measured in Tibet were used with either pure dung or wood fuel.
339
The differences in the emissions between these stove groups have been discussed above, but
340
in summary: dung fuel produced higher PM2.5 , similar CO, and lower EC emission factors
341
compared to wood fuel. Dung fuel also produced higher AAE than wood fuel, suggesting
342
OC that is more absorbing when dung fuel is used. However, only differences in EC were
343
statistically significant (p = 0.009).
344
Comparison with previous measurements and emission inventories
345
In order to estimate emission fluxes to the atmosphere, inventory developers typically select
346
a single emission factor or range for a particular type of fuel and use by examining pub-
347
lished measurements. Figure 1 and Table S6 summarize the body of knowledge available to
348
inventory developers.
349
Emission factors from South Asian chulhas were similar to those previously reported in
350
field tests for PM2.5 and CO (Figure 1), and therefore, as expected, also similar to recommen19
ACS Paragon Plus Environment
Environmental Science & Technology
351
dations for emission inventories. The PM2.5 emission factor (7.0±4.0 gkg−1 ) was within the
352
range proposed by Akagi et al. (2011) (6.6±1.7 gkg−1 ), 49 and similar to the GAINS (Green-
353
house gas - Air pollution Interactions and Synergies) inventory choice of about 8 gkg−1 (510
354
mgMJ−1 ) for biofuel-burning traditional stoves. 3 On the other hand, EC emission factors
355
(1.4±0.8 gkg−1 ) were higher than those used in the GAINS inventory (1.1 gkg−1 ) and
356
tended to be higher than those reported previously (0.9± 0.5 gkg−1 , p = 0.02).
357
Emission factors from combustion in Tibet have not been published previously, and the
358
values reported here diverge from inventory recommendations. PM2.5 emission factors in
359
Tibet were high (25 gkg−1 ±21 gkg−1 ) compared to the GAINS inventory and to emission
360
factors from field tests in the literature (10 gkg−1 ±6 gkg−1 , p = 0.03, Table S6). CO emission
361
factors from Tibetan stoves (120±35 gkg−1 ) were higher than a recommendation of Akagi
362
et al. (2011) of 77 gkg−1 for wood and 105 gkg−1 for dung. 49 EC emission factors in Tibet
363
(0.7±0.7 gkg−1 ) were lower than other published values, although this difference was not
364
significant, p > 0.05. 3,50
365
There are several reasons emissions from stoves in Tibet could differ from other stoves
366
in other areas, such as fuels used, use of stoves for heating, atmospheric conditions or a
367
combination of these factors. The dung used in Tibetan stoves was not mixed with biomass
368
as was typical in other places and yak dung was used, whereas cow or buffalo dung is more
369
typical in South Asia. In addition the stoves were used for both cooking and space heating
370
which may result in higher emission factors because heating stoves may have longer periods
371
of unattended smoldering, while cooking stoves are more likely to be tended to maintain
372
high temperatures and flaming conditions (smoldering discussed in SI section 13). Tissari
373
et al. (2008) found that PM2.5 emissions from smoldering combustion were about 10 times
374
higher than “normal combustion” at 11.1 gkg−1 PM1 , largely due to elevated organic carbon
375
emissions. 51
376
Tibetan stoves were partially heating stoves and previous measurements have also sug-
377
gested higher PM2.5 emission factors for heating stoves compared to cooking stoves. The
20
ACS Paragon Plus Environment
Page 20 of 30
Page 21 of 30
Environmental Science & Technology
378
US EPA compilation of emission factors (AP-42) uses a PM10 emission factor (comparable
379
to PM2.5 because combustion particles are typically smaller than 2.5µm 22 ) of 15.3 gkg−1 for
380
conventional wood heating stoves and 17.3 gkg−1 for fireplaces. 52 The wood heating stove
381
in Tibet was in the range of the AP-42 PM emission factor for heating stoves and is about
382
double that of cooking stoves in this study (7±4 gkg−1 ). Broderick et al. (2005) reviewed
383
fireplace emission factors in the United States and observed a log-normal distribution with
384
a long tail, where about 8% of PM emission factors were above 20 gkg−1 . 53
385
Finally, high altitude combustion may result in lower combustion efficiencies as oxygen
386
concentrations are only about 60% of those at sea level. The atmospheric pressure in Tibet,
387
at the location of the dung stove measurements, was 567 hPa, and for the wood stoves was 687
388
hPa. In comparison, stoves measured in Nepal were measured at 989 hPa and in Honduras,
389
around 1000 hPa. 14 One other high elevation combustion test was reported previously where
390
a heating stove measured at 731 hPa produced high PM (28.3 gkg−1 ) and CO (219 gkg−1 )
391
emission factors. 54 However, the altitude effect was not replicated in a controlled test at
392
834 hPa. 54 Lower flame temperatures and slower burning rates have been observed in high
393
altitude combustion, 48 but effects on emission factors are not well documented.
394
Overall, the PM2.5 emission factor from Tibetan stoves was two times higher than emis-
395
sion inventory estimates. The average EC emission factor for Tibetan stoves was 25% lower
396
and for South Asian chulhas was 55% higher than the GAINS inventory estimate. High
397
PM2.5 emission factors from smoldering combustion may occur in many global cooking en-
398
vironments and are likely high emitting , low probability events (the tail of a log-normal
399
distribution). These events are a challenge for inventory estimates where few measurements
400
are available and many stoves and cooking conditions have not been measured. The same
401
situation does not appear to apply for EC, which did not have the same variability observed
402
for PM2.5 . EC may be constrained to fairly low emission factors over a wide range of com-
403
bustion conditions and inventory estimates may be less likely to need significant adjustments
404
for EC.
21
ACS Paragon Plus Environment
Environmental Science & Technology
405
The EC emission factor is likely sufficiently constrained that a global average emission
406
factor for biomass stoves is appropriate. However, emissions of PM2.5 show regional variabil-
407
ity and a highly skewed log-normal distribution, such that a global value is prone to higher
408
uncertainty. For PM2.5 , regionally determined and applied emission factors in inventories is
409
likely more appropriate than a global average.
410
Acknowledgement
411
The authors thank Dr. Kirk Smith his contributions to original project design and site con-
412
nections. This research was supported by EPA STAR 83503601 Characterization of Emis-
413
sions from Small, Variable Solid Fuel Combustion Sources for Determining Global Emissions
414
and Climate Impact. The contents are solely the responsibility of the authors and do not
415
necessarily represent the official views of the US EPA. US EPA does not endorse the purchase
416
of any commercial products or services mentioned in the publication.
417
Supporting Information Available
418
Description of stoves and fuels, metric calculations, OCEC parameter file, details about
419
thermal-optical reflectance and transmittance, compilation of literature emission factors,
420
figures showing non-carbon components of PM2.5 , figure showing organic carbon aborption,
421
and figures showing additional probability density functions. Dataset of emission character-
422
istics for individual events.
423
References
424
(1) Zhang, Q.; Streets, D.; Carmichael, G.; He, K.; Huo, H.; Kannari, A.; Klimont, Z.;
425
Park, I.; Reddy, S.; Fu, J.; Chen, D.; Duan, L.; Lei, Y.; Wang, L.; Yao, Z. Asian
22
ACS Paragon Plus Environment
Page 22 of 30
Page 23 of 30
Environmental Science & Technology
426
emissions in 2006 for the NASA INTEX-B mission. Atmos. Chem. Phys. 2009, 9,
427
5131–5153.
428
(2) Ohara, T.; Akimoto, H.; Kurokawa, J.; Horii, N.; Yamaji, K.; Yan, X.; Hayasaka, T.
429
An Asian emission inventory of anthropogenic emission sources for the period 1980 -
430
2020. Atmos. Chem. Phys. 2007, 7, 4419–4444.
431
(3) Klimont, Z.; Kupiainen, K.; Heyes, C.; Purohit, P.; Cofala, J.; Rafaj, P.; Borken-
432
Kleefeld, J.; Sch¨opp, W. Global anthropogenic emissions of particulate matter including
433
black carbon. Atmos. Chem. Phys. 2017, 17, 8681–8723.
434
(4) Sen, A.; Mandal, T.; Sharma, S.; Saxena, M.; Gupta, N.; Gautam, R.; Gupta, A.;
435
Gill, T.; Rani, S.; Saud, T. Chemical properties of emission from biomass fuels used in
436
the rural sector of the western region of India. Atmos. Environ. 2014, 99, 411–424.
437
(5) Saxena, M.; Sharma, S. K.; Tomar, N.; Ghayas, H.; Sen, A.; Garhwal, R. S.;
438
Gupta, N. C.; Mandal, T. K. Residential biomass burning emissions over northwestern
439
Himalayan region of India: chemical characterization and budget estimation. Aerosol
440
Air Qual. Res. 2016, 16, 504–518.
441
(6) Global Alliance for Clean Cookstove partners, The Water Boiling Test Version 4.2.3
442
Cookstove Emissions and Efficiency in a Controlled Laboratory Setting. 2014; http:
443
//cleancookstoves.org/technology-and-fuels/testing/protocols.html.
444
(7) Jetter, J.; Zhao, Y.; Smith, K. R.; Khan, B.; Yelverton, T.; DeCarlo, P.; Hays, M. D.
445
Pollutant emissions and energy efficiency under controlled conditions for household
446
biomass cookstoves and implications for metrics useful in setting international test
447
standards. Environ. Sci. Technol. 2012, 46, 10827–10834.
448
(8) Just, B.; Rogak, S.; Kandlikar, M. Characterization of ultrafine particulate matter
449
from traditional and improved biomass cookstoves. Environ. Sci. Technol. 2013, 47,
450
3506–3512. 23
ACS Paragon Plus Environment
Environmental Science & Technology
451
(9) Jayarathne, T.; Stockwell, C. E.; Bhave, P. V.; Praveen, P. S.; Rathnayake, C. M.;
452
Islam, M. R.; Panday, A. K.; Adhikari, S.; Maharjan, R.; Goetz, J. D.; DeCarlo, P. F.;
453
Saikawa, E.; Yokelson, R. J.; Stone, E. A. Nepal Ambient Monitoring and Source
454
Testing Experiment (NAMaSTE): Emissions of particulate matter from wood and dung
455
cooking fires, garbage and crop residue burning, brick kilns, and other sources. Atmos.
456
Chem. Phys. Discuss. 2017, 2017, 1–51.
457
(10) Stockwell, C. E.; Christian, T. J.; Goetz, J. D.; Jayarathne, T.; Bhave, P. V.;
458
Praveen, P. S.; Adhikari, S.; Maharjan, R.; DeCarlo, P. F.; Stone, E. A.; Saikawa, E.;
459
Blake, D. R.; Simpson, I.; Yokelson, R. J.; Panday, A. K. Nepal Ambient Monitoring and
460
Source Testing Experiment (NAMaSTE): Emissions of trace gases and light-absorbing
461
carbon from wood and dung cooking fires, garbage and crop residue burning, brick
462
kilns, and other sources. Atmos. Chem. Phys. Discuss. 2016, 2016, 1–57.
463
464
(11) Li, X.; Wang, S.; Duan, L.; Hao, J.; Nie, Y. Carbonaceous aerosol emissions from household biofuel combustion in China. Environ. Sci. Technol. 2009, 43, 6076–6081.
465
(12) Smith, K.; Uma, R.; Kishore, V.; Lata,; Zhang, J.; Joshi, V.; Rasumussen, R.; Khalil, M.
466
Greenhouse gases from small-scale combustion devices in developing countries Phase
467
IIa: Household Stoves in India. 2000; EPA-600/R-00-052.
468
(13) Smith, K.; Khalil, M.; Rasmussen, R.; Thorneloe, S.; Manegdeg, F.; Apte, M. Green-
469
house gases from biomass and fossil fuel stoves in developing countries: A Manila pilot
470
study. Chemosphere 1993, 26, 479–505.
471
(14) Roden, C. A.; Bond, T. C.; Conway, S.; Pinel, A. B. O. Emission factors and real-
472
time optical properties of particles emitted from traditional wood burning cookstoves.
473
Environ. Sci. Technol. 2006, 40, 6750–6757.
474
475
(15) Johnson, M.; Edwards, R.; Frenk, C. A.; Masera, O. In-field greenhouse gas emissions from cookstoves in rural Mexican households. Atmos. Environ. 2008, 42, 1206–1222. 24
ACS Paragon Plus Environment
Page 24 of 30
Page 25 of 30
Environmental Science & Technology
476
(16) Christian, T.; Yokelson, R.; Crdenas, B.; Molina, L.; Engling, G.; Hsu, S.-C. Trace gas
477
and particle emissions from domestic and industrial biofuel use and garbage burning in
478
central Mexico. Atmos. Chem. Phys. 2010, 10, 565–584.
479
(17) Grieshop, A. P.; Jain, G.; Sethuraman, K.; Marshall, J. D. Emission factors of health-
480
and climate-relevant pollutants measured in home during a carbon-finance-approved
481
cookstove intervention in rural India. GeoHealth 2017, 1, 222–236.
482
(18) Wathore, R.; Mortimer, K.; Grieshop, A. P. In-use emissions and estimated impacts of
483
traditional, natural-and forced-draft cookstoves in rural Malawi. Environ. Sci. Technol.
484
2017, 51, 1929–1938.
485
(19) Du, W.; Zhu, X.; Chen, Y.; Liu, W.; Wang, W.; Shen, G.; Tao, S.; Jetter, J. J.
486
Field-based emission measurements of biomass burning in typical Chinese built-in-place
487
stoves. Environ. Pollut. 2018, 242, 1587–1597.
488
(20) Fleming, L. T.; Weltman, R.; Yadav, A.; Edwards, R. D.; Arora, N. K.; Pillarisetti, A.;
489
Meinardi, S.; Smith, K. R.; Blake, D. R.; Nizkorodov, S. A. Emissions from village
490
cookstoves in Haryana, India, and their potential impacts on air quality. Atmos. Chem.
491
Phys. 2018, 18, 15169–15182.
492
(21) Sun, J.; Zhi, G.; Jin, W.; Chen, Y.; Shen, G.; Tian, C.; Zhang, Y.; Zong, Z.; Cheng, M.;
493
Zhang, X. Emission factors of organic carbon and elemental carbon for residential coal
494
and biomass fuels in China-A new database for 39 fuel-stove combinations. Atmos.
495
Environ. 2018, 190, 241–248.
496
(22) Eilenberg, S. R.; Bilsback, K. R.; Johnson, M.; Kodros, J. K.; Lipsky, E. M.;
497
Naluwagga, A.; Fedak, K. M.; Benka-Coker, M.; Reynolds, B.; Peel, J. Field measure-
498
ments of solid-fuel cookstove emissions from uncontrolled cooking in China, Honduras,
499
Uganda, and India. Atmos. Environ. 2018, 190, 116–125.
25
ACS Paragon Plus Environment
Environmental Science & Technology
500
(23) Roden, C. A.; Bond, T. C.; Conway, S.; Osorto Pinel, A. B.; MacCarty, N.; Still, D.
501
Laboratory and field investigations of particulate and carbon monoxide emissions from
502
traditional and improved cookstoves. Atmos. Environ. 2009, 43, 1170–1181.
503
(24) Li, C.; Bosch, C.; Kang, S.; Andersson, A.; Chen, P.; Zhang, Q.; Cong, Z.; Chen, B.;
504
¨ Sources of black carbon to the Himalayan–Tibetan Plateau Qin, D.; Gustafsson, O.
505
glaciers. Nat. Commun. 2016, 7, 12574. doi:10.1038/ncomms12574.
506
507
(25) Ch` ylek, P.; Ramaswamy, V.; Srivastava, V. Albedo of soot-contaminated snow. J. Geophys. Res.: Oceans 1983, 88, 10837–10843.
508
(26) Bonasoni, P.; Laj, P.; Marinoni, A.; Sprenger, M.; Angelini, F.; Arduini, J.; Bonaf`e, U.;
509
Calzolari, F.; Colombo, T.; Decesari, S.; Di Biagio, C.; di Sarra, A. G.; Evangelisti, F.;
510
Duchi, R.; Facchini, M. Atmospheric Brown Clouds in the Himalayas: first two years of
511
continuous observations at the Nepal Climate Observatory-Pyramid (5079 m). Atmos.
512
Chem. Phys. 2010, 10, 7515–7531.
513
(27) Johnson, M. A.; Pilco, V.; Torres, R.; Joshi, S.; Shrestha, R. M.; Yagnaraman, M.;
514
Lam, N. L.; Doroski, B.; Mitchell, J.; Canuz, E.; Pennise, D. Impacts on household fuel
515
consumption from biomass stove programs in India, Nepal, and Peru. Energy Sustain-
516
able Dev. 2013, 17, 403–411.
517
(28) Xiao, Q.; Saikawa, E.; Yokelson, R. J.; Chen, P.; Li, C.; Kang, S. Indoor air pollution
518
from burning yak dung as a household fuel in Tibet. Atmos. Environ. 2015, 102, 406–
519
412.
520
(29) Weyant, C. L.; Athalye, V.; Ragavan, S.; Rajarathnam, U.; Lalchandani, D.;
521
Maithel, S.; Baum, E.; Bond, T. C. Emissions from South Asian brick production.
522
Environ. Sci. Technol. 2014, 48, 6477–6483.
523
(30) Kirchstetter, T.; Corrigan, C.; Novakov, T. Laboratory and field investigation of the
26
ACS Paragon Plus Environment
Page 26 of 30
Page 27 of 30
Environmental Science & Technology
524
adsorption of gaseous organic compounds onto quartz filters. Atmos. Environ. 2001,
525
35, 1663–1671.
526
527
(31) Dillner, A.; Phuah, C.; Turner, J. Effects of post-sampling conditions on ambient carbon aerosol filter measurements. Atmos. Environ. 2009, 43, 5937–5943.
528
(32) Bond, T. C.; Anderson, T. L.; Campbell, D. Calibration and intercomparison of filter-
529
based measurements of visible light absorption by aerosols. Aerosol Sci. Technol. 1999,
530
30, 582–600.
531
(33) Zhang, J.; Smith, K.; Uma, R.; Ma, Y.; Kishore, V.; Lata, K.; Khalil, M.; Ras-
532
mussen, R.; Thorneloe, S. Carbon monoxide from cookstoves in developing countries:
533
1. Emission factors. Chemosphere: Global Change Sci. 1999, 1, 353–366.
534
(34) Keene, W. C.; Lobert, J. M.; Crutzen, P. J.; Maben, J. R.; Scharffe, D. H.; Land-
535
mann, T.; H´ely, C.; Brain, C. Emissions of major gaseous and particulate species during
536
experimental burns of southern African biomass. J. Geophys. Res.: Atmos. 2006, 111,
537
D04301. doi:10.1029/2005JD006319.
538
(35) Bergstrom, R. W.; Pilewskie, P.; Russell, P.; Redemann, J.; Bond, T.; Quinn, P.;
539
Sierau, B. Spectral absorption properties of atmospheric aerosols. Atmos. Chem. Phys.
540
2007, 7, 5937–5943.
541
(36) NIOSH, Elemental carbon (diesel particulate ): method 5040. 2003; NIOSH Manual of
542
Analytical Methods, 4th Edition (Issue 3). National Institute for Occupational Safety
543
and Health.
544
(37) Chow, J. C.; Watson, J. G.; Chen, L.-W. A.; Arnott, W. P.; Moosm¨ uller, H.; Fung, K.
545
Equivalence of elemental carbon by thermal/optical reflectance and transmittance with
546
different temperature protocols. Environ. Sci. Technol. 2004, 38, 4414–4422.
27
ACS Paragon Plus Environment
Environmental Science & Technology
547
548
(38) Welch, B. On the comparison of several mean values: an alternative approach. Biometrika 1951, 38, 330–336.
549
(39) Field, A. Discovering statistics using SPSS ; Sage publications, 2009.
550
(40) Hobbs, P. V.; Reid, J. S.; Kotchenruther, R. A.; Ferek, R. J.; Weiss, R. Direct radiative
551
forcing by smoke from biomass burning. Science 1997, 275, 1777–1778.
552
(41) Pereira, S.; Wagner, F.; Silva, A. M. Scattering properties and mass concentration of
553
local and long-range transported aerosols over the South Western Iberia Peninsula.
554
Atmos. Environ. 2008, 42, 7623–7631.
555
(42) Reid, J. S.; Eck, T. F.; Christopher, S. A.; Koppmann, R.; Dubovik, O.; Eleuterio, D.;
556
Holben, B. N.; Reid, E. A.; Zhang, J. A review of biomass burning emissions part III:
557
intensive optical properties of biomass burning particles. Atmos. Chem. Phys. 2005, 5,
558
827–849.
559
(43) Turpin, B. J.; Lim, H.-J. Species contributions to PM2. 5 mass concentrations: Revisit-
560
ing common assumptions for estimating organic mass. Aerosol Sci. Technol. 2001, 35,
561
602–610.
562
(44) Bond, T.; Doherty, S.; Fahey, D.; Forster, P.; Berntsen, T.; Deangelo, B.; Flan-
563
ner, M.; Ghan, S.; Krcher, B.; Koch, D.; Kinne, S.; Kondo, Y.; Quinn, P.; Sarofim, M.;
564
Schultz, M. Bounding the role of black carbon in the climate system: A scientific as-
565
sessment. Journal of Geophysical Research D: Atmospheres 2013, 118, 5380–5552.
566
567
568
569
(45) Andreae, M.; Gelencs´er, A. Black carbon or brown carbon? The nature of lightabsorbing carbonaceous aerosols. Atmos. Chem. Phys. 2006, 6, 3131–3148. (46) Lack, D.; Langridge, J. On the attribution of black and brown carbon light absorption using the ˚ Angstr¨om exponent. Atmos. Chem. Phys. 2013, 13, 10535–10543.
28
ACS Paragon Plus Environment
Page 28 of 30
Page 29 of 30
Environmental Science & Technology
570
(47) Guofeng, S.; Siye, W.; Wen, W.; Yanyan, Z.; Yujia, M.; Bin, W.; Rong, W.; Wei, L.;
571
Huizhong, S.; Ye, H. Emission factors, size distributions, and emission inventories of
572
carbonaceous particulate matter from residential wood combustion in rural China. En-
573
viron. Sci. Technol. 2012, 46, 4207–4214.
574
575
(48) Li, Z.-H.; He, Y.; Zhang, H.; Wang, J. Combustion characteristics of n-heptane and wood crib fires at different altitudes. Proc. Combust. Inst. 2009, 32 II, 2481–2488.
576
(49) Akagi, S.; Yokelson, R. J.; Wiedinmyer, C.; Alvarado, M.; Reid, J.; Karl, T.;
577
Crounse, J.; Wennberg, P. Emission factors for open and domestic biomass burning
578
for use in atmospheric models. Atmos. Chem. Phys. 2011, 11, 4039–4072.
579
(50) Bond, T. C.; Streets, D. G.; Yarber, K. F.; Nelson, S. M.; Woo, J.-H.; Klimont, Z. A
580
technology-based global inventory of black and organic carbon emissions from combus-
581
tion. J. Geophys. Res.: Atmos. 2004, 109, D14203. doi:10.1029/2003JD003697.
582
(51) Tissari, J.; Lyyr¨anen, J.; Hyt¨onen, K.; Sippula, O.; Tapper, U.; Frey, A.; Saarnio, K.;
583
Pennanen, A.; Hillamo, R.; Salonen, R. Fine particle and gaseous emissions from normal
584
and smouldering wood combustion in a conventional masonry heater. Atmos. Environ.
585
2008, 42, 7862–7873.
586
587
588
589
(52) United States Environmental Protection Agency, AP 42, Fifth Edition, Volume I Chapter 1: External Combustion Sources. 1997. (53) Broderick, D. R.; Houck, J. E.; Crouch, J. Development of a fireplace baseline particulate emission factor database. OMNI and HBPA report. 2005.
590
(54) Peters, J. A.; DeAngelis, D. High altitude testing of residential wood-fired combustion
591
equipment. United States Environmental Protection Agency. EPA-600/S2-81-127. 1981.
29
ACS Paragon Plus Environment
Environmental Science & Technology
ACS Paragon Plus Environment
Page 30 of 30