-------
Residential Wood Combustion Study
Executive Summary
Final Report Prepared by:
Barbara A. Burton
Dr. Alan J. Senzel
Del Green Associates, Inc.
Environmental Technology Division
Contract Number: 4D2304NASA
Prepared for:
U.S. Environmental Protection Agency
Region 10
Seattle, Washington 98101
Task Manager
Wayne Grotheer
July, 1984
Disclaimer
This report was prepared for the U.S. Environmental Protection Agency by Del Green Associates, Inc.
Environmental Technology Division in fulfillment of contract number 4D2304NASA.
The opinions, findings, and conclusions in the report are those of the authors, and not necessarily those of
the U.S. Environmental Protection Agency.
Acknowledgements
This report provides a summary of eight tasks performed under contract No. 68-02-3566 to Del Green
Associates, Inc. The authors of this report acknowledge the work of the following individuals or companies
in completing the original task reports:
Tasks 1,2a, and 7 • NEA, Inc.
Tasks 2b, 4, collaboration on Task 5, and overall project management • Del Green Associates, Inc.
Tasks 3 and 6 -William Greene and Dr. Robert Gay
Task 5 • OMNI Environmental Services (with Del Green Associates, Inc.)
The efforts of numerous individuals are also gratefully acknowledged: the staff of the Oregon Department of
Environmental Quality (especially Barbara Tombleson), the Washington Department of Ecology, the Spokane
and Puget Sound Air Pollution Control Agencies, the Idaho Department of Health and Welfare, and of course
staff members of Region 10, U.S. Environmental Protection Agency.
Funding of this Executive Summary was provided by the U.S. Department of Energy as part of the Pacific
Northwest and Alaska Bioenergy Program.
-------
Table of Contents
List of Tables
Executive Summary
Taskl
Ambient Air Quality Impact Analysis ..
Task2a
Current & Projected Air Quality Impacts
Task 2b
Household Information Survey
Tasks
Wood Fuel Use Projection
Task 4
Technical Analysis of Wood Stoves....
Tasks
Emissions Testing of Wood Stoves
Task6
Control Strategy Analysis
Task 7
Indoor Air Quality
.11
.13
.17
.23
.27
References 29
List of Figures
Figure 1
Trends of Wood Use and Air Quality Impacts.
Figure 2
Test Site Configuration
Figure 3
Paniculate Emissions Results:
Fuel Moisture Tests
page page
. . 1 Table 1
Sampling and Analytical Protocol 3
. .3 Table 2
Impact of Residential Wood Combustion 5
. . 7 Table 3
Average PAH Concentration for Residential Sites 5
. .9 Table 4
Comparison of Wood Burned to Fine Particulate
Measured at Three sites in February, 1981 7
Table 5
Estimated Future RWC Fine Particulate Impacts 7
Table 6
Summary of Survey Results 9
Table 7
Major Factors Included in Wood Use
Trend Projection Using Marshall's Modified Model 11
Table 8
Comparison of Wood Use Trends in Portland,
Assuming Different Rates of Wood Cost Increases .... 12
Table 9
Best Estimate Projections of
Residential Wood Fuel Use 12
Table 10
Residential Wood Use and Total Suspended
Particulate Trends, Assuming 2% per Year
Real Increase in Price of Wood 12
Table 11
Emissions Summary 18
Table 12
Simplified Test Procedures Summarized 19,20,21
Table 13
Descriptions of Improved Technology Stoves and
Add-on Devices Tested 22
Table 14
Criteria and Weight Factors Used in
Keppner-Tregoe Analysis 23
Table 15
Summary of Estimated Costs and Particulate
Emissions Reduction Benefits for Fifteen RWC
Emission Control Strategies 24
Table 16
Residential Wood Combustion
IndoorSampling Program 27
Table 17
Residential Wood Combustion
IndoorSampling Program
- Summary of PAH Composite Results 27
Table 18
.18 Comparison of This Survey with Other Surveys 28
page
.17
M.S.
77 West Jackson Boulevard,
L 60604-3590
-------
Executive Summary
Residential Wood Combustion Study for Region 10, U.S. Environmental Protection Agency
Between 1900 and 1970, this country
saw a massive shift in home heating
from coal and wood to the newer, cleaner
sources of heat such as oil and
electricity. With the large price increases
and supply uncertainties of the 1970s for
the newer energy sources, however,
many homeowners began to look again
at the use of wood for home heating.
This renewed interest in residential
wood combustion has been particularly
intense in the Pacific Northwest, where
firewood is readily available. Wood
burned in woodstoves in Portland,
Oregon, for example, was estimated to
have increased from less than 50,000
cords per year in 1970 to over 150,000
cords per year in 1980.
During these same years, major air
pollution control efforts by industries
have substantially reduced the emissions
from these sources in the Pacific
Northwest. Despite these efforts, levels
of particulate air pollution in the ambient
air for many Pacific Northwest cities are
still of concern. Residential wood
combustion has been identified in
several cities as the major contributor to
these high wintertime particulate levels.
(A Medford study showed residential
wood combustion produced up to 86%
of the fine particulate during 24-hour,
worst case time periods.)
These increased levels of residential
wood combustion are of considerable
concern to the public and to air pollution
control agencies for a number of
reasons. Smoke from wood combustion
is relatively rich in carcinogens, toxic
pollutants, and substances irritating to
the respiratory tract. In addition,
particulate from woodsmoke tends to
consist largely of fine or respirable
particulate, which is recognized as being
much more of public health concern than
more coarse particulate. Fine particulate
can lodge within the lungs, whereas
more coarse particulate is trapped in the
upper respiratory tract and is expelled.
The wood smoke problem is
compounded by the fact that the smoke
is emitted close to the ground in
residential areas, as opposed to
emissions through tall industrial stacks.
In addition to the obvious public
health concerns about residential wood
combustion, visibility reduction and
limits on the airshed capacity are
important. Under current environmental
laws, each state is required to have a
plan for making sure all areas are within
acceptable ambient air pollution levels,
and for ensuring that the air quality is
not allowed to deteriorate. For some of
those cities having air pollution
problems, extensive studies have been
performed to establish an "airshed
capacity."1 Residential wood combustion
is now responsible for a large percentage
of this airshed capacity in several cities,
limiting the amount of industrial growth
that can occur without very expensive air
pollution control measures to further
reduce those industrial emissions. As
most cities wish to have the ability to
attract new industry, this lack of an
adequate margin for growth is of major
concern.
Because of public health, esthetic, and
economic growth considerations,
residential wood combustion has
become a focus of air pollution control
agency interest in the last few years.
This study was commissioned by the
U.S. Environmental Protection Agency,
Region 10, to provide some of the
necessary information for interested
agencies. It was conducted in 1980-81,
and was designed to answer these
questions for the Pacific Northwest:
• How serious is the current air
pollution problem due to residential
wood combustion?
• Will this problem increase over the
next 20 years, and if so, how much?
• What are the most effective and
realistic strategies to reduce the air
pollution impact of residential wood
combustion?
Seven individual areas of study were
completed to help answer these
questions for the Pacific Northwest.
These were:
Task 1—Ambient Air Quality Impact
Analysis. Evaluates current impact of
residential wood combustion on
ambient air quality.
Task 2a—Current and Projected Air
Quality Impacts: Projects impact on
ambient air quality by residential
wood combustion through the year
2000.
Task 2b—Household Information
Survey. Relates findings of household
surveys in three metropolitan areas:
Portland, Oregon and Spokane and
Seattle, Washington.
Task 3—Wood Fuel Use Projection.
Projects likely residential wood use in
Portland, Spokane, and Seattle
through year 2000.
Task 4—Technical Analysis of Wood
Stoves. Evaluates existing literature
and test results on promising wood
stove designs and operating
procedures, which could reduce the
air pollution emissions from
residential wood combustion.
Task 5—Emissions Testing of Wood
Stoves. Relates results of emissions
tests on "state-of-the-art" wood
stoves and add-on devices, evaluates
alternative test procedures, and
evaluates the effect of wood moisture
content on emissions.
Task 6—Control Strategy Analysis.
Evaluates and ranks possible control
strategies, including expected
emission reductions, projected cost,
and other significant advantages and
disadvantages of each proposed
control strategy.
Task 7—Indoor Air Quality. Tests and
evaluates results of indoor air
sampling for several homes having
wood stoves.
Significant findings for this project are:
• Residential wood combustion
contributed an average of 75%, 84%,
and 81% of the fine (less than 2.5
micron diameter) particulate in the
three Portland, Seattle, and Spokane
neighborhoods, respectively. These
neighborhoods were chosen for
special testing because of apparently
high levels of woodburning activity.
The sampling was done in February,
1981, and is considered typical of
reasonable worst-case impacts.
• The contribution of residential wood
combustion to fine particulate at nine
other monitoring sites in Pacific
Northwest cities ranged from 20% to
93%. This sampling was done from
October, 1980 to March, 1981. It should
be noted that these values are based
Airshed capacity is the amount of emissions possible in an area without exceeding ambient air standards.
1
-------
on a small number of samples and
may not represent the worst possible
fine paniculate levels from RWC.
• Average ambient air levels of benzo (a)
pyrene, a known carcinogen, were
measured at levels approximately
equivalent to that of smoking two to
six cigarettes/day. The highest level of
benzo (a) pyrene measured was
approximately equivalent to that of
smoking four to sixteen cigarettes/
day.
• The household survey in three small
neighborhoods in Portland, Oregon,
the Seattle, Washington, metropolitan
area (Bellevue) and Spokane, Wash-
ington showed 48%, 85%, and 80%
respectively of the households had
used a wood burning device (wood
stoves or fireplaces) in the previous 12
months. The average wood use per
woodburning household varied from
2.0 cords/year (Portland) to 1.2 cords/
year (Seattle), with Spokane residents
burning 1.9 cords/year.
• Without a major effort to control the
levels of woodburning, it is projected
that 24-hour worst case fine
particulate will increase from the 1981
levels, 53%, 27%, and 21% in Port-
land, Seattle, and Spokane,
respectively, by the year 2000. Much of
this increase will result from an antici-
pated shift from fireplace use to
dirtier but more energy efficient wood
stoves.
• By the year 2000, wood use is
projected to increase by 41%, 19%,
and 7% in the Portland, Seattle, and
Spokane metropolitan areas,
respectively over 1981 levels. These
projections are based on a computer
model developed by Norman Marshall,
Dartmouth College. The model is
largely driven by the relative cost of
wood heat to oil, gas, and electricity.
These projections should be used
with care, as they are based on
projected fuel costs which are
difficult to estimate accurately.
• State-of-the-art wood stoves and add-
on devices were not shown to emit
substantially less pollutants under the
chosen operating conditions than the
standard, box type stove, except for a
ceramic stove in this 1981 testing.
However, the ceramic stove was
operated at a much higher burn rate
(per manufacturer's direction), and
high burn rates are associated with
lowered emissions. Since the time of
this testing, numerous advanced
designed wood stoves have been
developed, some of which have
demonstrated very low emissions in
testing by other researchers.
• Other test results showed that wood
dried to a moisture content of 25-35%
on a dry basis (approximately equal to
that from six to nine months air
drying) had the least emissions. Of
the simplified test procedures
evaluated, only the carbon monoxide
and total hydrocarbon tests had a
reasonable correlation with EPA
method 5 for particulate, which was
used as the standard. However,
extensive additional data would be
needed to confirm these correlations
before such methods could be used in
an ongoing test program. The carbon
monoxide and total hydrocarbon tests
cost about $4,200 for a nine-test
series, as opposed to $15,000 for EPA
method 5.
• Because of the lack of standardized
test procedures nationwide and the
relatively few tests done, it was not
possible at the time of the study to
definitely identify any significantly
cleaner stoves or devices from avail-
able literature and test results.
However, several methods are
currently available to reduce
emissions from existing woodstoves.
The include proper sizing of the stove,
drying wood before burning, and
burning hot fires.
• The two types of control strategies
that appear to be the most promising
are limiting new stoves sold to
cleaner burning models, and
increased public education as to
proper stove choice, wood storage
practices, and stove operating
procedures. The particulate emissions
reductions from these types of control
strategies are estimated at from 39%
(mandatory stove certification, only
clean-burning stoves sold) to 6.2%
(encourage burning of dry firewood).
• With proper stove maintenance,
indoor pollution levels should not
increase when wood stoves are used.
One home with a leaky stove did have
higher indoor pollution levels, with
exposure equivalent to smoking 10 to
38 cigarettes per day for benzo (a)
pyrene. However, the testing was done
during relatively mild weather and may
not be typical of worst case, heavier
burning.
An executive summary for each task is
presented in the following sections. Each
summary includes a brief discussion of
study methods and the major findings.
These summaries are based on the full
reports, which are referenced at the end
of this report.
-------
Task 1
Ambient Air Quality Impact Analysis
The increasing use of wood as a
source of residential heat is a
phenomenon common to many states in
the Pacific Northwest. Residential Wood
Combustion (RWC) carries a significant
potential for adverse health effects to
large segments of the population. The
impact of residential wood combustion
emissions is especially severe because
plume impacts typically occur at ground
level very near the source. In addition,
the areas of highest RWC emission
density often coincide with the areas of
maximum population density, the highest
RWC emission rates occur at times
when most people are in those
residential neighborhoods, and most
particulate emissions are within the size
range deposited within the lungs. RWC
emissions are relatively rich in carcino-
genic organics, toxic pollutants, and
respiratory irritants. For all of these
reasons, wood smoke represents an
important problem that is of growing
public concern.
RWC emissions are becoming
increasingly important as a major
contributor to violations of current
particulate air quality standards and are
implicated in issues related to visibility
reduction, odors, and public health. New
attention being focused on fine
particulate with the proposal of an
Inhalable Particulate National Ambient
Air Quality Standard by EPA on March 20,
1984, also has caused concern about the
RWC impact on 24-hour standard attain-
ment. The continuing economic
pressures to expand the use of wood for
residential heating, and the limited
regulatory pressures restricting the use
of wood, may cause additional concern
about the impact of RWC emissions on
public health, esthetics, and the future
"livability" of many communities.
The purpose of this task was to
evaluate the current impact on ambient
air quality by RWC in the Pacific North-
west. The testing occurred in the 1980-
81 heating season in eight cities, and
included evaluation of fine and total
suspended particulate as well as seven
carcinogenic compounds. Samples were
selected to correspond to the maximum
impact from RWC, and represent "worst
case" ambient conditions measured
during the field program. Elemental and
carbonaceous components were
analyzed to assist in the Chemical Mass
Balance methods used in identifying the
fraction of the total sample attributable
to RWC and other sources.
A detailed household wood use survey
was conducted around three of the
monitoring sites, under Task 2B of this
study. Further interpretation of the RWC
impact estimates as they relate to
projected increases in wood use is
included in Task 2A.
Sampling Methodology
A comprehensive ambient air sampling
and analysis program was conducted in
eight cities in Oregon, Washington, and
Idaho during the October, 1980-March,
1981 space heating season to provide an
assessment of current maximum 24-hour
RWC impacts on particulate air quality.
Seventy-seven selected fine particle
samples were analyzed for 35 trace
elements, carbon, polynuclear aromatic
hydrocarbon (PAH) compounds, and
Carbon-14. The quantitative impact of
RWC on ambient air was then evaluated.
Site Selection. Data from 12 historical
ambient air data collection sites and
three new sites were included in the
study, with the three new sites located in
residential areas with apparently
significant RWC activities (as evidenced
by wood piles, stove pipes, and wood
smoke). For each of these three new
monitoring sites, a companion site in a
nonresidential area was established to
provide background information if no
adequate background site was already in
operation as part of the existing
monitoring network. The other nine sites
already were in operation for routine
ambient air monitoring, and were located
largely in commercial or industrial areas.
The cities providing data were Seattle,
Spokane, Tacoma, Longview, and Yakima,
Washington; Boise, Idaho; and Portland
and Medford, Oregon.
Timing and Selection of Samples. The
three special ambient sites were
sampled between January 31,1981 and
March 10,1981. Samples analyzed from
the other sites were gathered between
October, 1980 and March, 1981. Samples
were collected on the weekends (when
RWC emissions are expected to be the
greatest) and on one day in the middle of
the week.
Not all samples collected were
analyzed because of cost constraints.
Seventy-seven samples were selected for
analyses from those days when the
impact from RWC would be expected to
be the greatest because of cold temper-
atures, an inversion, and low wind speed
conditions. High nephelometer readings,
which indicate high fine particulate
levels, also were used to select samples
for analysis.
Sampling Equipment. Five types of
samplers were used for this project. Brief
descriptions of each sampler, the
monitoring site having the device, and
the analytical use of the collected
samples are given in Table 1.
Table 1
Sampling and Analytical Protocol
Sampling Device Information Provided
Name by Sampling Device
Hi-Vol Collects total suspended
particulate « 30 urn)
Sierra Model 235 Collects respirable par-
High Volume ticulate ( < 2 um) on
Cascade Impactor glass fiber filter
Dichotomous Collects respirable par-
sampler ticulate ( < 2 5 um) and
coarse particulate
(25-15 um)on teflon
filter
Size selective Collects particulate
inlet sampler < 15 um on glass fiber
(SSI) filter
Integrating Measures the light
nephelometer scattering character-
istics of ambient air
(associated with fine
particulate)
Site Locations Having
Sampling Device
All 15 sites except Lake Sammamish
(Seattle area) and County Health
(Spokane area)
Country Homes (Spokane), Turnbull
(Spokane), Newport Way (Seattle
area), Lake Sammamish (Seattle
area), Marcus Whitman School
(Portland), Carus (Portland),
County Courthouse (Medford)
All 15 sites
Fairview & Liberty (Boise)
Fire Station 12 (Tacoma),
County Courthouse (Yakima)
Analyses Performed
on Sample
Weighing (to deter-
mine TSP), gas
chromatography/ mass
spectroscopy
(PAH analysis),
Carbon-14
Weighing (to deter-
mine fine particul-
late concentration),
elemental/organic
carbon (chemical
mass balance and
PAH)
Weighing (fine par-
ticulate), X-ray
fluorescence and
neutron activation
analysis (trace ele-
ments for chemical
mass balance)
Gas chromatography/
mass spectroscopy
(PAH), Carbon-14
Not used except to
indicate days of high
fine particulate to
assist in sample
selection
-------
Analyses Performed
Analyses of the chosen samples were
performed to measure the concen-
trations of (1) certain carcinogens
associated with wood smoke, (2) total
suspended and respirable particulates,
and (3) various other elements and
carbon forms so that the contribution of
RWC could be determined.
Fourteen carcinogens have been
associated with wood smoke. Seven
polynuclear aromatic hydrocarbons (PAH)
known to be carcinogens were tested.
Samples for PAH analyses were
collected using Sierra Cascade
Impactors, Hi-Vols, and the Size Selective
Inlet sampler. Gas chromatography/mass
spectroscopy was used for the analysis.
Total suspended particulate levels
were collected with the Hi-Vols. Fine
particulate levels were measured using
the Sierra Cascade Impactor (< 2.0 /^m)
and the dichotomous sampler (< 2.5
The concentrations of 35 trace
elements, needed for the Chemical Mass
Balance analysis, were determined using
X-ray fluorescence and neutron
activation analysis. In addition to the
ambient samples collected by the dicho-
tomous samplers, resuspended local soil
samples were analyzed for the trace
elements. Soil samples were collected in
Boise, Seattle, and Spokane and were
used to help determine the ambient
sample fraction attributable to airborne
soils or road dust. Fine particulate
source emission chemistry from other
emission sources (such as RWC and
transportation) were determined from
other studies, such as the Portland
Aerosol Characterization Study.
Samples were evaluated for elemental
and organic carbon, both for the PAH
analyses and for the Chemical Mass
Balance evaluation. Carbon-12 and
Carbon-14 were analyzed as an
independent means of validating the
Chemical Mass Balance estimates of
RWC impacts, as described below.
Fingerprinting RWC Emissions
Two methods were used for
determining the fraction of the ambient
fine particulate that is attributable to
RWC and other sources. The Chemical
Mass Balance (CMB) Model served as the
primary tool, and the newly developed
Carbon-14 Assessment Method was used
to verify the results of the CMB method.
The CMB Model starts with emission
samples of all significant emission
sources, including geologic (road dust),
transportation, residual oil combustion,
major industries, residential wood
combustion, and other significant
sources. These emission samples from
the different sources are analyzed for a
long list of trace elements and other
distinguishing compounds, in an effort to
uniquely identify each source category.
Ambient samples collected are
similarly analyzed. Using the Chemical
Mass Balance Model, the contribution of
the different emission sources can be
determined for any specific ambient
monitoring site, provided the individual
source categories have been properly
identified.
Measurements of Carbon-14 were
used to validate CMB-derived RWC
impact estimates. The radioisotope
Carbon-14 recently has been identified as
a unique trace of contemporary carbon
sources (such as RWC or slashburning),
since fossil fuels have essentially no
Carbon-14 (the fossil fuels are old
enough that the Carbon-14 has stabilized
to Carbon-12). Because of the timing and
location of the ambient sampling,
interferences with other contemporary
carbon sources such as slashburning
were eliminated, leaving RWC as the
most likely source of all the Carbon-14
measured. Representative firewood
samples from Portland, Boise, Spokane,
and Seattle were obtained and analyzed
to determine the Carbon-14 levels of the
wood burned in each community—an
important factor in calculating the C-14
content of the ambient aerosol. Results
of the Carbon-14 analysis were found to
be consistent with RWC impact
estimates developed by the CMB tech-
nique within the limits of experimental
error.
Results
Ambient air quality studies conducted
during the 1980-81 space heating season
in eight Pacific Northwest communities
clearly indicate that RWC emissions are
the most important contributor to the
fine particle mass less than 2^m. Since
the program design sought to determine
maximum RWC impacts, the following
conclusions reflect reasonable worst
case impact conditions rather than, for
example, annual average source impacts
representative of each community's
airshed.
Key findings of Task 1, then include
the following:
1) RWC emissions typically account for
66% to 75% of the fine particle mass
in residential areas, while trans-
portation sources contribute 5%,
secondary sulfate, 5.6%, and all
industrial sources less than 0.5% (see
Table 2).
2) Background RWC impacts were found
to range from 3-12 ^g/m3, 24-hour
average—a factor of ten lower than
the urban sites, suggesting that
70-80% of the RWC impact is related
to local sources.
3) Maximum 24-hour impacts (fine
particle mass) exceeded 60^g/m3, at
residential sites located in Seattle,
Spokane, Portland, Medford, and
Boise. Impacts at industrial sites in
Longview, Seattle, and Tacoma were
significantly lower.
4) The highest impacts measured in this
study (Boise—128 ^g/m3, 8 hour
average (fine particle mass)) must be
considered as upper limit estimates
requiring further verification.
5) PAH concentrations measured at
urban sites were a factor of ten higher
than those measured at the rural sites
(see Table 3). Although the measured
PAH concentrations should be of
concern, there are no direct dose-
response relationship currently
available in the literature upon which
to base a quantitative assessment of
public health risk. However, some
qualitative comparison may be
possible from the benzo (a) pyrene
(B(a)P) measured at the sites. This
carcinogen is also present in cigarette
smoke. The levels of B(a)P measured
at the sites were equivalent to that
from two to six cigarettes smoked per
day for the average B(a)P concen-
trations recorded at the sites, and that
from four to sixteen cigarettes
smoked per day for the highest (B(a)P)
level measured at the monitoring
sites.
-------
Table 2
Impact of Residential Wood Combustion
Based on Ambient Monitoring and Chemical Mass Balance
October, 1980 - March, 1981
City
Portland, OR
Medford, OR
Seattle, WA4
Seattle, WA
Seattle, WA
Spokane, WA
Spokane, WA
Spokane, WA
Tacoma, WA
Yakima, WA
Longview, WA
Boise, ID
Site Name
Whitman School1
Courthouse
Newport Way1
South Park
Georgetown
Country Homes1
County Health
Crown Zellerbach
Fire Station #12
Courthouse
City Shops
Fairview & Liberty Streets
Land Use
of Site
Area
Residential
Commercial
Residential
Residential
Industrial
Residential
Commercial
Industrial
Industrial
Commercial
Commercial
Commercial
Number
of
Samples
12
2
10
2
3
15
1
4
4
4
4
9
Average
Total Fine
Paniculate
( ug/my
454
—
361
—
397
550
53.1
370
470
538
41 8
121 8*
Average RWC
Percent of
Fine Panic-
ulate9
75 1 %
20 3%
838%
65.8%
734%
81 0%
64.7%
453%
749%
93 1 %
61 4%
69.5%*
RWC Contribu-
tion-24 hour
Maximum Fine
Paniculate
61 7±157
622±190
488±138
68 2 ±19.9
355±243
681 ±160
—
191 ±173
44 3 ±32.2
55 1 ± 32 4
40.5 ± 24 1
127.9 ±29.9*
RWC Contribu-
tion-24 hour
Minimum Fine
Paniculate
( ug/m')'
13.7-3.9
531 ±18 8
82±26
65.9+196
204+153
19.9±126
—
149±123
203±146
480±377
14.2 + 92
50.0 ±13.4*
1 Monitoring sites specifically established for this task
2Fine participate is defined as less than 2 5 urn
'Determined by means of chemical mass balance analysis
•Based on fraction < 15 urn Estimate unvahdated
'Actually located in Bellevue, Washington, which is in the Seattle area
Tab/e3
Average PAH Concentration for Residential Sites (ng/m3)*
(1980 - '81 Heating Season)
Arithmetic
PAH Group Mean
Benzc
-------
-------
Task 2a
Current & Projected Air Quality Impacts
This task brings together the results of
Task 1 (Ambient Air Quality Impact
Analysis), Task 2B (Household Informa-
tion Survey), and Task 3 (Wood Fuel Use
Projection).
The impact of the fine particulate
generated by residential wood
combustion (RWC) on visibility reduction
and the 24-hour worst case fine
particulate levels were projected to the
year 2000, for the cities of Portland,
Oregon and Seattle and Spokane,
Washington.
Worst Case 24-Hour Impact
In each of these three cities, a one
mile square residential area was chosen
for study. Each area clearly had extensive
RWC activity, as evidenced by many
woodpiles, chimneys, and woodstove
stacks. An ambient air monitor was sited
in the middle of each of the three areas
with samples collected in February, 1981.
Eight hundred households in each
square mile area were surveyed as to
their woodburning practices in general
and specifically in February, 1981.
Between 36% and 58% of the surveyed
households responded.
Using the household survey results
and the ambient monitoring data, a rela-
tionship between the amount of wood
burned and the fine particulate (< 2.5 p.m
diameter) measured at the ambient
monitoring stations was established.
These ratios were different for each city,
since each city has different meteor-
ological and pollution dispersion
characteristics. The results are briefly
summarized in Table 4 for a 24-hour,
worst case condition.
In order to project the most likely
worst-case 24-hour fine particulate
levels, the results shown in Table 4 were
combined with the wood use and
emissions projections through year 2000
made in Task 3 for the three metropolitan
areas. The projected 24-hour worst case
fine particulate levels are shown in Table
5. These results show a 53%, 27%, and
21% increase from 1981 fine particulate
levels (due to RWC and non-RWC) by the
year 2000 for Portland, Seattle, and
Spokane respectively, for a typical neigh-
borhood with heavy RWC use. However,
much of this increase is expected to
occur by 1985. Increases from 1981 in
total fine particulate levels are projected
at 28%, 27%, and 12% for Portland,
Seattle, and Spokane respectively by
1985.
A number of assumptions were made
to extend the results of the three square
mile survey areas to the three metro-
politan areas. These are:
• Wood use projections for the metro-
politan areas can be used to predict
the wood use in the survey neighbor-
hoods.
• The survey neighborhoods are typical
of neighborhoods with heavy RWC
use.
• The meteorology and pollution disper-
sion characteristics of the survey
areas are typical of the cities in which
they are located.
• The background or non-RWC fine
particulate levels will remain the same
through year 2000.
• The meteorology, pollution dispersion,
and wood burning practices during
February, 1981 were typical of a
reasonable worst-case RWC month.
The projected particulate emissions,
cords of wood burned, and fine particu-
late impacts are shown graphically in
Figure 1.
Table 4
Comparison of Wood Burned to Fine Particulate Measured at Three Sites in
February, 1981
24 Hour Worst Case Condition
Fine Particulate*, Ambient Air
City/Site Particulate Emissions Attributable to RWC Total
Portland/ 1 2 2 tons/month 149 ug/m3 198 ug/m3
Marcus Whitman
Seattle/ 6 5 tons/month 25 2 ug/m3 34 7 ug/m3
Newport Way1
Spokane/ 7 1 tons/month 33 2 UG/m3 39 7 ug/m3
Country Home
Fine Particulate
Ambient Air Impact/
Ton Emissions
12 ug m3
ton
38 ug/m3
ton
4.6 ug/m3
ton
* 2 5 um diameter
1 Actually located in Bellevue, Washington, within the Seattle Metropolitan area
Table 5
Estimated Future RWC Fine Particulate
(24 Hour Reasonable Worst Case)
Impacts
Fine Particulate, Ambient Air - ug/m3
City/Site Year
Portland 1981*
(Marcus Whitman) 1985
1990
1995
2000
Seattle 1981*
(Newport Way') 1985
1990
1995
2000
Spokane 1981*
(Country Home) 1985
1990
1995
2000
"Actual Measured Data
'Actually located in Bellevue,
Attributable to RWC
149*
204
199
231
253
252*
347
347
360
347
332*
381
393
41 7
41 1
Total
198*
253
24.8
280
302
347*
442
442
455
442
397*
446
458
482
482
Projected % Change in
Total Ambient Al( Fine Particulate
Over Next Five Years
+ 28%
-2%
+ 21 %
+ 8%
+ 27%
0%
3%
-3%
+ 12%
+ 3%
+ 5%
0%
Washington, within the Seattle Metropolitan area
-------
Visibility Reduction
The reduction of visibility is of
particular concern to Pacific Northwest
communities with scenic views. Fine
particulate is primarily responsible for
the reduction in visibility in urban areas.
Research cited in Task 2 has shown a
definite link between visual range
(measured in kilometers) and light
scattering or Bscat (measured in kilometer
-'). Further, a high correlation between
Bscat (measured by a device called an
integrating nephelometer) and fine mass
has been found in a Portland study.
Using these two relationships, it is
estimated that in Portland, by the year
2000, there will be an additional decrease
in visual range of about 4 miles in
average 24-hour worst case conditions.
Figure 1
Trends for Wood Use and
Air Quality Impacts
LU
16-1
15-
14-
13-
12-
11-
10-
9-1
5001-
400 -
Portland, Oregon
CO
Q
DC
O
O
300 -
200
1981
31
29
RWC
Impact*
27 (ug/m3)
Average
25 24 hr
Worst
Case
23
-21
2000
1st order approximations
60-i
50-
M
§40-
30-1
200
150
100
Seattle, Washington
Air Quality Impact
-145
1981
'85
'90
Year
'95
40
35
RWC
Impact*
(ug/m3)
Average
24 hr
Worst
Case
30
2000
*1st order approximations
40-1
35-
I 30-
LLJ
Q_
CO
25->
130r
120
CO
Q
CC
O
O
110
100
Spokane, Washington
Cords/Yea
1981
'85
'90
Year
'95
50
45 RWC
Impact*
(ug/m3)
Average
24 hr
Worst
40 Case
35
2000 '
*1st order approximations
-------
Task 2b
Household Information Survey
Surveys on residential wood
combustion practices were conducted in
Portland, Oregon, and Seattle and
Spokane, Washington in the Spring of
1981. Each survey area was selected to
include a one-mile square area around a
temporary air monitoring station that had
been operating during Feburary 1981 so
that the survey information could be
correlated to the ambient data. Results
of the surveys will be used to help
predict citywide residential wood use
trends and, in turn, to help develop
possible control strategies to reduce air
pollutants from woodburning devices.
The survey results are summarized in
this report by means of data presented in
chart form.
The surveys were mailed to a random
selection of 800 households for each of
the three cities by an industry trade
association in Portland and by state
university associations in Seattle and
Spokane. The response rates were 58
percent for Spokane, 48 percent for
Seattle, and 36 percent for Portland. All
responses were sent to Del Green
Associates, Inc., for evaluation.
Data from the surveys were coded,
keypunched onto computer cards, and
entered into a computer using the Statis-
tical Package for the Social Sciences
(SPSS). A quality assurance check of
every tenth survey revealed an error rate
of 19 out of 11,340 (0.2%) for recording of
responses.
Summary statistics for the responses
were calculated separately for each of
the three survey areas, all of which were
chosen because they were in residential
areas with evidence of substantial wood-
burning activities. In interpreting the
results, which are summarized in Table 6,
one must remember that the survey
areas were one square mile each within
much larger urban areas; thus, the
results cannot necessarily be assumed
to be representative of their respective
citywide areas.
The Portland survey area (Multnomah
County, immediately adjacent to South-
east Portland but outside the city limits)
was a working-class community with an
average annual household income of
$18,400 and production work as the
predominant occupation of the head of
the household. The homes were relatively
old (average age of 36 years) and small
(1400 square feet), with single family
residences prevailing. Fireplaces were
the most common type of woodburning
device (36.9% of homes), closely
followed by wood stoves at 31.7%. More
than half the wood burned was in wood
stoves, however.
Seattle's survey area (Bellevue, a
community four miles due east of Seattle
across Lake Washington) was an upper
middle-class neighborhood with an
average annual household income of
$35,000 and predominantly newer (1962),
large (2300 square feet), single family
residences owned by the occupants.
Fireplaces were present in 97.1% of the
homes, with wood stoves in only 13.5%.
The Spokane survey area (located just
north and outside the city limits) was a
community characterized by newer (1961),
medium-sized (1700 square feet), single
family homes. The average annual family
income was $27,000. Fireplaces were by
far the predominant woodburning
devices, with 97% of the households
surveyed having one or more. Only 11.7%
of the homes had wood stoves, but
22.8% of the wood used was burned in
wood stoves.
Results and Conclusions
The surveys show (see Table 6) that
although relatively few households use
wood as their primary heat source (5-
14%), most families now use wood at
least as a secondary source of heat (1.0-
1.6 cord/year/household). Based on the
respondent's anticipated future wood
use and plans to purchase new wood-
burning units, it appears likely that wood
use will continue to increase substan-
tially, at least for the short term.
Of major concern to regulatory
agencies considering pollution control
Table 6
Summary of Survey Results
Parameter
Number of households
returning questionnaire
Number of questionnaires
sent out
Total number of households
in survey area (excluding
apartments without chim-
neys, and trailer parks) -
based on homes with
listed or unlisted phones
Households using a wood
burning unit within last
12 months
Households using wood as
a primary source of heat
Households having:
Wood stove
Fireplaces
Woodburning furnace
Any woodburning device
No woodburning device
Estimated total number
of cords of wood burned
in last 12 months (all
households in survey
area)
% of wood burned in each
type of device:
Woodstove
Fireplace
Woodburning furnace
Average wood use per
household (all surveyed)
Average wood use per
woodburning household
Households planning to
buy woodburning unit
Plans for 1981-82 wood
burning compared to
1980-81 (all households)
More
Same
Less
Portland
288
800
2082
497%
142%
31 7%
369%
1 6%
587%
41 3%
2070 cords/yr
565%
393%
4.2%
1 0 cords/yr
2 0 cords/yr
165%
31 9%
599%
90%
Seattle
380
800
1429
845%
58%
135%
97 1 %
06%
1 3%
1450 cords/yr
253%
746%
01%
1 0 cords/yr
1 2 cords/yr
239%
278%
607%
11 5%
Spokane
443
800
930
797%
50%
11 7%
970%
1 3%
34%
1470 cords/yr
228%
765%
08%
1 6 cords/yr
1 9 cords/yr
204%
277%
629%
93%
-------
strategies is how wood for home heating wood burn significantly more wood than and demographic variables were found
is obtained and how it is seasoned and those who purchase their firewood. The for the three cities surveyed. Comparison
stored before use. This study shows that surveys showed no significant difference of the results of this study with those of
50-70% of the wood burned was in wood storage and aging practices other wood burning surveys conducted
chopped by homeowners, 65% was aged between the three metropolitan areas. in the Pacific Northwest in 1979 and 1980
over one year, and 75% was stored under was difficult because of wide differences
cover. In all three cities, the surveys No other consistently reliable in the survey questions and populations
show that those who chop their own fire- correlations between wood use practices sampled.
10
-------
Wood Fuel Use Projection
Under this task, wood use projections
using 1980 as a base line through the
year 2000 were determined for the Port-
land metropolitan area, the City of
Seattle, and the City of Spokane. An
estimate of the projected total
suspended particulate (TSP) contribution
from residential wood combustion also
was made. The short-term wood use
projection (through 1983) was based on
recent trends in wood use. The long-term
trends were determined using a resi-
dential wood use trend model developed
by another researcher.
Short-Term Trend Methodology
The short-term trends were
determined using household wood use
surveys and firewood cutting permits in
the recent past and projecting these
values into the near future. Ambient air
monitoring data was examined as an
indirect indicator of wood use, but had
limited usefulness in the trend analysis
either because the data had not been
collected consistently and completely
over the years needed, or because the
measured parameter could not be linked
solely to residential wood combustion.
The Light Scattering Coefficient (Bscat)
was the most useful ambient data
available. The Bscat measures the degree
to which small particles in the air scatter
light (wood smoke particulate consists of
90% small particles). The trends in Bscat
values were very close to those indicated
by firewood cutting permits. It should be
noted that Bscat values will be affected
by other sources of fine particulate
emissions, however.
Firewood cutting permits issued for
publicly owned forests near the three
cities were used as the best data avail-
able. Records have been kept on an
annual basis for the number of permits
issued, although not the actual amount
of wood cut. The firewood taken from the
public lands consists largely of branches
and cull logs left behind after logging
operations have been completed. Other
sources of firewood, such as lumber mill
scraps and wood from private lands,
normally are not known as to quantity.
Household wood use survey data is more
accurate than the permit data, but is not
available on a frequent enough basis for
use in trends analysis. The survey data
was used with the firewood cutting
permit data to quantify the wood burned
(i.e., if 10,000 permits were issued in a
given year when a survey showed 500,000
cords of wood were burned, then if
20,000 permits were issued in another
year about 1,000,000 cords of wood were
burned in that year).
Long-Term Trend Methodology
After a review of existing trend
analyses for residential wood use, a
model developed by Norman Marshall
(Dartmouth College) was determined to
be the best available. This model was
modified and used with input data
mostly from the Pacific Northwest. The
major assumption that drives the model
is that as conventional fuel sources
increase in cost, wood use will increase
subject to such factors as cost of wood
stove installation and the inconvenience
of using wood compared to conventional
fuels.
The major factors included in
Marshall's model are listed in Table 7,
along with the principal source of data.
The effect of possible regulations can be
included in the model if desired for
control strategy development, but have
been excluded for this evaluation.
A detailed check (or calibration) of the
model was conducted for the Portland
data and less detailed calibrations for
Seattle and Spokane. Portland received
the most detailed check since there was
much better survey data available to
conduct the evaluation. Briefly, the cali-
bration consisted of comparing the
actual wood use trends between 1970
and 1980 with what the model predicted,
using 1970 input data. The model was
verified in all three cities as being
reasonably accurate.
One major area of uncertainty is the
projected costs of fuel, particularly
wood. The availability of wood is
expected to drop in the next twenty
years as timber harvests level off in
terms of board feet of lumber, while the
harvest shifts to second growth timber
with its 70-90% less residue than old
growth. Another factor of unknown
dimension will be the competition for the
diminishing wood residues by other
users, including industrial boilers and
Major Factors Included in Wood Use Trend Projection Using Marshall's Modified
Model
Factor
Source of Input Data/Basis of Assumption
1 Initial number of households
and projected growth
2 Heating requirements per household
3 Historical conventional fuel
prices and fuel usage split
4 Historical wood prices
5 Future conventional fuel prices
and fuel usage split
6 Future wood prices (after inflation)
7 Efficiencies of home heating
devices over time
8 Change in wood stove purchases
as market is saturated
9 Cost of wood stove installation
based on fraction of house to be heated
10 Effect of self-cut wood on price
11 Effect of payback period
12 Fireplace use
Local planning agencies
Utilities' estimates of Btu's/1000 square feet
for various fuels in 1970, plus average home
size from Portland Real Estate Report, were
used for Portland This data was adjusted for
Seattle and Spokane based on different climates.
Utility assumptions of 25% decrease in heating
requirements by 2000 based on conservation
Utilities
Classified ads in each city
Bonneville Power Administration, utilities and
Oregon Department of Energy projections
Assumed to be 2% (the same as conventional fuels)
since no better estimates were available
Several papers by researchers
Factor developed by Marshall
Factor developed by Marshall
Marshall's factor modified by forest resources available
per capita
Marshall's factor modified by different climate and
heating requirements
Assumed to decrease proportionally with wood price in-
crease
11
-------
particleboard producers. This increasing
competition for diminishing logging resi-
dues is expected to be at least partially
relieved by private woodlots increasing
firewood production. Since no wood cost
projections from knowledgeable sources
were available, a factor of + 2%/year real
cost increase (after inflation) was chosen
for wood, to coincide with the utilities'
projected cost increases of the
conventional fuels. Table 8 shows the
effect on the modelled results if this
value is changed to 0%/year change and
+ 5%/year change. As indicated, the
projected change in residential wood
combustion between 1980 and 2000
varies from +80% if real wood prices
remain constant, to -31% if real wood
prices increase 5%/year. Any significant
departures of fuel costs for the
conventional fuels can be expected to
similarly impact wood use trends.
Significant Findings
For all three cities, residential wood
combustion increased rapidly in the late
1970s, but now is leveling off and is
expected to remain at a relatively
constant level until 1990. Between 1990
and 2000, Portland wood use will
increase by 17%, Seattle wood use will
drop by 5%, and Spokane wood use will
increase by 2%. Between 1980 and 2000,
wood use will increase by 37% in Port-
Table 8
Comparison of Wood Use Trends in Portland, Assuming Different Rates of Wood
Cost Increases
Years 1980 - 2000
Cords/Year Wood Burned % Change in
Annual Change in Real Wood Wnorl Iko
Cost (after Inflation)1 1980 2000 1980-2000
+ 2%/Year2 350,000 460,000 +31%
0%/Year 350,000 630,000 +80%
+ 5%/Year 350,000 240.0003 —31%
'1970 to 1980, average cord wood prices increased 217% compared to a 110% increase in the
consumer price index, equivalent to a real price rate increase of 4 2%/year, compounded
This rate of price increase was used in the detailed wood use trend analyses, and approximately
equals the projected rate of conventional fuel price increases (2 08%/year)
'This approximates the amount of wood burned in 1970
Tab/8 9
Best Estimate Projections of Residential Wood Fuel Use for Portland, Seattle,
Spokane (1980 - 2000) and Corresponding Particulate Emissions
Stove/Furnace Fireplace Total Total Particulate
Number of Wood Usage Wood Usage Wood Usage Emissions
Year Households (1000 cords/yr.) (1000 cords/yr) (1000 cords/yr.) (1000 tons/yr.)
PORTLAND METROPOLITAN AREA
1980 471,850 150 190 340 93
1985 537,800 240 190 430 128
1990 603,750 240 170 410 125
1995 669,700 300 150 450 145
2000 735,650 340 140 480 159
CITY OF SEATTLE
1980 220,000 45 110 155 37
1985 246,180 85 100 185 51
1990 269,720 85 100 185 51
1995 294,360 90 100 190 53
2000 323,180 85 100 185 51
CITY OF SPOKANE
1980 70,920 28 93 121 27
1985 77,940 42 84 126 31
1990 83,960 45 81 126 32
1995 90,910 51 78 129 34
2000 98,860 54 75 129 34
land, 17% in Seattle, and 7% in Spokane
For all three cities, the most rapid
growth is projected to occur in 1980-85
(26% for Portland and Seattle, 4% for
Spokane). These results are summarized
in Table 9.
Another projected shift in burning
practices is away from inefficient fire-
places towards wood stoves and wood
furnaces. Fireplaces are so inefficient at
heating (-10 to +20%), that their use is
considered primarily for esthetic
purposes rather than for heating. As fuel
costs increase, it is expected that fire-
place use will drop sharply both because
more fireplaces are converted to more
efficient burning devices and because
wood will become increasingly too
expensive to burn in an open fireplace
only for esthetic reasons. Wood burning
in open fireplaces is projected to drop in
all three cities.
This shift from open fireplaces to
wood stoves will have a significant
impact on total suspended particulate
from residential wood combustion. Unde
current operating practices, wood stoves
emit almost twice as much particulate as
fireplaces. In Portland, for example,
wood use is projected to increase by
37% by year 2000, but the total
suspended particulate emissions are
projected to increase by 76% due to the
switch from fireplaces to wood stoves.
This data and the wood use by wood
stoves and furnaces, and fireplaces is
listed in Table 10.
Table 10
Residential Wood Use and Total Suspended Particulate Trends, Assuming 2% per Year Real Increase in Price of Wood
(Years 1980 - 2000)
Wood Stoves and Furnaces Fireplaces Total Cords/Year Burned % Increase in
Cords/Year Cords/Year TSP from RWC
City 1980 2000 % Change 1980 2000 % Change 1980 2000 % Change 1980-2000
Portland 150,000 340,000 +125% 200,000 140,000 —30% 350,000 480000 +37% +76%
Metro
Area
CUV Of 45,000 85,000 + 89% 110,000 100,000 —9% 155,000 185,000 +19% +40%
City of 28,000 54,000 + 93% 93,000 75,500 —19% 121,000 129000 + 7% +26%
Spokane
12
-------
Task 4
Technical Analysis of Wood Stoves: Combustion Principles, Design Considerations, Operating
Techniques
Design and operation of residential
wood combustion devices influence both
performance and emissions. Important
design considerations include
mechanisms to increase thermal
efficiency and improve combustion effi-
ciency. Both these efficiencies must be
relatively high to have an overall efficient
residential wood combustion (RWC)
device. Until the last five years or so,
levels of residential wood combustion
were low enough that there was no real
demand for improved stove designs
which increase efficiencies and decrease
emissions. There remains considerable
room for improvement in the design of
stoves. Since this study was conducted,
many improved units have appeared. It is
expected that in the next few years the
emerging stove technology will result in
substantial emission reductions,
possibly by as much as 75%.
Task 4 consisted of a thorough litera-
ture search and evaluation of existing
data on woodburning devices and
operating procedures. Major areas of
investigation were:
• Emission rates from various types of
woodburning devices, particularly fire-
places and wood stoves.
• Changes in emission rates when
different add-on devices or stove
design modifications are made.
• Evaluation of emerging stove
technology.
• Effects on emission rates of such
operating variables as fuel type, fuel
moisture content, combustion air, and
firing rate.
Several difficulties were encountered
in compiling this information. The
interest in emissions from residential
wood combustion was relatively new at
the time of this report, and little data
existed. Much of this data is from stove
manufacturers in support of their
marketing efforts. The sampling methods
and operating conditions during testing
(wood type, size, moisture content, firing
rate, etc.) usually were not consistent,
making comparisons between devices
very difficult. Little or no hard data was
available on new technology. The data
therefore has been presented in general-
ized terms.
Summary of Significant Findings
General Combustion Principles
and Emissions
The term "efficiency" when used in
conjunction with a wood-burning device
is the measure of how much net energy
is available and useful per energy unit
(such as British Thermal Unit or Btu) con-
tained in the wood. The overall efficiency
is a combination of the "combustion
efficiency" and the "thermal efficiency."
Modifications that affect either type of
efficiency will affect the emissions.
Combustion efficiency refers to the
percent of the potential energy available
in the wood that is actually released
during burning. Final combustion
products are carbon dioxide and water.
Products of incomplete combustion
include such pollutants as carbon
monoxide, unburned or partially burned
particles of wood, hydrocarbons and ash.
Incomplete combustion and the asso-
ciated pollutants can result from
insufficient oxygen, low temperatures in
the combustion zone, and insufficient
time or mixing of oxygen and fuel to
allow complete combustion. Many of the
modifications proposed by stove
manufacturers are designed to improve
the combustion efficiency, which
reduces air pollution per unit of wood
while increasing the amount of heat per
unit of wood burned.
A measure of thermal efficiency is the
percent of the heat generated by
combustion that is useable and released
to the room(s) to be heated. Hot exhaust
gases comprise the heat that is lost. For
maximum thermal efficiency, the exhaust
gases should be as cool as possible
while still allowing a proper draft (to pull
the smoke out of the house) and not
having excessive condensation of water
or creosote deposits in the chimney.
Creosote is another name for hydro-
carbons in the exhaust gases that have
condensed onto the cooler stove pipe or
chimney forming a sticky residue which
can ignite and cause a fire hazard. An
example of a thermally efficient
operation would be a banked, slow-
burning, oxygen starved fire in a wood
stove, since most of the heat generated
is released to the room. Such burning
practices are thermally efficient, but the
combustion efficiency is very low.
Smoldering fires release large amounts
of pollutants and are undesirable from an
air pollution standpoint.
Many of the modifications, add-ons,
and new stove technologies are designed
to improve the combustion efficiency,
the thermal efficiency, or both. In
general, the following features have been
investigated or included:
• Limit combustion air to the minimum
required to sustain combustion, and
thereby minimize the amount of cold
air pulled into the house (thermal
efficiency).
• Combustion air introduced at those
points where combustion is occurring
(thermal and combustion efficiency).
• Pre-heating combustion air to aid
combustion (combustion efficiency).
• Require the hot gases to follow a
longer path, allowing more complete
combustion and better heat capture
(combustion and thermal efficiency).
• Afterburning of exhaust gases, usually
with the aid of a catalyst (combustion
and thermal efficiency).
These features will be discussed in
more detail in later sections.
Fireplaces
Fireplaces generally are inefficient,
and can actually cause a net drop in
house temperatures under some circum-
stances. As the outside temperature
approaches 10 °F, more energy is lost (as
cold air is pulled into the house) than
energy released by the fire. Depending
on the outside temperature, the overall
efficiency of fireplaces varies from -10
to 20%.
The rate at which wood is burned
appeared to be the most important vari-
able affecting the emissions from fire-
places. A hot full fire burns the cleanest.
Glass doors which are closed during
fireplace use, limiting the amount of
combustion air, are net energy losers,
since the glass reduces the gross heat
output by 50-55%. Glass doors are effec-
tive in reducing the loss of warm room
air up the chimney when the fire has
burned down.
Air circulation or heat transfer
systems have a limited effect on the
efficiency of fireplaces. Various studies
report from a 2.5% (without fans) to
8.6% improvement in efficiency from
heat transfer systems.
13
-------
Wood Stoves
Wood stoves come in many configur-
ations, with an overall efficiency of from
40 to 70% depending on the design.
Fireplace inserts are very similar to wood
stoves, except that the existing chimney
is used rather than a stove pipe. The
following discussion includes fireplace
inserts.
One of the more common variations
among wood stoves is the air flow path.
The types of stoves in use are the
updraft, downdraft, cross draft, diagonal
flow, and "S" flow. Studies by the State
University of New York of the efficiency
of stoves with these configurations
found no significant differences among
them.
Regulating the draft does improve the
efficiency of stoves. These air controls
can be manual or automatic. A baro-
metric damper in the stack or a thermo-
static damper at the stove air inlet can
be used to regulate the air into the
combustion chamber. By regulating the
air, enough oxygen can be introduced to
sustain adequate combustion without
allowing excess air which cools the fire,
inhibits combustion, and draws too
much cold air into the house.
Sizing a wood stove to the intended
area to be heated is very important in
minimizing air emissions. The complete
report includes detailed instructions on
sizing a stove for a particular application.
The cleanest, most efficient fires are
those that use a fairly high burn rate.
Many stoves are oversized for the
intended use, so that brisk fires cannot
be maintained without overheating the
room. Instead, homeowners with over-
sized stoves tend to burn wood in their
stoves at a slow rate by starving the fire
of oxygen. These slow, smoldering fires
waste wood, create pollution, and can
cause fire hazards through an excessive
build-up of creosote in the stove pipes.
Wood-Fired Furnaces
Wood furnaces are not widely used at
this time and were not investigated in
depth. The efficiency of furnaces is
similar to wood stoves, about 40 to 75%.
Catalysts and Afterburners
Afterburners are devices for
introducing a secondary fuel into the
exhaust stream and igniting to burn up
the incompletely combusted portions of
the exhaust gas. Such devices are not in
common use, and no data was found on
the emissions from afterburners on wood
stoves.
Catalysts also allow the combustion of
exhaust gases, by lowering the ignition
temperature to a level that can be found
in stack gases. At least 400 °F is required
for a noble metal catalyst to work. In
theory, the catalyst promotes the
chemical reaction of oxidation (or
burning) without itself being used up.
Adequate oxygen and unburned
combustion products (such as hydro-
carbons or carbon monoxide) are
required in addition to the catalyst to
sustain this secondary combustion.
The use of catalysts is relatively recent
and little data existed at the time of the
study. It does appear that proper stove
operation can be very important to
satisfactory catalyst operation, as can
placement of the catalyst within the
stove. The stack gases have to be kept
hot, which may cause overheating of the
room if the stove is oversized. The
catalyst can be "poisoned" by metals
present in magazine inks and pressed
wood resins. The catalyst can be fouled
by creosote formed when the stove is
not burning hot enough for the catalyst
to function. There also are questions
regarding possible hazardous or toxic
materials resulting from the catalytic
combustion process that have yet to be
answered.
Add-On Pollution Control Devices
The only system identified as being
specifically for pollution control was a
stainless steel wire mesh inserted into
the exhaust stack. This device is
designed to accumulate condensible
hydrocarbons when the gas stream is
cooler, and burn the accumulated creo-
sote when gas temperatures rise. A
study by the Oregon Department of
Environmental Quality showed a 50%
reduction in particulate levels, but source
tests conducted in Task 5 of this study
showed no decrease in emissions.
Heat Storage Systems
In theory, a combustion device that
can store and slowly release heat will be
more efficient, more comfortable, and
require less operator attention. Many
stoves now include fire-brick linings of
the combustion chamber to accomplish
this. No technical data was available that
furnished a comparison of either
emissions or stove efficiencies, however.
14
Wood Selection and Preparation
Proper selection and preparation of
the wood to be burned can have a signi-
ficant impact both on the overall
efficiency and the emissions. High
moisture content wood have a lower
effective Btu content than dry wood,
since energy Is required to evaporate the
moisture during combustion. The
moisture also may interfere with the
combustion process by cooling the fire.
Emissions are greater from wet wood
both because more wood is required for
a given Btu output and because of the
decreased combustion efficiency.
Overly dried wood, such as kiln dried
scraps from lumber mills, also can cause
excessive emissions. The dried wood
burns too fast for complete combustion,
with the result of unburned components
escaping as pollutants. Optimal moisture
content is 10-20%, which corresponds to
air drying for 3 to 12 months (depending
on the wood species and climate).
Properly dried wood has significantly
more heating value per unit of wood
burned than green wood.
Log size affects the emissions, largely
through the rate of combustion. Large
logs burn too slowly to generate hot
enough temperatures for complete
combustion, whereas small logs burn too
quickly with unburned combustion
products escaping up the stack. The
optimal log size is 31/2 to 5" in diameter
both for combustion efficiency and mini-
mizing emissions.
Wood species selection has an effect
on emissions. Although most wood
species have about the same Btu value
per pound, other burning characteristics
will vary. Hardwoods are denser and have
a higher Btu value per volume than soft-
woods, and therefore require less
frequent fire charging. The Btu value
varies from 24 to 31 million Btu's per
cord for other species such as
ponderosa pine and western red-cedar.
Douglas Fir, commonly used in the
Northwest, has a heating value in
between these extremes. Woods also
vary in their tendency to form ash and
creosote. For example, at least one study
has found that creosote formation is
higher when soft wood is burned than
when hard wood is burned.
-------
Operation
Proper operation of the woodburning
device affects the amount of wood
burned, the comfort of the heat
produced, and the pollutants emitted. In
general, operating practices which
promote complete combustion will
minimize the wood used and the
pollutants emitted. The information on
the effect of burn rate on emissions from
wood stoves in conflicting. Wood stoves
do burn cleanest when the firing
chamber is filled 30% by volume with
wood logs should be laid so that air can
be circulated freely during combustion. A
charging rate of about once per two
hours appears to be the optimum in
maintaining consistent burning
conditions, and minimizing emissions,
without being too demanding on the
operator.
The stove must be properly sized for
the specific installation; the bigger is not
necessarily better. Overnight banking of
the fire should not be used; air dampers
should be left open to promote complete
combustion. A stack or surface thermo-
meter should be used to monitor stove
operation and ensure a hot fire without
excessive stack heat loss.
Proper operation is also important for
fireplaces. Due to their very low (possibly
negative) efficiency, fireplace use should
be avoided in extremely cold weather.
Hot, full fires should be burned to reduce
emissions. Fireplaces with glass doors
should be operated with the doors
opened during burning, then closed
during the burn down period and when
the fireplace is not in use.
15
-------
-------
TaskS
Emissions Testing of Wood Stoves
Under this Task, a number of different
woodstoves were tested for particulate
and gaseous pollutants. The three major
objectives were to: 1) identify the effect
of wood moisture on emissions; 2) eval-
uate several simpler, less expensive test
methods that might be used as an alter-
native to the relatively expensive
particulate measurement test; and 3) test
the emissions from state-of-the-art,
improved stoves and add-on devices.
Test Procedures
The tests were conducted by OMNI
Environmental Services, Portland,
Oregon between June and October 1981,
under laboratory conditions. It was
known from previous research that many
variables can affect the emission rate
from woodstoves, including burn rate,
wood moisture content, type of wood,
wood size, and size of each batch of
wood burned. A standard burn procedure
was therefore established, to minimize
the effects of these variables as much as
possible. The single exception was the
series of tests for the effect of wood
moisture on emissions, where the wood
moisture content was deliberately varied.
The standard burn procedure was:
• A constant heat output was main-
tained, at a low to moderate burn rate
typical of Pacific Northwest burning
practices. This burn rate was about 2.5
kg wood/hour (5.5# wood/hour). The
constant heat output was maintained
by adjustments of air inlets and/or
damper, based on the combustion
chamber temperature.1
• One single charge of wood was fed to
a hot bed of coals. The test period ran
from when the wood was well lit until
it had been reduced to 10% of the
original weight.
• Seasoned Douglas fir wood (except
for wood moisture tests) of 12" to 16"
girth was used.
• The door to the stove was opened for
30 seconds to three minutes at the
start of each run, to ensure the fire
was well lit. After that, the door was
normally only opened once, towards
the end of the burn cycle, to re-
distribute the wood.
• The stove was located on a weighing
platform, so that it could be deter-
mined when 90% by weight of the
wood had been burned.
The tests conducted for each run were
particulate, carbon monoxide, carbon
dioxide, oxygen, opacity, total hydro-
carbon, creosote deposition, and smoke
spot density. The particulate test was
modified EPA Method 5, with an
unheated filter after the third impinger to
collect condensible hydrocarbons. The
test procedures for the other parameters
measured are discussed in the section
on simplified test methods. Figure 2
shows the stove and testing
configuration.
Figure 2
Test Site Configuration
A • Length of reduction is 0.6 ft. when D = 8 in.
and is 1.3 ft. when D = 6 in.
D - 8 in. for all runs except runs 14 and 15 when
D = 6 in.
Rain cap removed /
during testing /
Roof
Transmissometer
3ft.
_ 10 inch
diameter
A
A
T
Gaseous
Sample-
Point
9.5 ft.
Floor
0.5 ft.
Particulate
Sample Point
D
12ft.
11.75ft.
8ft.
Scale
1. The one exception was the ceramic stove. A much higher burn rate was used for this stove, as recommended by the manufacturer. Also, the burn rate could
not be adjusted for this stove because there were not adjustable air inlets.
17
-------
Effect of Wood Moisture Content on
Stove Emissions
An airtight box stove, typical of those
common in the Pacific Northwest, was
used for this test series. Two test series
were run for the following moisture
levels: low—12% moisture content on a
wet wood basis 04% on dry basis)1,
medium—20 to 21% moisture content
on a wet wood basis (25 to 26% on dry
basis), and high—56% moisture content
on a wet wood basis (126% on dry basis).
The results are shown in a tabular form
in Table 11 (summary of all test results),
and graphically in Figure 3.
The lowest paniculate emissions were
measured for the medium moisture
content level. The dryer wood tended to
burn faster, and to maintain a constant
heat output when the air inlets were
restricted. It is believed that the high
emissions from burning the dry wood
were due to the incomplete combustion
resulting from the combustion air
restriction and the extremely rapid
burning rate of the low moisture wood.
Figure 3
Paniculate Emissions Results:
Fuel Moisture Tests
g/kg mg/m'
50 1000
40 800
LJJ
£ 30 600
1
20
10
400
200
[%j Total Paniculate
| Front Half Paniculate
| | Creosote
x.x Burn Rate, kg/hr.
2.4
-1.9-
I
2.1
— 2.9-
1.7
Run
Low Moisture
12% Wet Basis
(14% Dry Basis)
1 7
Medium Moisture
20% Wet Basis
(25% Dry Basis)
High Moisture
56% Wet Basis
(126% Dry Basis)
Table 11
Emissions Summary
Paniculate Emissions
Test STOVE TYPE/Test
Run
1 Box-21 % moisture content9
2 Box-12% moisture content9
3 Box-56% moisture content9
4 Box-12% moisture content9
cold start
5 Box-12% moisture content9
6 Box-19% moisture content9
cold start
7 Box-20% moisture content9
8 Box-56% moisture content9
cold start
9 Box-56% moisture content9
10 Box/Catalytic Add-on
1 1 Box/Catalytic Add-on
12 Box/Non-Catalytic Add-on
13 Box/Non-Catalytic Add-on
14 Catalytic Box
15 Catalytic Box
16 Catalytic/Secondary Air
17 Catalytic/Secondary Air
18 Ceramic
19 Ceramic
1 Dry Basis
'Oregon DEQ Method 7
3EPA Method 5
Bum
Rate
(Ibftir)'
19 (41)
24 (53)
21 (47)
8 1 (17 8)
2 7y (60)
5 8 (12 8)
29 (64)
20 (45)
17 (37)
21 (46)
26 (58)
24 (54)
22 (49)
17 (37)
22 (49)
30 (67)
21 (47)
6 4 (14 2)
39 (87)
g/dscm
1 3
34
1 5
36
50
38
1 5
09
2
3
3
8
8
22
1 5
06
1 2
02
0 1
Total1
Ib/hr
009
028
0 16
072
037
054
0 12
011
008
0 10
009
021
017
0 14
0 11
009
0 14
002
001
g'kg'
22
54
34
40
62
42
19
24
22
22
17
38
35
38
23
14
30
1
2
Front
Half1
g/kg' <%)•
6 3 (29)
140 (27)
110 (33
9 6 (24)
11 0 (17)
8 3 (20)
4 4 (23)
6 1 (25)
4 8 (22)
37 (17)
33 (20)
7 4 (20)
61 (17)
59 (16)
48 (21)
3 5 (25)
57 (19)
0 63 (53)
0 97 (49)
Cresote' Carbon Gaseous
Monoxide HC
mg/m'-kg g/kg4
969
917
218
216
592
291
568
109
240
223
190
273
337
379
273
88
317
56
27
•Mass Emissions per Mass Dry Fuel Consumed
5Percent of Total Emissions
'As Hexane
190
189
160
210
220
170
160
190
110
110
90
200
160
120
50
80
150
20
50
g/kg4 '
13.8
16.9
10.5
11 9
12.1
9.2
88
11 1
66
80
6.2
11.8
9.3
10.7
73
39
58
0 4
06
Opacity*
%
36
30
20
40
37
46
28
20
11
24
16
22
23
10
10
5
13
0
0
Stack Gas
Flow
SCFM
(DSCM)
30 (1060)
38 (1350)
50 (1750)
91 (3230)
34 (1200)
66 (2320)
36 (1270)
55 (1950)
32 (1130)
37 (1320)
32 (1130)
54 (1890)
42 (1190)
30 (1060)
34 (1200)
71 (2500)
56 (1960)
52 (1850)
49 (1720)
Excess
Air
%
230
160
180
140
180
140
170
410
320
380
240
320
280
140
172
347
498
63
97
•Visual Observer
9Wet Basis «
(Weight of moisture)
(Total weight of wood including moisture x 100%)
1. Dry wood basis refers to the percentage of the moisture weight compared to the weight of the dry wood (0% moisture content). This can be calculated as
follows:
% moisture content dry basis =
weight of moisture
weight of wet log — weight of moisture
x 100%
18
-------
Evaluation of Simplified Test
Procedures
The standard test method for total
particulate, EPA Method 5, is expensive
both in the initial price of the equipment
and to run each test. For the series of
nine tests recommended for certification
of an individual brand of stove, for
example, using EPA Method 5 would
cost about $15,000. This high cost makes
it very difficult to conduct the necessary
basic research into ways of reducing
emissions from woodstoves, as well as
making it difficult for stove manu-
facturers to experiment with stove
design. This task evaluated the more
likely alternate test methods, and how
well the results compared with the EPA
Method 5 particulate tests. The pollu-
tants and/or alternative methods
evaluated were: carbon monoxide, total
hydrocarbon, creosote deposition,
opacity, and smoke spot density.
The carbon monoxide and total hydro-
carbon tests both showed a reasonable
correlation to particulate emissions,
once a correction for excess air was
made. The linear correlation coefficient
was 0.80. For both tests, a non-
dispersive infrared (NDIR) analyzer was
used, since both tests involve measuring
concentrations rather than total
emissions by weight. The amount of air
going up the stack will very much
influence the concentration values
(amount of particulate per unit volume of
air) recorded even though the total
pounds of emissions remain the same.
To correct for this and to allow a
comparison with the total particulate by
weight, the concentration values were
corrected to their values when the air in
excess of the theoretical minimum
required for complete combustion
(excess air) was considered zero. The
correction factor is based on the
theoretical CC>2 concentration at
complete combustion and the actual
measured
These tests have several advantages.
They are relatively inexpensive— about
$4,200 for a nine-test series, compared to
$15,000 for the EPA Method 5 particulate
tests. They also give immediate results,
and allow the emissions to be quantified
in different parts of the combustion
cycle rather than just one value for the
entire cycle. One disadvantage in using
the NDIR to measure total hydrocarbon
is that the instrument may have a varying
response to different types of hydro-
carbons. The NDIR should produce
useful results under similar testing
conditions, however.
Creosote deposition was measured by
weighing small steel plates before and
after suspending in the stove flues
during a burn cycle. These plates, each
three inches by five inches, were placed
in the flue at two different levels. The
weights of creosote deposits were then
compared to each other and to the
particulate measured by EPA Method 5.
The precision between the pairs of
values for creosote was poor, with a
correlation coefficient of only 0.62. The
average of the pair values had a good
correlation (correlation coefficient of
0.82) with the measured particulate, but
only if two of the 19 tests were thrown
out. If all 19 tests were used, the
correlation coefficient drops to 0.52,
indicating a poor match between
creosote and particulate emissions.
Opacity* as determined by a trained
observer or by a transmissometer located
in the stack correlated very well with
particulate concentration, (0.80) but not
as well with total particulate (0.69-0.76).
There are no simple correction factors for
opacity to account for the excess air.
Since we are interested in the relation of
opacity to total particulate, opacity was
not shown to be a good alternative
method.
The smoke spot density measure-
ments were not useful as a predictor of
particulate emissions. In this test
method, a given volume of stack gases
are pulled through a filter paper. The spot
left on the filter paper is then measured
for darkness. The test was not sensitive
enough, and showed little or no relation-
ship to particulate emissions.
In summary, the carbon monoxide and
total hydrocarbon measurements offer
the best alternative method. They are
much less expensive than the EPA
Method 5. Since they do not directly
measure particulate, however, their
usefulness is greater as a screening tool
for research and stove design experi-
mentation, rather than for certification
where greater precision is required. Table
12 is a summary of the simplified test
procedures.
Table 12
Simplified Test Procedures Summarized
Estimated Cost2
Method/
Description
Correlation to
Total Particulate1
Advantages
Disadvantages
Capital
Investment
Per Series
Per Run (9 tests)
Total Particulate/EPA
Method 5, modified
to include filter for
condensible hydro-
carbons
1 Accurate measurement
of emissions, including
condensibles
> Integrated sample
over entire burn cycle
Expensive
1 Long sample time
required (1-hr min-
imum); not ideal
for measuring
discrete periods
within burn.
1 Experience with
method required
$20,5003
$2,100
$15,000
(Continued on Page 20)
Footnotes on Page 21
"The opacity of a smoke plume is a measure of how much a background object, such a tree or building, is visually obscured by the smoke plume.
19
-------
Table 12 (Continued)
Simplified Test Procedures Summarized
Method/ Correlation to
Description Total Particulate1
Filterable 20-50% of total •
Participate participate mass
basis. Correla-
tion coefficient •
not determined
•
High Volume/ — •
Not Tested
Carbon monoxide/ 08 •
Orsat and Non-
dispersive In-
frared (NDIR)
•
•
Total Hydro- 08 •
carbon/NDIR
Creosote/2 small 0 8 (1 7 of 1 9 tests) •
steel plates 0 5 (1 9 of 1 9 tests)
weighed before •
and after, averaged.
•
•
Advantages
Accurate measurement
of emissions
Integrates sample
over entire burn
cycle.
Slightly less expen-
sive than total par-
ticulate method.
Short sample time en-
ables measurement of
discrete periods
within burn
Provides instantane-
ous and continuous
output, excellent for
monitoring burn cycle
Inexpensive to use
once capital invest-
ment incurred.
Suitable for screen-
ing method, using
cheaper, less accur-
ate instrumentation
Suitable for field
monitoring
Same as Carbon
monoxide
Uncomplicated to use.
Inexpensive (low level
of effort)
No significant capi-
tal investment
Capable of measuring
discrete periods with-
in burn cycle.
Disadvantages
• Does not measure
condensible par-
ticulates
• Expensive
• Long sample time
required (1-hr min-
imum), not ideal
for measuring dis-
crete periods with-
in burn
Experience with
method required.
• Expensive
• Collection effi-
ciency for conden-
sible organics un-
known.
• Multiple samples
required to obtain
measurements for
entire burn cycle.
• No direct measure-
ment of particulate.
• Same as Carbon
monoxide
• Potential for variable
response to different
HC species
• Not direct measure-
ment of particulate
emitted to atmos-
phere
• Results likely de-
pendent upon numer-
ous variables such
as stack tempera-
ture, and excess
air.
Estimated Cost2
Capital Per Series
Investment Per Run (9 tests)
20.5003 " 1,900 12,500
16,500 $1,900 $12,600
$13,5005 $560 $4,200
600-26006 8 280 —
23,500s 560 4,200
4500-67007 e 280 —
1,000 250 2,800
(Continued on Page 21)
Footnotes on Page 21
20
-------
Table 12 (Continued)
Simplified Test Procedures Summarized
Estimated Cost2
Method/
Description
Correlation to
Total Paniculate1
Advantages
Disadvantages
Capital
Investment
Per Run
Per Series
(9 tests)
Opacity/Trans-
missometer or
visible emission
inspector.
08 • Inexpensive to use.
• Little or no capital
investment.
• Suitable for field
monitoring (visible)
emission observer)
Does not directly
measure particu-
late.
1 Results highly de-
pendent upon excess
air levels
$0 -1,000"
$280
$2,800
Smoke spot den-
sity/Bacharach
Smoke Spot Tester
06
Inexpensive to use
Low capital investment.
Easy to use
Results dependent
upon excess air
levels
75 00"
280
2,800
• Large number of
measurements re-
1 Very short measurement quired over entire
time, may be used to monitor burn cycle.
discrete periods within burn
cycle • Little correlation to
total particulate
1 Convenient for field
monitoring.
'Correlation coefficient determined from this study, using EPA Method 5 as the basis of comparison
2See Appendix D of the full report for basis of cost estimates.
Includes laboratory quality, CO, CO2, and O2 monitors for accurate and continuous determination of stack gas composition for determining stack gas volumetric
flow by stoichiometry. Subtract $12,500 for monitors and add $3,000 for orsat and low velocity flow measurement instrumentation equipment (net change—
$9,500) if orsat/velocity methods to be used.
"Add $5,000-$10,000 for flame ionization detector if total hydrocarbons analysis is desired
tion adjusted for excess air.
6lf used as a screening test with orsat or less accurate instrumentation to determine emissions (CO concentration adjusted for excess air).
7lf used as a screening test with orsat or less accurate instrumentation (e.g., NDIR instead of heated FID) to determine emissions (hydrocarbon concentration
adjusted for excess air)
'Add $1,000 if platform balance to be used to monitor fuel consumption rate
Improved Technology Stoves and
Add-on Devices
Based on a literature review, three
areas of stove technology were chosen
for testing. These were catalytic, modi-
fied technology, and combined tech-
nology woodstoves. The two add-on or
retrofit devices tested were catalytic and
non-catalytic units. These units are
described in Table 13. (overleaf)
The tests were run as explained
previously in the "Test Procedures"
section. That is, one load of wood with
25-30% moisture content was introduced
to a hot bed of coals and a low to
moderate burn rate was maintained.
Where catalysts were present, the manu-
facturer's recommended minimum
temperatures were also maintained. The
one exception to this was the ceramic
stove. This stove did not have an
adjustable air inlet and the burn rate
could not be regulated. The burn rate
was significantly higher (5.4 versus
2.5 kg/hour).
The emission testing results are
shown in detail in Table 11. These results
indicate that none of the improved units
except the ceramic stove tested better
than the standard box stove under
burning conditions similar to those that
typically occur in Northwest homes. The
ceramic stove did have substantially
lower emissions, however the very high
burn rate compared to the other units
may have contributed to some or all of
this emissions reduction. The burn rate
used was considered marginal for proper
catalyst operation for those units having
catalysts, although they were typical for
Northwest homes.
A literature review showed that other
researchers have found similar improved
technology devices result in lower
emissions. The test results are limited,
however, and are difficult to compare to
this study because of different burn
rates, wood species and moisture
content, and other different operating
parameters.
21
-------
Table 13
Descriptions of Improved Technology Stoves and Add-on Devices Tested
Type of Stove/
Device
Catalytic Stove
Catalytic/modified
combustion stove
Ceramic stove
(modified
technology)
Catalytic retrofit
device
Non-catalytic
retrofit device
General Description Firebox Stove Surface
of Stove/Device Volume Area
Modified welded steel 2 6 ft1 25 4 ft2
box stove. Catalyst-
precious metal on
ceramic honeycomb
support
Catalyst-precious 3 0 ft1 35 ft2
metal on flat
ceramic plates
Cast ceramic stove, 3 2 ft' 22 3 ft2
spherical combustion
chamber Air space
between chamber
and outer shell No
controls on air flow
Heat exchanger
Placed in flue just
above stove collar.
Precious metal
catalyst on ceramic
honeycomb Heat
exchanger
Filters particulate,
either burns on pad
or is removed and
cleaned
Secondary
Tertiary Air Baffles
Preheated secondary After catalyst
combustion air-
introduced at
catalyst
Secondary-preheated, Before and
above grate after catalyst
Tertiary-preheated,
at catalyst plates
None None
Secondary air
prior to catalyst
None
'Weight of wood is on a dry basis For comparison, the box stove tests (no add-ons) under similar wood moisture and firing conditions
2Much higher burn rate used, as recommended by the manufacturer
Particulate
Emissions/
Weight Wood
Burned1
30 5 g/kg
22 g/kg
1 5 g/kg2
22 g/kg
27 5 g/kg
had 20.5 g/kg emissions
22
-------
Task6
Control Strategy Analysis
The growth in residential wood
combustion has been identified as a
significant contributor to non-attainment
of total suspended particulate (TSP)
ambient standards in several Pacific
Northwest cities. Under the Clean Air Act,
each state is required to prepare a legal
implementation plan for bringing each
non-complying geographical area into
compliance. Traditionally, these state-
prepared control strategies for particulate
have focused on industrial emissions and
have been successful in substantially
reducing the industrial contributions.
Attention is turning now towards possible
control of remaining major particulate
contributors, including residential wood
combustion. Task 6 examines and
evaluates possible control strategies to
reduce residential wood combustion
emissions.
Seventy-five possible control strategies
were selected, including those strategies
that have been implemented somewhere
in the world, those suggested by know-
ledgeable air pollution control agency
personnel, and those suggested by
project members. A systematic ranking
system was developed using the Keppner-
Tregoe evaluation process (see Table 14)
and each of the 75 control strategies
ranked. The fifteen highest ranking
strategies were further evaluated as to
costs, projected emissions reductions,
and significant advantages and
•disadvantages.
The ranking and evaluation process
necessarily included two major sources of
uncertainty. These are the assumptions
required to be made because of lack of
data (costs and projected emission
reductions, for example), and the
somewhat subjective values assigned
during the ranking process. Such factors
Table 14
Criteria and Weight Factors Used in Keppner-Tregoe Analysis
CRITERIA
MUST Criteria
1 Reduce air pollution impacts from RCW
2 Meet legal requirements
3. Widely applicable to RWC equipment or
operating practices
4 Must not increase safety hazard
5 Can be implemented within five years,
unless long-term benefits great
WANT Criteria
1 Reduce average RWC emissions/household
2 Reduce number of RWC households
3 Widely applicable
4 Maximum public acceptance
5 Discourage worst appliances/practices
6 Minimum consumer cost
7 Uses proven technology
8 Minimum circumvention of control measure possible
9 Maximum agency administrative feasibility
10. Encourages innovative technology
11 Minimum free market interference
12. Promotes conservation/use of renewable resources (except wood)
WEIGHT FACTOR*
Mandatory
Mandatory
Mandatory
Mandatory
Mandatory
13
13
10
9
9
6
5
4
3
2
2
1
•These weight factors were calculated using an analytic tool called "paired comparison" For further
details, see Appendix B of the complete report on Task 6.
as the response to public education
programs, public acceptance of control
strategies, and the political feasibility in
passing the necessary laws, were
particularly difficult to assess quantita-
tively. Any specific control strategy
development would require in-depth
analysis of local conditions and re-
evaluation of these control strategies.
However, this study should assist control
agency personnel in the initial selection
of possible control strategies for further
evaluation. The fifteen highest ranking
strategies are briefly described in the
following pages. A summary of the costs
and expected particulate reduction for
each strategy is shown in Table 15.
(overleaf)
23
-------
Table 15
Summary of Estimated Costs and Particulate Emissions Reduction Benefits for
Fifteen RWC Emission Control Strategies
Control Strategy
1 Mandatory testing/certification,
tax credit
2 Mandatory testing/certification,
no tax credit
3 Encourage use of larger pieces of firewood
4 Mandatory testing/labeling, tax
credit for cleaner units
5 Encourage use of other fuels
and energy sources
6 Government funded research and
development
7 Promote proper sizing of
wood stoves
8 Mandatory testing/labeling, no
tax credit
9 Mandatory weathenzation — all
households Cost effective
10 Stove testing/rating by Trade
Association
1 1 Mandatory cost effective weathenzation —
New or replacement RWC households
12 Encourage burning of dry firewood
13 Curtailment of RWC during air
pollution episodes
14 Annual inspection/maintenance
of installed units
15 Require underfire air for new
fireplaces
* Assume 1980 as base year
'Assume 5000 new wood stoves/year sold, 90%
'Assume 5000 new wood stoves/year sold, 60%
Start-up
$ 50,000
50,000
50,000
UNKNOWN
UNKNOWN
50,000
50,000
25,000
Costs
AdmlnJYear
$ 130,000
130,00
45,000
160,000
45,000
130,000
150.000
130,000*
50,000
45,000
8,000
1,000,000
% Particulate Reduction
Other 1985*
$400/unit 3
tax credit
2
115
$400/unit 1
tax credit
1
$1450 per
household
1
$1450 per
household
62
335
UNKNOWN
1100/fireplace'
3Fuel savings exceed cost of weathenzation
are certified
are certified
1990 1995 2000
11 30 39
8 22 30
115 115 115
5 15 21
7 105
3 8 11
35 164
3 8 11
35 115
62 62 62
33s 335 33s
1 2
Annual costs only for
administrative expenses
•Costs paid by private sector
Cost per % Particulate Reduction
1985*
$ 600,000/%'
65,000/%
3.900/%
1,330,000/%!
140,000/%
140,000/%
7,300/%
400/%5
1990
$ 175,000/%'
16,2007%
3,900/%
266,000/%*
6,400/%
45,000/%
9,100/%3
45,000/%
14,300/%3
7,300/%
4001%'
200,0001%'
1995
$ 64,300/%'
5,900/%
3,900/%
88,700/%2
16,7007%
21,400/%3
16,700/%
7,100/%3
7,300/%
400/%s
5Reduces particulate emissions only during episodes
2000
I 49,500/%'
4,300/%
3900/%
63.300/%2
4,300/%
12,0007%
12,000/%
7,300/%
400/%s
100,000/%«
Effect on
annual TSP levels unknown
8Assume 2000 new fireplaces each year
Fifteen highest ranking strategies:
Mandatory certification of wood-
stoves. Only clean stoves can be sold.
Tax credit of $400 to purchaser.
Strategy rank: #1
Expected % paticulate reduction
by year 2000: 39%
Expected cost: $130,000 year
administrative costs plus $4007
stove
This strategy would require that all
models of wood stoves and furnaces
be tested, and that only units capable
of emitting less than 5 grams particu-
late/kilogram of wood burned could be
sold. This level of emissions is
approximately one-fourth as great as
emitted by the average stove in 1982.
A tax credit or rebate of $400 would
be given to help defray the cost to the
consumer of these cleaner units,
expected to be about $1200 each. It is
assumed that an average woodstove
has a 10-year life, and therefore about
10% of the woodstoves will be
replaced each year. This strategy is
expected to be very effective in the
amount of particulate reduction. The
cost of the program for the tax
credits, and the resistance of
consumers to the extra cost and limit
on their choice of stoves are major
drawbacks. Depending on the size of
the geographical area, a larger or
smaller percentage of new or replace-
ment stoves may be "bootleg," i.e.,
cheaper, dirtier stoves. If a relatively
small area is chosen for mandatory
certification, it is expected there will
be a larger percentage of dirty stoves.
Overall, it was assumed that about
90% of new or replacement stoves
would be the cleaner units, with the
remaining 10% uncertified dirtier
stoves. The cost of the tax credit
program for cleaner woodstoves
seems reasonable, since it is compar-
able to the amount of money currently
spent in Oregon on tax credits to
reduce particulate emissions from
industries.
Another drawback in some juris-
dictions is the legal prohibition from
regulating residential heating. Where
such laws exist, they would have to
be changed before mandatory certifi-
cation could occur.
24
Mandatory certification of wood-
stoves. Only clean stoves can be sold.
No tax credit.
Strategy rank: #2
Expected % particulate reduction
by year 2000: 30%
Expected cost: $130,000/year
This strategy is the same as #1
above, but without a tax credit. It is
considerably cheaper, but less effec-
tive in particulate emission reduction
as it is expected that more consumers
will buy dirtier, uncertified stoves. It
was estimated that only about 60% of
the new or replacement woodstoves
purchased would be certified clean-
burning stoves, with the remaining
stoves bought from other nearby
areas where uncertified stoves are
sold. This strategy would likely face
some public opposition because of
the high cost of the certified stoves,
with no off-setting tax credit.
-------
Encourage use of larger firewood
piece size through public education.
Strategy rank: #3
Expected % participate reduction
by year 2000:11.5%
Expected cost: $45,000/year
Based on very limited test data, it
has been demonstrated that increas-
ing the log diameter by two inches
decreases the emissions by 32-36%
for log diameter sizes commonly used
in woodstoves (two to six inches).
Researchers speculate that this
phenomena exists because the vola-
tile organic fraction is released too
quickly from small logs, and escapes
up the stack before it can be
combusted. Since this method of
emission reduction actually results in
less work and inconvenience for the
woodstove user, it is expected to be
readily accepted. However, these test
results should be confirmed prior to
such a public education program
being started.
Mandatory testing and labeling of all
new woodstoves, but all stoves can be
sold. Tax credit for cleaner units.
Strategy rank: #4
Expected % particulate reduction
by year 2000: 21%
Expected cost: $130,000/year
administrative costs plus $400/
stove tax credit
It was assumed that even with a tax
credit, only about 40% of the wood-
stoves purchased would be the
cleaner units. Dirtier units would
continue to be purchased because of
their much lower cost.
Encourage use of other fuels and
energy sources.
Strategy rank: #5
Expected % particulate reduction
by year 2000: unknown
Expected cost: unknown
This strategy assumes that the use
of conventional energy sources such
as oil, gas, and electricity, and alter-
native energy sources such as solar
energy, would be encouraged. Whether
or not such a strategy is practical is a
major drawback, particularly where
increasing oil usage is involved. Wood
use would also be discouraged under
this strategy by restricting the time of
year and amount of wood removed
from public lands, and by encouraging
alternative uses for the wood, such as
using the wood for fuel in industrial
boilers.
Government-financed research and
development.
Strategy rank: #6
Expected % particulate reduction
by year 2000: unknown
Expected cost: unknown
This strategy assumes that the
government would encourage and
offer financial support for a major
research effort. Areas to be
researched include developing less
polluting woodburning units, improv-
ing operator practices, and developing
a better and cheaper standardized
emissions test procedure. The support
to be offered could include a staffed
emissions test facility free for
promising research, tax credits for
research, and substantial awards for
the designer of exceptionally clean-
burning units.
Quantifying the benefits of such an
effort is not possible. However,
research is clearly an important factor
in reducing RWC emissions over the
next 20 years.
Promote downsizing of stoves
through public education.
Strategy rank: #7
Expected % particulate reduction
by year 2000:10.5%
Expected cost: $45,000/year
Many woodstoves now in operation
are too large for the space to be
heated, with users shutting down the
air supply to the stoves to slow down
the fire and to prevent overheating the
area. These smouldering, slow fires
result in very high emissions. By
encouraging properly-sized stoves,
the average charge size is reduced,
brisker fires produce an equivalent
amount of heat, and emissions are
reduced.
There are a number of advantages
to the woodstove owner in having a
smaller stove: the unit itself is
cheaper; less wood will be burned for
the same amount of heat output; and
the safety problem from creosote
accumulation will be reduced. The
major disadvantage will be that more
25
frequent stoking with less wood will
be required, which is an inconven-
ience to the user.
Mandatory testing and labeling of all
new woodstoves, but all stoves can be
sold. No tax credits for cleaner units.
Strategy rank: #8
Expected % particulate reduction
by year 2000:11%
Expected cost: $130,000/year
This is the same as #4 above,
except without tax credits for cleaner
units. It was assumed that without a
tax credit only about 20% of the new
units purchased would be clean-
burning.
Mandatory weatherization of all
households to cost effective level.
Strategy rank: #9
Expected % particulate reduction
by year 2000:16.4%
Expected cost: $150,000/year plus
$1450/household weatherized
It is assumed under this strategy
that about half of the households with
woodstoves would weatherize. Of
those houses weatherized, there
would be 15% fewer burn days
(marginally cool days), and that 40%
less wood would be burned on days
when the woodstove is used. It is
further assumed that financing would
be no- or low-interest loans, to be
made by the government or local
utility.
Such a weatherization program has
obvious benefits in reducing the
conventional energy usage, such as
oil or gas, that are commonly used to
supplement wood heat in homes
having woodstoves. Some areas
already have such financing assist-
ance available. The cooperation of
government or the utilities in helping
to finance each home's weatherization
obviously is a key element in this
strategy.
-------
Stove testing and rating by trade
associations.
Strategy rank: #10
Expected % particulate reduction
by year 2000:11%
Expected cost: $130,000/year (paid
by trade associations)
This strategy assumes that a trade
association would voluntarily test and
rate all woodstoves, and further, that
the association would widely publicize
such results. There are some prece-
dents for such a program: the testing
of electrical equipment by the Under-
writers' Laboratory, and the testing of
refrigeration units by the Air Condi-
tioning and Refrigeration Institute.
However, whether or not this program
would be instituted for woodstoves is
unknown. The level of particulate
reduction and cost are assumed to be
the same as for the government test-
ing and labeling strategy, #8,
described above.
Mandatory weatherization for house-
holds buying new or replacement
woodstoves.
Strategy rank: #11
Expected % particulate reduction
by year 2000:11.5%
Expected cost: $50,000/year plus
$1450 per household insulated
This is the same as strategy #9
above, except that not all houses
would have to be weatherized. This
has the advantage of affecting fewer
households, which will reduce the
cost. It is expected that more people
will be tempted to circumvent this
strategy by installing woodstoves
without weatherizing, since they may
feel they are being unfairly singled out
for substantial cost for insulation. The
estimated circumvention rate is 30%,
which reduces the percentage particu-
late reduction expected by that same
percentage over strategy #9, where all
houses are weatherized.
Encourage burning of dry firewood.
Strategy rank: #12
Expected % particulate reduction
by year 2000: 6.2%
Expected cost: $45,000/year
Some tests have shown that reduc-
ing the moisture content of firewood
to 25% results in better combustion
efficiency and less emissions. This
strategy assumes that through public
education, the number of households
properly covering woodpiles can be
increased. Vendors selling more than
10 cords/year of wood would also be
required to state the moisture content
of wood sold. An additional reduction
in emissions is possible if fall cutting
of firewood on public lands was
prohibited, since the firewood would
then have been seasoned at least six
months prior to use. However, it was
expected that forestry officials may
object to this as their goal is to have
the extra wood removed as soon as
possible to allow reforestation.
Curtailment of residential wood
combustion during air pollution
episodes.
Strategy rank: #13
Expected % particulate reduction
by year 2000: 33% (only during
episodes)
Expected cost: $8,000/year
Voluntary curtailment of RWC
would be requested whenever a
specific 24-hour particulate level was
exceeded. Mandatory curtailment of
RWC would be implemented if the
voluntary approach was not effective.
Enforcement would be by visual
opacity checks during the day. A
strong public education program
would be included to encourage the
public's cooperation.
Annual inspection/maintenance of
installed units.
Strategy rank: #14
Expected % particulate reduction
by year 2000: unknown
Expected cost: $1,000,000/year
This strategy, if implemented,
would be most effective in reducing
emissions from stoves with catalysts
(with a limited life), and more sophis-
ticated stoves expected in the next
ten years (which may require frequent
adjustment and maintenance for
optimum performance). However, this
strategy would have high cost and
expected high public resistance.
Require underfire air for new
fireplaces.
Strategy rank: #15
Expected % particulate reduction
by year 2000: 2%
Expected cost: $100/fireplace, 1000-
3000 fireplaces to be constructed
per year.
Based on limited testing, a 40%
reduction in emissions was found
comparing similar fireplaces and
burning practices, but with one fire-
place having underfire air. If further
testing confirms these results, it will
be a relatively easy way to reduce fire-
place emissions. This strategy would
be implemented as part of the require-
ments for a building permit.
26
-------
Task?
Indoor Air Quality
Recent increases in the use of resi-
dential wood combustion appliances and
home weatherization have focused new
concern on public health risks asso-
ciated with indoor participate air
pollutants from wood stoves. Several
known carcinogens as well as substan-
tial fine participate emissions have been
identified with woodstoves. The purpose
of Task 7 was to develop a better
understanding of the indoor particulate
and polynuclear aromatic hydrocarbon
concentrations (PAH) during appliance
use, thereby providing a basis upon
which future indoor exposure levels can
be assessed.
Five homes in the Portland area were
chosen for indoor sampling. A range of
house ages and weatherization status
were chosen, along with one mobile
home. Sampling occurred over ten days
in May, 1981.
Sampling Methods
In order to separate impacts asso-
ciated with wood smoke from other
indoor sources of particulate and gases,
each home was tested for five days with
the wood stove operating, and five days
without the stove operating. A low
volume sampler with a 30 /^m inlet
restriction operated for 24-hour periods
at a flow rate of 70 liters per minute. The
sampler intake was located at least two
meters from the stove and at a height of
one meter.
Samplers were positioned outside of
each home. One 24-hour sample was
collected (concurrently with an indoor
sample) to measure lead levels. (A
comparison of the lead levels inside and
out allowed a qualitative evaluation of
the rate of air exchange into the house,
since lead very quickly settles out inside
a structure).
Analyses Performed
Each indoor sample was first weighed,
and then analyzed for seven PAH
compounds using gas chromatography/
mass spectroscopy. Lead was deter-
mined by X-ray fluorescence.
Results
No significant difference in four of the
five homes between burn and non-burn
days was found for either particulate
mass or PAH concentrations. The fifth
home did have much higher levels of
particulate and PAH on burn days. This
large difference was traced to fugitive
smoke leaks from the stove, particularly
during wood charging. The sample
results for total particulate mass and
lead are shown in Table 16. Table 17
includes the results for the seven PAH
compounds tested.
Table 16
Residential Wood Combustion
Indoor Sampling Program
Summary of Analytical Results for Mass and Lead
Particulate (ug/m3)
Home
No-Burn Burning
Number Home Type
1
2
3
4
5
Older Home
New Tract Home
Airtight Home
Mobile Home
Rural Home
1 Statistically insignificant at 95
2 5 day average based on 4, 24
(A)
505
165
18.7
32.9
insufficient data
% confidence interval
hour filters
(B)
73.6
23.0
195
38 62
77 42
Difference
(B-A)
23.1
6.51
081
571
—
Lead (ug/nn
Indoor (C)
5.05
3.83
754
2.42
282
X 10'2
X 10'2
X 10'2
X 10"2
X 10'2
Outdoor (D)
1.10
8.04
1.95
2.65
2.94
X
X
X
X
X
10-1
10'2
io-1
10'2
10'2
Ratio
C/D
.45
.47
.38
.91
96
Average Mass of
Wood Burned
Per Day (Kg/day)
18
19.5
17.5
5.8
107
Table 17
Residential Wood Combustion
Indoor Sampling Program
- Summary of PAH Composite Results (ng/m3) -
Home Fluoranthene Pyrene Beiu(a)8nthracene Benzofluoranthenes' Benzo(a)pyrene
Number No-Burn Burn No-Burn Burn No-Burn Burn No-Burn Burn No-Burn Burn
1 0.1 14 0.2 30 — * 413 0.3 51.3 — 26.3
2(A> 03 0.3 0.8 0.7 0.2 0.3 0.2 0.4 0.1 0.3
3 0.1 0.1 0.1 01 — 0.05 0.05 0.4 — 0.2
4 0.3 0.2 0.4 0.7 0.1 0.2 0.4 0.6 0.2 0.3
5 0.2 01 0.3 0.3 — — — — — —
* Blank values indicate specie concentration below minimum detection limit.
1 Benzo(b)fluoranthene and dibenz(a,h)anthracene were not completely resolvable from their isomers, and results
dibenzanthracenes.
-------
It should be noted that testing
occurred during relatively mild weather,
and does not reflect "worst case"
conditions. Average wood use during this
study was 50-60% of the wood use
expected in colder weather. Stoves were
operated an average of 5.6 hours/day on
days when the stoves were used.
The results clearly indicate that
improper stove maintenance or operation
can cause indoor particulate levels and
PAH compounds concentrations much
higher than under optimum stove
operation. The benzo(a)pyrene (B(a)P)
exposure levels seen in the one house
with a leaky stove approximate the equi-
valent of 10 to 38 cigarettes per day for
the inhabitants. B(a)P is a known
carcinogen. The house in question was
older and had an average air exchange
rate, indicating B(a)P levels could be
even higher for an "airtight" home.
Comparison With Other Studies
The results from this study were
comparable with other studies, consider-
ing the lack of uniform testing
conditions. Table 18 compares the results
of this study with five other studies.
Table 18
Comparison of This Survey with Other Surveys
Respirable Particulate
Benzo(a)pyrene (B(a)P)
Study
This survey
Spengler and Ju1
G. Benton et.al 2
GEOMET3
Butler, et al.4
Number
of
Samples
45
85
8
28
—
Concentration
Indoor On
Burn Day ug/m3
464
275
332
490
—
Concentration
Indoor On
Non-burn ug/m3
29.7
180
—
280
—
Number
of
Samples
45
—
2
—
Concentration
Indoor On
Burn Day ng/m3
5.4
—
11 4s
2.1
Concentration
Outside
ng/m3
01
06
29
1 J.D Spengler and C Ju, "Room-to-Room Variations in Concentration of Respirable Particulates in Residences", Environmental Science and Technology Vol 15
No 5, May 1981
2 G Benton, D Miller, M Reimold, and R. Sisson, "A Study of Occupant Exposure to Particulates and Gases from Woodstoves in Homes", Proceedings of the 1981
International Conference on Residential Solid Fuels, June 1981
3 D. Moschandraes, et.al, "Residential Indoor Air Quality and Wood Combustion", GEOMET Technologies, Inc , Rockville, MD
D. Moschandraes, et.al, "The Effects of Woodburning and the Indoor Residential Air Quality", Environmental International, Vol. 4, pp 463-468, 1980
4 J.D. Butler and P Crossley, "An Appraisal of Relative Airborne Suburban Concentrations of Polycycllc Aromatic Hydrocarbons Monitored Indoors and Outdoors",
The Science of the Total Environment, Elsevier Scientific Publishing Company, Amsterdam (1979)
5 Indoors on non-burn day
6 Maximum of 22 sampled residences
28
-------
References
1. Core, John E. et al., Residential Wood
Combustion Study—Task 1—Ambient
Air Quality Impact Analysis, Report
No. EPA 910/9-82-089a and EPA 910/
9-82-089b, U.S. Environmental Protec-
tion Agency, Seattle, Washington,
1982.
2. Core, John E. et al., Residential Wood
Combustion Study—Task 2A—
Current and Projected Air Quality
Impacts, Report No. EPA 910/9-82-
089c, U.S. Environmental Protection
Agency, Seattle, Washington, 1983.
3. Del Green Associates, Inc.,
Residential Wood Combustion Study
—Task 2B—Household Information
Survey, Report No. EPA 910/9-82-089d,
U.S. Environmental Protection Agency,
Seattle, Washington, 1982.
4. Green, William T. and Gay, Dr. Robert
L, Residential Wood Combustion
Study—Task 3—Wood Fuel Use
Projection, Report No. EPA 910/9-82-
089e, U.S. Environmental Protection
Agency, Seattle, Washington, 1982.
5. Del Green Associates, Inc.,
Residential Wood Combustion Study
—Task 4—Technical Analysis of
Wood Stoves, Report No. EPA 910/
9-82-089f, U.S. Environmental
Protection Agency, Seattle,
Washington, 1983.
6. Del Green Associates, Inc.,
Residential Wood Combustion Study
—Task 5—Emissions Testing of Wood
Stoves, Reports No. EPA 910/9-82-089g
and EPA 910/9-82-089h, U.S. Environ-
mental Protection Agency, Seattle,
Washington, 1982.
7. Gay, Dr. Robert L. and Green, William
T., Residential Wood Combustion
Study—Task 6—Control Strategy
Analysis, Report No. EPA 910/9-82-
089i, U.S. Environmental Protection
Agency, Seattle, Washington, 1982.
8. Core, John E. et al., Residential Wood
Combustion Study—Task 7—Indoor
Air Quality, Report No. EPA 910/9-82-
089J, U.S. Environmental Protection
Agency, Seattle, Washington, 1982.
29
------- |