EPA 910/9-82-089a
United States
Environmental Protection
Agency
Region 10
1200 Sixth Avenue
Seattle WA 98101
            Air & Waste Management Division February 1984
&EPA     Residential  Wood
            Combustion Study

            Task 1
            Ambient Air Quality  Impact Analysis
                            V^-i Pi tl H t=1JBU

-------
 RESIDENTIAL WOOD COMBUSTION STUDY




              TASK 1




AMBIENT AIR QUALITY IMPACT ANALYSIS

-------
 RESIDENTIAL WOOD COMBUSTION STUDY

              TASK 1

Ambient Air Quality Impact Analysis
           FINAL REPORT
           Prepared by:

           John E.  Core
        Dr. John A. Cooper
        Richard T.  DeCesar
        Dr. James E.  Houck
             NEA, INC.
         Beaver ton, Oregon
           Prepared for:

    DEL GREEN ASSOCIATES,  INC.
 ENVIRONMENTAL TECHNOLOGY  DIVISION
         Woodburn,  Oregon.

  Under Contract No.  68-02-3566

U.S. ENVIRONMENTAL  PROTECTION AGENCY
             Region X
     Seattle,  Washington   98101

           Task Manager
          Wayne Grotheer
          September,  1982

   U.S. Envtronmental PwUCtlon
   Region  V, Library
   230 South Dearttffll St?«H
   Chicago, ttUortL 60604

-------
THIS REPORT CONSISTS OF SEVERAL DIFFERENT PARTS.

THEY ARE LISTED BELOW FOR YOUR CONVENIENCE.


EPA 910/9-82-089a  Residential Wood Combustion Study
                   Task 1 - Ambient Air Quality Impact
                            Analysis

EPA 910/9-82-089b  Task 1 - Appendices

EPA 910/9-82-089c  Task 2A - Current & Projected Air Quality
                             Impacts

EPA 910/9-82-089d  Task 2B - Household Information Survey

EPA 910/9-82-089e  Task 3 - Wood Fuel Use Projection

EPA 910/9-82-089f  Task 4 - Technical Analysis of Wood Stoves

EPA 910/9-82-089g  Task 5 - Emissions Testing of Wood Stoves
                            Volumes 1 & 2

EPA 910/9-82-089h  Task 5 - Emissions Testing of Wood Stoves
                            Volumes 3 & 4  (Appendices)

EPA 910/9-82-089i  Task 6 - Control Strategy Analysis

EPA 910/9-82-089J  Task 7 - Indoor Air Quality

-------
                      DISCLAIMER




This report has been reviewed by Region 10,  U.  S.  Environmental




Protection Agency, and approved for publication.   Approval




does not signify that the contents necessarily reflect the




views and policies of the U. S. Environmental Protection Agency,




nor does mention of trade names or commercial products constitute




endorsement or recommendation for use.

-------
                           TABLE OF CONTENTS





                                                                  Page






     ABSTRACT	    ii



     EXECUTIVE SUMMARY   	   iii



     ACKNOWLEDGEMENTS 	    iv



     LIST OF TABLES	    v



     LIST OF FIGURES	    vi



1.0  INTRODUCTION  	     1



2.0  PROGRAM METHODOLOGY   	    2



     2.1  Air Sampling Program  	    2



     2.2  Analytical Protocol	•    6



          2.2.1  Gravimetric Analysis 	    6



          2.2.2  X-ray Fluorescence (XRF) 	    8



          2.2.3  Neutron Activation Analysis (NAA)   	    8



          2.2.4  Carbon Analysis	    10



          2.2.5  PNA Analysis   	    10



          2.2.6  Carbon-14 Analysis 	    10



     2.3  Quality Assurance 	



     2.4  Sample Selection  	    11



3.0  SOURCE APPORTIONMENT  	    12



     3.1  Chemical Mass Balance	    12



          3.1.1  Chemical Mass Balance Receptor Model 	    12



          3.1.2  Source Characterization  	    15



     3.2  Carbon-14	    16




          3.2.1  Fuel Ratio Measurements	    16



          3.2.2  Contemporary Carbon Source Activity  	    18
                                      •


          3.3.3  RWC Emission Carbon Content  	    18

-------
                                                                 Page







4.0  EXPERIMENTAL PROGRAM RESULTS	  .18




5.0  DISCUSSION OF RESULTS	    21




6.0  CONCLUSIONS	- .  .  .  .    25




7.0  REFERENCES	    27




8.0  APPENDICES




     1.  Air Monitoring Site Description




     2.  Source Emission Composition




     3.  Field Program Sample Summary




     4.  Source Impact Summaries by Site




     5.  CMB Source Apportionment Calculations




     6.  Quality Assurance Program Elements

-------
                              ABSTRACT
               A Study of Residential Wood Combustion
            Task 1:  -Ambient Air Quality Impact Analysis
     The ambient air quality impact of Residential Wood Combustion  (RWC)
emissions in eight Pacific Northwest cities was studied during  the  1980-
81 space heating season.  Seventy-seven worst case fine particle ambient
samples collected in Seattle, Spokane, Ya'kima, Tacoma and Longview,
Washington; Boise, Idaho; Portland and Medford, Oregon were selected
and analyzed for Polynuclear Aromatic Hydrocarbon compounds, 35 trace
elements and carbon.  The impact of RWC and other sources was determined
by Chemical Mass Balance calculation and validated against independent
carbon-14 measurements.
     This document describes the project design, methodology and results
of the RWC impact monitoring program.  The findings of this task form the
basis for projections of future RWC impacts discussed under task 2A of
this series.  They also provide the first comprehensive survey of RWC
impacts within Pacific Northwest communities during a period likely to
be typical oi maximum RWC emissions.

-------
                          EXECUTIVE  SUMMARY

     Recent increases in  the use of  wood as a source of  residential
space heating within Pacific Northwest states has  led  to  concerns
focusing on public health risks and  degrading air  quality within many
communities.  A comprehensive survey of eight cities in  Oregon, Washington
and Idaho was conducted during the October, 1980 - March, 1981  space
heating seasons to provide a better  assessment of  current maximum  24 hour
RWC impacts on particulate air quality.  Seventy-seven selected fine
particle samples from Seattle, Spokane, Tacoma, Longview and Yakima,
Washington; Boise, Idaho; Portland and Medford, Oregon were analyzed for
35 trace elements, carbon, PNA compounds and carbon-14.  RWC impacts were
then quantified by Chemical Mass Balance and validated by independent
carbon-14 measurements.
     Source apportionment results typically assign from 66% to 84% of the
fine particle mass (less than 2.5 urn) and 36-60% of the TSP mass measured
in residential land use areas to RWC emissions during selected, worst case
woodsmoke conditions.  Twenty-four hour, worst cast impacts exceeding 60
ug/m3 were measured at residential sites in Seattle, Spokane, Portland,
Medford and Boise.  A maximum impact of 128 pg/m3  (upper limit estimate)
was recorded at the Boise residential site.  Benzo(a)pyrene concentrations
as high as 22 ng/m3—a factor of ten higher than background—were
measured in Spokane.
     Results from the ambient monitoring program clearly document  the
importance of RWC emissions as Che most important contributor to fine
particle mass concentrations in many Northwest communities during worst
case winter particulate air quality conditions.
                                 iii

-------
                            LIST OF TABLES


Table                                                          Page


  1     RWC Ambient Sampling Network .............   3

  2     Program Analytical Design  ..............   7

  3     XRF Minimum Detectable Concentrations  ........   9

  4     Relative Wood Type Use
        Data Used to Form 1'*C Fuel Samples ..........  21

  5     Carbon-14 Analysis Results for Typical Wood Fuels  .  .  21

  6     CMB Estimates of RWC Impacts (October, 1980-March,
        1981)
        Fine Particle Fraction ................   4
  7     PNA Analysis Results for Selected RWC Samples  ....  26

  8     Average PNA Concentration for Residential Sites
        (ng/nr) (1980-81 Heating Season)  ...........  27

  9     Comparison of. RWC Impacts Based on CMB and
        Radiocarbon Measurements ...............  28

-------
                           LIST OF FIGURES
Figure                                                           Page
  1     Source Contributions, Fine Particle Fraction  	  25




  2     1'*C Validation of CMB Impact Estimates	29

-------
   1.0  INTRODUCTION

      The increasing use of wood (and coal) as a source of residential heat  is  a
 phenomenon common to many states in the Pacific Northwest.  For a variety of
 economic and political reasons, the public's interest in the use of wood as an
 important heating fuel has led to a resurgence in concerns focusing on wood smoke
 visibility reduction,  odors,  public health risks and the "livability" of many
 communities.
      The use of wood for residential heating, while aesthetically pleasing,
 carries a significant  potential for adverse health effects to large segments of
 the  population.  The impact of Residential Wood Combustion (RW.C) emissions is
 especially severe because plume impacts typically occur at ground level very near
 to the  source.   In addition,  the areas of highest RWC emission density often
 coincide with the areas of maximum population density and the majority of the RWC
 particulate  emissions  are within the size range deposited within the lungs.  RWC
 emissions are relatively rich in-carcinogenic organics,  toxic pollutants and
 respiratory  irritants.   For all of  these reasons,  wood  smoke
 represents  an important problem that is of growing public concern.
      The chemical products formed  in the combustion of  wood  have recently been
 shown to contain  17  priority  pollutants,  14 carcinogenic  compounds  and 6 toxic or
 mucus coagulating agents  which,  when considered in addition  to  toxic gaseous
 emissions and  respiratory  irritants,  collectively  represent  an  important health
 risk. :
      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  the adoption  of an  Inhalable  Particulate National  Ambient  Air
Quality  Standard  has also  caused concern  about  the  RWC impact on  24  hour standard
 attainment.  The  continuing economic pressures  to  expand  the  use  of  wood and coal
 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,  aesthetics and  the future "livability" of many communities.
     A multi-faceted study of  current and  projected air quality impacts,  fuel  use

-------
trends, the current status of wood stove emission testing, RWC control  strategy
alternatives and indoor exposure has been conducted in response  to  the  current
concern over the RWC problem.  This document reports on Task  1 in this  series;
the air quality impact of RWC emissions on ambient air quality.  This  report
focuses on the impact of particulate emissions, although carbon  monoxide,
aldehydes, phenols, volatile organic compounds and oxide of nitrogen emissions
are also of potential concern.
     The objective of Task 1 was to estimate the maximum 24 hour impact  of  RWC
emissions in Portland, Oregon, Seattle and Spokane, Washington during  the  1980-81
heating season.  Historical samples' collected in Longview and Yakima, Washington,
Boise, Idaho and Medford, Oregon were included to provide a broad-based  study
representative of Pacific Northwest communities.  Other objectives  of  the  program
were to (a) determine the ambient concentration of polynuclear aromatic
hydrocarbons (?NA) compounds associated with RWC emissions (b) validate  RWC
impacts calculated by Chemical Mass Balance methods with estimates  obtained by
Carbon-14 measurements and (c) develop a data base upon which future projections
of RWC impacts can be based given fuel use trend information.

2.0  PROGRAM METHODOLOGY

     The program design of the RWC Ambient Impact Analysis includes six  elements,
each of which are important to the program success.  Data analysis  methods
focusing on the Chemical Mass Balance (CKB) model and Carbon-14  techniques, the
air sampling program, sample selection, analytical protocol,  source
characterization and data interpretation methods.  Each -are described  in this
section.

2.1  Air Sampling Program

     The RWC air sampling program was established and operated to define the
ambient particulate impact of residential wood heating emissions in three
Northwest cities; Portland, Oregon, Spokane and Seattle, Washington.   Sampling
was conducted at rural, residential, commercial and industrial land use  areas  .
listed in Table 1.  The sampling network can be subdivided into  two groups:

-------
                                   Table  1




                         RWC Ambient  Sampling Network
Site //
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15





City Location
•
Spokane Country Homes
Spokane Turubull
Seattle Newport Way
Seattle Lake Samamish
Portland Marcus Whitman School
Portland Cams
Medford County Courthouse*
Seattle South Park (Dallas)*
Tacoma Fire Station 12*
Seattle Georgetown*
Yakima County Courthouse*
Longview City Shops*
Spokane County Health*
Spokane Crown Zellerbach*1
Boise Fairview & Liberty St.*
Notes:
i- Located about 1 mile from an
2- Cyclone Sampler (< 1 urn)
3. Integrating Nephalometer
* Historical Samples
+ Located in near proximity to
Land Use/
Site Type
Residential *
Background
Residential
Background
Residential
Background
Commercial
Residential
Industrial
Industrial
Commercial
Commercial
Commercial
Industrial
Commercial
Aluminum Reduction




Sampling Number of Monitoring Instrument
Period (1981) Samples HV SSI Sierra DiclioL
1/31-3/10
10/30/80-3/1/81
2/6-3/8
3/9/81
1/30-3/5
2/6-3/5
12/17/80- 1/16/81
1/15-1/16
1/4-2/9
10/18/80-2/3/81
11/17/80-2/21/81
12/11/80-1/16/81
10/31/80
1/17-3/11
12/9/80-2/23/81
Plant




15 x
5 x
10 x
1
8 x
6 x
1 x
2 x
4 x
3 x
4 x
4 x
1
4 x
9 x





X X
X X
X X
X X
X X
X X
X X
X
X
X
X
X
X
X
X X





Residential development

-------
 Sites 1-6 active during the October,  1980 co March,  1981 period and sites 7-15
 from which historical samples were selected.
      Sites active during the 1980-81  heating season  were operated either on a
 daily or every other day schedule  on  a 24-hour  basis.   All residential sites were
 selected on the basis of the apparent level  of  wood  burning activity as indicated
 by  visible plumes,  wood stove stacks  and  firewood  storage piles.-  Final siting
 was contingent upon availability of power, security  considerations,  sampler
 exposure and adherence to  established siting criteria  (40 CRF Part 58) .
 Characteristics of  specific sites  are noted  in  Appendix 1.
      Sampling was planned  to concentrate  on  weekends,  when greater use of wood
 stoves and fireplaces was  assumed  likely  to  occur.   Flexibility was,  however,
 provided to sample  continuously during  any air  stagnation episodes with cold
 temperatures.   Sampling was planned for the  Friday-Monday period,  plus a mid-week
 day,  provided weather conditions were sufficiently cold and stable.   Weather
 forecasts  were monitored and sampling decisions were made on a daily  basis.   The
 following  criteria  were adopted as guidelines in determining whether or not
 sampling would be conducted:
      1.   Predicted  minimum  temperature of less  than  40°F and preferably less than
          35°F
      2.   Daily average  wind  speeds less than  5 mph
      3.   Presence of  atmospheric inversions
 Background  site  samples  were  not planned  Co  correspond  to each sample  collected
 at  each  residential  site.   Instead, 3-6 background samples  were selected to
 represent  a  range of  background levels.  Most background  site  samples  did
 correspond  to  a  residential  site sample, however, providing  a  number  of days for
 each  city  for  which  CMS  analysis can  include  comparison  of  simultaneous in-town
 and background  samples.

      Sampling  Protocol

     The air sampling protocol was  intended  to capture particulate  samples  (TSP,
respirable and  inhalable) for subsequent chemical analysis,  to  estimate
residential wood combustion  (RWC)  impacts.  At each residential  and background
site, the following  instruments were operated for this study:

-------
      1.   A standard high volume sampler,  to measure total suspended particulate
          (TSP),  on glass fiber filter media  (hi-vol).
      2.   A standard high volume sampler equipped  with  a Sierra Model 235 High
          Volume Cascade Impactor  to collect "respirable" (< 2 microns in
          particle diameter) aerosols on glass  fiber filters (Sierra).
      3.   A dicholomous sampler to collect a "respirable" or fine fraction (<2.5 urn)
          and a coarse (2.5-15 ym) fraction on  teflon filters.
      4.   A size  selective inlet sampler designed  to collect inhalable  particles
          less than 15 urn on glass fiber filters was used at the Boise  site (SSI) .
Although samplers from EPA Region X, Che Oregon Department of Environmental
Quality and NEA, INC. were used in the network, most of the sampling
instrumentation required was already in regular use at che background monitoring
sites in the Portland area (at Carus) and Spokane area (at Turnbull) , as part of
the regular particulate monitoring network operated by DEQ and the Spokane County
Air Pollution Control Agency (SCAPCA),  respectively.  A new background site was
established temporarily in the Seattle area for this study.  The dichotomous
sampler deployed there was provided by the State of Washington Department of
Ecology (DOE) Laboratories.
     Field checks of instruments for Quality Assurance purposes were performed in
Portland by NEA,  in Seattle by DOE staff,  and in Spokane by SCAPCA and DOE
personnel.  Standard operating procedures were developed and provided to field
technicians.  Filters for all instruments were preweighed by NEA, logged and
distributed to field technicians in all three sites.  Glass fiber filters were
maintained in cold storage from soon after retrieval until final weighing and
analysis.   Following sampling, used filters were logged in and weighed by NEA
staff to determine sample mass, then converted to standard units.  Corrections
for fine particle deposition on the coarse mode filter were made to gravimetric
and elemental data.

-------
 2.2  Analytical  Protocol

     The  analytical program consisted of gravimetric analysis of aerosol
 mass, elemental  analysis by x-ray fluorescence, organic/elemental carbon  by
 thermal combustion, Polynuclear  Aromatic Hydrocarbons (PNA) by gas chroma-
 tography/mass  spectroscopy and Carbon-14 by low level B particle counting.
 Each of these  methods is described below.  The sampling and analytical
 techniques used  were designed to provide aerosol chemistry information needed
 to identify RWC  impacts by Chemical Mass Balance and Carbon-14 source
 apportionment  methods discussed in sections 2.4 and 2.5, respectively.  Table 2
 summarizes the sampling and analytical design used as well as the intended
 purpose of the data collection.

 2.2.1  Gravimetric Analysis

       Hi-vol  Sierra sampler filters used at sites 1-6 were preweighed,
 conditioned and  final weights determined at NEA,  INC.  using a Torval balance.
 Replicate filter weights were routinely conducted on 12% of the filters to
verify that weights were maintained with ± 10% relative precision.   Dichotomous
 filters were weighed on a Cann Model 27 electro balance with replicate
weighings of 12% of che filters co also insure that weights were held to
within t  10% relative precision.   Filter handling,  logging, shipping and
calculations were conducted in accordance with NEA standard operating procedures,
Because of the potential loss  of  volatile organic hydrocarbons,  all hi-vol
and Sierra samples were shipped and stored  under  refrigeration.  .

-------
                                               Table 2

                                      Program Analytical  Design
 Sampling Equipment
Sample Provided
                                               Analytical
                                               Method*
                                        Provides Data Input to
             CMir
Carbon-1A
PNAD
NAAQS
 Sierra 235 Impactor
Fine Fraction -
Glass Substrate
                     Combustion
 Dichotomous Impactor
 Ili-vol Sampler
 Size Selective Inlet
Fine Fraction -
Teflon Substrate

TSP Fraction -
Glass Fiber

IP Fraction -
Glass Fiber
XUF,
NAA6
C-142
GC/MS

GC/MS3
C-14
* Gravimetric analysis completed on all  samples
  Combustion oxidation for Elementa I/Organic  carbon
2 Carbon-1A
J PNA compounds by Gas Chromatography-Mass  spectrometry
'' Chemical  Mass Balance
5 Polynuclear AromaLic Hydrocarbons
6 X-ray Fluorescence/Neutron Activation  Analysis

-------
2.2.2  X-rav Fluorescence  (XRF)
     Fine particle samples collected on dichotomous filters were  analyzed  for
trace elements using x-ray fluorescence (XRF).  The technique  is  based  on  the
fact that atoms within  the sample are excited  from their ground state  to higher
energy levels by x-radiation from an x-ray tube.  The excited  atoms  emit discrete
energy x-rays as they return to their normal  energy level.  The energy  of  the
emitted x-rays is characteristic of the emitting element and is used to
qualitatively identify  the element while  the  number of x-rays  detected  is
proportional to the number of atoms present and provides a quantitative measure
of a specific element's concentration through  a direct comparison with  standards.
Thirty-three elements are commonly measured by this technique, provided their
minimum detectable concentrations are greater  than those noted in Table 3.
Intermethod comparison  of XRF and neutron activation analysis  for about 7
elements is conducted routinely.  A detailed  description of potential
interferences, calibration calculations and quality assurance  procedures is
documented in NEA standard operating procedures (Appendix 6) .

2.2.3  Neutron Activation Analysis  (NAA)

     NAA is used (a) to measure the concentration of several elements  (Na, Mg)
for which quantitative  measurements cannot be  obtained by XRF  analysis  and  (b)  to
provide intermethod quality assurance comparisons for 7 other  elements  measured
by both XRF and NAA.  The technique requires  that the sample be placed  in  a
neutron flux from a nuclear reactor.  Elements within the specimen  absorb
neutrons and undergo nuclear transmutations forming radioactive elements which,
in turn, emit radiations characteristic of the specific radionuclides.  Since  the
number of these radioactive elements formed are proportional to the  number  of
original stable element atoms -present, measurement of the number  of  specific
radionuclides radiation is an indirect quantitative measure of the  stable  element
concentrations.  Counting was done with 30% and 17% ORTEC HpGe gamma-ray
spectrometers.

-------
                                    Table  3

                             X-RAY FLUORESCENCE
                   MINIMUM DETECTABLE CONCENTRATIONS*
                               STANDARD  ANALYSES-i-
Element
AI
Si
P
s
Cl
s
C*
Ti
V
Cr
«n
Fe
Hi
Cu
Zn
Ga
A3

Se
3r
Rb
Sr
1
Zr
Pd
AS
Cd
In
Sn
Sb
Sa
La.
Kg
?b
I
10
3
10
30
20
6
6
2
1
1
1
2
1
1
1
1
1
1
i_
2
4
4
3
30
3
12
14
13
13
15 ;
100 |
100
4
5
Three sigiaa incerrerence-free MDC  for particulaces deuosiced on filters  «ich 1 jig/ca-
densities sucn as Nuclepore and  some cellulose and :eflon  filters XDC '-  'ciae)"*"-
(filter aensicy)**-
Tor scretcned teflon  filters
Custom analysis for specific elements at special >35C's available upon request.

-------
2.2.4  Carbon Analysis

       Organic and elemental  carbon  (OC/EC) analysis of a sample  aliquot  is
conducted by heating  samples  collected on  glass or quartz fiber  in  an  oxygen-
free atmosphere at a  temperature of  600CC  to oxidize organic carbon to C02
which, in turn, is reduced  to CHi,  in a Mn02 bed heated to 900°C.  The  methane
is then measured by a flame ionization detector (FID).  Elemental carbon  is
then measured by introducing 0% into the sample oven and combusting remaining
carbon to C02 which is also reduced  to methane and measured by FID.  Artifact
EC formed from pyrolysis of OC is  detected  (and corrections made) through the
use of an optical filter reflectivity system.

2.2.5  PNA Analysis

       Selected Sierra and hi-vol  filters were submitted for PNA  analysis by
gas chromatography/mass spectroscopy.  Each sample was extracted  for 8 hours
with dichloromethane followed by removal of polar compounds on an alumina
column and solvent evaporation.  Internal standards were added to each sample
prior to GC/MS performed as outlined elsewhere.3

2.2.6  Carbon-14 Analysis

       Sample analysis for Carbon-1.4 consists of mounting the particle filter
directly on the inner wall of a combustion tube and converting the  sample
carbon to high purity C02 by combustion in an oxygen stream over  CuO and
Ag-wool at 800°C and 400°C, respectively.  Final purification of  the sample
COa is accomplished at low temperature (-78°C)  distillation and exposure  to
Cu-foil at 800°C.  Sample mass is  then determined by standard gas volumetric
techniques.
     '  The L1*C content is determined by directly counting the C02 at 1-4  atm
in a miniature low-level gas proportional counter using massive shielding,
meson cancellation and pure counter  construction materials.  The  1UC decay
rate is based on 8  particle counting.
                                   10

-------
2.3  Quality Assurance
     The project quality assurance program consisted of four elements each of
which are briefly described below:
     2.3.1  Ambient Monitor Siting
          Each of the ambient monitor sites selected for this study was
     reviewed with respect to EPA siting criteria described in 40 CFR 58.
     Each of the sites is described in Appendix 1.
     2.3.2  Analytical Program
          A brief description of quality assurance elements associated
     with the analytical program are described in section 2.2 and in greater
     detail in Appendix 6.   Since elemental analysis is central to the CMB
     source apportionment method, the XRF elemental analysis quality assurance
     plan is briefly discussed below.
          The XRF QA program includes all aspects of the ambient sampling
     and filter handling protocols.   Special  precautions were taken in the
     field and in the laboratory to  avoid sample contamination,  loss of
     particles during filter shipment and potential for filter  identification
     problems.   Clean room  techniques,  for example,  were used during loading
     (and unloading)  filters from their holders  and specially designed sample
     shipment containers were  used to minimize the potential for particle
     losses  due to  sample tipping.   During XRF analysis, quality control
     standards  and  blank samples were analyzed for each group of 10 samples
     analyzed and results were compared  to known standard values to maintain
     precision  to within ±  2%.   Sample  groups  outside  of this  range were
     reanalyzed.  Analytical results  for  several elements  (K,  Ca,  Fe,  As,
     Br and Pb)  were  intercompared  between excitation conditions  and calibration
     standards  have  been verified by  extensive interlaboratory  and intermethod
     analysis.
          Similar procedures have  been  followed  with respect to  organic  and
     elemental  carbon analysis results to  insure that  analytical precision and
     accuracy is  held to within  i  15%  relative to  the  sample concentration.
     Both internal  quality  control  (replicate  analysis  of a  "standard" filter)
     and  external standards  (filters  containing  known  amounts of  carbon)  were
     routinely  analyzed.
                                 11

-------
      Gravimetric  analysis  of  the ambient samples  included  replicate
analysis  (12%   audit  level) of  the  filters  to maintain weighing pre-
cision and  consistency of  tare  and  balance  calibration checks  during
each  weighing  set  to  sithin ± 0.004 mg of true value.
2.3.3  Field Monitoring Program
      Standard  operating procedures  were developed for Hi-vol and dichotomous
sampler calibration, maintenance, filter handling and flow  audits to insure
consistency within the data set.  Each procedure is described  in Appendix
6.  In addition, notes were maintained to reflect meteorological
circumstances  in the near vicinity of the sampling site.  These notes
are included in Appendix 3.
2.3.4  Data Management
     Two levels of data validation QA were maintained; (a)  verification
of the transcribed data against the source documents and (b) examination
and resolution, where possible,  of inconsistencies in the data  sets.
For example, bivariate plots of elemental measurements (Pb/Br,  Al/Si,
etc) were made prior to CMB analysis of the data to identify and
validate outliers.
                             12

-------
 2.4  Sample Selection


      Samples submitted for chemical analysis were selected on the basis of (a)

 fine fraction (< 2.5 ym)  mass loading, (b) light scattering (Bscat) average

 where available, (c) relative meteorological stability (wind speed and air

 stagnation advisories) and (d)  daily minimum temperatures.

      The above criteria were used to estimate the probable relative RWC impacts

 represented by sample filters.   As noted earlier, filters likely to reflect

 maximum impacts  were selected.   Other criteria which influenced selection and

 ranking of filters  recommended  for analysis included:
      1.   Completeness of  data for best CMS analysis
          Ideally, respirable particulate samples for both inorganic and carbon
          analysis are desirable at both the urban monitoring site,  and a suit-
          able background  site during the same 24 hour  sampling  period.  However,
          this ideal  situation devotes considerable resources to analysis of
          background  samples,  and was often not possible,  due to lack of (a)
          respirable  particle size samples,  (b)  background samples taken on same
          day (c) pairs of  samples suitable for organic and inorganic analysis.

      2.   Available Background Samples
          Analysis of enough  background samples to describe a range  of  potential
          background  RWC impact  on days of  expected high RWC impact  at  urban
          sites.

      3.   Preference  for Analysis of  Respirable Particulate Samples
          RWC emissions are assumed to be  almost entirely  in the respirable size
          range (_<_ 2.5  microns).   Accordingly,  preference  was given  to  inorganic
          analysis of dichotomous sampler  fine fraction filters  and  companion
          carbon  analysis of  Sierra filters  (<  2 urn).   If,  however,  a SSI  sample
          (<  15 urn) was used  for  carbon analysis,  companion elemental analysis
          was completed on both  dichotomous  fine and  coarse filters  and summed
          together  to obtain comparable  sets of data  for an inhalable size  fraction.
      4.   Consistency of Mass  Loadings  on Co-located  Samplers
          In  some cases, (Boise)  samples with  possible  high RWC  impacts  were not
          selected for  analysis because  of  apparent discrepancies  in  mass
          loadings on filters  from co-located  samplers,  e.g.,  (a)  too large a
          difference  in dichotomous fine +  coarse  mass  and  hi-vol  (SSI)  mass,  or;
          (b)  standard  hi-vol TSP  mass  loadings  significantly lower  than
          respirable  or inhalable  samples taken  during  the  same  24 hour  period at
          the  same site.

      5.   Availability  of Recommended  Filters  for  Analysis
          Availability  of both the  urban and background  samples  for analysis was
          a prerequisite to sample  selection.


     The  process of   selecting filters  for analysis was  proceeded by  first

assembling available data on the mass  loadings of respirable, inhalable and

total suspended particulate  (TS?)  samples collected during the October,  1980-

March, 1981 period.   Since dichotomous samplers were  operating in a number  of


                                   13

-------
cities in Oregon, Washington and Idaho, it was feasible to concentrate on
respirable particle size samples in the following cities:
                         Washington
                         Seattle
                         Spokane
                         Yakima
                         Longview
Idaho
Boise
Lack of dichotomous samplers, and other considerations, ruled out selection
of filters from some cities of potential interest (i.e., Olympia, Washington).
Second, samples associated with higher 24 hour average Bscat values were
assumed to correspond to higher RWC impacts.  Peak Bscat values which occurred
in the evening and early morning hours also were assumed to indicate potential
RWC activity.  The third consideration involved selecting samples associated
with maximum atmospheric stability and cold weather, as represented by low
surface wind speeds and low daily minimum temperatures, respectively.  The
fourth and primary criteria for selection were the fine particle mass
concentrations.
     Eighteen Sierra and SSI -samples from five Northwest cities were selected
for PAN analysis to (a) indicate potential human exposure to hazardous
chemicals and (b) to determine if  PAN concentrations were correlated with RWC
impacts determined by CMS and Carbon-14 analysis.  In addition,  13 samples were
selected for Carbon-14 analysis to provide a second, independent estimate of
RWC impact. Both sets of samples were selected in accordance with the criteria
described above.

3.0  SOURCE APPORTIONMENT

     Identification of RWC particulate impacts requires the application of
methods capable of apportioning aerosol mass among possible sources.  Chemical
receptor models have been selected for use in this study because these methods
do not rely on emission inventory or meteorological data as a basis  ;or
                                   14

-------
 quantification of source impacts.   Two independent,  yet complementary
 receptor model approaches have been included in this program design.   Each
 are  discussed in the following sections.

 3.1   Chemical Mass Balance

      The chemical mass balance model matches source  particle size and chemical
 "fingerprints" to those measured at the receptor  to  back-calculate the impact
 of specific  source classes of  similar chemical composition.  Given
 data  on  ambient concentrations of  several  chemical species  and  the percent by
 weight of those species in the source emissions,  a set  of equations is
 prepared and  solved to determine source impacts.

 3.1.1  Chemical Mass  Balance  Receptor Model

       The starting point for the  Chemical mass balance model is  essentially
 the same as the source (dispersion)  model.   The source  model states that  the
 contribution  of a source to a receptor is  dependent  upon the emission rate
 and a dispersion factor.   Limitations of dispersion  modeling, however,  arise
 from difficulty in quantifying dispersion  factors in areas  of complex terrain
 and/or meteorology,  from difficulties in identifying emission rates,  and  other
 operating parameters  associated with emission sources.   The  advantage of  the
 CMS, and other receptor approaches,  is that  an exact knowledge  of  these
 parameters is  unnecessary.
     The basic assumptions inherent  in the CMS approach are  the same  as those
associated with other  receptor and  source  (dispersion)  models:

     The particulate  mass  measured  at  the receptor is a direct, linear
 sum of the mass contribution  from a  number of specific  sources
                    P
                m = £   M.           .                (1)
                    j-1  J

where
      m  =  measured  concentrations of  particles from  all  "p"  sources  (ug/nT)
      M.   = the  impact  at  the  receptor of source "j"  (yg/m")
                                   15

-------
 Similarly,  the concentration of a given chemical species is assumed to be the
 sum of the  contributions from each individual source emitting that species.
           C'.  = E    F. . '  X..  M.                    (2)
             >•    .   13    ^   3
where
             C'.  =  measured  concentration  of  species  "i"  from  all  sources
                   "p"  (yg/m3)
             F..  =  the  fraction of  species  "i"  in  the emissions  of  source
                   "j"  as measured  at  the  source
             X..  =  the  coefficient  of  fractionation of species "i"  in  the
                   emissions of source "j".   Fractionation  is  a  systematic
                   change in relative  species abundances  as  the  aerosol
                   ages due, for example,  to  gravitational  settling  or
                   condensation of  volatile organic hydrocarbons
             M.   -  the  impact at the receptor of source "j"  (yg/m3)

     The F..' values associated with major emission  sources are measured or
adapted from literature values, and the C. data are  obtained  through field
monitoring programs.  One can ensure a fractionation coefficient which approaches
unity by (a) using stack, gas dilution techniques to  ensure capture  of condensible
                                                                              •
aerosols and (b) obtaining size-resolved source composition and ambient aerosol
measurements.  By  conducting fine and coarse mode CMS calculations, source
composition  (F..')  errors due to gravitational settling of large particles can
be minimized.  If no further fractionation or source emission composition
changes are  assumed to occur,, then X.. is assigned a value of unity and F..'
can be considered  to remain unchanged in time and distance downwind when
measured at the receptor (F..' = F..).
                                   16

-------
      By dividing both sides of equation (2)  by the particulate mass
 collected at the receptor,  all of the variables in the equation are
 expressed as a ratio to the aerosol mass:
                p
           C.  - Z   F..  S.  (i=l to n)               (3)
            1   1 = 1   13   J
      where
           C.  = concentration of  species "i"  expressed  as a ratio to
                                     P
                mass  (i.e.,  C± -  C'±/l  M.)

          F..  " the  fraction of species  "i"  in  the  emissions from source
                "j"  at the receptor,  expressed  as ratio to mass
           S.  = the  impact at the  receptor of source  "j" expressed as a
                ratio  to  mass emitted  from all  sources  impacting the
                                        P
                receptor(i.e.,  S  .  = M./Z   M.)

            n  = the number of chemical species  considered
            p  = the number of sources  impacting  the receptor.
     Since  S.  is  the quantity  to  be determined, a  set  of simultaneous equations
can be written for all  "p"  sources of species  "i", where "i"  is defined as  a
"fitting" element.  Once the set  of equations  are  formed,  the  value of S. can
be solved by  standard matrix manipulation techniques using  linear least squares
or effective  variance solutions utilizing commonly available matrix mathematics.
     The effective variance  least  squares fitting method used  by  NEA,  INC.  in
this study  weights the  solution in accordance with uncertainties  in both  the
source composition  (F..) and the  ambient aerosol species (C.).  The procedure
ensures that  those components  with the  greatest uncertainty receive less  weight
in the fit  and  that both sources  of uncertainty are propagated  through to the
source impact  (S.) determination.2
     Uncertainties in the source  impact estimates are  calculated  by an error
propagation algorithm incorporating (a)  the imprecision  in  the  ambient aerosol
chemical species  (C.) analysis and (b)  the variability  in the  source  composition
species (F..) at  the one sigma level.
                                   17

-------
3.1.2  Source Characterization





     The data sets referred to below and included in the Appendix have



been developed by comprehensive analysis of representative source test



samples, or adapted from the literature cited.  In all cases, source



emissions have been collected using size segregated (fine, coarse) sampling



systems.   Hogged fuel boiler and residential wood combustion samples were



taken with a size segregating source testing procedure for aerosols.  Following



sample collection, the filters were analyzed for organic and elemental


                       -2           -   -
carbon content,  ions (SO,  ,  NO  ,  Cl ,  p )  and trace metals.  Source emissions



are then characterized in terms of weight percent.   Further detail on the



sampling protocol may be  found in the references cited.   The source signature



described in this section forms an essential element of  the CMB source



apportionment analysis.   References to  the  origin of the emission chemistry



used are noted below.
                                     18

-------
Mneumonic

RDUST
MARIS


CDUST


KRAFT
    Source      Znission  Chemistry Origin                    Reference
    Geologic    Local  soil  samples were collecced in the
               vicinity  of the  Boise sice (5  samples),
               Spokane-Country  Hones (3  samples) and
               Seattle-Newport  Hay (4 sices) .   Following
               laboratory  processing, resuspension and
               preliminary XRF  analysis,  composite soil
               chemistries vere formed for  comprehensive
               XRT  analysis.  City specific soil chemistry
               was  Chen  used  in association with samples
               from each community.   Although  Che uneumoaic
               RDUST  is  used  in Che  CMS  calculations,  Che
               soil chemistry used cannot distinguish
               between native soils,  paved  road dust,
               unpaved road dust or  other sources of
               similar aerosol  chemistry.
               .Medford Aerosol  Characterization Scudy            i
                         Medford Aerosol  Characterization  Scudy

                         Bedford Aerosol  Characterization  Scudy


                         Portland Aerosol Characterization Scudy

                         Portland Aerosol Characterization Scudy
              Based on excess sulfur calculated as sulfate
              Portland Aerosol Characterization Scudy
              Based on excess zinc
operacions (or
other zinc
rich source)
Marine        Portland Aerosol Characterization Scudy
Aerosol
Continental   Portland Aerosol Characterization Scudy
Ousc
SESWD


TRAKS

HOCFU
ALUMR

SDOIL


SECSO
GALVN
Residential
'flood
Comouscion
Transoor-
tation
Hog cuel
Boiler
Aluminum
Reduction
Residual
Oil
Combustion
Secondary
Sulfate
Galvanizing
           Kraft
           Recovery
           Furnace
              Portland Aerosol Characterization Study
                                   19

-------
3.2  Carbon-14

     Measurements of radioisotope carbon-14/carbon-12-ratios have  been  used  by
Cooper and Curie to distinguish between "modern" and fossil fuel-derived
carbonaceous aerosol, providing an important new method of RWC  impact assessment.
The method's ability to resolve these two classes of sources is based on  the fact
that the isotopic ratio of lllC/12C in plant and animal tissue is in  equilibrium
with the same ratio in atmospheric COa, whereas fossil carbon contains
essentially no radiocarbon (1J*C) because of its old age compared to  the half
life of the  1I4C isotope (5730 years).  Consequently, carbon 14  is  a  unique
tracer of carbonaceous aerosol from contemporary carbon sources (such as  wood
combustion) and can be used to apportion the carbon content of  the aerosol.
     Application to RWC impact analysis requires knowledge of (a)  ll*C/12C
ratio in the fuel being burned, (b) contemporary carbon source  activities
impacting the receptor (i.e., structural fires, forest fires, open burning of
residential debris, incineration)  and (c)  the carbon content of RWC  emissions.
Samples submitted for analysis must be size fractionated to minimize inter-
ference from large particle contemporary carbon sources, such as.leaf fragments.
Each of the three requirements are addressed below:

3.2.1  Fuel Ratio Measurements

       Interpretation of ambient carbon 1UC/12C ratios requires a  knowledge
of the isotope ratio in the fuel burned in the community.  If the  ratio is
significantly greater than 100%, corrections in the assumptions are  required.
Four samples of typical wood fuel burned in Northwest communities  were
analyzed by the ll(C/12C method.  Fuel samples were formed following  completion  of
a telephone survey of wood suppliers, U.S. Forest Service and university
extension agents in Boise, Seattle, Portland and Spokane.  Representative logs  of
each of the species were obtained from each city and composite  source samples
formed based on their relative use shown in Table 4.  Cuts were made
perpendicular to the concentric growth rings to insure that sawdust  samples
were integrated in correct proportion to the age of the "fixed" carbon  contained
in the logs.

-------
                                 Table 4

                         Relative Wood Type Use
                   Data Used to Form 1UC Fuel Samples
      City
                                                            Percent  in
                                Species                  Composite Sample
 Seattle                         Alder                           70%
                                 Douglas  Fir                      30%

 Portland                        Alder                           20%
                                 Douglas  Fir      .                60%
                                 Hemlock                          20%

 Spokane                         Douglas  Fir                      30%
                                 Western  Larch                    40%
                                 Ponderosa Pine                   15%
                                 Lodge Pole Pine                  15%

 Boise                            Douglas  Fir                      40%
                                 Ponderosa Pine                   30%
                                 Lodge Pole Pine                  30%


     The results of  the  1  C  analysis  of  these  samples are listed below in
terms of percent of  modern carbon and are  in very close agreement to previously
                                                                                  7
published values (109%)  measured relative  to international radiocarbon standard.

The Seattle result  (133  ±  2%)  is substantially greater than the mean of the other

three samples  (111 ± 4.6).  No explanation for this discrepancy was  found.


                                Table 5

                      Carbon-14 Analysis Results
                        For Typical Wood Fuels

              Airshed                % Modern Carbon

              Portland     .               115 t 1*
              Boise                       106 ± 1
              Spokane                     112 ± 1

              Seattle                     133 ± 2
                                   Mean    116.5 ±  11.6%

* Analytical errors exclusive of  sampling uncertainty
                                   21

-------
     As noted  in  section  1.0,  the objective of this program was  to  quantify
the maximum  24 hour  impact of  RWC emissions in Pacific Northwest  communities
during the 1980-81 heating season.  Results presented in  this document,
however, largely  represent air quality impacts measured during the  month
of February, 1981—a period during which the average heating load demand  was
less than that which typically occurs during the period of peak  residential
space heating  demand (November, December and January).  Data reported herein
therefore represents high impact days selected from samples collected during
a period of  moderate weather  conditions.  Higher impacts may occur  during
periods of greater space heating demand and poorer atmospheric dispersion.
     Program results for Medford, Seattle  (South Park and Georgetown),
Spokane (County Health) and other sites where 4 or less samples were analyzed,
cannot provide a  comprehensive understanding of the actual magnitude of RWC
emission impacts  at  these sites.  Instead, results for these communities
should be viewed  as preliminary estimates of likely worst case impacts
representative of the sampling period.
     Table 6 summarizes, by site, CMS estimates of RWC air quality  impacts
measured (a)  during the field monitoring program and (b) from selected
historical samples.  Figure 1 illustrates average source contributions to the
fine particle mass in Spokane, Seattle and Portland, as well as inhalable
mass fraction in Boise.  Table 7 includes results of Polynuclear Aromatic
Hydrocarbon  (PNA) analysis for selected samples from five Northwest  cities,
and Table 8 lists average TSP, fine mass and RWC impacts for concurrent
samples (only)  by site.  Results of the carbon-14 analysis are presented  in
Table 9 and Figure 2.

5.0  DISCUSSION OF RESULTS

     The Chemical Mass Balance results reported herein are often associated
with relatively high levels of uncertainty based on (a) imprecision  in
analytical measurement of ambient samples and (b) uncertainty in the RWC
emission chemistry reported in Appendix 2.   Since the RWC source composition
values in this  analysis were developed during the Medford Aerosol Character-
ization Study,  the impact estimates reported assume a fireplace/wood stove
ratio  similar to Medford (0.63).
                                   23

-------
                                       Table 6

                             CMB Estimates of HWC Impacts
                              (October. 1980-Marcli, \98l)
                                 Flue Particle Fraction

g,.i,l,,aio,.
Si>ukuiic

bt*uk.«ii«*

ii^tok U IIV *

b u U 1 I \ «
Scull U*
Su.ilt le«
TtM I.H..I

1 ilk l.i.a"
l.uiim/leu'
1 Jjl.l.
du 1 »•-••"

Ur«n"".
I'Ot 1 t dllj
Hcil lord*
Sice
Ho.

1

13

14

3
8
10
9

1 1
12

15


5
7
No. of
S a mp lea

15

1

d

10
2
3
<,

t,
i.

y


8
!•
Averagi HWC
Purctnt of
Site Untie' Vine Ha*t

Country
llouea (R)
County
llaalth (C)
Crown
Zellerbach (1)
Newport Way (»)
South Park (R)
Georgetown (I)
Fire
Station 112 (I)
Courthouse (C)
City Shops (C)

Folrvlew 4
Liberty St. (C)


81

64

45

8}
65
73
74

93
61

69


Whitman Sclioo 1(R) 6B
Courthouse (R)


.0

.7

.3

.B
.8
.4
.9

. 1
.4

.5


.5
-
Average
Flue Hitm

55.

53.

37.

36.
-
39.
47.

53.
4 1 .

121,


40


0

1

0

1

7
0

8
,8

,8


.0

t <|||>.|C t

44.5 i II .1

j<. J t 21 . J

16.7 t 14.9

30.2 i U.t.
-
29 . 1 : 20. 2
35.2 i 25. U

50.0 i 34. y
25.6 t 1 7. «

64 . 7 1 22. J1


il .'< i 10.3
-
.'•, II. Mil
H.ixlii.iim

6«. 1 1 10.0



19.1 1 1 1 . J

', I) li : 1 J . a
0(1.2 •- 19.9
J5 5 : 24 3
'. <. . i : 3 J . 2

j i . 1 : il .1,
-,(> •> '. 1:. 1

U'7 9 ! <"J.9' '



-------
                                                      Figure 1
                                             Average Ik  Hour Worst  Case
                                                Source Contributions
                                              -Fine Particle Fraction-
                                             (October,  1980-March,  1981)
RESIDENTIAL WOOD COHDU3T10H
                               ALUMINUM PROC. (0.4X)
                                RESIDUAL OIL <0.03X>
                                 TRAHBPORTATIOH <3X>

                                 GEOLOGICAL 

                                 SECONDARY 8ULFATE t3.6X>
                               UNEXPLAINED (9 .7K>
   SPOKANE COUHTHY  HOMES 3ITE I
                                    *E310£NT.AL  WOOO CUnUUST.O^  RESJOUftL  01L  <„_„,


                                                                     TRANSPORTATION (5  BX. )



                                                                     8ECOIIOAKY SULFATE  ( 7*. )


                                                                   MARINE < i . i>. >

                                                                  GEOLOGICAL ( I . y>. y
                                                                             lLE  lll;M|'l)K1  UfiY

                                                                             (Ui:l 1 evuo )
KE6IOEMTIAL  WOOD COKQUSTIOH
     (GBX)
RESJDUftL OIL <0.6X>
HAR1HE (0.6K)
 GEOLOGICAL ( I .8>:>

TRANSPORTATION (H. IX >

8ECOHO(\RY 8ULFATE  («  2X )


 UHEXPLAIMEO < I 8 . 7 >i )
RESIDENT \M.
UOOO
COrtOUSTI OH
 ( 70>. )
                                                                        \
                                                                                                             < H
                                                                                                     TRAHSPORTftI I Oil (3  ?/: )

                                                                                                     SECOHOARY SULFnlt  ( .!  I

                                                                                                     MARINE  (0.0^.)


                                                                                                    IJIIEXPLAIHCO (0  (,'<)
      PORTLAND  MAHCU9  WIUTHAH
                                                                       I) 0 I 'J t  I II I W >.' I I l<
                                                                                         I I 11 i k I

-------
                                                                 Table 7

                                                          PNA Analysis Results
                                                        For Selected RUC Samples
                                                                  (ng/m1)
City
Seiilt le
Se.ilt In
Seat I It
Sesit L le
SJP.JI 1 In
Sual 1 le
Spokane
Spok.me
Spokdite
Spok.me
Port 1 .in
Purl l,n,
I'ort l.iii
F'orl Ifin
llolse
llolse
Medlonl
Hcdtord

Sltet
Newport Way (R)
Newport Way (H)
1 ' Newport Way (R)
South I'ark (It)
South Park (R)
1 dke Samamlsh (fl)
Country Home (R)
Country Home (R)
Country Home (R)
Tvirnbull (B)
l Marcus Whitman Sell (R)
.1 Carus (B)
t'alrvlew & Liberty (C)
Falrvlew & Liberty (C)
Courthouse (C)
Courthouse (C)
* Sleira Model 235 linpactor
** Size Selective Inlet
Date
2/6
2/7
2/8
1/15
1/16
J/9
2/1
2/6
3/1
2/6
1/31
•ill
3/5
2/6
1/8
12/12
12/17
1/16

Minimum Detectlble Concent rat Ions
1 Slli
/ [
; key
i
Sampler
S*
S
S
111 Vol
111 Vol
S
S
S
S
S
S
S
S
S
SSI**
SSI
S
S

benz(a)
anthracene
8.9
5.2
0.6
48.2
7.B
--
38.6
50.7
8.6
0. 1
13.6
43.9
14.3
0.4
1 .3
0.9
27.6
22.5

: f luoranthene, pyre
benzof luoranthenes
benzo(ghl)perylene
benzo-
f luoranthenes
3.0
9.6
0.3
13.6
1.8
0.8
18. 1
6.0
13.2
0.3
3.2
1 1.6
5.4
1.5
3.6
1.6
12.4
8.3

ne, benz(a)anthrace
t benzo(a) pyrene
, dlbenzanthracene
benzo(a) fluor-
pyrene anthene
1.8
4.2
0.3
6.5
1.0
0.3
11.1
3.6
3.7
--
1.5
5.4
2.7
0.5
2.3
0.8
8.5
8.3

ne. )
1
3
0
11
3
0
42
39
5
0
2
23
2
0
0
0
12
10

0.05 ng/m
0.1 ng/m
.9
.0
.5
.9
.2
. 1
.2
.0
.4
.3
.8
.6
.9
.4
.7
.3
.4
.4

i
i
dlbenz-
pyrene anthracenes
2.
5.
1 .
16.
4.
0.
69.
45.
9.
0.
3.
27.
3.
0.
0.
0.
15.
9.


9
3 0.7
0
2 0.9
7
1
9 1 .0
0 0.2
3 1.4
4
5
9 0.2
9
6
8
3
9
0


benzo(ghl)
perylene
0.9
3.3
0.2
7.1
1.2
0.3
4.5
1 .8
6.6
—
1 .3
2.9
2.1
0.7
1.5
2.3
3.4
3.1


    (It) Residential   (1)  Industrial
I  l.or.i led In llellevue. Wash Ing ton

-------
                                                 Table 8
City/Site*
No. of
Obs.
                                       TSP,  Fine Particle and RWC
                                             Impact Averages
                                       (Concurrent Samples, Only)
    TSP Mass
Mean ± Std.  Dev.
     (lag/m3)
    Fine Mass
Mean ± Std. Dev.
     (pg/m3)
	RWC  Impact	
Mean i  Std. Dev.      %  of  TSP
     (ug/m3)            Mass
Spokane

  Country Home(R)     13
  Crown Zellerbach(I)  3
            118.9 ±  50.9
            232.0 ± 156.9
                       53.0 ± 17.7
                       16.4 ±  0.5
                       44.8 ± 14.9
                       37.6 ±  2.4
                        36.5
                        11.1
Seattle

  Newport Way(R)
             71.7 ±  22.8
                       33.8 ± 10.5
                       29.4 ± 11.0
                                                                            42.0
Tacoma
  Fire Station 12(1)   4        159.0 ±  56.7         47.0 t  5.5

Portland

  Whitman School(R)    8         62.1 ±  19.7         43.3 ± 15.1
                                                       35.2 t  9.7
                                                       27.3 ± 10.3
                                                                 29.7
                                                                 43.9
Yakima

  Courthouse(C)        3
            221.6 ±  87.1
                       55.3 ± 12.4
                       50.6 ±  4.0
                                                                            25.6
Boise
  Fairview & Liberty(C)9        148.7 ±  59.3        121.6 ± 43.5**        84.6 ± 26.9*-'
                                                                            60.4
*  Other city  sites have less than 3 concurrent observations
** Inhalable particulate mass (< 15pm)
C = Commerical       I = Industrial      R = Residential

-------
                                                                Table 9

                                                  Comparison of KWC Impact Estimates
                                               Based on CMB and Radiocarbon Measurements
City/Site
Spokane
Count ry
Homes


Seattle
Newport
U.iy


South
Park

Portland
Marcus
Whi tman


Boise
Date


3-1-81
2-6-81
2-1-81


2-6-81
2-8-81
2-7-81

1-16-81
1-15-81


3-5-81
1-31-81
2-7-81

Z Modern
Carbon in
uuod (B)


112
112
112


133
133
133

133
133


115
115
115

I Modern
Carbon in
Aerosol (A)


97
107
118




96
100

63
53


96
97
90

Z RHC
(C-A/B)


86.
95.
(100)




72.
75.

47.
39.


83.
84.
78.



6
5





2
2

4
8


5
3
3

Total
Carbon
(l'g/m')(D)


40
34
48




24
35

53
52


32
35
44



.1
.8
.3




.2
.4

.7
.4


.5
.9
.9

RUC Carbon RUC 12)G

'"C/CMB
Ratio


0.99
0.96
1.18




0.75
0.91

0.62
0.53


0.92
0.92
0.95

Falrvlew
& l.lber

Medford
County
house

ty 1-8-81
12-12-80

Cour t-
12-17-80
1-16-81
106
106


111
111
64
24


104
114
60.
22.


93.
(100)
4
6


7

54
94


45
44
.5
.1


.8
.1
32
21


42
44
.9
.3


.9
.1
54.9
35.4


71.5
73.5
79
123


53
62
t 21
1 33


i 19
i 19
0.69
0.29


1.35
t .18
Notes
  (B) Calculated In relation to  International radiocarbon  standard.   Values are greater tlian IOOZ because of  '""C  generated  during nuclear atmoRplier Ic
      test Ing.
  (A) X of modern carbon In atmospheric C02. adjusted  to age  of  wood  burned.
  (C) 7. of modern carbon from RWC  sources.   100*  is  an upper  limit.
  (I)) Measured total carbon concentration  (sura of elemental,  organic  carbon).
  (K) Concentration of TWC carbon  (|ig/in ).
  (K) RUC impact based on '""C, corrected  lor associated II.  0.  N.
  (f;) ONI! 'estimates from filters of  uncertain total  sample  volume.

-------
                                                Figure 2
                          1'*C Validation of  CHB  Impact  Estimates  (yg/m3)
                              (with associated uncertainty  error bars)
    90
    80
    70
4   60 -
    50
                          (i)
                                    (2)
    :o
                                                          _L
     20           30           40            50            oO
                                          IUC RWC Imoacc (ug/m3)

          ''CMS sstiaaces  from filters of  uncertain :ocal volume
          (1) Seattle Soucn ?ark  16 Jan.  1981
          (2) Seattle Soucn ?ark  15 Jan.  1981
                                                                       70
                                                                                     30
                                                                                                  90
                                                29

-------
     Although the fuel use survey  (task 2B) results indicate different
fireplace/wood stove emission ratios for Portland, Seattle and Spokane
(0.45, 1.94 and 1.77, respectively), an insufficient body of source emission
chemistry for fireplaces and wood  stoves is available to clearly differentiate
between these two sources.  Apportioning impacts between fireplace and wood
stove emissions is further complicated by the high degeee of variability
in the composition of the emissions.  Given the level of effort that had
previously been committed to development of the Medford RWC emission
composition data, the difficulties noted above and the presence of
independent validation of CMB-derived RWC impacts during the Medford Aerosol
Study (based on C-14 results), further attempt to modify the Medford emission
chemistry data were not attempted.
     RWC impact estimates derived by the CMB methods have been compared to
independent measurements made by radiocarbon techniques.  Impact estimates
obtained from these two methods (Table 9)  show relatively good agreement
for samples collected at most sites.  Discrepancies noted at the Seattle
South Park and Boise Fairview and Liberty locations may be caused by.a
number of factors including (a) deviations from the fuel use mix assumption,
(b) the presence of important contemporary carbon sources other than
residential wood combustion or (c) errors in the CMB calculations.  Each are
discussed more fully below.
     A.  Wood Fuel Mix
          Since the 1UC estimates are based on an average age (ll*C content)
     of firewood in each community, deviations from the wood specie
     weightings assumed for each city (Table 4) will result in an error
     in the percent of modern carbon in the ambient aerosol.  For example,
     if residents in the near vicinity  of  the Seattle South Park site
     burned a lesser proportion of alder to Douglas fir than estimated on
     a city-wide basis,  the radiocarbon estimates would be higher than
     estimated.  If the mean percent of modern carbon in Seattle fuels is
     assumed to be nearer to the average value for other Pacific Northwest
     (116%), RWC estimated by lkC would increase by about 13%.
     B.  Other Contemporary Carbon Sources
          The radiocarbon method quantifies the total amount of contemporary
     carbon in the ambient aerosol whereas the CMB estimates of RWC impact
                                  30

-------
      are calculated  on  the  basis  of -numerous  aerosol  components.   An
      exact match between  estimates developed  by  both  methods  requires that
      essentially all  of the contemporary  carbon  measured by the  '^C
      technique be associated with RWC  sources—an assumption  that  may not
      be valid where  (a) the upper size  cut  of the ambient  sampling system
      allows capture  of  carbonaceous particles of large aerodynamic diameter
      (> 2.5 urn) such  as plant  fibers or  (b) industrial sources utilizing
      wood waste (e.g., Hogged  fuel boilers) are  major carbon  emission
      sources.
      C.  Errors in CM3 Calculations
          Close examination of  the CMS  calculations for several of the
      sampling sites  indicates  the potential for  inclusion  of  distillate
      oil combustion  impacts  in  the fitting  process.  Calculations  for the
      South Park site  (site  8, Appendix  5) indicates that,  by  reason of the
      underestimate of sulfur, distillate  oil  may be an important component.
      If so, it would  contribute to a 30%  overestimation of the RWC impact.
      The rationale for excluding distillate oil  from Che CMS  fitting  process
      is explained below.
      Site average RWC impacts   ranged from  16.7  (Spokane-Crown Zellerbach)  to
83.7  ug/m3 (Boise-Fairview  & Liberty St.) and  represented  from 20% (Medford)
to 93% (Yakima) of the fine  particle mass.  The highest 24 hour impact
(128  ug/m3) was found at the Boise site.  The  lowest average  impacts  (16-30 ug/mi>)
were  measured at the background and industrial sites (Tacoma,  excepted)  while
residential sites varied from 30 to 45 ug/m3.
      RWC impact estimates obtained for Boise  are significantly higher  than those
obtained in other Northwest  cities.   The maximum impact measured in Boise
occurred on January 18,  1981 at the Fairview & Liberty site,   127.9  ±  29 .'9  ug/m3,
8 hour average.  These source impact estimates are based on inhalable  particulate
samples captured on Teflon and glass fiber filters ra.ther than fine particle
(< 2.5 um)  samples.   As  a result, CMS estimated RWC impacts must be considered
to be upper limit values subject to  considerable uncertainty.   The uncertainty
is associated with (a) the RWC source composition which is  based on measurements
of fine particle emissions  (b)  RWC impact estimates  are highly dependent  on
measurements  of K,  Cl, elemental and organic carbon,  components which  in  the
size  fraction greater than 2.5 urn, can become  associated  with  sources  other
than wood burning emissions.

                                  31

-------
 Discrepancies between SSI and dichotomous sampler mass  measurements
 reported for co-located, concurrent samples further  complicates data
 interpretation.  For these reasons, further verification  of  the Boise
 results is recommended.
      The maximum  industrial source  impacts, while minimal, were  noted at the
 Longview site  (Kraft  recovery furnace and aluminum reduction), Tacoma Fire
 Station 12  (aluminum  reduction) and the Spokane-Crown Zellerbach site (Hog
 fuel  boiler  and aluminum reduction  emissions).
      In addition, the chemistry of  a large number of samples could be explained
 by  inclusion of a distillate oil source rather than a secondary  sulfate source.
 This  would tend to reduce the impact of RWC sources by about 30%.  However,
 since  the quality of  the CMB fits were (a) quite similar in either case and
 (b) distillate oil combustion impacts at the background sites were similar to
 residential  sites, distillate oil combustion was not routinely included as a
 source.   It  is likely, however, that oil combustion associated with residential
 space  heating is an important contributing source to which a portion of the
 measured  sulfur should be assigned.
     The Chemical Mass Balance calculation performed successfully apportioned
 greater  than 90% of the measured mass with reasonable agreement between measured
 and model-predicted element concentrations.   Occasional overpredictions of the
 mass,  however, occur due to uncertainty in the gravimetric analysis or in the
 source  assignments.   In most cases, the calculated and measured mass values are
 within  the estimated mass uncertainty.
     Analysis of selected samples for PNA compounds was completed because of
 concern about risks to public health associated wich  these potentially-
 carcinogenic compounds.  Review of  RNA analysis  results  associated with this
 study  suggests that mean concentrations within residential areas are an order
of magnitude higher than similar measurements  at  background  sites.  Table 10
 presents a summary of average PNA concentrations  from residential sites during
 the 1980-81 heating season.  Since  the samples selected  for  analysis were
intended to reflect worst case RWC  impacts, the values reported  do not indicate
typical conditions.
                                   32

-------
                                 Table 10

                        Average PNA Concentration
                     For Residential Sites (ng/m3)*
                        (1980-'81  Heating Season)
Arithmetic
Mean
19
7
4
10
14
0
2
.5
.4
.1
.6
.4
.3
,7
Standard
Deviation
17
5
3
13
18
0
1
.3
.2
.1
.2
.9
.4
.9
Maximum
Value
50
18
11
42
69
1
6
.7
.1
.1
.2
.9
.4
.6
Arithmetic
Minimum Mean
Value Background
0
0
0
0
0
0
0
.6
.3**
.3**
.3**
.3**
.05
.2
0
0
0
0
0
0
0
.2
.9
.4
.3
.4
.05
.3
 PAH  Group
 Benzo(a)Anthracene
 Benzofluoranthenes
 Benzo(a)Pyrene
 Fluoranthene
 Pyrene
 Dibenzanthracenes
 Benzo(ghi)perylene
  *  Average of 15 selected 24 hour samples
 **  One-half minimum detectable  concentration
      PNA measurements made during this study can be compared to annual  average
measurements of B(a)P during the 1960 "s in Portland, Eugene and Medford which
ranged from 2.3 to 4.8 ng/m3.6  Quarterly average maximum values of 8.2 ng/m3
were recorded in Medford in 1968.  Although annual and quarterly average B(a)P
concentration data are generally not available after 1970, periodic data from
other Northwest locations (i.e., Seattle) suggests that concentrations  may
                                                                           7, a
have decreased during the 1970-76 period and then increased following 1977.
Routine PNA concentration measurements are no longer conducted by EPA.
     The average B(a)P concentrations noted in Table 10 are similar to  the annual
and quarterly mean values discussed above, but direct comparison is confounded
by differences in averaging times.
     There are no published studies available which adequately compare  the
carcinogenic activity of all ten PNA compounds under similar experimental
conditions.  Similarly,  there is no data available concerning human responses
                                    33

-------
to individual PNA  compounds,  since  environmental  exposures  to  PNA invariably
involve contact with  complex  and  usually  undefined  PNA mixtures.   Therefore,
an assessment of the  cancer risk  related  to  PNA compounds emitted from RWC
sources can only be considered when placed in  perspective with other  sources
of exposure.  Conversion, however,  of  ambient  PNA concentrations  to in-
halation intake requires consideration of the  volume  of air breathed  per
day  (15 m3 for a 10 year old  child  and 23 m3 for  an adult male) and a 24
hour average assumed  concentration  of  B(a)P.9  If an  average daily volume
of 19 in3 and an average B(a)P concentration  of 4.1  ng/m3 is assumed,  a
daily dosage rate  of  0.078 yg/day is obtained.  This  compares  to  an estimated
total B(a)P intake of 0.16-1.6 yg/day  associated  with food consumption.  As
a further comparison, Bridbord10 estimates that smoking 20 cigarettes per  day
provides an exposure  of 0.4 ug/day  of  B(a)P, suggesting that the  average
B(a)P concentration identified in this study is equivalent to  dosage
received by smoking about 4 cigarettes/day.  At the maximum B(a)P  concen-
tration measured (11.1 ng/m3, 24 hour  average), the dosage is  equivalent to
11 cigarettes/day.
     The impact of RWC emissions on total suspended particulate (TSP)  air
quality was found  to  range from 11% in industrial areas (Spokane—Crown
Zellerbach site) to nearly 44% in residential areas (Portland—Marcus Whitman
School).  In Boise, the RWC impact, calculated on the basis of the inhalable
particulate fraction, may contribute as much as 60% to the TSP mass during
worst case RWC impact conditions.  Although  this  data clearly  documents
the importance of wood combustion emissions  to TSP air quality under  the
conditions of this study, direct inference as  to  the  importance of RWC
emissions on TSP air quality standard non-attainment  is difficult  because
(a) the RWC impacts cannot be calculated in  terms of  an annual geometric
mean averaging period and (b)  the 24 hour worst case  RWC impacts  included
in this study may not represent the 24 hour worst cast TSP concentrations
upon which standard attainment is based.
                                  34

-------
6.0  CONCLUSIONS

     Ambient air quality studies conducted during the 1980-81 space heating
season in eight Pacific Northwest communities clearly indicate  that
residential wood combustion (RWC) emissions, as estimated by validated
chemical mass balance analysis, are the most important contributor to the
fine particle mass less than 2 ym.  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:
     •  RWC emissions typically account for 66% to 75% of the fine particle
        mass, while transportation sources contribute 5%, secondary sulfate,
        5.6% and all industrial sources less than 0.5%.
        Background RWC impacts were found to range from 3-12 yg/m , 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.
     •  Maximum 24 hour impacts exceeded 60yg/m3 at  residential sites
        located in Seattle,  Spokane, Portland and Medford.   Impacts at
        industrial sites in Longview, Seattle and Tacoma were significantly
        lower (15-30 yg/m3).
        The highest impacts measured in this study were in  Boise,  Idaho
        (128 yg/m3, 8 hour average).  These measurements must, however,
        be considered as upper limit estimates requiring further verification.
     •  PNA concentrations measured at urban sites were a factor of ten
        higher than chat measured at the rural background sites.  Although
        the measured PNA concentrations should be of concern, no direct
        dose-response relationship are currently available  in the literature
        upon which to base a quantitative assessment of  public health risk.
     Further interpretation of the results of Task 1 are included in the
Task 2A report (Analysis of  Air Quality  Impacts).
                                  35

-------
                                REFERENCES

  x°   J.A.  Cooper,  "Environmental  Impact  of  Residential  Wood  Combustion
      Emissions  and its  Implications",  Journal Air Poll. Control Assoc. 30  (8),
      1980.

  2>   J.A.  Cooper,  LiA.  Currie,  G.A.  Klouda,  "Application  of  Carbon-14
      Measurements  to  Impact Assessment of Contemporary  Carbon  Sources  on
      Urban Air  Quality", Environ.  Science &  Technology, September,  1981.

  3 *   "Performance  Test  for the  Evaluation of Computerized GC/MS Equipment
      and Laboratories", USEPA,  EMSL, ORD. 1980.

  "'   R.T. DeCesar  and J.A. Cooper, "Medford  Aerosol Characterization Study",
      Final Report  to  the State  of Oregon Department of  Environmental
      Quality, February, 1981.

  s'   J.A. Cooper and J.G. Watson, "Portland Aerosol Characterization Study",
      Final Report  to  the State  of Oregon Department of  Environmental
      Quality, July, 1979.

      "Scientific and Technical Assessment Report on Particulate Polycyclic
      Organic Matter" USEPA # EPA-600/6-75-001, 1975.

 7'   "Health Assessment Document for Polycyclic Organic Matter" US EPA
      // EPA-600/9-79-008- 1979.

 ,8 *   "Coal Health Effects Review Committee;  Summary Report to the Oregon
     Department of Environmental Quality",  State of Oregon Dept. of Env.
     Quality. April, 1971.

 9"   International Commission of Radiological Protection (Report of the
     Task Group on Reference Man), Pergamon  Press,  New York,  1974.

10'  K. Bridbord, et.  al.,  "Human Exposure  co Polynuclear  Aromatic Hydrocarbons'
     Carcinogenesis, Volume I, Raven Press,  NY,  NT, 1976.
                                    36

-------