Formaldehyde Indoor Air Model – Pressed Wood Products, Version 2.0
(FIAM-pwp v2.0): 

Model User Manual and Documentation 

Exposure Assessment Branch

Economics, Exposure and Technology Division

Office of Pollution Prevention and Toxics

Office of Chemical Safety and Pollution Prevention

U.S. Environmental Protection Agency

1200 Pennsylvania Avenue, N. W.

	Washington, D.C.  20460

July 2012

EPA acknowledges the analytical and draft preparation support of Versar,
Inc. of Springfield, Virginia in the preparation of this report,
provided under Contract No. EP-W-10-005.

  TOC \o "1-3" \h \z \u    HYPERLINK \l "_Toc332143174"  1.	INTRODUCTION
  PAGEREF _Toc332143174 \h  1  

  HYPERLINK \l "_Toc332143176"  2.	USER INTERFACE	  PAGEREF
_Toc332143176 \h  3  

  HYPERLINK \l "_Toc332143177"  2.1	Background – Model Versions	 
PAGEREF _Toc332143177 \h  3  

  HYPERLINK \l "_Toc332143178"  2.2	House Screen	  PAGEREF _Toc332143178
\h  4  

  HYPERLINK \l "_Toc332143179"  2.2.1	Inputs and Options	  PAGEREF
_Toc332143179 \h  4  

  HYPERLINK \l "_Toc332143180"  2.2.2	Basis for Defaults	  PAGEREF
_Toc332143180 \h  8  

  HYPERLINK \l "_Toc332143181"  2.3	Source Screen	  PAGEREF
_Toc332143181 \h  14  

  HYPERLINK \l "_Toc332143182"  2.3.1 	Types of Inputs	  PAGEREF
_Toc332143182 \h  14  

  HYPERLINK \l "_Toc332143183"  2.3.2 	Adding Default Inputs	  PAGEREF
_Toc332143183 \h  15  

  HYPERLINK \l "_Toc332143184"  2.3.3 	Adding Custom Inputs	  PAGEREF
_Toc332143184 \h  16  

  HYPERLINK \l "_Toc332143185"  2.3.3 	Basis for Defaults	  PAGEREF
_Toc332143185 \h  16  

  HYPERLINK \l "_Toc332143186"  2.4	Exposure Screen	  PAGEREF
_Toc332143186 \h  21  

  HYPERLINK \l "_Toc332143187"  2.4.1 	Inputs and Options for Age of
Sources and Formaldehyde Level of Interest	  PAGEREF _Toc332143187 \h 
21  

  HYPERLINK \l "_Toc332143188"  2.4.2 	Inputs and Options for Exposure
Groups	  PAGEREF _Toc332143188 \h  22  

  HYPERLINK \l "_Toc332143189"  2.4.3 	Inputs and Options for
Concentrations Outside the Home	  PAGEREF _Toc332143189 \h  23  

  HYPERLINK \l "_Toc332143190"  2.4.4 	Basis for Defaults	  PAGEREF
_Toc332143190 \h  24  

  HYPERLINK \l "_Toc332143191"  2.5	Result Screen	  PAGEREF
_Toc332143191 \h  27  

  HYPERLINK \l "_Toc332143192"  2.5.1 	Concentration Results	  PAGEREF
_Toc332143192 \h  27  

  HYPERLINK \l "_Toc332143193"  2.5.2 	Exposure Results	  PAGEREF
_Toc332143193 \h  28  

  HYPERLINK \l "_Toc332143194"  2.5.3 	Printing Inputs and Results	 
PAGEREF _Toc332143194 \h  28  

  HYPERLINK \l "_Toc332143195"  2.5.4 	Archiving Inputs and Results for
a Run	  PAGEREF _Toc332143195 \h  28  

  HYPERLINK \l "_Toc332143196"  2.6	Model Access and Additional Features
  PAGEREF _Toc332143196 \h  29  

  HYPERLINK \l "_Toc332143197"  2.6.1 	Accessing the IGEMS Website	 
PAGEREF _Toc332143197 \h  29  

  HYPERLINK \l "_Toc332143198"  2.6.2 	Saving Inputs for Later Access	 
PAGEREF _Toc332143198 \h  29  

  HYPERLINK \l "_Toc332143199"  2.6.3 	Context-sensitive Help	  PAGEREF
_Toc332143199 \h  29  

  HYPERLINK \l "_Toc332143200"  2.7 	Example Applications	  PAGEREF
_Toc332143200 \h  29  

  HYPERLINK \l "_Toc332143201"  2.7.1 	Example 1 – Modeling with Model
Default Values	  PAGEREF _Toc332143201 \h  29  

  HYPERLINK \l "_Toc332143202"  2.7.2 	Example 2 – Modeling a
Hypothetical Chamber Test for CARB Compliance	  PAGEREF _Toc332143202 \h
 30  

  HYPERLINK \l "_Toc332143203"  2.7.3 	Example 3 – Demonstrating an
Equilibrium Concentration	  PAGEREF _Toc332143203 \h  31  

  HYPERLINK \l "_Toc332143204"  2.7.4	Example 4 – Examining the
Incremental Contribution of a Product	  PAGEREF _Toc332143204 \h  32  

  HYPERLINK \l "_Toc332143205"  2.7.5	Example 5 – Varying the Decay
Rate (Half Life)	  PAGEREF _Toc332143205 \h  33  

  HYPERLINK \l "_Toc332143206"  3.	MODEL CALCULATIONS AND ASSUMPTIONS	 
PAGEREF _Toc332143206 \h  35  

  HYPERLINK \l "_Toc332143207"  3.1	Generalized Mass-Balance Equation	 
PAGEREF _Toc332143207 \h  36  

  HYPERLINK \l "_Toc332143208"  3.2	Steady-State Model for Initial
Formaldehyde Concentration in One Compartment	  PAGEREF _Toc332143208 \h
 37  

  HYPERLINK \l "_Toc332143209"  3.3 	Extension of the Steady-State Model
to Two Compartments	  PAGEREF _Toc332143209 \h  39  

  HYPERLINK \l "_Toc332143210"  3.4 	Matthews Model Compared to HBF
Model	  PAGEREF _Toc332143210 \h  40  

  HYPERLINK \l "_Toc332143211"  3.5 	Other Potentially Useful Modeling
Constructs	  PAGEREF _Toc332143211 \h  42  

  HYPERLINK \l "_Toc332143212"  3.6 	Discussion of Indoor Sinks	 
PAGEREF _Toc332143212 \h  45  

  HYPERLINK \l "_Toc332143213"  3.7	Decay of Indoor Concentrations	 
PAGEREF _Toc332143213 \h  45  

  HYPERLINK \l "_Toc332143214"  3.8	Default Value for Half Life	 
PAGEREF _Toc332143214 \h  48  

  HYPERLINK \l "_Toc332143215"  3.9	Temperature and Humidity Adjustments
  PAGEREF _Toc332143215 \h  51  

  HYPERLINK \l "_Toc332143216"  3.10	Exposure Calculations	  PAGEREF
_Toc332143216 \h  53  

  HYPERLINK \l "_Toc332143217"  4.	 EVALUATION OF MODEL ALGORITHMS AND
PREDICTIONS	  PAGEREF _Toc332143217 \h  55  

  HYPERLINK \l "_Toc332143218"  4.1	Mathematical Correctness of Model
Algorithms	  PAGEREF _Toc332143218 \h  55  

  HYPERLINK \l "_Toc332143219"  4.2	NIST Assessment of Matthews Model
Predictions	  PAGEREF _Toc332143219 \h  55  

  HYPERLINK \l "_Toc332143220"  4.3	Assessment of FIAM-pwp Predictions
– EPA Pilot Study	  PAGEREF _Toc332143220 \h  56  

  HYPERLINK \l "_Toc332143221"  4.3.1	Initial Baseline Period	  PAGEREF
_Toc332143221 \h  57  

  HYPERLINK \l "_Toc332143222"  4.3.2	First Loading	  PAGEREF
_Toc332143222 \h  58  

  HYPERLINK \l "_Toc332143223"  4.3.3	Air-out Period after First Loading
  PAGEREF _Toc332143223 \h  59  

  HYPERLINK \l "_Toc332143224"  4.3.4	Second Loading	  PAGEREF
_Toc332143224 \h  60  

  HYPERLINK \l "_Toc332143225"  4.3.5	Air-out Period after Second
Loading	  PAGEREF _Toc332143225 \h  61  

  HYPERLINK \l "_Toc332143226"  4.3.6	Third Loading	  PAGEREF
_Toc332143226 \h  62  

  HYPERLINK \l "_Toc332143227"  4.4	FIAM-pwp Predictions for Recent
Residential Field Monitoring Studies	  PAGEREF _Toc332143227 \h  63  

 

Appendix A – Summary of EPA Formaldehyde Pilot Study

Appendix B – Model Sensitivity

Appendix C – Basis for Emission Classes

Appendix D – Historical Testing of PWPs to Derive Slope and Intercept
Values

Appendix E – Derivation of Two-zone Steady-state Solution

  TOC \h \z \t "Figure Caption" \c    HYPERLINK \l "_Toc332144055" 
Figure 2-1. FIAM-pwp House Screen.	  PAGEREF _Toc332144055 \h  4  

  HYPERLINK \l "_Toc332144058"  Figure 2-2. Popup Screen – Create Your
Own Structure Type.	  PAGEREF _Toc332144058 \h  5  

  HYPERLINK \l "_Toc332144059"  Figure 2-3. U.S. Climate Zones.	 
PAGEREF _Toc332144059 \h  5  

  HYPERLINK \l "_Toc332144060"  Figure 2-4. Popup Screen – Create Your
Own Climate Zone.	  PAGEREF _Toc332144060 \h  6  

  HYPERLINK \l "_Toc332144061"  Figure 2-5. FIAM-pwp Source Screen.	 
PAGEREF _Toc332144061 \h  14  

  HYPERLINK \l "_Toc332144062"  Figure 2-6. Source Screen after Adding
Default PWP Types	  PAGEREF _Toc332144062 \h  15  

  HYPERLINK \l "_Toc332144063"  Figure 2-7. Source Screen after Adding
Custom Values	  PAGEREF _Toc332144063 \h  16  

  HYPERLINK \l "_Toc332144064"  Figure 2-8. Exposure Screen	  PAGEREF
_Toc332144064 \h  21  

  HYPERLINK \l "_Toc332144065"  Figure 2-9. Preparing to Add a New
Exposure Group	  PAGEREF _Toc332144065 \h  23  

  HYPERLINK \l "_Toc332144066"  Figure 2-10. Result Screen with First
Two Exposure Groups Selected	  PAGEREF _Toc332144066 \h  27  

  HYPERLINK \l "_Toc332144067"  Figure 2-11. Emission-rate Parameters
for FIAM-pwp Default PWPs in an Apartment.	  PAGEREF _Toc332144067 \h 
32  

  HYPERLINK \l "_Toc332144068"  Figure 2-12. Formaldehyde Concentration
Decline with Fast and Slow Decay Periods.	  PAGEREF _Toc332144068 \h  34
 

  HYPERLINK \l "_Toc332144069"  Figure 3-1. Representation of Declining
Emission Rate by Power-law Function vs. Exponential-decay Model.	 
PAGEREF _Toc332144069 \h  44  

  HYPERLINK \l "_Toc332144070"  Figure 3-2. Representation of Declining
Emission Rate by Power-law Function vs. Exponential-decay Model with
Periods of Fast and Slow Decay.	  PAGEREF _Toc332144070 \h  44  

  HYPERLINK \l "_Toc332144071"  Figure 4-1. FIAM-pwp House-Screen Inputs
Used to Model Study of New California Homes.	  PAGEREF _Toc332144071 \h 
65  

  HYPERLINK \l "_Toc332144072"  Figure 4-2. FIAM-pwp Source-Screen
Inputs Used to Model Study of New California Homes.	  PAGEREF
_Toc332144072 \h  66  

  HYPERLINK \l "_Toc332144073"  Figure 4-3. FIAM-pwp Results for
Modeling of New California Homes.	  PAGEREF _Toc332144073 \h  66  

  HYPERLINK \l "_Toc332144074"  Figure 4-4. FIAM-pwp Source-Screen
Inputs Used to Model CDC Study of Travel Trailers.	  PAGEREF
_Toc332144074 \h  67  

  HYPERLINK \l "_Toc332144075"  Figure 4-5. FIAM-pwp Results for
Modeling of Travel Trailers.	  PAGEREF _Toc332144075 \h  67  

 

  TOC \h \z \t "Table Caption" \c  \* MERGEFORMAT    HYPERLINK \l
"_Toc332143577"  Table 2-1. House Dimensions and Calculated Volume for
Five Housing Types	  PAGEREF _Toc332143577 \h  9  

  HYPERLINK \l "_Toc332143578"  Table 2-2. Summer and Winter Indoor
Temperatures from RECS 2005, by Climate Zone	  PAGEREF _Toc332143578 \h 
10  

  HYPERLINK \l "_Toc332143579"  Table 2-3. Year-round Indoor
Temperatures Derived from RECS 2005, by Climate Zone	  PAGEREF
_Toc332143579 \h  10  

  HYPERLINK \l "_Toc332143580"  Table 2-4. Modeled Year-round Indoor RH
Level, by Climate Zone	  PAGEREF _Toc332143580 \h  11  

  HYPERLINK \l "_Toc332143581"  Table 2-5. Alternative
Analytical/Regulatory Options for EPA Formaldehyde Rule	  PAGEREF
_Toc332143581 \h  17  

  HYPERLINK \l "_Toc332143582"  Table 2-6. Default Slopes and Intercepts
for Six Default PWP Types	  PAGEREF _Toc332143582 \h  18  

  HYPERLINK \l "_Toc332143583"  Table 2-7. Default Exposed Surface Areas
(in m2) for Six PWP Types, by Structure Type – New-home Case	  PAGEREF
_Toc332143583 \h  20  

  HYPERLINK \l "_Toc332143584"  Table 2-8. Default Exposed Surface Areas
(in m2) for Six PWP Types, by Structure Type – Renovation Case	 
PAGEREF _Toc332143584 \h  20  

  HYPERLINK \l "_Toc332143585"  Table 2-9. Default Annual Hours at Home
and Non-home Locations, by Exposure Group.	  PAGEREF _Toc332143585 \h 
24  

  HYPERLINK \l "_Toc332143586"  Table 2-10. Default Concentration Values
for Non-home Locations	  PAGEREF _Toc332143586 \h  25  

  HYPERLINK \l "_Toc332143587"  Table 2-11. Derivation of Baseline
Values for Daycare, School and Work Locations	  PAGEREF _Toc332143587 \h
 26  

  HYPERLINK \l "_Toc332143588"  Table 2-12. Concentration Values for
Selected Non-home Locations, by Emissions Scenario	  PAGEREF
_Toc332143588 \h  26  

  HYPERLINK \l "_Toc332143589"  Table 2-13. Run Notes for Example 1	 
PAGEREF _Toc332143589 \h  30  

  HYPERLINK \l "_Toc332143590"  Table 2-14. Run Notes for Example 2	 
PAGEREF _Toc332143590 \h  31  

  HYPERLINK \l "_Toc332143591"  Table 2-15. Run Notes for Example 3	 
PAGEREF _Toc332143591 \h  31  

  HYPERLINK \l "_Toc332143592"  Table 2-16. Run Notes for Example 4	 
PAGEREF _Toc332143592 \h  33  

  HYPERLINK \l "_Toc332143593"  Table 2-17. Run Notes for Example 5,
Part 1 (fast decay period)	  PAGEREF _Toc332143593 \h  34  

  HYPERLINK \l "_Toc332143594"  Table 2-18. Run Notes for Example 5,
Part 2 (slow decay period)	  PAGEREF _Toc332143594 \h  34  

  HYPERLINK \l "_Toc332143595"  Table 4-1.  Model Inputs (Sources and
Sinks) for Initial Baseline Period	  PAGEREF _Toc332143595 \h  57  

  HYPERLINK \l "_Toc332143596"  Table 4-2.  Modeling Results for Initial
Baseline Period	  PAGEREF _Toc332143596 \h  58  

  HYPERLINK \l "_Toc332143597"  Table 4-3.  Model Inputs (Sources and
Sinks) for First (Medium) Loading	  PAGEREF _Toc332143597 \h  58  

  HYPERLINK \l "_Toc332143598"  Table 4-4.  Modeling Results for First
Loading	  PAGEREF _Toc332143598 \h  59  

  HYPERLINK \l "_Toc332143599"  Table 4-5.  Model Inputs for Air-out
Period after First Loading	  PAGEREF _Toc332143599 \h  59  

  HYPERLINK \l "_Toc332143600"  Table 4-6.  Modeling Results for Air-out
Period after First Loading	  PAGEREF _Toc332143600 \h  60  

  HYPERLINK \l "_Toc332143601"  Table 4-7.  Model Inputs (Sources and
Sinks) for Second (High) Loading	  PAGEREF _Toc332143601 \h  60  

  HYPERLINK \l "_Toc332143602"  Table 4-8.  Modeling Results for Second
Loading	  PAGEREF _Toc332143602 \h  61  

  HYPERLINK \l "_Toc332143603"  Table 4-9.  Model Inputs for Air-out
Period after Second Loading	  PAGEREF _Toc332143603 \h  61  

  HYPERLINK \l "_Toc332143604"  Table 4-10.  Modeling Results for
Air-out Period after Second Loading	  PAGEREF _Toc332143604 \h  62  

  HYPERLINK \l "_Toc332143605"  Table 4-11.  Model Inputs (Sources and
Sinks) for Third (High) Loading	  PAGEREF _Toc332143605 \h  62  

  HYPERLINK \l "_Toc332143606"  Table 4-12.  Modeling Results for Third
Loading	  PAGEREF _Toc332143606 \h  63  

  HYPERLINK \l "_Toc332143607"  Table 4-13. Summary of Recent
Formaldehyde Field Monitoring Studies	  PAGEREF _Toc332143607 \h  64  

 

1.	INTRODUCTION

The model described in this document, the Formaldehyde Indoor Air Model
- Pressed Wood Products, Version 2.0 (FIAM-pwp v2.0), is intended to
assist the user in estimating human inhalation exposure to airborne
formaldehyde emissions from composite wood products (CWPs) installed in
new or existing residences, including smaller structures such as camper
trailers. The model estimates indoor formaldehyde concentrations and
associated human inhalation exposures for such situations.

FIAM-pwp is an adaptation of a model developed during the 1980s by Dr.
Thomas Matthews and colleagues, at the Oak Ridge National Laboratory.
The “Matthews model” was designed to estimate the steady-state
indoor formaldehyde concentration due to emissions from wood products in
a single indoor compartment or zone. Product emission rates in the model
are dependent on the formaldehyde concentration in the vapor phase –
as the concentration increases the emission rates decrease, other things
being equal. The initial (“steady-state”) indoor concentration
calculated by the model is assumed to decrease over time, as the
formaldehyde “reservoirs” in various sources (or sinks) are
gradually depleted. The decrease over time is assumed to follow a
first-order exponential process, at a decay rate that corresponds to an
assumed half life for the collective formaldehyde emissions.

In the mid-1990s, an EPA-sponsored pilot study was undertaken by
Versar/GEOMET to assess the contribution of UF-bonded wood products to
formaldehyde levels in homes and to evaluate current EPA exposure
models. Shortly after this study was completed, Versar was tasked by EPA
to develop a modified version of the Matthews model. The primary
modifications were (1) estimation of steady-state formaldehyde
concentrations for the case of a two-zone indoor environment, and (2)
incorporation of reversible (re-emitting) indoor sinks. The modified
model also included the option to be run in a single-zone mode. The
model was not finalized due to shifting priorities at the time of its
development. 

In 2006, formaldehyde from pressed wood products received national
attention when the U.S. Federal Emergency Management Agency (FEMA)
provided travel trailers and mobile homes for habitation by residents of
the U.S. gulf coast who were displaced by Hurricane Katrina and
Hurricane Rita. Some individuals who moved into these temporary homes
complained of breathing difficulties, nosebleeds, and persistent
headaches. In December 2007 and January 2008, the Centers for Disease
Control and Prevention measured formaldehyde levels averaging 77 parts
per billion, (ppb; geometric mean) in a random sample of FEMA-supplied
occupied travel trailers, park models, and mobile homes. In February
2008 U.S. health officials announced that potentially hazardous HCHO
levels were found in both travel trailers and mobile homes provided by
FEMA.

In April 2008 the California Air Resources Board (CARB) formally adopted
an Airborne Toxic Control Measure (ATCM) to reduce formaldehyde
emissions from CWPs sold, offered for sale, supplied, used, or
manufactured for sale in California. On July 7, 2010, the Formaldehyde
Standards for Composite Wood Products Act (FSCWPA) was signed into law.
This statute, which adds a Title VI to TSCA, establishes formaldehyde
emission standards for hardwood plywood, particleboard, and
medium-density fiberboard that are identical to the CARB ATCM Phase II
standards. TSCA Title VI directs EPA to promulgate implementing
regulations by January 1, 2013 that address: sell-through dates for
products; stockpiling; third-party testing and certification; auditing
and reporting of third-party certifiers; recordkeeping; chain of
custody; labeling; enforcement; products made with no-added formaldehyde
(NAF) and ultra-low emitting formaldehyde (ULEF) resins; laminated
products; finished goods; hardboard; and products containing de minimis
amounts of composite wood products.  

As a result of the activity described above, EPA decided to revise and
update, when possible, the formaldehyde model developed in the mid-1990s
to facilitate an assessment of exposures before and after the EPA
formaldehyde rule goes into effect. FIAM-pwp v1.0 was the initial result
of this effort, with FIAM-pwp v2.0 developed in response to peer review
and workgroup comments. Version 2.0 adds exposure estimates to the model
but indoor sinks have been excluded, for reasons discussed later in this
document (see Section 3). Section 2 describes the user interface for the
model and Section 3 describes the mathematical basis, underlying
calculations and related assumptions. Model evaluation, including
comparisons to research-house and field-monitoring data, is provided in
Section 4. The EPA pilot study mentioned above is summarized in Appendix
A. Analyses related to model sensitivity are presented in Appendix B.
The basis for emission classes used on the Source Screen is presented in
Appendix C. Historical chamber-test data, conducted primarily during the
1980s but extending into the 1990s and providing part of the basis for
emission-rate parameters used in FIAM-pwp, are described and discussed
in Appendix D. Appendix E describes the mathematical derivation of the
two-zone implementation of the steady-state model. 

2.	USER INTERFACE

2.1	Background – Model Versions 

In 2009 an on-line, user-friendly version of the formaldehyde model
(FIAM-pwp v1.0) was developed along with documentation to guide its use.
Later that same year a peer review of the model was facilitated for EPA
by Eastern Research Group, Inc. (ERG). The purpose of the peer review
was to review and comment on three main aspects of the FIAM-pwp model:
the soundness of the algorithms and exposure assumptions; the ease of
use of the graphical interface; and the completeness and clarity of the
model documentation. Unedited written comments submitted by five
reviewers in response to the peer-review charge were organized in a
report prepared for EPA by ERG in October 2009. 

Version 2, named FIAM-pwp v2.0 and the subject of this document, was
developed to run under the Internet Geographical Exposure Modeling
System (IGEMS) platform. Some additions to or subtractions from the
previous version are related, at least in part, to the peer reviewers’
comments. For example, a decision was made to drop certain FIAM-pwp
features that provided particular challenges or difficulties: (1) inputs
for indoor sinks, an area of considerable modeling uncertainty; and (2)
inputs for emission characteristics of cabinet components, which were
very demanding. In addition, exposure groups are being handled
differently in the current version, for consistency with the approach
used in  EPA’s 2012 formaldehyde exposure assessment.

The user interface for the current version includes three input screens
(House, Source and Exposure Screens) and an output screen (Result
Screen) that displays the modeling results; these results are
continually updated, such that any time the Result Screen is accessed
the results displayed reflect all current inputs, much like a
spreadsheet. First-time users, in particular, are advised to go through
the input screens in sequence (i.e., House then Source then Exposure),
examining results either “along the way” or after all inputs have
been entered, at the user’s discretion.

The House Screen is described in Section 2.2, the Source Screen in
Section 2.3, and the Exposure Screen in Section 2.4. For each of these
screens a description of inputs and options is provided first, followed
by the basis for model defaults. The Result Screen is described in
Section 2.5. Instructions for accessing the model via the IGEMS portal
are provided in Section 2.6, along with documentation of ancillary model
features such as saving runs for later access, printing inputs and
results for a run, archiving inputs and results in Excel, and accessing
the context-sensitive help embedded in the model. Example applications
are provided in Section 2.7.

2.2	House Screen

This screen (see Figure 2-1) is intended to collect information on the
house volume and internal conditions (e.g., temperature, airflows) and
to provide certain options pertaining to model calculations. The inputs
here can be divided into three screen portions: (1) documentation of the
run and selection of structure type and climate zone (upper portion),
(2) editable information on the number of zones, zone volumes and
airflows (middle portion); and (3) indoor conditions and formaldehyde
decay parameters (lower portion).

 

Figure 2-1. FIAM-pwp House Screen.

Many of the inputs on this screen are optional, in the sense that
default values are provided by the model. The choice of structure type
on the upper portion of the screen provides default values for zone
volumes and airflow rates, which are displayed on the middle portion of
the screen. The choice of climate zone on the same line provides default
values for indoor temperature and humidity, which are displayed on the
lower portion of the screen. Other inputs on the lower portion (e.g.,
background concentration) are not linked to any user choice but do have
default values. Although the model can be run using default values, care
should be taken in choosing the type of structure to be modeled and in
choosing appropriate values for inputs such as the background
concentration, emissions half life, and indoor temperature and relative
humidity. 

2.2.1	Inputs and Options

Run Title and Run Notes. Text inputs for these two items are intended to
assist the user in documenting the purpose of the model run along with
certain choices made or options exercised.  

Structure Type and Climate Zone. Both of these items have pick lists
from which the user can choose to obtain default values, which can
either be used “as is” or edited. The user also has the option of
creating additional structure types or climate zones and saving them for
later use.

There are five default structure types in FIAM-pwp: single-family (SF)
detached homes; SF attached homes; apartments; manufactured/mobile
homes, and camper trailers. The SF detached/attached homes are two-story
structures with two zones (upstairs and downstairs) whereas the others
are single-zone structures.

The values associated with FIAM-pwp default structure types are
“untouchable” (i.e., they cannot be directly edited), but the
choices on the pick list include one called Create Your Own Structure
Type – if this choice is made, then the popup screen shown in Figure
2-2 will appear. At that point, the user assigns a structure type name
and enters zone descriptions, volumes and airflows, and then clicks the
Save button. The MAKEACOPY button (see Figure 2-1, above) is used to
copy and then edit/save any of the default choices. Guidance on entering
volume and airflow values is provided later in this section, under the
heading Zones, Volumes and Air Flows.

Figure 2-2. Popup Screen – Create Your Own Structure Type.

FIAM-pwp uses the U.S. climate zones defined by the Department of Energy
(see Figure 2-3) There are five climate zones, numbered 1 though 5, with
1 being the coldest and 5 the warmest.

Figure 2-3. U.S. Climate Zones.

Source: U.S. Energy Information Administration (  HYPERLINK
"http://www.eia.doe.gov/emeu/cbecs/climzonenew.gif" 
http://www.eia.doe.gov/emeu/cbecs/climzonenew.gif )

As with default structure types, the values associated with default
climate zones in FIAM-pwp are “untouchable.” Other values can be
assigned and saved via a choice called Create Your Own Climate Zone,
which is included on the climate-zone pick list – if this choice is
made, then the popup screen shown in Figure 2-4 will appear. At that
point, the user assigns a name, enters the desired temperature and
humidity values, and clicks the SAVE button. The MAKEACOPY button on the
main screen can be used to copy and then edit/save any of the default
choices. Regardless of the climate zone chosen, the user can override
the associated temperature and humidity values by changing them directly
in the lower portion of the House Screen.

Figure 2-4. Popup Screen – Create Your Own Climate Zone.

Air Exchange Rate. This item is “grayed out” because no user entry
is allowed here; the rate is displayed for informational purposes only.
The value displayed is calculated by FIAM-pwp using volume and airflow
information and, thus, can be changed by editing those inputs. 

Zones, Volumes and Air Flows. Houses in FIAM-pwp can have either one or
two zones. Among the default structure types, single-family detached and
attached homes are conceptualized as having two separate zones (upstairs
and downstairs) whereas manufactured homes, apartments and camper
trailers are conceptualized as having a single zone. The maximum number
of zones allowed is two. Thus, for example, to represent a house with
three stories, at least two of the three stories would need to be
combined into one zone. The indoor-outdoor airflow rates determine the
air exchange rate. For example, for an apartment with a volume of 261
m3, the default indoor-outdoor airflow rate of 52.2 m3/h results in an
air exchange rate of 0.2 ACH (i.e., 52.2 m3/h / 262 m3 = 0.2/h). 

If the user selects Create Your Own Structure Type from the pick list or
edits one of the default choices after making a copy, then the popup
screen previously shown in Figure 2-2 will appear. The required
information includes the name of the structure type, the description and
volume for each zone, and airflow rates. For a one-zone house only the
first line needs to be completed and only two air flows are needed –
Flow from Outside and Flow to Outside. These flows must be equal to
provide an overall flow balance for the house. For a two-zone house six
airflows are needed – the flow for each zone to and from outdoors plus
the flows in both directions between the two zones. Individual flow
pairs (e.g., flow from and to outside for zone 1) do not necessarily
need to be balanced, but the total flow out of a zone should match the
total flow into it. On the main screen, the program updates the total
flow per zone as the user provides or edits inputs. If the total flows
for any zone are not balanced, then the program raises a flag by
assigning a red font to the zone totals. The model still will run in
such cases, but the user is advised of the possibility of
counterintuitive or misleading results.

The default air flows are balanced, that is, the airflow rate from
outdoors to indoors is equal to that from indoors to outdoors. For
two-zone structures – SF detached/attached homes – the house volume
is equally apportioned between upper/lower stories and identical
indoor-outdoor airflow rates are assumed per story. Further, as an
expedience, it is assumed that the airflow rate between the two indoor
zones is balanced and equal to the indoor-outdoor airflow rate. Thus,
for example, for a SF detached home with a volume of 811 m3 or 405.5 m3
per story, both the indoor-outdoor airflow rate per story and the
interzonal airflow rate between stories are set at 81.1 m3/h.  

If one of the default two-zone homes is selected but the user prefers to
treat the home as a single, well-mixed zone, then FIAM-pwp will add the
zone volumes and will “collapse” the airflow rates to the
single-zone case, meaning that the only flow information needed is the
airflow rates from outdoors to indoors and vice versa. For the
single-zone case these two airflow rates should have the same value. If
the number of zones is changed back to two, then the program will revert
to the volumes and airflow rates originally associated with these house
types. 

Background Concentration. This concentration – the first of eight
inputs on the lower portion of the House Screen – is intended to
reflect the contribution to the indoor formaldehyde concentration of
both the outdoor concentration and indoor sources other than PWPs (e.g.,
tobacco smoke, permanent press fabrics, paints, cosmetics). The FIAM-pwp
default value for the background concentration is 7.5 ppb. If a value of
zero is chosen for the background concentration, then the modeling
results will reflect only the contributions of the indoor sources
selected for the model run via inputs on the Source Screen. 

Emissions Half Life. This input governs the rate at which indoor
formaldehyde concentrations decay over time. The FIAM-pwp default value
is 1.5 years; with this half life, the emission rate 10 years after new
construction would be about 1% of the initial value. With a half life of
3.0 years it would take 20 years (i.e., twice as long) to reach 1% of
the initial value.

Temperature and Humidity. Model algorithms to predict initial indoor
formaldehyde levels include temperature and humidity adjustments for the
emission rate from PWPs. Default values for indoor temperature and
relative humidity are assigned based on the user’s choice (or
creation) of a climate zone. These defaults can be edited directly on
the lower portion of the screen, but if that is done then these new
values would only be associated with the saved run name. Alternatively,
values can be saved for access in any run by creating a new climate zone
or by copying and editing an existing one (see Structure Type and
Climate Zone, earlier in this section).  

Time from ‘Initial Concentration’ and ‘Decay Concentration to'. By
default, FIAM-pwp “decays” initial indoor concentrations that are
calculated under a steady-state assumption to time points that are 3, 6
and 12 months later than the initial point in time. The input labeled
‘Time from Initial Conc’ (see Figure 2-1) enables the user to
specify an additional point in time (in months) over which the model
decays the initial concentration values; this input is set to 24 months
by default. The input labeled ‘Decay Concentration to’, with a
default of 10 ppb, allows the user to specify a concentration for which
the model will calculate the time to decay from the initial modeled
concentration to that value. For a two-zone structure, the model will
calculate this decay time based on the higher of the zone-specific
initial modeled concentrations.

Temperature and Humidity Coefficients. The model initially calculates
indoor concentrations at an indoor temperature of 23 (C and an indoor
relative humidity of 50 percent, and then adjusts the results to the
temperature and humidity conditions selected by the user. The
temperature and humidity coefficients specified here are used by the
model to make these adjustments.

2.2.2	Basis for Defaults

Structure Type. The five default structure types in FIAM-pwp are the
same as those used in the EPA’s Formaldehyde from Composite Wood
Products: Exposure Assessment dated March 2012. Collectively, they
account for the major types reported by the American Housing Survey
(AHS) (  HYPERLINK "http://www.census.gov/hhes/www/housing/ahs/ahs.html"
 http://www.census.gov/hhes/www/housing/ahs/ahs.html ) and the
Residential Energy Consumption Survey (RECS) (  HYPERLINK
"http://www.eia.doe.gov/emeu/recs/"  http://www.eia.doe.gov/emeu/recs/
): single-family (SF) detached homes; SF attached homes; apartments; and
manufactured/mobile homes or trailers. For the purpose of estimating
structure volume, a further distinction was made for manufactured homes
due to size differences – single-wide (SW) versus double-wide (DW).
Camper trailers and park models also are of interest because, like
manufactured homes, these housing types have been used for some
residents displaced by natural disasters such as Hurricane Katrina.
Camper trailers were chosen to represent the small-volume temporary
housing for displaced residents, as this housing type accounted for 80
percent of the smaller-volume structures that were tested as part of
FEMA’s formaldehyde monitoring in temporary housing,.

For SF detached/attached homes, apartments, and manufactured homes,
default structure volumes were calculated based on statistics resulting
from the U.S. Department of Energy’s 2005 RECS, using the reported
average floor area for the most recently built category (years
2000-2005) and assuming a ceiling height of 8.5 feet. For camper
trailers, a typical size used for displaced hurricane victims was
assigned, with an assumed ceiling height of 7 feet. 

The following floor areas and associated dimensions were assumed for the
five structure types:

SF detached home – 3456 sq ft (36 ft x 48 ft, 2 stories)

SF attached home – 2240 sq ft (28 x 40 ft, 2 stories)

Apartment – 1120 sq ft (28 ft x 40 ft)

Manufactured home – 1216 sq ft (16 ft x 76 ft) for SW; 1680 sq ft (28
ft x 60 ft) for DW

Camper trailer – 328 sq ft (8 ft x 35 ft with 4 ft x 12 ft
“slide-out”).

The dimensions chosen for SF homes and apartments were consistent with
their respective average floor areas as reported in the RECS 2005 survey
results.

The dimensions listed above are exterior dimensions; for interior volume
calculations, 0.5 ft was subtracted from both the length and width to
account for the area occupied by exterior cladding, sheathing, studs and
wallboard. For SF homes a simplifying assumption was made that the upper
and lower floors are of identical dimensions. For manufactured homes, a
weighted average of SW and DW volumes was used; based on statistics from
the U.S. Census Bureau’s Manufactured Home Survey (MHS) for years
2000-2009 (  HYPERLINK "http://www.census.gov/const/www/mhsindex.html" 
http://www.census.gov/const/www/mhsindex.html ), 70 percent of the
manufactured homes were assumed to be the double-wide type. The
resultant house volumes are listed, along with the house dimensions, in
Table 2-1.

Table 2-1. House Dimensions and Calculated Volume for Five Housing Types

Housing Type	Number of Stories	Length, 

ft	Width, 

ft	Ceiling Height, ft	Volume, 

ft3 (m3)

SF Detached	2	48	36	8.5	28,666 (811)

SF Attached	2	40	28	8.5	18,466 (523)

Apartment	1	40	28	8.5	  9,233 (261)

Manufactured

   Single-wide

   Double-wide

   Averagea	

1

1

1	

16

28

--	

76

60

--	

8.5

8.5

8.5	

  9,947 (282)

13,908 (394)

12,720 (360)

Camper Trailer

   Main body

   Slide-out	

1

1	

35

12	

8

4	

7

7	

2,147 (61)

a Weighted average of single-wide (30 %) and double-wide (70 %); see
text.

Climate Zone / Temperature and Humidity. Default indoor
temperature-values were determined for each of five U.S. climate zones
using responses to RECS 2005 survey questions about usual temperatures
maintained in the residence. One of the questions – “At what
temperature does your household usually keep your home in the winter?”
– was asked under three conditions: (1) during the day when someone is
home; (2) during the day when no one is home; and (3) during sleeping
hours. The same set of questions also was posed with reference to the
summer season. 

As described in the exposure assessment report cited in Section 1 of
this document, the RECS database was analyzed for these questions to
estimate the year-round average temperature in each climate zone. Table
2-2 summarizes responses to the RECS questions on daytime temperature
when someone is home and nighttime (sleeping hours) temperature by
climate zone. The columns labeled “Overall” are weighted averages of
the daytime and nighttime average temperatures for summer and winter,
with weights of 2/3 for daytime and 1/3 for nighttime. The overall
summer and winter temperatures, in turn, were averaged to derive a
year-round average temperature, by weighting the season-specific
averages by the number of months associated with each. The resultant
year-round temperature estimates are listed in Table 2-3.

Table 2-2. Summer and Winter Indoor Temperatures from RECS 2005, by
Climate Zone

U.S.

Climate

Zone	Average Summer Temperature	Average Winter Temperature

	Day	Night	Overall	Day	Night	Overall

1 (coldest)	73.4	73.3	73.4	69.1	65.8	68.0

2	73.0	73.0	73.0	69.6	67.6	68.9

3	73.2	73.2	73.2	69.8	68.1	69.3

4	73.9	73.2	73.7	70.9	68.6	70.1

5 (warmest)	75.1	74.8	75.0	72.4	70.6	71.8

Table 2-3. Year-round Indoor Temperatures Derived from RECS 2005, by
Climate Zone

U.S.

Climate

Zone	Summer Temperature	Winter Temperature	Year-Round

Temperature

	Average Temperature	Number of Months	Average Temperature	Number of
Months

	1 (coldest)	73.4	3	68.0	9	69.4

2	73.0	4	68.9	8	70.3

3	73.2	4	69.3	8	70.6

4	73.7	5	70.1	7	71.6

5 (warmest)	75.0	7	71.8	5	73.6



The RECS does not ask about humidity levels maintained in the residence.
Consequently, a mass-balance modeling approach was applied to estimate
the year-round average indoor relative humidity (RH) level for each
climate zone. A representative city was chosen for each zone:

Climate Zone 1 – Milwaukee, WI

Climate Zone 2 – Chicago, IL

Climate Zone 3 – New York, NY

Climate Zone 4 – Charlotte, NC

Climate Zone 5 – Houston, TX.

Hourly outdoor temperatures for a typical meteorological year (TMY) were
obtained for each city as inputs to a mass-balance model. The modeling
approach, described in the formaldehyde exposure assessment report cited
in Section 1, is consistent with methods described in American Society
for Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE)
Standard 160P (Criteria for Moisture-Control Design Analysis in
Buildings). In the mass-balance model, moisture in the indoor air is
determined by moisture in the outdoor air, indoor moisture generation by
the occupants, and moisture removal by a central air-conditioning (AC)
system. 

Modeled indoor RH levels for each climate zone are listed in Table 2-4
for two cases – with and without an AC unit. The effect of the AC unit
is most pronounced in Houston, due primarily to the higher outdoor
temperature and, hence, the greater AC run time leading to greater
moisture removal. Because the vast majority of the U.S. housing stock
has air conditioning, the estimated RH values with AC were used in
assigning default humidity values. 

Table 2-4. Modeled Year-round Indoor RH Level, by Climate Zone

Climate Zone	RH Level with AC	RH Level without AC

1 (coldest)	59.1%	63.8%

2	58.1%	64.7%

3	56.3%	66.4%

4	59.8%	72.6%

5 (warmest)	61.4%	85.3%



Air Exchange Rate and Airflows. The default indoor-outdoor airflows in
FIAM-pwp were chosen such that the air exchange rate for each of the
default structure types is 0.2 air changes per hour (ACH). For the
exposure assessment report cited in Section 1, the focus is on PWPs in
new housing; thus, a default value was chosen that lies toward the lower
end of the distribution, as newer structures tend to be better sealed
and have lower air exchange rates. For residences, the Exposure Factors
Handbook (EFH) suggests 0.45 ACH as a central value and 0.18 as a
lower-end value, based on statistics developed by Koontz and Rector
(1995) from a database of perfluorocarbon tracer (PFT) measurement
results. The EFH also cites Murray and Burmaster (1995), who used the
same PFT database to summarize distributions for data subsets defined by
climate region and season; from that analysis, the 10th percentile value
ranged from 0.1 to 0.3 ACH across climates and seasons, averaging about
0.20 ACH.

In a more recent study by Offerman et al. (2008) on ventilation and
formaldehyde in new California homes, the 25th percentile value for air
exchange rate was 0.2 ACH for the two larger subsets of homes that were
studied. In a study by Maddalena (2008) concerning chemical emissions in
four FEMA temporary housing units (camper trailers), the measured air
exchange rate varied from 0.15 to 0.39 ACH across the trailers,
averaging 0.25 ACH. Persily et al. (2010) modeled air infiltration rate
distributions for the U.S. housing stock; for single-family homes built
in 1990 or later, the 25th percentile value was 0.15 ACH and the 50th
percentile value was 0.26 ACH. Given the above statistics, 0.2 ACH was
chosen as the value to represent new homes, including detached/attached
homes, apartments, manufactured homes, and camper trailers. 

For the renovation case that also was examined as part of the
formaldehyde exposure assessment, a central value such as 0.45 ACH, as
suggested in the EFH, may be more appropriate. However, the data on
which that suggested value is based are somewhat dated and there has
been a general tendency toward increasing airtightness of U.S. homes
over time; for example, the median value reported for the Offerman et
al. (2008) study was 0.33 ACH. Thus, a value of 0.40 ACH – double that
used for new homes – is suggested for the renovation case. 

As noted above, the indoor-outdoor airflow rates determine the air
exchange rate. For example, for an apartment with a volume of 261 m3,
the default indoor-outdoor airflow rate of 52.2 m3/h results in an air
exchange rate of 0.2 ACH (i.e., 52.2 m3/h / 262 m3 = 0.2/h). The default
air flows are balanced, that is, the airflow rate from outdoors to
indoors is equal to the rate from indoors to outdoors. For two-zone
structures – SF detached or attached homes – the house volume is
equally apportioned between the upper and lower stories/floors and
identical indoor-outdoor airflow rates are assumed for each story.
Further, as an expedience, it is assumed that the airflow rate between
the two indoor zones is balanced and equal to the indoor-outdoor airflow
rate. Thus, for example, for a SF detached home with a volume of 811 m3
or 405.5 m3 per story, both the indoor-outdoor airflow rate for each
story and the interzonal airflow rate between stories are set at 81.1
m3/h.  

Background Concentration. Although no studies have been conducted with
the explicit purpose of characterizing the background concentration,
such a value conceivably can be inferred from concentration
distributions characterized via field monitoring studies. For example,
if a field study were to focus on existing (as opposed to new) homes,
then those toward the lower end of the concentration distribution
presumably would be older homes with aged PWP sources and, thus,
indicative of the background formaldehyde concentration. As noted in
Section 3.3 of the exposure assessment report cited in Section 1 of this
document, the results of field studies, when examined in this manner,
indicate that an appropriate value for the residential formaldehyde
background concentration would lie between 5 and 10 ppb. The midpoint of
this range – 7.5 ppb – was chosen as the FIAM-pwp default value for
the background concentration. 

Emissions Half Life. As described in Section 3.8 of this document,
suggested values for the formaldehyde emissions half life, reflecting
its decay indoors over time, range from 1.5 to 3 years. An estimate of
2.92 years was based on a cross-sectional analysis of homes of varying
ages, as a proxy for time-series data from one or a few homes. If none
of the materials or furnishings in such homes were changed by the
occupants, then this approach might provide an unbiased estimate of the
half life. However, as a house ages the occupants will tend to add or
replace materials and furnishings, some of which may emit formaldehyde.
The addition of these new sources will result in higher concentration
than in the case where no new sources were introduced. As a result, the
cross-sectional analysis will yield an artificially low estimate of the
decay rate, which means an artificially high estimate of the half life.
Thus, 2.92 years is best viewed as an upper-bound estimate. A more
likely value, based on chamber studies of collections of pressed wood
products as they age, is within the range of 1.5 to 2 years. A half life
of 1.5 years was chosen as the FIAM-pwp default value; with this half
life, the emission rate 10 years after new construction would be about
1% of the initial value.

Temperature and Humidity Coefficients. The default values for these
coefficients – 9799 for temperature and 0.0175 for humidity – are
based on tests by Berge et al. involving two Norwegian particleboard
specimens. These default values were used by HUD in setting the 1984
standards for particleboard and hardwood plywood paneling, and also have
been used in ASTM, ANSI and Composite Panel Association (CPA) standards
or guidelines. The chamber tests underlying these coefficients were
conducted at two temperature settings (22 and 28 (C) and two relative
humidity settings (30 and 60 %). 

An alternative set of coefficients – 8930 for temperature and 0.0195
for humidity – derives from work by Myers, who reviewed the extant
literature and based his recommended temperature coefficient on chamber
testing of about 40 specimens (particleboard and hardwood plywood
paneling) from 11 laboratories. The temperature range of these tests was
from 20 to 40 (C. He also developed a humidity coefficient, but
expressed little confidence in the number; the relative humidity in the
tests on which the correction factor was based ranged from 20 to 90
percent. Thus, his work encompassed more products and wider
temperature/humidity ranges than the Berge et al. study. The HUD
standard was promulgated before Dr. Myers completed his work. 

2.3	Source Screen 

This screen (see Figure 2-5) enables the user to specify the types and
amounts of wood products used or installed in the house, the zone(s) in
which each type is located, and input parameters (slope and intercept)
that govern the formaldehyde emission rate. Following a description of
each type of input, guidance is given on choosing default inputs and/or
adding custom inputs on this screen.

Figure 2-5. FIAM-pwp Source Screen.

2.3.1 	Types of Inputs

PWP Type. There are six default PWP types in FIAM-pwp: 

Oriented strand board (OSB) or softwood plywood (SWPW), typically used
as underlayment

Particleboard, often used as the platform to which a countertop laminate
is affixed

Medium density fiberboard (MDF), often used in cabinet drawers and/or
shelves

Coated CWPs (e.g., interior doors that are primed/painted, vinyl-covered
molding)

Hardwood plywood (HWPW), often used for wall paneling and sides/backs of
cabinets or entertainment centers

HWPW laminate, with uses similar to HWPW but made by affixing a wood
veneer to a particleboard, medium-density fiberboard, or veneer-core
platform.

A distinction is made between HWPW and HWPW laminate because EPA may
regulate these two PWP types differently.

Emission Class. Default values for FIAM-pwp emission-rate parameters
(slope and intercept) have been developed for four emission classes:
Baseline; CARB1; CARB2; and NAF. As described below, these emission
classes correspond to analytical or regulatory options that have been
included by EPA as part of the considerations for a formaldehyde rule.

Slope and Intercept. These two parameters govern the emission rate from
PWPs, including dependence of the emission rate on the indoor
formaldehyde concentration. The slope is a mass transfer coefficient
that reflects the “backpressure” effect of the indoor formaldehyde
concentration on the emission rate, whereas the intercept reflects the
hypothetical emission rate when the indoor formaldehyde concentration is
zero.

Equilibrium Concentration. The equilibrium concentration (mg/m3)
displayed for each source is a calculated quantity – intercept
(mg/m2-hr) divided by slope (m/hr) – that cannot be edited by the
user; it is the theoretical maximum concentration that would occur if
that were the only indoor source, with an arbitrarily large exposed
surface area. The equilibrium concentration is indicative of the
relative strength of emissions from each source. Lower-emitting sources,
such as oriented strand board (OSB) or softwood plywood (SWPW), can act
as net “absorbers” (i.e., indoor sinks) in the presence of
higher-emitting sources, due to the backpressure effect.

Location (Zone) and Exposed Surface Area. If a two-zone structure is
being modeled, then sources can be assigned to either or both of the two
zones. As discussed later (see Section 2.3.2 and Section 2.3.3), one
input line is required for each source type in each zone. The default
amounts of exposed surface area for sources in FIAM-pwp are the same as
those used in EPA’s Formaldehyde from Composite Wood Products:
Exposure Assessment dated March 2012. The exposed surface area (in m2)
for a PWP can be viewed as its loading rate (in m2/m3) multiplied by the
structure volume (in m3).

2.3.2 	Adding Default Inputs

Each input line or row entered on the Source Screen refers to a specific
type of source in a specific zone. For any of the five default structure
types, the simplest way to add a group of sources is using the Add
Default PWP Types button – taking this action loads each of the six
default PWP types in each applicable zone (zone 1 for apartments,
manufactured homes and trailers; zones 1 and 2 for SF detached/attached
structures). Figure 2-6 shows the Source Screen after this action has
been taken. In loading these defaults, there is a choice of Emission
Class (baseline, CARB1, CARB2, or NAF) and Case (new-home or
renovation). The example shown in the figure is for apartments, with
baseline emissions and the new-home case. As indicated in the figure,
user selections are made, or inputs are provided, in the lower portion
of the screen; these choices/entries are then displayed in the upper
portion.

 

Figure 2-6. Source Screen after Adding Default PWP Types

Once added, any default can be removed or edited using the Remove/Edit
buttons to the right. Alternatively, sources can be “turned off” by
un-checking the corresponding check-box to the left. The column to the
left of PWP Type indicates whether a given line or row in the table is a
FIAM-pwp default (D), revision (R) of a default, or newly created (C).
For example, when a default source is edited, the character in this
column will change from D to R. If the line is later changed back to the
default values for area, source and intercept (either by re-editing
values or by using the Restore button), then the character displayed in
the corresponding column will revert to D.

Sources also can be added, one at a time, using the Add button. In
addition to Emission Class and Case, choices need to be made for PWP
Type and Area. The default area for any combination of PWP and zone,
displayed to the right, can be changed before clicking the Add button;
after the button is clicked, the area (or any other value on that line)
can be changed using the Edit button. In either case, any source for
which default values have been changed will be marked with an R in the
second column. All sources chosen or added can be removed at once using
the Clear List button. The maximum number of sources allowed by FIAM-pwp
is 20.

2.3.3 	Adding Custom Inputs

Beyond adding default sources, the user also can add new sources and
associated values using the Add Custom Values button; this is the only
way to add sources if a structure type other than one of the defaults
has been chosen. Inputs are required for the name of the PWP Type, the
Emission Class, the Zone (1 or 2), and values for Area, Slope and
Intercept. Figure 2-7 shows the appearance of the Source Screen after
using this button to add one source line.

Figure 2-7. Source Screen after Adding Custom Values

2.3.3 	Basis for Defaults

As noted above, the default PWP types available in FIAM-pwp are those
used in the EPA 2012 formaldehyde exposure assessment. The four emission
classes in FIAM-pwp, which correspond to analytical or regulatory
options under consideration by EPA for a forthcoming formaldehyde rule,
are described briefly in Table 2-5.  EPA is considering some options
beyond those listed, such as exclusion of HWPW laminates from the CARB1,
CARB2 and NAF options. The basis for these classes is described in some
detail in Section 4.4.8 of the report on EPA’s formaldehyde exposure
assessment, which is reproduced as Appendix C in this document. In
brief, within each emission class or regulatory option, a certain
percentage of each product type is assumed to be in compliance with an
existing standard such as CARB1; these assumed percentages vary across
the emission classes. The procedure for estimating the emission rate for
each product type, assuming that it meets a given standard, is described
below.

Table 2-5. Alternative Analytical/Regulatory Options for EPA
Formaldehyde Rule

Option a	Description

Baseline	Estimated emission levels (year 2013) in the absence of an EPA
rule.

CARB1, Including Laminated Products	All CWPs, including laminated
products, are at emission levels that meet CARB b Phase 1 standards.
Products that meet CARB Phase 1 standards at baseline are assumed to
remain at that emission level.

CARB2/Statutory Option,

Including Laminated Products	All CWPs, including laminated products, are
at emission levels that meet CARB Phase 2 standards. Products that meet
CARB Phase 2 standards at baseline are assumed to remain at that
emission level.

NAF c, Including Laminated Products	All CWPs, including laminated
products, are at emission levels that meet NAF standards. Products that
meet NAF standards at baseline are assumed to remain at that emission
level.

a Corresponds to emission class used in FIAM-pwp

b California Air Resources Board

c No added formaldehyde

Slope and Intercept. Appendix D of this document lists emission-rate
parameters for various types of PWPs that were derived from sets of
chamber tests conducted during the 1980s and up through the mid-1990s.
Although the results from these tests are somewhat dated, an important
distinguishing feature is that all such tests were conducted at multiple
air exchange rates and/or product loading ratios, enabling estimation of
model parameters – slope and intercept – that reflect the dependence
of the emission rate on the indoor formaldehyde concentration. 

Equation 3-6 in Section 3.2 of this document is used as a basis for
determining the intercept that is associated with a product meeting a
certain emission standard, under the assumption that the air entering
the chamber is formaldehyde-free:

	b = CSS * (1 + m * Area / Q) * (Q / Area)	(Eqn. 2.1)

where:

b	=	intercept – the emission rate at zero concentration in the air
(mg/m2-hr)

CSS	=	steady-state formaldehyde concentration in the chamber (mg/m3)

m	=	slope – the mass transfer coefficient (m/hr)

Area	=	exposed surface area of the product (m2)

Q	=	airflow rate into / out of the chamber (m3/hr)

	=	air exchange rate (1/hr) x chamber volume (m3)

ASTM E1333-10 is a commonly used method for demonstrating compliance
with formaldehyde emission standards. The method prescribes an air
exchange rate for testing of 0.5 ACH and product loading ratios of 0.43
m2/m3for PB, 0.95 m2/m3for HWPW, and 0.26 m2/m3for MDF. Using these
rates along with an assumed slope (m) for a given PWP type, the
corresponding intercept can be determined for a hypothetical source that
just meets a given emissions standard. The assumed slope for each PWP
type is the average of the slopes across all sets of emission tests for
that type (after excluding statistical outliers in certain cases), as
listed in Appendix D.

For example, the CARB2 emission standard for MDF is 0.11 ppm or 0.135
mg/m3. An arbitrary chamber volume of 100 m3, which together with an air
exchange rate 0.5/hr yields a value of 50 m3/hr for Q, was used in the
calculation. The volume does not matter here, as the exposed surface
area scales to volume via the loading ratio; with a volume of 100 m3 the
corresponding surface area is 26 m2. A slope of 1.06 m/hr is assumed;
this value is the average slope from about 30 sets of chamber tests of
different MDF specimens. 

Substituting the assumed values given above into Eqn. 2-1, the intercept
(b) that corresponds to the CARB2 emission standard for MDF is
calculated as follows:

b = 0.135 mg/m3 * (1 + 1.06 m/hr * 26m2 / 50 m3) * (50 m3 / 26 m2)

= 0.40 mg/m2-hr

Table 2-6 lists the assumed slopes and calculated intercepts for each
default PWP type, using the procedure described above and assuming a mix
of products meeting different standards within each emission class, as
described in Appendix C.

Table 2-6. Default Slopes and Intercepts for Six Default PWP Types

PWP Type	Slope	Intercept by Emission Class:



Baseline	CARB1	CARB2	NAF

OSB/SWPW a	0.61	0.030	0.030	0.030	0.030

Particleboard	0.70	0.131	0.124	0.122	0.030

MDF	1.06	0.281	0.271	0.269	0.128

Coated CWP	0.52	0.082	0.074	0.071	0.022

HWPW	0.27	0.042	0.025	0.021	0.013

HWPW Laminate	0.27	0.042	0.023	0.021	0.013

		a This relatively low-emitting PWP is not subject to any emissions
standard.

Exposed Surface Area. The default amounts of exposed surface area for
sources are the same as those used in EPA’s  2012 formaldehyde
exposure assessment. As described in Section 4.4.6 of the EPA exposure
assessment report, the following philosophy was adopted in determining
appropriate loading rates for different types of PWPs in various
structure types – technically the specific source (cabinets,
countertop, paneling, etc.) of each PWP type does not matter, as the
model treats all sources in the same way; thus, for any loading scenario
it is sufficient to determine and then sum all exposed surface areas per
PWP type. Floor plans were used as a basis for developing the loadings.
Spreadsheets were developed to document the loadings of PWP components
in items such as cabinets, doors and trim, and to then sum across like
components to develop the resultant loadings for each of six PWP types
listed in the table above.

The default loading rates differ for the new-home vs. renovation cases.
For the renovation case, it was assumed that an entire kitchen was
renovated, including cabinets, countertop and ancillary areas such as a
pantry. It was further assumed that the vinyl covering the kitchen floor
and adjacent hallway floor was replaced with engineered wood. The
interior doors, underlayment and trim throughout the house were assumed
to remain as before the renovation, as were PWPs in any areas beyond the
kitchen other than the adjacent hallway floor. Table 2-7 lists the

FIAM-pwp default values for exposed surface area (in m2), for each of
the six PWP types listed above in each of the five default structure
types, for the case of new construction. Table 2-8 lists the same
information for the renovation case. 



Table 2-7. Default Exposed Surface Areas (in m2) for Six PWP Types, by
Structure Type – New-home Case

Structure Type	OSB/SWPW	Particleboard	MDF	Coated CWP	HWPW	HWPW Laminate

	Zone	Area	Zone	Area	Zone	Area	Zone	Area	Zone	Area	Zone	Area

Apartment	1	  71.480	1	3.255	1	4.645	1	78.165	1	18.137	1	  7.773

Camper	1	  15.675	1	1.020	1	2.090	1	13.005	1	29.575	1	12.675

Mobile Home	1	  98.478	1	3.906	1	4.366	1	65.786	1	26.610	1	11.404

SF Attached Home	1	  71.645	1	1.390	1	2.175	1	52.910	1	  6.2685	1	 
2.687

	2	  71.645	2	5.205	2	4.790	2	57.010	2	15.278	2	  6.548

SF Detached Home	1	110.965	1	1.080	1	2.615	1	65.495	1	  5.474	1	  2.346

	2	110.965	2	3.095	2	4.355	2	65.620	2	16.464	2	  7.056



Table 2-8. Default Exposed Surface Areas (in m2) for Six PWP Types, by
Structure Type – Renovation Case

Structure Type	OSB/SWPW	Particleboard	MDF	Coated CWP	HWPW	HWPW Laminate

	Zone	Area	Zone	Area	Zone	Area	Zone	Area	Zone	Area	Zone	Area

Apartment	1	0	1	3.255	1	4.645	1	39.775	1	15.365	1	  6.585

Camper	1	0	1	1.020	1	2.090	1	  4.470	1	17.017	1	  7.293

Mobile Home	1	0	1	3.906	1	4.366	1	23.153	1	34.710	1	14.876

SF Attached Home	1	0	1	1.390	1	2.175	1	10.340	1	  2.286	1	  0.980

	2	0	2	5.205	2	4.790	2	33.205	2	18.680	2	  8.006

SF Detached Home	1	0	1	1.080	1	2.615	1	11.755	1	  2.286	1	  0.980

	2	0	2	3.095	2	4.355	2	31.790	2	27.342	2	11.718

2.4	Exposure Screen

This screen (Figure 2-8) enables the user to provide inputs related to
the age of indoor sources and a formaldehyde level of interest (LOI)
that is used in certain exposure calculations as well as activity
patterns (time spent in different locations or environments) for
exposure groups (e.g., infants, workers, retires) and formaldehyde
concentrations in environments outside the home. As with the House
Screen, the inputs here can be divided into three screen portions: (1)
Age of Sources and Formaldehyde LOI (upper portion); (2) Exposure Groups
(middle portion); and (3) Concentrations Outside the Home (lower
portion).

 

Figure 2-8. Exposure Screen

2.4.1 	Inputs and Options for Age of Sources and Formaldehyde Level of
Interest

Age of Sources. The assumed age of indoor sources at the start of the
exposure period (i.e., at the time when an individual moves into the
modeled structure) affects the modeled indoor formaldehyde
concentrations to which an individual is exposed – in general, the
lower the age of sources the higher the indoor concentrations at the
start of the exposure period. With a source age of zero years – the
model default – FIAM-pwp begins exposure calculations with an initial
indoor concentration in a newly constructed home, which is intended to
reflect an equilibrium condition shortly (typically within 30 days)
after a house has been loaded with PWPs. A source age of zero years also
would apply to new materials/furnishings added as part of a major
remodeling effort. A model limitation here is that all sources are
assumed to be the same age; FIAM-pwp does not allow different ages to be
specified for different sources.

Formaldehyde Level of Interest (LOI). One of the exposure calculations
is the percent of time an individual is exposed to formaldehyde
concentrations at or above some level of interest. By default, the LOI
is set to 10 ppb; this is easily changeable by entering a different
value.

2.4.2 	Inputs and Options for Exposure Groups

Default Groups. As shown above in Figure 2-8 (middle portion), FIAM-pwp
has six default exposure groups:

Infants (0 to < 2 years of age)

Pre-school/school children (age 2 to < 16 years of age)

Non-industry workers (age 16 to <64)

Fabrication-industry workers (age 16 to <64)

Retirees (age 64 or greater)

Part-time workers (age 16 to <64)

The choice of age breaks to distinguish infants and pre-school/school
children from adults was driven by the possible age dependency of cancer
slope factors. EPA has developed age-dependent adjustment factors
(ADAFs) to address the potential for differential potency associated
with exposure during early life (below 16 years of age). There is a
10-fold adjustment for ages 0 to <2 years, a 3-fold adjustment for ages
2 to <16, and no adjustment for ages 16 and older. Workers were split
into full-time and part-time subgroups; full-time workers were further
divided to distinguish non-industry workers from wood-industry segments
where higher occupational formaldehyde exposures can occur. In terms of
activity patterns, the retiree group also can be used to represent
individuals of age 16 to <64 who are unemployed or working at home.

A subset of the exposure groups can be selected for a model run by
checking or un-checking the check-boxes on the left for each line. The
default values for any group can be modified using the Edit button to
the right; the original/default values can be restored using the Restore
button. As discussed below, user-defined exposure groups also can be
added. The third column from the left on each line indicates whether the
inputs for each exposure group are default (D) values, revisions (R) of
defaults, or newly created (C).

The inputs for exposure groups are annual hours spent at each of five
locations or environments:

In zone 1 at home

In zone 2 at home

At work, school or daycare

In a vehicle

At all other locations

The sum of the default hours at these locations is 8,760, that is, the
number of hours in one year; if the edited values do not sum to 8,760,
then FIAM-pwp displays the sum in red font to indicate that corrections
are in order. For two-zone structures, Zone 1 is conceptualized as
upstairs or the sleeping area and Zone 2 is conceptualized as downstairs
or the living area. For one-zone structures the sleeping and living
areas are grouped together in a single zone.

Custom Groups. User-defined exposure groups can be added via the Add
button, in the screen area below the listing of the default groups.
Figure 2-9 shows illustrative inputs, just before adding them. Inputs
are the name of the group and the annual hours spent in each of the five
locations/environments, along with the formaldehyde concentration at the
work/school/daycare location (see further discussion below, in Section
2.4.3). The hours by location should sum to 8,760; if not, after the
exposure group has been added FIAM-pwp will assign a red font to Total
Hours to indicate that some adjustments are in order. Once added, newly
defined exposure groups can be later removed by the user. A maximum of
three newly defined exposure groups can be added.

For concentrations outside the home, the formaldehyde concentration at
the work/school/daycare location can vary by exposure group. Thus, when
a new exposure group is added the required inputs include not only the
annual hours at each location but also the concentration at the
work/school/daycare location.

Figure 2-9. Preparing to Add a New Exposure Group

2.4.3 	Inputs and Options for Concentrations Outside the Home

Locations outside of, or away from, the home are included in FIAM-pwp to
provide a total or “around-the-clock” exposure perspective. The
concentration values assigned to these locations are displayed on the
lower portion of the Exposure Screen (see Figure 2-8 on previous page).
The default values shown in the figure can be modified by the user. The
changes can be made in either the ppb or µg/m3 column; FIAM-pwp will
automatically convert in either direction using the relationship 1 ppb =
1.23 µg/m3. The model defaults can be restored at any time using the
Load Default button at the bottom of the screen.

2.4.4 	Basis for Defaults

Exposure Groups. The 2009 Exposure Factors Handbook (EFH) update
(External Review Draft) was the primary source of information on
activity patterns, that is, how an individual’s time is allocated
among residential and non-residential locations. Estimates of annual
hours at home and non-home locations are provided for each exposure
group in Table 2-9. The total time spent at home was determined from EFH
tables that indicate the time spent at home in all rooms combined. The
time at home in Zone 1 was defined as the sum of time spent in the
bedroom or bathroom. EFH statistics indicate that infants spend, on the
average across the population, one hour per day at daycare facilities.
Children of school age were assumed to spend 6.5 hours per day at
school, for 180 days per year. Full-time workers were assumed to spend 8
hours per day at work, for 250 days per year. The hours at work or
school for part-time workers and retirees were based on EFH statistics,
as were the hours in a vehicle for each exposure group. The time spent
in “all other” environments was determined by subtraction from the
total time budget of 8,760 hours per year.

Table 2-9. Default Annual Hours at Home and Non-home Locations, by
Exposure Group.

Exposure Group	Location

	Home

 Zone 1 a	Home

 Zone 2	Work or School or Daycare	In

 Vehicle	All

 Other	Total, all Locations

Infants	4,958	1,652	    365	252	1,533	8,760

Pre-school/School 	4,253	1,292	1,170	356	1,689	8,760

Non-industry Workers	3,327	2,032	2,000	590	   811	8,760

Fabrication Workers	3,327	2,032	2,000	590	   811	8,760

Retirees	3,607	3,538	   107	372	1,136	8,760

Part-Time Workers	3,935	2,038	1,000	401	1,386	8,760

a Zone 1 – upstairs (sleeping area); Zone 2 – downstairs (living
area).

Concentrations Outside the Home. Unlike the residential concentrations
calculated by

FIAM-pwp, which are decayed over time, these concentrations are treated
as constant over time. The default concentrations (see Table 2-10)
assigned to non-home locations, which assume a baseline emissions
scenario, were derived from several sources. In-vehicle and
outdoor/other concentrations were based on average values reported from
EPA’s air monitoring network (see Section 3.1 of the formaldehyde
exposure assessment report). The value assigned to vehicles (6 ppb) is
the average reported for the “mobile” land use category; the value
assigned to outdoors/other (3 ppb) is the average of values reported for
“residential” and “commercial” land use categories, with
consideration also given to outdoor formaldehyde levels measured in
residential monitoring studies. The value assigned to the
fabrication-industry workplace was taken from EPA’s occupational
exposure assessment for formaldehyde. The values assigned to daycares,
schools and non-industry workplaces involved calculations that are
described below.

Table 2-10. Default Concentration Values for Non-home Locations

Location	Formaldehyde Concentration

	ppb	µg/m3

Daycare	9.8	12.0

School	8.7	10.7

Work, Non-Industry	10.0	12.3

Work, Fabrication	199.5	244.9

Vehicle	6.0	7.4

Outdoors/Other	3.0	3.7

In developing the baseline concentrations for daycare, school and
non-industry work locations, it was assumed that 10 percent of the
buildings in each category were of recent construction or were recently
renovated. As with the fabrication-industry workplace, the value for
newer buildings associated with the non-industry workplace was taken
from the EPA’s 2011 Assessment of Occupational Exposure to
Formaldehyde from Composite Wood Products report referenced above. Newer
schools were assigned the 95th percentile formaldehyde concentration,
thought to represent newer or recently renovated buildings, from a
relatively large monitoring study of portable and traditional classrooms
in California schools. The value assigned to newer daycare facilities is
an average of the concentration for non-industry work and averages from
field studies, of newer homes. For older daycare, school and
non-industry buildings, the 25th percentile for formaldehyde levels
measured in the EPA BASE study was used. Table 2-11 shows the derivation
of the age-weighted baseline concentration values for these building
types. 

Table 2-11. Derivation of Baseline Values for Daycare, School and Work
Locations

Location	Formaldehyde Concentration, ppb

	Older Buildings	Newer Buildings	Composite a of Older & Newer

Daycare	  7.0	35.0	  9.8

School	  7.0	24.0	  8.7

Work, Non-Industry	  7.0	36.5	10.0

a Composite value = 90% x (value for older buildings) + 10% x (value for
newer buildings).

The values for vehicle and outdoor/other concentrations are assumed to
be the same for different  analytical options (baseline, CARB1, etc.)
whereas those for the other locations can vary (see Table 2-12), because
the formaldehyde concentrations in the fabrication workplace, as well as
newer/renovated daycares, schools and non-industry workplaces, could be
affected by a new rule for formaldehyde. As with the baseline values,
suggested concentration values for other analytical options for newer
non-industry workplaces were taken from the recent EPA report referenced
above. For daycare facilities, the percent reductions across analytical
options used for fabrication and non-industry work were applied. A
similar approach was used for schools, except that the percent
reductions were lowered slightly (i.e., by 10 percent) to account for
generally higher prevailing air exchange rates in schools, which would
tend to temper the effect of formaldehyde emission reductions.

Table 2-12. Concentration Values for Selected Non-home Locations, by
Emissions Scenario

Location	Formaldehyde Concentration, ppb

	Baseline a	CARB1	CARB2	NAF

Daycare b	9.8	9.5	9.4	8.0

School b	8.7	8.1	8.0	7.3

Work, Non-Industry b	10.0	9.4	9.3	7.9

Work, Fabrication	199.5	192.0	190.8	93.0

a FIAM-pwp provides baseline values as the defaults for non-home
locations.

b The values assigned to these locations are a weighted average of those
for older and newer buildings; see text .

In calculating total exposure for the exposure groups, FIAM-pwp combines
the time spent in various locations with concentrations encountered at
those locations. For the work/school/daycare location, FIAM-pwp uses the
daycare concentration for infants, the school concentration for
children, and the work concentration for workers and retirees. When a
new exposure group is added (see Figure 2-9 above), in addition to the
time spent at each location the user also needs to enter an appropriate
value for the concentration at the work/school/daycare location.

2.5	Result Screen

This screen (Figure 2-10) lists modeled indoor formaldehyde
concentrations (upper portion of the screen) together with exposure
estimates (lower portion). Further details are provided below.

 

Figure 2-10. Result Screen with First Two Exposure Groups Selected

2.5.1 	Concentration Results

FIAM-pwp lists the concentrations calculated for Zones 1 and 2 under the
temperature and humidity conditions entered by the user; the Zone 2
results are set to the background concentration if the user has
specified a one-zone structure. Section 3.9 describes the procedures and
gives the equations that are used to adjust modeled concentrations to
the user-specified conditions. Concentrations are reported both in
volume/volume (ppb) and mass/volume (µg/m3) units. 

Initial concentrations – those listed for 0 months under Time from
Initial Concentration – reflect the condition shortly after a house
has been loaded with PWPs. As described in Section 3.7, estimates of
initial indoor concentrations due to PWPs are decayed exponentially by
the model, in accordance with a user-specified emissions half life
(default of 1.5 years), to develop estimates for subsequent points in
time. Results are provided for time points that are 3 months, 6 months
and 12 months later than that for the initial concentration values, plus
a user-specified point in time (24 months by default). The
user-specified value for length of time over which to decay the initial
concentration is entered on the House Screen, where the default
emissions half life also can be changed.

The results also include the time required for the highest initial
indoor concentration to drop below a user-specified concentration (an
input on the House Screen). The equation used for this calculation is
given in Section 3.7. By definition, the indoor concentration cannot
decay to a value that is lower than the background concentration (an
input on the House Screen). If the value to which the indoor
concentration is to be decayed is less than or equal to the background
concentration, then the calculation result is undefined. In such cases,
FIAM-pwp reports 0 months as the time to decay.

As described in greater detail in Section 3, the model calculates the
initial indoor concentration(s) under an assumed steady-state condition.
In real life, a true steady-state is never achieved, due to the dynamics
of formaldehyde source emissions and indoor sinks. Formaldehyde sources
typically do not emit at a constant rate, but rather at a rate that
tends to decline over time as the “reservoir” of initial free
formaldehyde, coupled with that formed through hydrolysis, is gradually
depleted. Although a true steady-state is never reached, it has been
observed in field studies that the combined actions of sources and sinks
will result in an initial rise in the indoor concentration, followed by
a “leveling off” period that may be fairly brief and then a longer
period of gradual decline in concentration. The initial
“steady-state” concentrations estimated by FIAM-pwp are intended to
approximate the indoor concentrations during the leveling-off period,
which typically would occur within 30 days or less following the
installation of PWPs.

2.5.2 	Exposure Results

An average daily concentration (ADC) is reported for each zone of the
house, or other modeled structure, and for each default or added
exposure group selected by the user. The percent of time during which
the concentration in each zone is greater than a user-specified level of
interest (LOI) is also reported. For each exposure metric, annual
averages are reported for the first modeled year and for the next ten
years thereafter. ADC calculations for exposure groups are based on time
spent, and concentrations encountered, in the two zones of the house
together with times and concentrations when the individual is out of the
house. Concentrations in the house are calculated by the model and decay
over time, whereas out-of-house concentrations are input by the user on
the Exposure Screen and are assumed to be constant over time. Further
details on exposure calculations, including equations, are provided in
Section 3.10.

2.5.3 	Printing Inputs and Results

An option at the top of the Result Screen – Generate pdf – enables
the user to create a pdf file that contains both the inputs and results
for a model run. The pdf mimics the appearance of the input screens and
creates a table listing the same model outputs as shown on the Result
Screen. 

2.5.4 	Archiving Inputs and Results for a Run

Another option at the top of the Result Screen – Download Excel –
enables the user to save a file in Excel format that contains all inputs
and results for the current model run. The saved Excel file has three
sheets or tabs, named Combined, Inputs and Results. FIAM-pwp creates an
Excel file for each run name; if multiple runs are made under the same
run name and the download button is pressed after each run, then
FIAM-pwp will append each set of inputs and results to the Excel file.
The Inputs sheet has one column for each input and creates one Excel row
for each model run. The Results sheet has columns for each output and
creates 11 rows for each model run, because ADC results are provided for
Zone 1, Zone 2, and exposure groups for each of 11 years (year 0-1, year
1-2, etc.). FIAM-pwp concentration results, which relate to single
points in time, are duplicated across the 11 rows. Similarly, the
Combined sheet containing inputs and results has 11 rows and duplicates
both the inputs and the modeled concentrations across the 11 rows that
are generated for each run.

2.6	Model Access and Additional Features

2.6.1 	Accessing the IGEMS Website

The IGEMS website that houses FIAM-pwp can be accessed at   HYPERLINK
"https://ofmpub.epa.gov/igems-jsp/"  https://ofmpub.epa.gov/igems-jsp/ .
After the Privacy Act Notification has been acknowledged, a user can log
in if he/she has an account (otherwise set up a new account). After
logging in, a user can access FIAM-pwp by selecting it benath
Formaldehyde Indoor Air Model from the choices to the left.

2.6.2 	Saving Inputs for Later Access

The first time FIAM-pwp is used, the Run FIAM button at the bottom will
not be available until inputs have been provided or choices have been
made on the House Screen, Source Screen and Exposure Screen. When the
run button is pressed, FIAM-pwp will prompt for a Run File Name. Up to
20 characters can be used in naming the file. After the SAVE button is
pressed, FIAM-pwp will show the Result Screen. A new run file can be
created, or a current one saved, at any time – creating, opening and
saving run files works much like similar functions in Microsoft Windows.
Once a run file has been named and saved, it will be included on the
list that can be accessed by pressing the Open File button. If the SAVE
button is pressed using a file name that already exists, FIAM-pwp will
ask whether the existing run file should be overwritten. 

2.6.3 	Context-sensitive Help

FIAM-pwp provides context-sensitive help via a series of buttons with a
“?” icon. When the icon is pressed a dialog box appears with help
specific to the input or option with which it is associated. Pressing
the Close button returns the user to the input or output screen. There
is also an Overview button for each of the input screens that; when this
button is pressed FIAM-pwp displays a brief description of the purpose
of the screen and the types of inputs required. All of the help dialogs
provided by FIAM-pwp can be saved by pressing the Save FIAM Help button
that is accessed on the Result Screen.

2.7 	Example Applications

Six examples are provided in this section to illustrate certain FIAM-pwp
features and options as well as the insights that can be obtained from
making model runs, particularly a series of related runs.

2.7.1 	Example 1 – Modeling with Model Default Values

This example illustrates how FIAM-pwp results can be obtained quickly
when using default values for items such as type of structure, climate
zone, indoor sources, and exposure groups. As described in previous
sections, the default values reflect decisions or choices made in the
EPA 2012 formaldehyde exposure assessment (see footnote 10 in Section
1). The example further illustrates how minor changes from the defaults
can be made to view complementary results.

The key choices for this example are indicated in the run notes provided
in Table 2-13 for Part 1. First, on the House Screen, Apartment (a
one-zone structure) was selected as the structure type and Zone 5 was
selected as the climate zone. Second, default PWP types were loaded on
the source screen with the choices of baseline emissions and new
construction (as opposed to renovation). Third, default exposure groups
were chosen on the Exposure Screen. Model defaults were used for all
other inputs not mentioned above. The modeling results indicated an
initial indoor formaldehyde concentration of 78.6 ppb.

Table 2-13. Run Notes for Example 1

Screen	Notes

Part 1

House	Select Apartment and Climate Zone 5; defaults otherwise

Source	Select "Add Default PWP Types" with Emission Class = baseline,
Case = new-home (six sources will appear, all checked)

Exposure	Select default exposure groups (no action needed; all checked
by default)

Result	Initial concentration – 78.6 ppb (97.1 µg/m3) for Zone 1, N/A
for Zone 2

Part 2

House	Keep Apartment and Climate Zone 5

Source	Select "Add Default PWP Types" with Emission Class = CARB2, Case
= new-home (Note: clear previous list before adding new sources)

Result	Initial concentration – 68.5 ppb (84.6 µg/m3) for Zone 1, N/A
for Zone 2

Part 3

House	Keep Apartment and Climate Zone 5; change air exchange rate to 0.4
ACH by changing flows to and from outside from 52.26 to 104.52 m3/h

Source	Select default PWP types – Emission Class = baseline, Case =
renovation (Note: clear previous list before adding new sources) 

Result	Initial concentration – 49.0 ppb (60.5 µg/m3) for Zone 1, N/A
for Zone 2

Part 4

House	Select SF Detached; keep Climate Zone 5

Source	Program will issue a warning due to change in structure type;
click OK.

Select default PWP types – Emission Class = baseline, Case = new-home 

(Note: clear previous list before adding new sources; 12 sources will
appear)

Result	Initial concentration – 57.1 ppb (70.5 µg/m3) for Zone 1, 

59.9 ppb (74.0 µg/m3) for Zone 2



Next (see Part 2), the emission class was changed to CARB2, which has
lower emissions than baseline. All other inputs were kept the same as
for Part 1. The modeled initial concentration in this case was 68.5 ppb.
Then (Part 3) the emission class was changed back to baseline while
choosing the loading rates for renovation instead of new construction
and changing the air exchange rate from 0.2 to 0.4 ACH, by doubling the
airflow rates to/from outdoors; the modeled initial concentration in
this case was 49.0 ppb. Last (Part 4), baseline emissions and new
construction were chosen while changing the structure type to SF
detached (a two-zone structure). The initial modeling results in this
case were 57.1 ppb for Zone 1 (sleeping area) and 59.9 ppb for Zone 2
(living area).

2.7.2 	Example 2 – Modeling a Hypothetical Chamber Test for CARB
Compliance

This example is drawn from the 2012 EPA exposure assessment report
previously cited in Section 1. ASTM E1333-10, Standard Test Method for
Determining Formaldehyde Concentrations in Air and Emission Rates from
Wood Products Using a Large Chamber (ASTM 2010), is commonly used to
demonstrate compliance with formaldehyde emission standards. The method
prescribes an air exchange rate for testing of 0.5/hr and product
loading ratios of 0.43 m2/m3for PB, 0.95 m2/m3 for HWPW, and 0.26
m2/m3for MDF. Using these rates and an assumed slope, the intercept can
be determined for a hypothetical source that just meets a given
emissions standard. The CARB2 emission standard for MDF is 0.11 ppm (110
ppb) or 0.135 mg/m3. Per the 2012 EPA exposure assessment, a MDF
specimen complying with the standard would have an intercept value of
0.40 mg/m2-hr, assuming a slope of 1.06 m/hr based on historical chamber
tests of MDF specimens. 

FIAM-pwp was used to model a hypothetical chamber test with an MDF
specimen that just meets the standard (see notes for this example in
Table 2-14). An assumed air exchange rate of 0.5/hr and chamber volume
of 100 m3 equates to an airflow rate of 50 m3/hr into and out of the
chamber. The volume assumed does not matter, as the exposed surface area
scales to the volume via the loading ratio – with a volume of 100 m3
the corresponding MDF surface area is 26 m2. The assumed slope was 1.06
m/hr, the average value from ~ 30 historical tests of various MDF
specimens (see Appendix D of this report). The modeling result – an
initial concentration of 108.5 ppb – confirms compliance of the
hypothetical MDF specimen with the standard (110 ppb). 

Table 2-14. Run Notes for Example 2

Screen	Notes

House	Use “Add Your Own Structure Type” to create new type named
Chamber Zone 1 (Chamber) – volume = 100.0 m3; flow to/from outside =
50.0 m3/hr

Check box next to “Check here if one-zone case”

Create climate zone named Standard Conditions, 23 °C and 50 % RH

Change background concentration to 0.0 ppb

Source	Add custom values for PWP type MDF, Emission Class Meet CARB2 110
ppb

Area = 26.0 m2, Slope = 1.06 m/hr; Intercept = 0.4 mg/m2-hr

Exposure	No exposure group selected (uncheck all)

Result	Initial concentration = 108.5 ppb (134.1 µg/m3)

2.7.3 	Example 3 – Demonstrating an Equilibrium Concentration

The default MDF source in FIAM-pwp has a baseline emission rate
characterized by an intercept of 0.28122 mg/m2-hr and a slope of 1.06
m/hr. The formaldehyde equilibrium concentration, obtained by dividing
the intercept by the slope, is 0.2653 mg/m3 or 265.3 µg/m3. Attainment
of an equilibrium concentration can be demonstrated via modeling.
FIAM-pwp was run with a hypothetical chamber volume of 100 m3 and an
airflow rate of 50 m3/hr, as in the previous example. An arbitrarily
large exposed surface area of 100,000 m3 was assumed. The modeled
concentration in the chamber (265.2 µg/m3; see run notes in Table 2-15)
is consistent with the calculated value for the equilibrium
concentration.

Table 2-15. Run Notes for Example 3

Screen	Notes

House	Use created structure type named Chamber

Zone 1 (Chamber) – volume = 100.0 m3; flow to/from outside = 50.0
m3/hr

Check box next to “Check here if one-zone case”

Use created climate zone named Standard Conditions, 23 °C and 50 % RH

Change background concentration to 0.0 ppb

Source	Add PWP type MDF, Emission Class = baseline, Case = new-home

Area = 100,000 m2, Slope = 1.06 m/hr; Intercept = 0.28122 mg/m2-hr

Exposure	No exposure group selected (uncheck all)

Result	Initial concentration = 214.6 ppb (265.2 µg/m3)

2.7.4	Example 4 – Examining the Incremental Contribution of a Product

As described in Section 2.1, a prior online version of FIAM-pwp (v1.0)
was subjected to a peer review. In commenting on the backpressure
effect, one of the reviewers noted the following:

Assessment of acute/chronic exposure is associated with different
uncertainties and conceptual models. In the acute setting, the
backpressure effect may attenuate the contribution of a new product such
that the total acute indoor exposure concentration is not affected.
However, in a chronic or long-term timeframe, it is reasonable to expect
that the majority of available formaldehyde and precursors ultimately
will partition from the product into the indoor space. My recommendation
is that a conceptual framework be developed or documented to ensure that
users take these factors into consideration when investigating acute
versus chronic timeframes. For the acute timeframe, it is important to
consider the total collection of assembled indoor products. For chronic
exposure, a more accurate characterization of dose attributable to a
single product is likely obtained by considering the decay profile of
that product in the absence of other products; the attenuation due to
the backpressure effect is temporary and, eventually, a source that is
initially attenuated will begin to emit formaldehyde as the indoor-air
concentration decreases. 

To illustrate some of the reviewer’s points, the incremental
contributions of illustrative CWPs were examined by modeling them both
as a single source and in the presence of other CWPs. As shown in Figure
2-11, FIAM -pwp has loadings of six default PWP types for apartments;
these were used in a series of model runs with an apartment as the
structure types (see run notes in Table 2-16). For these runs, the
background concentration was set to 0.0 ppb.

Figure 2-11. Emission-rate Parameters for FIAM-pwp Default PWPs in an
Apartment.

For the first run, only MDF was selected (by checking the corresponding
check-box), resulting in a modeled initial concentration of 23.4 ppb.
Next, the model was run checking all sources except MDF, with a
resultant modeled initial concentration of 70.7 ppb. The modeled initial
concentration when all sources, including MDF, were selected was 77.3
ppb – less than the sum of results from the first two runs and, thus,
illustrating the suppression of emissions in the presence of other
sources. A similar series of runs was made with OSB/SWPW as the focus.
The modeled concentration for OSB/SWPW only was 22.9 ppb – similar to
that for MDF, but with an exposed surface area about 15 times as large.
The modeled concentration with all sources except OSB/SWPW (88.3) ppb)
was higher than that obtained with all sources, including OSB/SWPW (77.3
ppb). This outcome illustrates that OSB/SWPW acts as a “net
absorber” in the presence of the other default PWP types used in the
example. The relative emissions strength of each PWP type is apparent
from its equilibrium concentration, which for MDF is more than five
times that of OSB/SWPW (equilibrium concentrations are listed in Figure
2-11).

Table 2-16. Run Notes for Example 4

Screen	Notes

Part 1

House	Select Apartment and Climate Zone 5; background = 0; defaults
otherwise

Source	Select default PWP types – Emission Class = baseline, Case =
new-home

Check only the MDF type

Exposure	No exposure group selected (uncheck all)

Result	Initial concentration – 23.4 ppb (28.9 µg/m3)

Part 2

House	Keep Apartment, Climate Zone 5, and background = 0

Source	Keep default PWP types – Emission Class = baseline, Case =
new-home

Check all PWP types except MDF

Result	Initial concentration – 70.7 ppb (87.4 µg/m3)

Part 3

House	Keep Apartment, Climate Zone 5, and background = 0

Source	Keep default PWP types – Emission Class = baseline, Case =
new-home

Check all PWP types, including MDF

Result	Initial concentration – 77.3 ppb (95.5 µg/m3)

Part 4

House	Keep Apartment, Climate Zone 5, and background = 0

Source	Keep default PWP types – Emission Class = baseline, Case =
new-home

Check only the OSB/SWPW type

Result	Initial concentration – 22.9 ppb (28.3 µg/m3)

Part 5

House	Keep Apartment and Climate Zone 5

Source	Keep default PWP types – Emission Class = baseline, Case =
new-home

Check all PWP types except OSB/SWPW

Result	Initial concentration – 88.3 ppb (109.0 µg/m3)



2.7.5	Example 5 – Varying the Decay Rate (Half Life)

As discussed later in this report (see Section 3.5), several peer
reviewers suggested multiple or changing decay rates to represent, for
example, “fast” (earlier) and “slow” (later) periods of
exponentially declining formaldehyde emissions. Such an undertaking,
although a “stretch” of FIAM-pwp’s capabilities, is possible
provided that adequate care is taken. The key to varying the decay rate
is having a means to predict the initial concentration for the point in
time at which this rate is changed.

For this example it was assumed that the fast decay could be represented
by a half life of one year and the slow decay by a half life of three
years. The fast-decay period was assumed to last one year, followed by a
slow-decay period of nine years. The resultant indoor-concentration time
series over the 10 years, calculated externally from the model, is shown
in Figure 2-12 together with the default background concentration (7.5
ppb) that was assumed for the example. The change in the decay rate can
be seen as the “kink” in the time series at an elapsed time of one
year.

 

Figure 2-12. Formaldehyde Concentration Decline with Fast and Slow Decay
Periods.

Inputs for the example were simplified by assuming a single emitter –
MDF. For the first run (Part 1; see Table 2-17), a half life of 1.0
years was assumed together with an MDF exposed surface area of 18.5 m2;
the modeled initial concentration was 63.7 ppb and the modeled
concentration after one year was 35.6 ppb. For the second run (Part 2;
see Table 2-18), the assumed half life was changed to 3.0 years. For
this run it also was necessary to adjust the source inputs such that the
modeled initial concentration would be 35.6 ppb. Through trial and error
it was determined that this outcome could be obtained by changing the
intercept for the MDF emission rate from the default value of 0.28122 to
0.1455 mg/m2-hr. The modeled initial concentration in this case was 35.6
ppb and the modeled concentration after 108 months (9 years, or 10 years
cumulative) was 11.0 ppb. The collective results from the two runs are
consistent with the time series in Figure 2-12.

Table 2-17. Run Notes for Example 5, Part 1 (fast decay period)

Screen	Notes

House	Select Apartment as the structure type

Use created climate zone named Standard Conditions, 23 °C and 50 % RH

Half life = 1.0 years

Source	MDF/baseline/new-home with surface area = 18.35 m2 and default
values for slope/intercept (add only this type with default surface
area; edit surface area)

Result	Initial concentration = 63.7 ppb (78.7 µg/m3)

Concentration at 12 months = 35.6 ppb (44.0 µg/m3)

Table 2-18. Run Notes for Example 5, Part 2 (slow decay period)

Screen	Notes

House	Change half life to 3.0 years

Change time from initial concentration to 108 months (to get to 10
years)

Source	Change intercept for MDF to 0.1455 mg/m2-hr (by editing)

Result	Initial concentration = 35.6 ppb (44.0 µg/m3) 

Concentration at 108 months = 11.0 ppb (13.6 µg/m3) 

3.	MODEL CALCULATIONS AND ASSUMPTIONS

This section describes the technical basis for the calculation routines
in the software, which consist of two major components. The first of
these predicts the initial “steady-state” formaldehyde
concentration(s) in one or two zones of the house due to the combined
actions of sources and sinks in a newly constructed home. The primary
input for this component is an empirical emission-rate algorithm (i.e.,
estimated initial emission rate as a function of formaldehyde
concentration and other environmental variables) for various types of
pressed-wood products. For more information regarding formaldehyde
emissions from pressed-wood products, the reader is referred to an EPA
report.

The second component of the model is a decay function that gradually
reduces the estimated initial steady-state concentration(s) due to
indoor sources and sinks over a specified period of time. The
concentrations are reduced in accordance with an exponential decay
function, using a decay rate that is determined from a user input for
the formaldehyde emissions half life. With this function, the
concentrations at any point in the “life” of a home can be
predicted, on the assumption that the materials in the house do not
change over time and the air exchange rate, indoor temperature and
indoor humidity also remain constant.

The first model component is based on the fundamental principle of
conservation of mass in an indoor environment, as described in Section
3.1. A model (see Section 3.2) was developed during the 1980s to predict
the formaldehyde concentration in a single indoor compartment under an
assumed steady-state condition, taking into account the dependence of
formaldehyde emission rates on the prevailing indoor concentration. This
model has been extended (see Section 3.3) to predict formaldehyde
concentrations in two indoor compartments under the same assumption.
Whether a one-chamber or two-chamber model is used in the calculations
depends on the type of structure selected by the model user (see Section
2.2).

An alternative formulation of the model, with an embedded calculation of
the equilibrium concentration, is described and discussed in Section
3.4. Other potentially useful modeling constructs are presented and
discussed in Section 3.5. Although the model, as currently formulated,
technically has the capability to represent indoor sinks, this model
feature has been dropped due to several uncertainties together with the
lack of supporting data. Some discussion of this topic is provided in
Section 3.6.

Section 3.7 describes the basis for the equation used to decay the
initial steady-state concentration(s) over time and provides examples to
illustrate calculations for concentrations at any point in time and for
the time to reach a pre-defined concentration such as 10 ppb. Section
3.8 documents the basis for selection of a default value for the
formaldehyde concentration half life, which is needed to decay the
concentration over time. Section 3.9 describes the equations used to
adjust indoor formaldehyde concentrations calculated by the model, under
baseline indoor conditions of 23 ( C and 50 % relative humidity, to the
user’s inputs for temperature and humidity conditions. Section 3.10
describes the exposure calculations used in the model.

3.1	Generalized Mass-Balance Equation

The formaldehyde concentration in an indoor environment is dependent on
two types of factors: (1) those that increase the concentration, such as
emissions from formaldehyde sources, and (2) those that decrease the
concentration, such as dilution with outdoor air assumed to be lower in
concentration than that indoors. Although the two types of factors would
appear to be distinct entities, they actually are interdependent because
of formaldehyde’s general behavior – emission rates from
pressed-wood products are dependent on the concentration in the indoor
volume, and the formaldehyde concentration, in turn, is dependent on
these emission rates as well as removal mechanisms.

The principle of conservation of mass in an indoor environment is
fundamental to the formaldehyde modeling process. Simply stated, this
principle means that the rate of change over time in the formaldehyde
mass indoors is determined by the rates of generation and removal. This
principle can be expressed mathematically for an indoor compartment or
zone as follows:

 	(Eqn. 3-1)

where:

V	=	indoor volume (m3)

C	=	formaldehyde concentration (mg/m3)

t	=	time (hr)

i, j	=	indoor compartments (0 signifies the outdoors)

Q	=	airflow rate between indoor zones or between indoor zone and
outdoors (m3/hr)

Ei,k	=	emission rate for formaldehyde source k in zone i (mg/hr)

Si,p	=	sorption rate for formaldehyde absorbent p in zone i (mg/hr).

Model results are presented in both volume/volume units (ppb) and
mass/volume units (µg/m3). Conversion between these unit conventions is
performed using the ideal gas law, assuming constant pressure (1
atmosphere) and the user-specified temperature.

3.2	Steady-State Model for Initial Formaldehyde Concentration in One
Compartment

A model developed in the 1980s by Matthews and colleagues was designed
to estimate the steady-state formaldehyde concentration resulting from
emission sources in a single indoor compartment. The following
discussion is in large part excerpted from Matthews et al. The
underlying model theory and its derivation are described in the same
document and in a paper by Matthews et al.

At steady state, the formaldehyde (CH2O) concentration in a single
compartment can be expressed as follows:

[CH2O]SS = [CH2O]out + CH2OER/(F * ACH * VOL)	(Eqn. 3-2)

where:

[CH2O]SS	=	steady-state formaldehyde concentration inside the
compartment (mg/m3)

[CH2O]out	=	formaldehyde concentration outside the compartment (mg/m3),
assumed to be constant over time

CH2OER	=	emission rate of a formaldehyde source inside the compartment
(mg/h)

F	=	fraction of air coming into the compartment that mixes within the
volume (i.e., the mixing factor)

ACH	=	air exchange rate with outdoors for the compartment (hr-1)

VOL	=	volume of the compartment (m3).

Assuming that F is equal to unity and using Q (airflow rate into and out
of the compartment, in m3/hr) to denote the product of ACH * VOL,
Equation 3-2 becomes:

[CH2O]SS = [CH2O]out + CH2OER/Q	(Eqn. 3-3)

Application of the model as expressed in Equation 3-3 is simplified by
assuming that all parameters in the equation remain constant (at steady
state) and that there are no permanent losses of formaldehyde due to
irreversible sorption to sinks.

Emissions from pressed-wood sources are area-dependent in that the
magnitude of the emission is a direct function of the surface area
(Area) of the source in the compartment. The equivalent of Equation 3-3
for an area-dependent source is:

[CH2O]SS = [CH2O]out + (CH2OER( * Area)/Q	(Eqn. 3-4)

with CH2OER( in units of mg/m2-hr and Area in m2.

The formaldehyde emission rate, CH2OER(, is formulated consistent with
a one-dimensional representation of Fick’s Law, describing the
emission rate as the release rate when the formaldehyde air
concentration is zero, reduced by the product of the mass transfer
coefficient and the vapor-phase concentration. This formulation can be
expressed as follows:

CH2OER( = -m * [CH2O]V + b		(Eqn. 3-5)

where:

m	=	the mass transfer coefficient (m/hr)

[CH2O]V	=	the CH2O concentration in the vapor phase (mg/m3)

b	=	a constant; the emission rate at zero CH2O concentration in the air
(mg/m2-hr).

As noted by Matthews, Hawthorne and Thompson (see footnote on this
page):

Thus, the modeled CH2O emission rates maximize at zero CH2O
concentration and decrease linearly to a net emission rate of zero at
some equilibrium CH2O concentration. If the CH2O concentration is raised
beyond the equilibrium level with the addition of new CH2O sources,
extrapolation of the model to higher CH2O concentrations indicates that
sorption of the CH2O by the original ‘emitter’ is expected to occur.
Experimental results of individual and paired emitters in small-scale
environmental chambers by Pickrel et al. and Godish et al. are
consistent with this theory. Paired source contributions of pressed-wood
products, insulation, carpeting and UFFI typically yielded CH2O
concentrations less than the sum of the concentrations for the
individual emitters. In addition, Godish found that the CH2O emission
strength of many weaker emitters in paired source combinations was
enhanced following brief exposure to a stronger emitter. This suggests a
temporary sorptive mechanism for CH2O followed by reemission of CH2O at
reduced CH2O concentrations.

The vapor-phase concentration, [CH2O]V, by definition is equal to
[CH2O]SS at steady state; therefore, substitution of [CH2O]SS and
Equation 3-5 into Equation 3-4 results in the following expression:

 	(Eqn. 3-6)

In the case of multiple pressed-wood sources, Equation 3-4 becomes:

[CH2O]SS = [CH2O]out + ((CH2OER(k* Areak) /Q	(Eqn. 3-7)

where the subscript k refers to the kth source. By extension, Equation
3-6 then becomes:

 	(Eqn. 3-8)

3.3 	Extension of the Steady-State Model to Two Compartments tc \l2
"3.3 	Extension of the Steady-State Model to Two Compartments 

The formulation for the single-compartment case is adapted for use with
a two-compartment case, by solving appropriate mass-balance equations.
The mass-balance equation for each compartment, neglecting sinks (see
Section 3.6), is as follows:

 	(Eqn. 3-9)

where:

Vi	=	volume of compartment i (m3)

Ci	=	concentration in compartment i (mg/m3)

t	=	time (hrs)

i,j	=	indoor compartments (0 signifies the outdoors)

Qij	=	flow rate from compartment i into compartment j (m3/hr)

Ei,k	=	emission rate for formaldehyde source k in zone i (mg/hr)

It is acknowledged here that there are formaldehyde sources other than
pressed-wood products in the indoor environment; however, such emitters
(e.g., tobacco smoke, permanent press fabrics, paints, cosmetics)
typically are comparatively weak. Together, the outdoor concentration
and these weaker emitters contribute to what can be described as a
"background" concentration. As an expedience, for purposes of this
model, the background concentration is treated as if all such
contributions can be represented by the outdoor concentration. Studies
of background concentrations indicate that the concentrations indoors
are generally offset from outdoors due to these other sources by
approximately 5 ppm (see Section 2.2.2). Although the emission rate of
these other formaldehyde sources can be impacted by the presence of new
sources such as PWPs, this impact is expected to minimal. Therefore,
treating these other sources as an outdoor contribution provides a
meaningful construct for isolating the impact of the new sources, by
providing a constant background offset.

With this representation, equation 3-9 can be alternately expressed as
follows:

 	(Eqn. 3-10)

where:

CB	=	background concentration, which is defined as the outdoor
concentration plus the contribution from indoor sources other than
pressed-wood products (assumed to be constant over time).

Considering the source equation as described in Equation 3-5 for
multiple pressed-wood products in either zone, the sum of these emission
sources in each compartment can be expressed as:

 	(Eqn. 3-11)

where:

CH2OER(i,k	=	unit emission rate of source k located in compartment i
(mg/m2-hr)

Areai,k	=	area of source k located in compartment i (m2)

mi,k	=	mass transfer coefficient of source k located in compartment i
(m/hr)

bi,k	=	emission rate at zero CH2O air concentration for source k located
in compartment i (mg/hr)

Ci	=	CH2O air concentration in compartment i (mg/m2).

Considering the combined Equations 3-10 and 3-11 for each compartment,
and solving simultaneously for the steady-state condition (see Appendix
A), the following equations are obtained:

 	(Eqn. 3-12)

 	(Eqn. 3-13)

where:

SS	=	indicates the steady-state solution

   for the k sources located in compartment 1

   for the k sources located in compartment 1.

   for the k sources located in compartment 2

   for the k sources located in compartment 2.

3.4 	Matthews Model Compared to HBF Model

An alternative to the Matthews model is the HBF model (name coined by
Myers), a model developed by Hoetjer for which mathematically identical
versions have been developed by Berge et al. and by Fujii et al. The
underlying theory and derivation of the model have been summarized by
Myers.

The HBF model formulation for the steady-state concentration for a
single source is as follows:

CSS = CEQ / (1 + N / kL) 	(Eqn. 3-14)

where:

CSS	=	the steady-state formaldehyde concentration (mg/m3)

CEQ	=	the formaldehyde equilibrium concentration (mg/m3)

N	=	the air exchange rate (1/hr)

k	=	mass transfer coefficient (analogous to m in Matthews model) (m/hr)

L	= 	product loading ratio (m2/m3)

It can be shown that the Matthews and HBF formulations are identical
models and that CEQ = b/m. The only difference between the two models is
in the regression equations that are used to derive the parameters, not
their theoretical constructs.

The Matthews formulation, under the assumption that the outdoor
concentration is zero, can be expressed as follows by substituting
Equation 3-5 into Equation 3-4:

CSS = (-m * CSS + b) * Area / Q	(Eqn. 3-15)

where CSS is as defined above and m, b, Area and Q are as defined
previously in Section 3.2. 

Substituting ACH * VOL for Q, Equation 3-15 can be expressed as:	

	CSS = (-m * CSS + b) * Area / (ACH * VOL) 	(Eqn. 3-16)

Equation 3-16 can be expressed using the same terms as in Equation 3-14,
by setting N = ACH and by setting L = (Area / VOL):

	CSS = (-m * CSS + b) / (N/L) 	(Eqn. 3-17)

Equation 3-17 can be solved for CSS as follows:

 	(Eqn. 3-18)

Setting the Matthews (Equation 3-18) and HBF (Equation 3-14)
formulations equal to each other yields the following equality:

 		(Eqn. 3-19)

Equation 3-19 can be solved for CEQ as follows:

	(Eqn. 3-20)

Therefore, setting CEQ = b/m leads to identical model formulations.

3.5 	Other Potentially Useful Modeling Constructs

One of the peer reviewers commented on the potential utility of the
saturation concentration (Csat) of formaldehyde in air over an emitter
of interest, as follows:

Equation 3-5 is reproduced below with slightly different notation along
with an equivalent equation from my previous work:

       (Equation 3-5 with different notation)

Gss 	= steady state emission rate (mg/m2-hr).  Equivalent to CH2OER’

Gmax = the maximum emission rate which occurs at zero backpressure; that
is, at zero concentration of formaldehyde in the air (mg/m2-hr). 
Equivalent to b.

m 	= the mass transfer coefficient (m/hr)

Css 	= vapor-phase concentration of formaldehyde at steady state
(mg/m3).  Equivalent to        [Ch2O]v

Equation 3-5 above is essentially identical to an algorithm that I have
often used for backpressure modeling, shown below:

Csat = the saturation concentration (mg/m3) of formaldehyde in the air
over the emitter of interest. That is the concentration that would be
expected in a volume (with high surface area/volume ratio) in which the
emitting walls were made up entirely of the emitter of interest and
there is zero ventilation. Note the equivalence of this last equation to
Equation 3-5 above and the identity of m:

 

Clearly, the relatively straightforward analytical measurement of Csat
could provide a potentially valuable extended methodology for the
evaluation and development of the model’s parameters relative to
specific emitters. 

Several peer reviewers suggested possible alternatives to the
exponential-decay model used in FIAM-pwp to represent declining
formaldehyde emissions over time. One suggestion was that a power-law
model may be more suitable and another suggestion was the use of an
expression with multiple or changing decay rates to represent, for
example, “fast” (earlier) and “slow” (later) periods of
exponentially declining emissions. Brown fitted a double-exponential
model to results from chamber experiments, noting that “the high but
decreasing emission rates in the first 3 months were consistent with
free formaldehyde and formaldehyde produced from easily hydrolysed
chemical bonds” and that  “longer-term, near-constant emissions may
be due to ongoing hydrolysis of resins used in the wood panels.” One
peer reviewer noted a conceptual model with three exponential-decay
compartments, attributed to three potential sources of formaldehyde
emissions:

Free formaldehyde (short-term compartment);

Decay of paraformaldehyde and other complex structures (medium-term
compartment); and

Hydrolysis of the resin (long-term compartment).

A single-exponential model for declining formaldehyde emissions over
time has the following form:

	ER(t) = ER(0) e -kt

	Where:

	ER(t) is the emission rate at time t (e.g., elapsed hours);

	ER(0) is the initial emission rate (t = 0); and

	k is the rate constant governing the exponential decline in the
emission rate.

The power-law function takes the following form:

	ER(t) = at -k

	Where a is a constant and k is an exponential index, dimensionless.

Although the power-law function is undefined when t is zero, an initial
emission rate can be “set” by assigning an arbitrarily small value
to t (the rate is highly sensitive to the value chosen for t).

The plots in Figure 3-1 illustrate the declining behavior of the
power-law function, with a set to 50 and k set to 0.25, in comparison to
the single-exponential model, with ER(0) set to 90 and k set to either
0.1 or 0.5. The power-law function generally has a sharper curvature
than the exponential model. This sharper curvature can be matched to an
extent with the exponential model, but only by choosing a more rapidly
declining emission rate (e.g., with k set to 0.5 as in the figure). The
shape of the time-related decline obtained using the power-law function
can be nearly matched by the exponential-decay model, however, when it
is modified to have an initial period of faster decline followed by a
second period of slower decline, as illustrated in Figure 3-2. If, for
example, the conceptual model above with three potential sources of
formaldehyde emission is a representative construct, then the approach
of using multiple exponentials as a semi-empirical approximation may be
preferable (recognizing the associated need for estimating multiple
parameters) because first-order processes are exponential in form. An
example application of FIAM-pwp with a varying decay rate is provided in
Section 2.7.5.

The above constructs, although potentially useful, currently share one
limitation – little existing data to provide a basis for assigning
appropriate parameter values. They should be kept in mind, however, for
future research efforts and modeling applications. By comparison, as
described in Section 3.8, there has been a fair amount of research on
the formaldehyde emissions half life, providing some basis for choosing
a value for the rate of decline for the simpler single-exponential decay
model.

Figure 3-1. Representation of Declining Emission Rate

by Power-law Function vs. Exponential-decay Model.

Figure 3-2. Representation of Declining Emission Rate by Power-law
Function

vs. Exponential-decay Model with Periods of Fast and Slow Decay.

3.6 	Discussion of Indoor Sinks tc \l2 "3.4 	Treatment of Indoor Sinks 

In concept, the treatment of indoor sinks in the calculation module
could be accomplished within the model framework by recognizing that the
emission term, as given by Equation 3-5, is comprised of two components.
The first component (-m*[CH2O]v) is in the same form as a first-order
sink representation. The second component (b) is a constant,
representing a constant emission rate. By setting b equal to zero and m
to a nonzero value, this “source” representation actually reflects
the behavior of a first-order sink. Alternatively, setting m to zero and
b to a nonzero value represents a product as a constant emitter.

Recognizing these behaviors, the source term could be used to represent
a simplistic sink by assigning a large enough value to m to result in a
net flux of formaldehyde into a material that acts as a sink.  However,
this simplistic approach does not allow a fully realistic representation
of the re-emission from sinks that has been observed in laboratory and
field studies. In the model, the concentration in a compartment is
assumed to decline from an initial “steady-state” concentration in
accordance with a user-specified half-life. This model representation
tends to diminish the effect of the sink over time, much like the
emission rates for materials that act as sources. In reality, sinks for
formaldehyde tend to accumulate mass when air concentrations are high,
and to re-emit this mass as the air concentration decreases. The
accumulation rate in the sink at any point in time is a function of the
air concentration and the mass residing in the sink at that point, as
well as other sink characteristics.

Because the complex relationships such as those described above cannot
be fully captured by using the current modeling framework to represent
sinks, it is recommended that users do not attempt to model sinks using
FIAM-pwp. Ignoring the potential effects of indoor sinks will tend to
result in a more conservative estimate, that is, higher modeled
concentrations. The magnitude of prediction error with exclusion of
indoor sinks is believed to be relatively small, on the order of 10 %
for most formaldehyde modeling situations.

3.7	Decay of Indoor Concentrations tc \l2 "3.5	Decay of Indoor
Concentrations 

The initial (steady-state) indoor formaldehyde concentration, as
calculated by the model described above, is assumed to decrease over
time as the reservoir of formaldehyde in various sources and sinks is
gradually depleted. The decrease in the initial value over time is
assumed to follow a first-order exponential process, expressed as
follows:

 	(Eqn. 3-21)

where:

Ct	=	formaldehyde concentration indoors (ppb) at time t (in years), due
to the combination of background concentration and emissions from PWPs
(Note: Ct = Ci at t = 0)

CB	=	background concentration (ppb), which is defined as the outdoor
concentration plus the contribution from indoor sources other than   
pressed-wood products (assumed constant)

Ci	=	initial concentration indoors (ppb) 

k	=	first-order rate constant for the exponential decline (years-1).

Note that, in Equation 3-21, only the contribution from the new PWPs
(i.e., Ci - CB) is decayed because the background contribution is
assumed to be constant, as discussed in Section 3.3. The value of k is
determined from the user-specified half life for the indoor formaldehyde
concentration, assumed to be dominated by an exponential decay in
formaldehyde source strength.  Solving for t results in the following
equation:

 	(Eqn. 3-22)

Solving for k results in the following equation:

 	(Eqn. 3-23)

Neglecting for the moment the background contribution to the indoor
concentration (i.e., assuming that the background concentration is
zero), Equation 3-23 reduces to:

 	(Eqn. 3-24)

By definition, Ct is equal to half of Ci when the half-life is reached:

 	(Eqn. 3-25)

where:

T = half-life (in years).

By substitution, Equation 3-24 becomes:

 	(Eqn. 3-26)

Substituting the default half life of 1.5 years (see Section 2.2) for T
in the above equation gives:

k = 0.462 years-1

Indoor concentrations at later points in time (i.e., subsequent to the
initial, steady-state values for each zone) are calculated by decaying
only the portion of the indoor concentration that is attributable to
indoor pressed-wood products. For example, to obtain the indoor
concentration 3 months (0.25 years) after the initial value, with an
assumed half-life of 1.5 years, Equation 3-21 would apply:

C(0.25 yr) = CB + (Ci - CB) e-0.462*0.25	(Eqn. 3-27)

If Ci were 50 ppb and the background concentration were zero, then the
indoor concentration three months later would be 44.5 ppb. With the
default background concentration of 7.5 ppb (see Section 2.2), the
indoor concentration would be 57.5 ppb initially and 52 ppb after three
months. In either case, the portion of the initial indoor concentration
due to indoor pressed-wood products (50 ppb) would be reduced by 5.5 ppb
over three months with a half life of 1.5 years.

The time required for the initial indoor concentration to decay to a
user-specified concentration (e.g., 10 ppb, the default value) can be
determined from the first-order exponential process using Equation 3-22
as follows, given an initial indoor concentration greater than the
specified concentration and assuming a background concentration of zero:

t = ln (Ci / 10) / k	(Eqn. 3-28)

Continuing with the above example, where k is 0.462 years-1 and Ci is 50
ppb, by substitution into Equation 3-28 we get:

t = ln (50 / 10) / 0.462	(Eqn. 3-29)

for which the value of t is 3.48 years, or 41.8 months.

The model uses the higher of the two initial indoor concentrations, when
a two-zone model is selected, for the calculation. Because the user is
allowed to specify a background concentration other than zero, which is
assumed to be constant over time, Equation 3-22 can be used to account
for the background contribution while decaying only the component due to
indoor sources and sinks: 

 	(Eqn. 3-30)

With the default background concentration of 7.5 ppb and an initial
concentration of 57.5 ppb, the value for t would be 6.48 years. The
result from Equation 3-30 is multiplied by 12 and reported in months. If
the highest initial indoor concentration is lower than the background
concentration, then zero months is reported as the time required for the
indoor concentration to decay to background. 

3.8	Default Value for Half Life

Ideally, an appropriate half-life value for representing the long-term
decay of formaldehyde emissions and concentrations in residences would
be determined by taking measurements in a house soon after it was loaded
with pressed-wood products and at periodic points in time over the next
several years. Continuing with this ideal situation, the products would
not change throughout the measurement period and factors such as indoor
temperature and humidity, as well as the indoor-outdoor air exchange
rate, would change only minimally in response to outdoor weather
conditions and occupant activities indoors. Further, the experiment
would be repeated in several houses. The data for each house would be
analyzed to fit a nonlinear (e.g., exponential) decay function to the
time-varying indoor concentrations, and the resultant estimates would be
compared across houses (or the data would be pooled across houses) to
determine an appropriate central value for the concentration half life.

The ideal data set has not been collected, however, most likely because
the complexities and costs of multiple measurements over time, coupled
with the need to restrict certain occupant activities over several years
(or to operate a research house to gain experimental control), would be
prohibitive.  There are two practical alternatives to the ideal
experimental situation, both of which have been attempted: (1) collect
or assemble point-in-time measurements from houses of varying ages, as a
proxy for the time-related decay in one or a series of houses; and (2)
conduct chamber tests at periodic points in time for pressed-wood
products, either individually or in combination.

Either alternative has significant limitations. For example,
cross-sectional data from many houses are, at best, a crude proxy for
the time-related behavior of a single house, or of a limited set of
well-characterized houses – the analysis is confounded by varying, and
generally unknown, construction features, occupant activities and
weather conditions at the times of measurement. Chamber tests, on the
other hand, have the advantage of careful experimental control. However,
for multiple tests over time, the material to be tested typically would
not be kept in the chamber-controlled environment throughout the testing
period, but rather during only the relatively brief occasions when
measurements were to be taken. Further, most such tests have been
conducted only for a single product or material. The time-related rate
of decline depends on the formaldehyde concentration; the rate of
decline is more rapid for a single product because other products are
not present to increase the formaldehyde concentration and thereby
retard its rate of formaldehyde release. Limited tests of
product/material ensembles have been conducted, but none have included
the full array of sources and sinks that typically would be present in
an occupied house. With these limitations in mind, some previous
investigations into the half life of formaldehyde emissions or
concentrations are summarized below.

As summarized in a Versar 1988 report, estimates of decay half lives for
formaldehyde concentrations in residences range from 2 to 44 years, with
most estimates below 5 years. These estimates have been based largely on
pooled cross-sectional data across a variety of residences, as a proxy
for time-series data from one or a few houses. It is noteworthy that the
higher estimates (20 to 44 years) are from data sets involving houses
aged from new to 50 years, whereas the lower estimates (2 to 4 years)
are from data sets where the houses generally were 10 years old at most.
If none of the materials or furnishings in a house were to be changed,
and if the occupants were to avoid using consumer products that emit
formaldehyde, then the indoor concentration could be expected to
eventually decay to some background level, similar to that if no sources
were present. As a house ages, however, occupants will tend to add or
replace materials and furnishings, some of which may emit formaldehyde.
Such changes, then, will introduce new sources, thereby raising the
formaldehyde concentration relative to the no-change case and
artificially inflating the half-life estimate.

The above conjecture is partly supported by limited studies involving
repeat measurements at different points in time from one or two houses,
for which the half-life estimate ranged from 2.0 to 2.6 years. A careful
analysis by Versar of a cross-sectional data set for 396 manufactured
homes, as described in the 1988 report referenced on the previous page,
resulted in a half-life estimate of 2.92 years. This estimate is
believed to be an upper bound for the true half life in a house that has
some significant formaldehyde sources from the outset of its
construction and occupancy history.

Chamber studies to investigate the decay of formaldehyde have focused
primarily on individual sources and, not surprisingly, have produced
shorter half-life estimates than most residential studies. As summarized
in the same (1988) Versar report, the estimated half lives from chamber
studies have ranged from 0.2 to 2.58 years. One noteworthy
investigation, conducted by Matthews et al., involved a combination of
sources (particleboard, hardwood plywood paneling and MDF) under
controlled environmental conditions (23 ( C and 50 % relative humidity).
The “slow” decay study, with an air exchange rate of 0.4 air changes
per hour (ACH) in the chamber, produced a half-life estimate of 1.52
years. The “fast” decay study was conducted with relatively high air
exchange rates designed to keep the chamber formaldehyde concentration
below 0.1 ppm. The lower background concentration was expected to
increase the emission rates and, thus, shorten the decay periods
relative to the “slow” decay study. Resultant half-life estimates
were 0.76, 0.62 and 1.08 years for particleboard, paneling and MDF,
respectively.  

Of the two decay studies by Matthews et al., the “slow” study is
believed to be more indicative of the formaldehyde decay behavior in
residences. However, the authors added the caution that care should be
taken in interpreting the study results. For example, the duration of
the study may have been inadequate to accurately reflect the emission
decay period for the collection of sources. In addition, the data may
reflect a relatively rapid decay of emissions from a few of the
strongest emitting sources, as opposed to the entire collection of
products, whereas in real life one might typically expect to find a
combination of relatively strong and weak emitters, or of sources and
sinks.

In another decay study for an individual source, Zinn et al. estimated
an apparent half-life of 216 days (0.6 years) using approximately one
year of large chamber data for particleboard and a natural logarithm
linear regression statistical analysis. As noted by one of the peer
reviewers:

The half-life reported in Zinn et al. (1990) is not directly comparable
to that adopted in the original and current Versar model because Zinn et
al. used the relationship HCHO = F + G x ln(time) with regression
constants F and G whereas the EPA half-life was based on a traditional
exponential decay model. The Zinn et al. logarithmic equation was used
by CARB to calculate the emissions inventory presented in the technical
basis for the formaldehyde airborne toxic control measure. In lieu of
the original logarithmic equation, the more generic two-compartment
(i.e., fast and slow decay) exponential decay model could be fit the
Zinn et al. data.

For the various reasons noted in the above review and discussion, there
is considerable uncertainty as to the most appropriate value to
represent the central tendency for the decay rate of formaldehyde, in a
residence where construction features, house contents and occupant
activities remain relatively stable over time. The cross-sectional data
from residential studies are believed to yield estimates erring on the
high side (i.e., longer estimated than actual half-life), and the
temporal data from chamber studies are believed to yield estimates
erring in the opposite direction. 

The collective evidence described above indicates that an appropriate
half-life value probably lies between 1.5 and 3.0 years. The value of
2.92 years from the cross-sectional analysis described above will tend
to err on the conservative side. That is, assuming that the initial
concentration in a house is predicted accurately, the 2.92-year value
likely will cause the modeled concentration to decay at a slower rate
than might be observed in a stable residential setting.

A 2005 paper by Groah focused on the topic of decay in formaldehyde
concentrations over time. The author notes that “newer sources of
formaldehyde are continually being brought into homes as they age.”
Examples would include the installation of new kitchen cabinets when
remodeling or the addition of new furnishings such as furniture or
shelving that may contain pressed-wood products. To the extent that such
situations exert an upward bias on half-life estimates that are based on
an analysis of homes of different ages, it could be argued that the
default half-life value of 2.92 years is too conservative, as mentioned
above. Two studies were noted by Groah that involved chamber testing of
a collection of pressed-wood products as they aged. One (by Matthews et
al., cited on the previous page) estimated the emissions half-life to be
1.52 years and the other (Groah and Gramp) indicated a half-life of 2.10
years. Based on these studies, a half-life on the order of 1.5 to 2.0
years is suggested should the user wish to apply a less conservative
value.  

3.9	Temperature and Humidity Adjustments

As noted in Section 2 and in the introduction to this section, the model
initially estimates the formaldehyde concentration(s) in a structure
under indoor conditions of 23 ( C temperature and 50 % relative humidity
– these are consistent with the conditions under which most chamber
tests have been conducted, in accordance with ASTM standard E 1333, for
estimation of emission rates or concentrations under an assumed
steady-state condition. The model then adjusts the initial concentration
estimates under “standard conditions” to the user’s inputs for
temperature and humidity. 

The temperature adjustment used in the model is of the form:

 	(Eqn. 3-31)

where:

CT	=	Temperature-adjusted formaldehyde concentration 

C0	=	Formaldehyde concentration at the baseline temperature 

R	=	Temperature coefficient for adjustment

TU	=	User-specified temperature, ( K

TB	=	Baseline temperature, 296 ( K (23 ( C).

The humidity adjustment, applied to the temperature-adjusted
concentration, is of the form:

   	(Eqn. 3-32)

where:

CTH	=	Temperature- and humidity-adjusted formaldehyde concentration

CT	=	Temperature-adjusted formaldehyde concentration 

A	=	Humidity coefficient for adjustment

HU	=	User-specified relative humidity, %

HB	=	Baseline relative humidity, 50 %.

Combining equations 3-31 and 3-32, the temperature- and
humidity-adjusted concentration can be expressed as follows:

 	(Eqn. 3-33)

where KTH is the combined temperature and humidity adjustment factor:

 	(Eqn. 3-34)

Thus, the adjustments require temperature and humidity coefficients,
for which the model provides two alternative sets of values. The first
set of coefficients – the model default -- was calculated by Berge et
al. for two Norwegian particleboard specimens. These coefficients were
used by HUD in setting the 1984 standards for particleboard and hardwood
plywood paneling, and also have been used in ASTM, ANSI and Composite
Panel Association (CPA) standards or guidelines. Chamber tests
underlying these coefficients were conducted at two temperatures (22 and
28 (C) and at two relative humidities (30 and 60 %). The coefficients
reported by Berge et al. were 9799 for temperature (assumes an Arrhenius
type of model) and 0.0175 for humidity (assumes a linear model).

The second set of coefficients was reported by Myers, who reviewed the
extant literature and developed his recommended temperature coefficient
based on chamber testing of about 40 specimens (particleboard and
hardwood plywood paneling) from 11 laboratories. The temperature range
of the product tests was from 20 to 40 (C. Dr. Myers also developed a
humidity coefficient, but expressed little confidence in the number; the
relative humidities in the tests on which the correction factor was
based ranged from 20 to 90 percent. Thus, his work encompassed more
products and wider temperature/humidity ranges than the Berge et al.
study. It is worth noting that the HUD standard was promulgated before
Dr. Myers completed his work. His reported correction coefficients were
8930 for temperature and 0.0195 for relative humidity, using the same
model assumptions as Berge et al.

The temperature and humidity adjustments modify the emission rate from
pressed-wood products and, therefore, the portion of the indoor
concentration due to indoor sources and sinks. That is, the temperature
and humidity adjustment is applied only to the portion of the
steady-state concentration contributed by pressed-wood products. The
adjustment applied to the two-zone concentrations calculated using
Equations 3-12 and 3-13 takes the following form:

 	(Eqn. 3-35)

 	(Eqn. 3-36)

Similarly, the adjustment applied to the one-zone concentration
calculated using Equation 3-8 takes the following form:

 	(Eqn. 3-37)

3.10	Exposure Calculations 

As noted in Section 2.5, the model calculates the following exposure
metrics for each zone of the house or other modeled structure:

Average daily concentration (ADC)

Percent of time concentration is greater than the user-specified level
of interest (LOI).

The ADC in each zone is calculated in yearly increments, by integrating
Equation 3-21. The value of this integral, between two times (t1 and t2)
that are exactly one year apart, is as follows:

 ADC = ((Ci  - CB)/-k)*e-k*t2 + CB*t2 - ((Ci  - CB)/-k)*e-k*t1 - CB*t1 
(Eqn. 3-38)

  where:

Ci	=	initial concentration indoors (ppb)

CB	=	background concentration indoors (ppb)

k	=	first-order rate constant for the exponential decline (years-1)

t1	=	starting time (years)

t2	=	ending time (years).

For a new house, the values of t1 and t2 for the first year are 0 years
and 1 year, respectively. The t1 and t2 values are 1 and 2 years for the
second year, 2 and 3 years for the third year, and so on. If, for
example, the starting concentration (i.e., at t1 = 0 years) were 57.5
ppb, the background concentration were 7.5 ppb, and the value of k were
0.462 years-1 (half life of 1.5 years), then the ending concentration
would be 39 ppb per Equation 3-21 and the value of the integral for the
first year would be 47.5 ppb per Equation 3-38.

The percent of time during which the concentration in a modeled zone is
greater than the LOI also is calculated in yearly increments, as
follows:

% Time =100 * (ln(Ci-CB) - ln(LOI-CB)) / k – t2)	(Eqn. 3-39)

where:

% Time 	=	percent of time concentration is greater than LOI

LOI 	=	level of interest (ppb).

Continuing with the example where Ci = 57.5 ppb, CB = 7.5 ppb, and t1 =
0 (i.e., new house), and with a value of 10 ppb for the LOI, the
calculation result for each of the first six years (i.e., for t2 ranging
from 1 to 6) is equal to 100%. For year 7 (t2 = 7) the calculation
result is 48.4%; thereafter, the calculation result is negative, which
the model resets to a value of 0%.

The ADC for any exposure group is calculated, in yearly increments, as a
weighted average of concentrations encountered in each zone of the house
or other modeled structure and concentrations encountered in non-home
locations: 

ADC = (HZ1*ADCZ1 + HZ2*ADCZ2 + HWSD*CWSD + HV*CV+ HO*CO) / 8760 	(Eqn.
3-40)

where:

H		=	hours spent per location

ADC	=	average daily concentration in Zone 1 or Zone 2 (ppb)

Z1				refers to Zone 1

Z2				refers to Zone 2

C		=	concentration at location away from home (ppb)

WSD 			refers to work, school or daycare location 

V	 			refers to in-vehicle location 

O				refers to other/outdoor location 

The ADC values per zone decrease with each passing year, per Equation
3-38, whereas concentrations in locations outside or away from the home
or modeled structure are assumed to be constant across the years. The
sum of hours across the various locations (Z1, Z2, WSD, V and O) must
equal 8760 for the calculation result to be valid. For a one-zone
modeled structure, Z2 is set equal to Z1 for the calculation.

4.	 EVALUATION OF MODEL ALGORITHMS AND PREDICTIONS

The evaluation of the model has been performed from two perspectives:
(1) independent checks on the correctness of model algorithms; and (2)
comparison of model predictions with measurements in real-world
situations.

4.1	Mathematical Correctness of Model Algorithms

An Excel spreadsheet has been developed as a tool for assessing the
mathematical correctness of algorithms used in the model. The Excel tool
includes individual sheets for inputs corresponding to model input
screens – House, Source, and Exposure – together with a separate
sheet that displays the calculation results (predicted HCHO
concentrations). Input areas within each screen are set up similarly to
those in the model, with the exception that no drop-down selections are
provided for sources, sinks or cabinets. Instead, input areas (cells)
are designated for direct entry of Slope, B and Area values. Results are
provided for both the one-zone and two-zone implementation of the model,
in concentration units of mg/m3 and ppm. 

Results in the Excel tool were compared with those reported by the model
for a variety of cases representing different combinations/values for
background concentrations, sources, and exposure groups, for both
one-zone and two-zone implementations. This exercise led to discovery of
minor errors in an earlier version in the handling of ppm vs. mg/m3
units for the background concentration (now ppb and µg/m3 units) and in
the temperature/humidity adjustment for indoor formaldehyde
concentrations. Other than those discrepancies, the results agreed
exactly in all cases. The errors discovered in the model’s calculation
routine have been corrected. 

4.2	NIST Assessment of Matthews Model Predictions

An effort was undertaken by researchers at the National Institute of
Standards and Technology (NIST) to validate the Matthews model on which
this model is based. Particleboard underlayment, hardwood-plywood
paneling and medium-density fiberboard (MDF) products initially were
characterized in chambers by measuring their HCHO surface emission rates
over a range of formaldehyde concentrations, air exchange rates and two
combinations of temperature and relative humidity. The products then
were installed in a two-room prototype house in different combinations,
and equilibrium HCHO concentrations were monitored in the house.

The following findings are excerpted directly from the authors’
abstract:

Particleboard underlayment and mdf, but not paneling, behaved as the
emission model predicted over a large concentration range, under both
sets of temperature and relative humidity. Good agreement was also
obtained between measured formaldehyde concentrations and those
predicted by a mass-balance indoor air quality model.

The mass-balance model referenced here is the Matthews model. The
following excerpt, taken directly from the authors’ conclusions,
provides some further insights on model performance:

In developing its models, ORNL assumed that there is relatively free
formaldehyde diffusion within the "bulk phase" of pressed-wood products.
Formaldehyde emission results from the difference in concentration
between this "bulk phase" and the atmosphere, in accordance with Fick's
first diffusion law. This assumption is apparently fairly descriptive of
particleboard and mdf, but not of hardwood plywood paneling. It is
hypothesized that internal diffusion is important in formaldehyde
emission by pressed-wood products, and layering of solid thin sheets of
plywood obstructs diffusion to a greater extent than do chips and pieces
in the other pressed-wood products. In other words, formaldehyde
emission may not be as diffusion-limited in plywood as it is in other
pressed-wood products. This would explain why formaldehyde emission by
plywood appeared relatively insensitive to temperature and RH used in
the present study. Higher temperature and RH increase resin degradation
to formaldehyde and increase the formaldehyde diffusion rate.
Apparently, the small changes of temperature and RH used in the present
study did not increase the "bulk phase" formaldehyde concentration or
the diffusion rate sufficiently to overcome the obstructing layers of
the relatively low-emitting plywood used in this study.

It is noteworthy here that particleboard and MDF generally are much
higher HCHO emitters than plywood or hardwood plywood paneling. Thus,
the NIST validation effort indicated accurate model predictions for the
major HCHO emitters among pressed-wood products.

4.3	Assessment of FIAM-pwp Predictions – EPA Pilot Study

An assessment undertaken by the author of this document involved
comparison of model predictions with concentrations measured in the EPA
pilot study house under various conditions, including before and after
loading the house with pressed-wood products. The assessment was
conducted by using known/measured factors such as zone volumes, airflow
rates, indoor temperature and humidity, and source/sink loading areas
together with estimates of emission model parameters for each type of
source or sink. Model inputs were developed for six cases (time periods
within the pilot study) at the conventional house – initial baseline,
first loading, air-out after first loading, second loading, air-out
after second loading, and third loading. Inputs supplied for model input
screens – House, Sources and Sinks – are described in a general
manner below. Further details specific to each of the six cases are
given in the subsections that follow.

The house screen requires inputs for the type of house, the zone volumes
and airflow rates, the background formaldehyde concentration, and the
indoor temperature and humidity conditions. The house used for the pilot
study – Conventional 1 – was selected. Airflow measurements with
PFTs on 14 different occasions during the pilot study averaged 0.59 ACH
and varied over a relatively narrow range of 0.5 to 0.7 ACH. The default
airflow rates provided with the model for this conventional house, which
correspond to an air exchange rate of 0.59 ACH, were used for all model
runs. 

Outdoor formaldehyde concentration measurements during the pilot study
varied over the narrow range of 0.001 to 0.003 ppm, averaging about
0.002 ppm. The average of 0.002 ppm was used for this input rather than
case-specific values. The model calculates formaldehyde levels in each
zone of the house under a base set of temperature/humidity conditions
(23 (C, 50 %) and a user-supplied set of conditions. The user-supplied
conditions were chosen to represent the average of the values measured
upstairs for each case at the conventional house. The major formaldehyde
sources all were located upstairs. The default temperature and humidity
coefficients provided with the model, for adjustment of modeled
formaldehyde concentrations from the base conditions to the
user-supplied conditions, were used for all model runs.

Model parameters for sources – Slope and B – have been estimated
previously from large-chamber emission tests conducted by industry or
the USEPA, and are supplied with the model. The major sources installed
in the conventional house – underlayment, kitchen/bathroom cabinets,
interior doors, and hardwood plywood – were chamber tested under three
different air exchange rates as part of the pilot study. Source loading
areas for the model runs are described later as they pertain to each
modeled case.

There are two major types of potential sinks in conventional house 1 –
carpet/padding and painted wallboard. Sinks can be represented in the
Matthews model using the same parameters as those for sources. The
Langmuir isotherm model described in the formaldehyde pilot study report
(p. 6-44) was used to develop parameter estimates for sinks, for use in
the Matthews model.

The modeling approach, rationale and assumptions for each of the six
modeled cases are described below. The results of this evaluation
exercise collectively indicate that the modified (two-zone) Matthews
model can predict measured values quite well for the conventional house,
but errs somewhat on the conservative side (i.e., it tends to
“over-predict”). The greatest area of uncertainty in applying the
model appears to be in choosing appropriate values for indoor sinks. One
additional area of uncertainty stems from the fact that humidity levels
were not measured downstairs; as noted previously, upstairs values for
temperature and humidity were used in constructing the model inputs.

4.3.1	Initial Baseline Period

Model inputs for this period are summarized in Table 4-1. The only
source present in the house during the initial baseline period was the
plywood subfloor, and the only sink was painted wallboard (carpet was
not yet installed). The area for plywood was taken as the product of the
upstairs volume (304 m3) and an assumed loading ratio of 0.43 m2/m3, or
131 m2. For wallboard the loading ratios for walls (0.95 m2/m3) and
ceilings (0.43 m2/m3) were used to estimate a total area of 420 m2.
Model source/sink parameters (intercept and slope) were taken from the
model defaults for softwood plywood. An intercept value of 0.04 mg/m2-h
was chosen for wallboard after running the model with plywood only and
obtaining an indoor formaldehyde concentration in the vicinity of 10
ppb.

Table 4-1.  Model Inputs (Sources and Sinks) for Initial Baseline Period

Source/Sink	

Slope (m/h)	

Intercept (mg/m2-h)	

Area (m2)



Softwood plywood	

-0.61	

0.03	

131



Painted wallboard	

-3.00	

0.04	

420



Modeling results for the initial baseline are shown in Table 4-2. With
plywood as the only source, and without including wallboard as a sink,
the modeled values for the conditions under which measurements were
taken (23.98 ( C, 23.6 % RH) were 8 ppb upstairs and 7 ppb downstairs,
close to the measured values of 9.5 and 9.1 ppb for the respective
zones. Addition of wallboard as a sink had no appreciable effect on the
modeled values. It is plausible that, at this point in time, the
wallboard had reached a state of equilibrium and, thus, acted as neither
a net source nor a net sink. When the intercept parameter for wallboard
was arbitrarily halved (to 0.02 mg/m2-h), to cause it to behave as net
sink, the modeled values were 5.0 ppb upstairs and 4.4 ppb downstairs
under the conditions when measurements were taken.

Table 4-2.  Modeling Results for Initial Baseline Period

Source/Sink Inputs	

Results (ppb) for Base/Measured Conditions

	

23 ( C, 50 % RH	

23.98 ( C, 23.6 % RH

	

Upstairs	

Downstairs	

Upstairs	

Downstairs



Softwood plywood as a source	

10.3	

9.0	

7.8	

6.9



Add wallboard as a sink	

10.7	

9.4	

8.2	

7.2



Measured values	

--	

--	

  9.5*	

9.1

*The measured value is an average across 3 monitoring sites (living
room, kitchen and bedroom).

4.3.2	First Loading

The modeled sources for the first (medium) loading included
underlayment, cabinets, interior doors and countertop (Table 4-3).
Plywood was not included as a source because it was covered by
underlayment or by floor tile. Loading areas for the sources were taken
from Section 3.5 of the pilot-study report, and slopes and intercepts
were taken from the pilot-study values provided with the model for
underlayment, doors and cabinets. For countertop, industrial
particleboard was selected to obtain slope/intercept parameters (average
values were used). The intercepts chosen for carpeting and wallboard
were predicated on an indoor-air concentration near 40 ppb.  

Table 4-3.  Model Inputs (Sources and Sinks) for First (Medium) Loading

Source/Sink	

Slope (m/h)	

Intercept (mg/m2-h)	

Area (m2)



Underlayment	

-1.27	

0.28	

  46



Cabinets	

-0.48	

0.08	

  59



Interior doors	

-0.50	

0.08	

  35



Countertop	

-0.70	

0.30	

   4



Carpeting	

-1.30	

  0.065	

 94



Painted wallboard	

-3.00	

0.15	

420



The modeling results for this case (Table 4-4) are compared with
measured values taken at run 4 (28 days after loading), by which time
the indoor-air concentrations appeared to have “leveled off” (see
Figure A-1 in Appendix A). The addition of carpeting lowered the modeled
concentrations slightly, and further addition of wallboard lowered the
modeled values by about 10 percent compared to estimates with no sinks.
The modeled concentrations for measured conditions, with carpeting and
wallboard included, were 5-10 ppb higher than measured values under
those conditions.

Table 4-4.  Modeling Results for First Loading

Source/Sink Inputs

	

Results (ppb) for Base/Measured Conditions

	

23 ( C, 50 % RH	

23.65 ( C, 50.1 % RH

	

Upstairs	

Downstairs	

Upstairs	

Downstairs



Underlayment, cabinets, doors and countertop as sources	

47	

40	

51	

43



Add carpeting as a sink	

46	

39	

49	

42



Add wallboard as a sink	

42	

36	

45	

39



Measured values (run 3, day 28)	

--	

--	

41	

28



4.3.3	Air-out Period after First Loading

For this air-out period, plywood again was the only source and wallboard
was the only sink. In this case, however, it was necessary to account
for the fact that the wallboard had adsorbed significant mass during the
first loading. Some of this mass would be lost during the air-out
period, which lasted 7 days, but the amount lost was uncertain.
Consequently, the intercept for the first loading (see Table 4-5) was
used as an upper bound when developing inputs and half this value was
used as a lower bound.

Table 4-5.  Model Inputs for Air-out Period after First Loading

Source/Sink	

Slope (m/h)	

Intercept (mg/m2-h)	

Area (m2)



Softwood plywood	

-0.61	

0.03	

131



Painted wallboard	

-3.00	

0.15/0.075*	

420

*Two runs were made, one with 0.15 as the intercept and one with 0.075.

As shown in Table 4-6, the modeling results with the alternative
intercept values for wallboard surround the measured values. The model
estimates with the smaller intercept value, to account for the likely
possibility that the wallboard had released a fair amount of mass during
the air-out period, are closer to the measurement results.

Table 4-6.  Modeling Results for Air-out Period after First Loading

Source/Sink Inputs 	

Results (ppb) for Base/Measured Conditions

	

23 ( C, 50 % RH	

23.44 ( C, 53.4 % RH

	

Upstairs	

Downstairs	

Upstairs	

Downstairs



Softwood plywood as a source	

10.3	

9.0	

11.5	

10.0



Add wallboard as a re-emitting sink (intercept = 0.15)	

34	

29	

38	

33



Change wallboard intercept to 0.075	

18	

16	

20	

18



Measured values	

--	

--	

23	

26



4.3.4	Second Loading

For the second (high) loading, the sources (Table 4-7) were similar to
those for the first loading, but with a larger area for underlayment and
with the addition of hardwood plywood paneling as another source. The
intercept values chosen for carpet and for wallboard were based on
concentration estimates obtained from running the model without any
sinks.

Table 4-7.  Model Inputs (Sources and Sinks) for Second (High) Loading

Source/Sink	

Slope (m/h)	

Intercept (mg/m2-h)	

Area (m2)



Underlayment	

-1.27	

0.28	

  94



Cabinets	

-0.48	

0.08	

  59



Interior doors	

-0.50	

0.08	

  35



Countertop	

-0.70	

0.30	

    4



Paneling	

-0.98	

0.16	

  36



Carpeting	

-1.30	

  0.095	

  94



Painted wallboard	

-3.00	

  0.225	

420



Modeled values (Table 4-8) were compared to measurements taken at run 2
(day 12), by which time it appeared that the indoor concentrations had
started to “level off” (see Figure A-1 in Appendix A). This earlier
time of leveling off, relative to that for the first loading, was likely
due to the wallboard sink being substantially “pre-loaded” with
formaldehyde mass as a result of the first loading. The modeling results
without sinks were 70-75 ppb upstairs and 60-65 ppb downstairs. The
addition of carpeting and wallboard as sinks lowered the modeled values
by about 10 percent. The modeled values with sources and sinks were
about 10 ppb higher than those measured upstairs (differences similar to
those for the first loading), and about 15 ppb higher than the measured
downstairs values.

Table 4-8.  Modeling Results for Second Loading

Source/Sink Inputs 	

Results (ppb) for Base/Measured Conditions

	

23 ( C, 50 % RH	

23.48 ( C, 50.1 % RH

	

Upstairs	

Downstairs	

Upstairs	

Downstairs



Underlayment, cabinets, doors, countertop and paneling as sources	

70	

60	

74	

63



Add carpeting as a sink	

68	

58	

72	

61



Add wallboard as a sink	

63	

54	

67	

57



Measured values (run 2, day 12)	

--	

--	

56	

40



4.3.5	Air-out Period after Second Loading

For the air-out period after the second loading, the only sources and
sinks again were plywood and wallboard (Table 4-9). As with the first
air-out period, it was necessary to account for the fact that the
wallboard had adsorbed mass during the second loading and then had lost
some of this mass during the air-out period. To accommodate uncertainty
in the amount of mass in the wallboard near the beginning of this
air-out period, an intercept close to that for the second loading was
used as an upper bound, and half that value was used as a lower bound. 

Table 4-9.  Model Inputs for Air-out Period after Second Loading

Source/Sink	

Slope (m/h)	

Intercept (mg/m2-h)	

Area (m2)



Softwood plywood	

-0.61	

0.03	

131



Painted wallboard	

-3.00	

0.19/0.095*	

420

*Two runs were made, one with 0.19 as the intercept and one with 0.095.

As shown in Table 4-10, the model results with alternative intercept
values for wallboard surround the measured values, and the model
estimates with the smaller intercept value are closer to the measurement
results. This pattern is the same as that observed for the comparison of
modeling and measurement results for the first air-out period.

Table 4-10.  Modeling Results for Air-out Period after Second Loading

Source/Sink Inputs 	

Results (ppb) for Base/Measured Conditions

	

23 ( C, 50 % RH	

22.83 ( C, 49.6 % RH

	

Upstairs	

Downstairs	

Upstairs	

Downstairs



Softwood plywood as a source	

10.3	

9.0	

10.0	

8.7



Add wallboard as a re-emitting sink (intercept = 0.19)	

43	

36	

41	

35



Change wallboard intercept to 0.095	

22	

19	

22	

19



Measured values	

--	

--	

27	

25



4.3.6	Third Loading

For this repeat of the high loading, the model inputs for sources and
sinks (Table 4-11) were the same as those used for the second loading,
including the intercepts for carpeting and wallboard.  

Table 4-11.  Model Inputs (Sources and Sinks) for Third (High) Loading

Source/Sink	

Slope (m/h)	

Intercept (mg/m2-h)	

Area (m2)



Underlayment	

-1.27	

0.28	

  94



Cabinets	

-0.48	

0.08	

  59



Interior doors	

-0.50	

0.08	

  35



Countertop	

-0.70	

0.30	

    4



Paneling	

-0.98	

0.16	

  36



Carpeting	

-1.30	

  0.095	

  94



Painted wallboard	

-3.00	

  0.225	

420



The measurement results for this loading indicated that concentrations
had “leveled off” (see Figure A-1 in Appendix A) by day 7; measured
values for run 1 (taken on day 7) therefore were used as a basis for
comparison with modeling results. The modeled values were about 10 ppb
higher than measured upstairs but about 25 ppb higher downstairs (Table
4-12). One possible explanation for the greater difference downstairs is
that the humidity level, which was not measured downstairs, may have
been lower than the value used in the model run.

Table 4-12.  Modeling Results for Third Loading

Source/Sink Inputs 	

Results (ppb) for Base/Measured Conditions

	

23 ( C, 50 % RH	

22.63 ( C, 64 % RH

	

Upstairs	

Downstairs	

Upstairs	

Downstairs



Underlayment, cabinets, doors, countertop and paneling as sources 	

70	

60	

89	

76



Add carpeting as a sink	

68	

58	

86	

74



Add wallboard as a sink	

63	

54	

80	

69



Measured values (run 1, day 7)	

--	

--	

72	

43

The results of this assessment collectively indicate that model
predictions typically agreed with measurement results within ± 25
percent. 

4.4	FIAM-pwp Predictions for Recent Residential Field Monitoring Studies

Recent (i.e., since year 2000) residential formaldehyde field studies of
note are summarized in Table 4-13. The first two studies – Offerman et
al. (2008) and Hodgson et al. (2000) – both were conducted on new
homes, with the Offerman study covering 100+ new single-family homes in
California and the Hodgson study covering fewer structures in the
eastern and southeastern regions of the U.S. but including both
manufactured and site-built homes. The average (geometric mean)
formaldehyde concentration indoors was between 30 and 35 ppb for both
studies, with maximum values of 117 and 58 ppb reported by the two
studies, respectively.

Studies of existing homes in urban U.S. areas by Liu et al. (2006) and
Sax et al. (2006) measured lower average (median) formaldehyde
concentrations, on the order of 10-15 ppb, and lower maximum values as
well, with the highest reported maximum (45 ppb) in a Los Angeles home
during the winter. Two Canadian studies (Gilbert et al 2006; Gilbert et
al 2005) had nearly identical results; in each case the average
(geometric mean) indoor formaldehyde concentration was around 25 ppb and
the highest measured concentration was slightly above 70 ppb.  

Some other residential environments of interest have been monitored on
occasion. For example, during December/January 2007-2008 the CDC
measured formaldehyde levels in 500+ structures of the types (travel
trailers, park models and mobile homes) used by FEMA for displaced
Hurricane Katrina/Rita victims (CDC 2008/2010; DHS 2009). The average
(geometric mean) formaldehyde concentration in these structures was 77
ppb (81 ppb in travel trailers) and the maximum measured concentration
was 590 ppb. Under an interagency agreement with the CDC, the Lawrence
Berkeley National Laboratory (LBNL; Maddalena et al. 2008) studied four
FEMA temporary housing units (camper trailers) to assess indoor
emissions of aldehydes and other VOCs. Measured formaldehyde levels in
these structures ranged from 269 to 753 ppb.

Table 4-13. Summary of Recent Formaldehyde Field Monitoring Studies

Study/Reference	Description (Location and Number/Types of Structures)
Formaldehyde Concentration, ppb



Central Value	Range

Offerman et al. 2008	108 new single-family homes in CA	31.1 (median)	3.8
to 116.7

Hodgson et al. 2000	New homes in eastern/SE U.S.:

4 new manufactured homes

7 new site-built homes	Geometric Means:

34

36	

21 to 47

14 to 58

Liu et al. 2006	234 homes in Los Angeles County, CA; Elizabeth, NJ; and
Houston, TX	16.3 (median)	10.2 to 26.4

(5th - 95th percentiles)

Sax et al. 2006	Inner-city homes:

NY City (46) – winter (W), summer (S)

Los Angeles (41) – Winter (W), fall (F)	Medians:

10 W, 15 S

15 W, 12 F	

4-18 W, 5-41 S

6-45 W, 6-26 F

Gilbert et al. 2006	96 homes in Quebec City, Canada	24.0 (geo. mean)	7.8
to 73.2

Gilbert et al. 2005	59 homes in Prince Edward Island, Canada	27.0 (geo.
mean)	4.5 to 71.1

Hodgson et al. 2004	4 new relocatable classrooms	8 (average
indoor-outdoor level)	4 to 12

(indoor-outdoor)

CDC 2008/2010	519 structures:

Travel trailers (360)

Park models (90)

Mobile homes (69)	77 (geo. mean)

81 (geo. mean)

44 (geo. mean)

57(geo. mean)	3 to 590

3 to 590

3 to 170

11 to 320

Maddalena et al. 2008	4 FEMA camper trailers	463 (average)	269 to 753

References Cited in Table 4-13

CDC. 2008/2010. Final Report on Formaldehyde Levels in FEMA-Supplied
Travel Trailers, Park Models, and Mobile Homes. Centers for Disease
Control and Prevention. Available at:

  HYPERLINK
"http://www.cdc.gov/nceh/ehhe/trailerstudy/assessment.htm#final" 
http://www.cdc.gov/nceh/ehhe/trailerstudy/assessment.htm#final . July
2008 (amended December 2010).

N.L. Gilbert, M. Guay, J.D. Miller, S. Judek, C.C. Chan, and R.E. Dales.
2005. Levels and Determinants of Formaldehyde, Acetaldehyde, and
Acrolein in Residential Indoor Air in Prince Edward Island, Canada.
Environmental Research 99(1): 11-17.

N.L. Gilbert, D. Gauvin, M. Guay, M.E. Héroux, G. Dupuis, M Legris,
C.C. Chan, R.N. Dietz, and B. Lévesque. 2006. Housing Characteristics
and Indoor Concentrations of Nitrogen Dioxide and Formaldehyde in Quebec
City, Canada. Environmental Research 102(1): 1-8.

A.T Hodgson, A.F. Rudd, D. Beal, and S. Chandra. 2000. Volatile Organic
Compound Concentrations and Emission Rates in New Manufactured and
Site-Built Houses. Indoor Air 10: 178-192.

A.T. Hodgson, D. G. Shendell, W.J. Fisk, and M.G. Apte. 2004. Comparison
of Predicted and Derived Measures of Volatile Organic Compounds inside
Four New Relocatable Classrooms. Indoor Air 14 (Supplement 8): 135-144.

W. Liu, J. Zhang, L. Zhang, B.J. Turpin, C.P. Weisel, M.T. Morandi, T.H.
Stock, S. Colome, and L.R. Korn. 2006. Estimating Contributions of
Indoor and Outdoor Sources to Indoor Carbonyl Concentrations in Three
Urban Areas of the United States. Atmospheric Environment 40: 2202-2214.

R. Maddalena, M. Russell, D.P. Sullivan, and M.G. Apte. 2008. Aldehyde
and Other Volatile Organic Chemical Emissions in Four FEMA Temporary
Housing Units – Final Report. LBNL-254E, Ernest Orlando Lawrence
Berkeley National Laboratory, Environmental Energy Technologies
Division, Berkeley, CA. 

F. J. Offermann, J. Robertson, D. Springer, S. Brennan, and T. Woo.
2008. Window Usage, Ventilation, and Formaldehyde Concentrations in New
California Homes: Summer Field Sessions. ASHRAE IAQ 2007 Conference,
Baltimore, MD.

S.N. Sax, D.H. Bennett, S.N. Chillrud, P.L. Kinney, and J.D. Spengler.
2004. Differences in Source Emission Rates of Volatile Organic Compounds
in Inner-City Residences of New York City and Los Angeles. Journal of
Exposure Analysis and Environmental Epidemiology 14 (Supplement 1):
S95-S109.

To further assess the predictive capabilities of FIAM-pwp, the model was
used to predict the average or central-value formaldehyde concentration
measured in two of the above field studies – the Offerman et al.
(2008) study of new California homes and the CDC (2008/2010) study of
recently constructed travel trailers. In both cases, the predictions
were made relying primarily on model defaults, occasionally supplemented
in cases where ancillary study measurements suggested that a model input
other than the default value might be more appropriate.

As the title of the Offerman et al. (2008) reference indicates, the
reported field measurements were conducted in new California homes
during the summer season. Neither the average age nor the age
distribution of the study homes was reported; they are simply described
as “new.” The average temperature and humidity in the homes during
monitoring similarly was not reported. The paper does note, however,
that these were single-family detached homes, most of which were further
described as having no mechanical outdoor air systems and no nighttime
ventilation cooling systems. The median air exchange rate in this subset
of homes was reported as 0.33 ACH by Offerman el al. (2008).

Given the above descriptions, the inputs/choices shown below in Figure
4-1 were made on the House Screen. Among those choices/inputs, the
following are changes to model default values:

“Standard Conditions” for temperature (23 °C) and relative humidity
(50 %)

Single-family detached home conceptualized as a single zone, with
airflows edited to yield an air exchange rate of 0.33 ACH.

Figure 4-1. FIAM-pwp House-Screen Inputs Used to Model Study of New
California Homes.

For the Source Screen (see Figure 4-2), model defaults were used for all
PWPs. The only change made was the reassignment of sources associated
with Zone 2 to be located in Zone 1, consistent with the single-zone
conceptualization of the house as noted above.

Figure 4-2. FIAM-pwp Source-Screen Inputs Used to Model Study of New
California Homes.

As shown in Figure 4-3, results of the model run for this case included
a predicted concentration of 39.2 ppb (48.4 µg/m3) for these homes
following construction. By comparison, the authors of the paper reported
a median concentration of 31.1 ppb (38.3 µg/m3; neither the arithmetic
nor the geometric mean was reported) for the subset of new homes modeled
here. As noted above, the age of the monitored houses was not reported
in the paper, other than their being “new.” If the average house age
was six months, for example, then the model prediction of 32.6 ppb
(40.3) µg/m3 would apply; this prediction agrees closely with the
reported median value. The model prediction for these homes at an age of
12 months was 27.4 ppb (33.9 µg/m3.

Figure 4-3. FIAM-pwp Results for Modeling of New California Homes.

Travel trailers accounted for nearly 70 % of the structures that were
monitored by CDC (2008/2010).

The structure type selected on the House Screen was camper trailer using
model defaults, including an air exchange rate of 0.2 ACH, with the
single exception of temperate and humidity, for which the same values
were used as for the single-family detached case. For the Source Screen
(see Figure 4-4), model defaults were used for all PWPs.

Figure 4-4. FIAM-pwp Source-Screen Inputs Used to Model CDC Study of
Travel Trailers.

The model prediction of 78.3 ppb (see Figure 4-5) agrees closely with
the geometric mean (81 ppb) reported by CDC. However, from the
information given in the CDC report, the trailers were approximately one
to two years old when monitored. Thus, the predicted value of 52.1 ppb
at an age of 12 months would be more appropriate. The 12-month predicted
value is about a third lower than the geometric mean of the field
measurements. One possible reason for the difference is that some of the
trailers monitored by CDC may have included import products (e.g., for
wall paneling or cabinets sides/back); the maximum monitored
concentration of 590 ppb among the CDC travel trailers supports this
speculation. If, for example, the intercept for HWPW were doubled to
account for this possibility, then the predicted concentration at 12
months would be 66.2 ppb.

Figure 4-5. FIAM-pwp Results for Modeling of Travel Trailers.

The modeling results indicate that FIAM works reasonably well for its
primary intended purpose – prediction of concentrations in new
residential structures shortly after their construction.

Appendix A

Summary of EPA Formaldehyde Pilot Study

A.1	STUDY OBJECTIVES AND DESIGN 

At a public meeting in January 1993, EPA representatives outlined a plan
for a proposed formaldehyde (HCHO) study that was to be conducted in two
phases – a main study and a pilot study. The main study was to involve
testing using passive monitors in 108 newly constructed conventional and
manufactured houses located throughout the United States. A smaller
pilot study was to be conducted first to evaluate the feasibility and
logistical considerations of the experimental design for the larger main
study.

In September 1994 the National Particleboard Association (NPA) signed a
Cooperative Research and Development Agreement (CRDA) with EPA to fund a
pilot study in newly constructed conventional and manufactured houses.
The general purpose of the pilot study was to evaluate methods used to
measure the contribution of UF-bonded building materials to indoor HCHO
concentrations in newly constructed conventional and manufactured
houses.

The specific objectives of the pilot study were to:

Test logistical considerations relevant to carrying out experimental
procedures for the testing program in a single conventionally-built
single-family house and in several manufactured houses.

Demonstrate that experimental variables or conditions likely to affect
formaldehyde concentrations in new houses (i.e., UF-bonded wood product
emission characteristics and loading rates, temperatures and indoor air
exchange rates) can be controlled, individually and jointly varied, held
sufficiently constant, and that the response can be measured to a
specified precision.

Demonstrate that test results can be obtained across a range of
different experimental conditions, similar to that which can be present
in new houses.

Estimate the extent of variability of the experimental results and the
variation with changes in experimental conditions.

Determine how to account for, or to eliminate or minimize, residual
formaldehyde carryover between test runs in the conventional house due
to effects of inherent sinks.

Evaluate the ability to control and vary the air exchange rate of houses
using an adjustable mechanical air handling system.

The study was guided by a Quality Assurance Project Plan (QAPP) with a
detailed description of the experimental design and variables,
monitoring protocols, and quality assurance/control procedures that
would be employed. According to the QAPP, the pilot study would include
one unoccupied, conventionally built, single-family house and four
unoccupied manufactured houses. Four manufactured houses were specified
in the pilot study because it was not practical to install and remove
products in a manufactured house. The conventional house was to be a
two-story house on a slab or crawlspace. The manufactured-house portion
of the pilot study eventually was dropped due to resource limitations.

The QAPP also specified that the following variables would be evaluated
in the pilot study:

Product emission characteristics: Only “medium” emitting products
were to be used. Emissions from particleboard and from plywood wall
paneling, as measured by ASTM method E 1333 (cited previously in this
document) were to be between 0.12 and 0.14 ppm. Because there were no
established protocols for measuring formaldehyde emissions from
kitchen/bathroom cabinets, commercially available cabinets constructed
with “medium” emitting materials (melamine-wrapped particleboard)
were to be used.

Product loading rate: Two different loading combinations of products –
Medium and High (defined later) – were to be installed in the house.

Environmental conditions in the house:

Temperature of 75 °F

Ventilation rate of 0.5 air changes per hour (ACH)

Relative humidity in the house, while not a controlled variable, was
targeted at 50 percent. Air leakage was to be determined through
blower-door tests, with efforts to make the house as airtight as
possible. Ventilation rates were to be controlled by a mechanical heat
recovery ventilator (HRV) attached to the heating and air conditioning
(HAC) system. Air exchange rates were to be measured by two tracer gas
methods – constant release of perfluorocarbon tracers (PFTs, EPA
Method IP-4A) and periodic injection/decay of sulfur hexafluoride (SF6,
EPA Method IP-4B). All USEPA methods cited on this page have been
compiled in a compendium.

Four product loadings were to be used – two Medium and two High. Prior
to loading the products in the house, a baseline value for the house
without any experimental UF-bonded wood products was to be determined.
The products then would be installed in the house and indoor
formaldehyde concentration measurements taken over approximately a
30-day period. At the completion of 30-day period, the products were to
be removed from the house. The house would be allowed to “air out”
for a period of time until a new equilibrium level (baseline) was
reached, after which the next set of products would be installed. The
order of the loading configurations in the house was to be randomized,
with the single constraint that Loadings 1 and 2 would be different
(i.e., one High and one Medium).

Indoor and outdoor formaldehyde concentrations (24-hour time-integrated
values) were to be measured by EPA method IP-6A (Solid Adsorbent
Cartridge) 7, 12, 28, and 33 days after products were installed in the
house. If necessary, additional testing could take place to confirm any
trends or unexpected results. Readings on days 12 and 33 were considered
statistical replicates of readings on days 7 and 28, respectively.
Indoor formaldehyde concentrations were to be measured in the kitchen,
living room, upstairs bedroom, basement, and outdoors. One duplicate
sample was to be collected on each sampling occasion, at a location to
be varied systematically.

Product emission tests were to be conducted in a large chamber in
accordance with ASTM E 1333, at three different air exchange rates.
Because there were no established protocols for large-chamber testing of
doors and cabinets, the following loading rates were be used:

Doors – five doors, at a total loading rate of 0.125 ft2/ft3

Kitchen cabinets/countertops – one base and one wall cabinet with
doors closed, at a loading rate of 0.133 ft2/ft3. A section of
countertop was to be placed on the base cabinet during testing.

The above loading rates were similar to those established by the
Department of Housing and Urban Development (HUD) for large-chamber
testing of particleboard and industrial panels.

For sink-effect testing of painted gypsum wallboard and carpet/padding,
products were to be placed in a small clean environmental chamber. A
known concentration of formaldehyde gas was to be injected into the
chamber. By accurately measuring time-varying concentrations in the
chamber, it could be determined whether the material was absorbing
formaldehyde gas and then re-emitting it after the source was turned
off.

A.2	STUDY METHODS

Securing a house builder or owner who was willing to participate in the
study proved to be more difficult and costly than anticipated, due
mainly to the constraint that the house could not be occupied during the
test period. A homeowner was located in Centreville, MD, who was
building a rental house and was willing to lease it for as long as
needed. The house was a conventionally built, 1326 ft2, two-story Cape
Cod style with a full basement (volume of 168 m3). The QAPP had called
for a house with crawl space, but efforts to find such a house in the
study area were unsuccessful. The total volume of the finished living
space was 10,746 ft3 (304 m3). The first floor of the house (712 ft2
floor area) had 5931 ft3 (168 m3) of living space that consisted of a
living room, kitchen, bathroom/laundry room, and bedroom. The second
floor of the house (614 ft2 floor area) had 4815 ft3 (136 m3) of living
space that consisted of two bedrooms and a bathroom. The door from the
first-floor finished living space to the unfinished basement was sealed
shut during the entire project. The house contained no furniture or
draperies. 

Industry representatives were responsible for procuring all pressed-wood
products that would be used in the pilot study. Sufficient quantities of
5/8” particleboard underlayment, 3/4” industrial particleboard for
countertops, 1/4” hardwood plywood wall paneling (3-ply birch face,
tropical hardwood back and core with 7 cut grooves along the length of
each panel to simulate random-width lumber planking), interior partition
doors, and kitchen and bathroom cabinets were obtained from member
companies or participating associations, or were purchased at local
building supply centers. Emission characteristics of the particleboard
and plywood wall paneling used in the study were determined prior to
selecting the materials for testing. 

Initial emission characteristics of the products used in the study (see
Table A-1) were determined by ASTM E 1333 at 0.5 ACH. All products were
stored in a controlled-access warehouse until they were installed in the
house. During storage, products were wrapped in 6-mil plastic to
minimize formaldehyde off-gassing. It was discovered during the study
that temperature control in the warehouse was marginal and that
temperatures ranged up to 85 ° F.

Products were installed in the house according to manufacturers’
recommended installation instructions, and conventional practices were
followed. The particleboard underlayment was installed with screws for
easy removal. The kitchen and bath cabinets were screwed tightly against
the wall. The plywood wall paneling was nailed to the wall to minimize
damage to the gypsum wallboard.

Table A-1. Initial Emission Characteristics of Products Used in the
Pilot Study

Product	Emissions, ppm	Loading Rate, ft2/ft3

Particleboard Underlayment	0.144	0.130

Plywood Wall Paneling	0.114	0.290

Cabinets	0.053	0.133

Interior Doors	0.052	0.125



Loading areas for pressed-wood products are shown in Table A-2. For the
Medium loading scenario, particleboard underlayment was installed on the
first floor, along with interior doors on both floors and a full set of
kitchen/bathroom cabinets and countertops. For the High loading
scenario, particleboard underlayment was installed on both the first and
second floors. In addition, twelve 4’ X 8’ sheets of plywood wall
paneling were installed. To establish a relatively uniform loading and
to avoid using partial panels, four full-size sheets were installed in
an upstairs bedroom, in the downstairs bedroom, and in the living room.
Interior doors and a full set of kitchen and bathroom cabinets and
countertops (same number as for Medium loading) also were installed. The
total exposed surface area for cabinets was calculated to be 631 ft2 (59
m2). Particleboard underlayment was not installed in the kitchen and
bathroom areas of the house, as those areas were covered by vinyl sheet
goods that had been permanently installed during final preparation of
the house before field testing was initiated.

Table A-2. Loading Areas for Pressed-wood Products at the Conventional
House

Component

	

Loading Area, ft2

	

1st Floor	

2nd Floor

	

Medium Loading	

High

Loading	

Medium

Loading	

High

Loading



Underlayment	

496.2	

496.2	

---	

519.3



Paneling	

---	

256.0	

---	

128.0



Doors	

203.6	

169.6



Countertop	

  37.9	

     5.65



Blower-door tests indicated that the house had numerous leaks in the
building shell and between the first floor and basement. Air leakage
sites were identified (service penetrations between the basement and
first floor and along the base plate) and sealed with caulk and foam.
Numerous leaks also were found in the forced-air HAC distribution
system. Sealing procedures reduced the leakage rate by more than 50
percent, from ~ 10 ACH at 50 Pa (corresponding to ~ 0.5 ACH under
moderate environmental conditions) to ~ 4 ACH at 50 Pa.

Initial study plans did not include measurement of formaldehyde
concentrations in the basement, because it was thought that the
combination of house sealing procedures and a closed interior door at
the top of the stairwell to the basement would minimize air
communication with the remainder of the house. However, during tracer
gas studies to characterize the air exchange rate for the house it
became obvious that tracer injected into the living area was being
transferred to the basement. Consequently, it was decided to include
formaldehyde measurements in the basement. In addition, PFT sources and
samplers were configured so that the average airflow rate between the
basement and living area could be quantified during each sampling event.

Following initial baseline testing, products for Loading 1 (Medium) were
installed and indoor concentration levels were measured in accordance
with the protocols outlined in the QAPP. It became evident during this
time that there was difficulty in controlling relative humidity levels
in the house, which reached over 70 percent. Humidity levels were
lowered to the desired range by adding additional dehumidification
equipment in the house. Because high humidity causes higher formaldehyde
emissions, the house was declared to be in an “upset condition”
during this time and it was decided that Loading-1 products should be
left in the house until conditions stabilized. The final readings for
Loading 1 were taken 78 days after product loading.

It also was determined during this time that resource constraints would
not permit the full implementation of the loading schedule established
in the study plan, which was modified for conduct of only three loadings
in the house – one Medium and two High loadings. In addition, the
air-out period between successive loadings was fixed at seven days
instead of waiting for formaldehyde levels to reach equilibrium.

Following the removal of Loading-1 products from the house and a
seven-day airing-out period, Loading-2 (High) products were installed.
Formaldehyde concentrations were measured on days 7, 12, 28, and 33.
Following removal of Loading-2 products from the house and a seven-day
airing-out period, Loading-3 (High) products were installed.
Formaldehyde concentrations again were measured on days 7, 12, 28, and
33 after loading. Following removal of Loading 3 products, the house was
returned to the homeowner. 

A limited amount of sink-effect testing was conducted as part of the
pilot study. From results of the laboratory sink-test results and
results from Loadings 1, 2 and 3, it appeared that the gypsum wall board
was acting as a significant sink. Because the exposed surface area of
the wallboard was greater than all of the UF-bonded products installed
in the house, it was considered critical to fully characterize the sink
effect behavior of painted gypsum wallboard. It also was considered
important to determine whether the carpet/padding used in the house was
acting as a sink or had any barrier effect on formaldehyde emissions
from the particleboard underlayment. 

The original protocol for “barrier testing” was modified to evaluate
the floor system with both particleboard underlayment and
carpet/padding, instead of just the carpet/padding. A sample of
particleboard underlayment was placed in a stainless-steel pan and
inserted in a small chamber. Formaldehyde concentrations in the chamber
were measured continuously with a real-time (Interscan) analyzer. As a
separate test, a second sample of particleboard with carpet/padding
directly above it was placed in a stainless-steel pan and inserted in
the chamber. Differences in the concentration profiles for the two tests
would reflect the barrier effect of the carpet/padding used in the pilot
study. To minimize the inherent variability within a piece of
particleboard, specimens used for these tests were prescreened via
small-chamber emission tests – the two pieces with the closest
emission rates were used for the tests.

A.3	STUDY RESULTS

The results of formaldehyde, temperature, humidity and air-exchange
measurements during the formal “runs” associated with each loading,
as well as their respective baseline periods, are compiled in Table A-3.
Formaldehyde concentrations measured in the upstairs rooms generally
were similar to one another and higher than those measured in the
basement. The highest concentration measured in the house – 76 ppb –
was in the kitchen during Loading 3 (High). The monitoring results
collectively indicate that, for all three loadings, indoor HCHO
concentrations had peaked by the time of measurements taken 33 days
after product loading. The formaldehyde results also indicate that
monitored levels for Loadings 2 and 3 generally were consistent with one
another and that these “high” loadings had higher HCHO levels (but
also a higher baseline) than the “medium” loading (Loading 1).

Although measured air exchange rates generally were in the target range
of 0.4 to 0.6 ACH, the PFT estimates consistently were higher than those
based on SF6. One possible reason for the discrepancy is the inherent
difference between the two measurement technologies. The SF6 method
involves periodic injection and mixing of the tracer through the HAC
system, followed by real-time analysis as the concentration declines.
The PFT method relies on a near-constant release of tracers from
multiple locations near the perimeter of the house, with time-integrated
sampling over a 24-hour period. 

Table A-3.  Summary of Monitoring Results for Loadings 1, 2, and 3

Runa	

Formaldehyde Concentration, ppb	

Temperature, (F	

RH, %	

Air Exchange, ACH

	

LR	

KIT	

BR	

BMT	

AMB	

LR	

KIT	

BR	

BMT	

AMB	

LR	

2ND	

AMB	

SF6	

PFT



Loading 1 (Medium)



Baseline	

9.1	

9.8	

9.5	

9.1	

1.3	

75.7	

75.8	

74.0	

63.4	

49.0	

22.8	

24.4	

55.2	

--	

0.50



1	

27.5	

29.9	

20.1	

--	

1.0	

73.8	

73.4	

73.2	

68.5	

56.9	

43.2	

41.0	

60.2	

0.44	

0.54



2	

29.6	

31.3	

29.9	

26.5	

7.0	

73.8	

73.6	

73.4	

66.7	

52.4	

46.2	

43.2	

76.8	

0.44	

0.54



3	

41.7	

43.9	

36.3	

28.2	

2.1	

75.1	

75.0	

74.9	

71.1	

69.6	

63.1	

60.2	

93.9	

0.40	

0.54



4	

36.8	

39.4	

31.1	

--	

1.7	

75.5	

74.8	

73.4	

73.1	

66.0	

50.4	

49.8	

60.6	

0.40	

0.57



Loading 2 (High)



Baseline	

23.0	

23.7	

22.1	

25.7	

2.8	

75.2	

73.1	

74.3	

77.0	

84.4	

54.5	

52.3	

80.3	

0.39	

--



1	

60.1	

60.8	

60.5	

39.7	

2.9	

75.3	

73.5	

74.5	

76.2	

81.5	

57.1	

55.5	

88.2	

0.38	

0.61



2	

55.8	

58.1	

54.7	

40.0	

2.3	

75.1	

73.4	

74.3	

75.4	

80.1	

50.9	

49.3	

74.1	

0.38	

0.71



3	

46.7	

48.5	

53.3	

38.0	

1.4	

74.4	

73.9	

75.6	

75.7	

75.4	

50.7	

43.8	

71.2	

0.36	

0.59



4	

39.0	

40.8	

48.1	

30.7	

0.9	

73.4	

73.2	

73.2	

73.6	

63.1	

51.6	

43.2	

71.6	

0.39	

0.60



Loading 3 (High)



Baseline	

26.0	

25.8	

27.7	

25.0	

<1.0	

73.3	

72.8	

73.3	

73.3	

61.4	

54.9	

44.3	

85.0	

0.41	

--



1	

71.9	

76.1	

65.6	

42.6	

1.1	

75.0	

73.2	

72.5	

73.6	

76.1	

62.7	

65.3	

95.1	

0.39	

0.69



2	

53.8	

55.2	

54.7	

29.9	

1.1	

72.3	

72.4	

72.0	

70.7	

60.3	

49.7	

47.1	

86.7	

0.36	

0.64



3	

43.2	

46.2	

39.5	

20.3	

1.1	

71.7	

71.8	

69.5	

66.6	

49.1	

43.8	

43.7	

87.2	

0.42	

0.62



4	

42.3	

40.4	

44.1	

22.8	

1.4	

71.7	

71.9	

70.7	

66.9	

55.4	

46.1	

44.9	

88.8	

0.35	

0.53

	Legend:						a Run 1 occurred 7 days after product loading;

	LR = living room; KIT = kitchen;			  Run 2 occurred 12 days after
product loading;

	BR = bedroom; BMT = basement;		  	  Run 3 occurred 28 days after
product loading;

	AMB = ambient;	2ND = second floor.		  Run 4 occurred 33 days after
product loading.

Upstairs measurements (average of kitchen, living room and bedroom
locations) for the three loadings are plotted in Figure A-1 against
elapsed time after UF-bonded wood products were installed (day
“zero” indicates baseline measurements). The curvature of the lines
is due to the spline fit that was applied and is not meant to imply that
values between measurement points can be readily interpolated. The
figure illustrates the general consistency across Loadings 2 and 3 and
their differences from Loading 1. Although concentrations for Loading 1
are similar to those for Loadings 2 and 3 toward the end of their
respective test periods, indoor humidity levels were relatively high
near the end of Loading 1.

Figure A-1. Upstairs Formaldehyde Concentrations over Time for Three
Loadings.

Differences in baseline levels and in the general shape of the time
series for Loading 1 versus Loadings 2 and 3 can be explained partly by
the wallboard sink effect. The wallboard sink was largely “empty” at
the outset of Loading 1 but was largely “pre-loaded” at the outset
of Loadings 2 and 3, as reflected in the higher baseline values for the
later loadings. The time series for Loadings 2 and 3 have a shape
consistent with that of an exponentially declining emitter, as would be
expected when the dominant sink (wallboard) has been pre-loaded and
becomes a “net zero emitter,” such that the declining rate of
emissions from aging wood products drives the concentration profile. For
Loading 1, however, this profile is dampened by the relatively large
mass being absorbed by the wallboard sink in an “unloaded” state.
Some of the difference in baseline values for Loading 1 versus Loadings
2 and 3 also may be due to the relatively low humidity level at the time
when baseline measurements were taken for Loading 1.

The key measurement parameter for the study, formaldehyde by the DNPH
method, was in control across all measurement periods. QA spikes
indicated that recoveries were consistently in the range of 90-110
percent, well within the accuracy goal of ±20 percent. Results of
duplicate DNPH samples indicated a combined sampling and analytical
precision of ±10 percent or better, again well within the corresponding
data quality objective. The standard deviation across the high loadings,
for measurements at the same location and at the same elapsed time since
loading of UF-bonded wood products, was not much larger than that for
duplicate measurements (i.e., at the same time and location within any
given loading). This finding points to the repeatability of high-loading
results for the conventional house, after the sinks were pre-loaded as a
result of the preceding medium loading.

Emission rates for UF-bonded wood products that were installed in the
pilot study house, based on large-chamber concentrations that were
adjusted to standard conditions of 77 ( F and 50% RH per ASTM E 1333,
are listed in Table A-4. Each product was tested at target air exchange
rates of 0.25, 0.5 and 1.0 ACH. Two patterns that were evident for all
products are noteworthy. First, as the air exchange rate is increased,
the emission rate also increases, but at a rate that is less than
proportional to the increase in air exchange. This trend is indicative
of a concentration “backpressure” effect, as applied in the Matthews
model. Second, for each product the emission rate was lower for products
used for later loadings (products were tested near the time of each
loading). These results indicate that some emissions decay likely
occurred during storage, even though products were wrapped in plastic.
The elevated temperature in the warehouse may have accelerated the decay
process somewhat.

Table A-4.  Computed Emission Rates for UF-Bonded Products

Loading	

Emission Rate, µg/m2-h

	

Underlayment	

Paneling	

Cabinets	

Doors



At Target ACH of 0.25a



1

2

3	

104.9 (0.25 ACH)

99.2 (0.25 ACH)

90.4 (024 ACH)	

Not Used

42.9 (0.25 ACH)

32.9 (0.25 ACH)	

45.6 (0.17 ACH)

45.2 (0.24 ACH)

45.7 (0.25 ACH)	

52.0 (0.24 ACH)

Not Tested

49.9 (0.25 ACH)



At Target ACH of 0.50



1

2

3	

178.3 (0.50 ACH)

162.6 (0.50 ACH)

149.9 (0.51 ACH)	

Not Used

72.3 (0.50 ACH)

56.8 (0.50 ACH)	

77.1 (0.51 ACH)

69.9 (0.50 ACH)

68.5 (0.50 ACH)	

74.2 (0.50 ACH)

Not Tested

68.4 (0.51 ACH)



At Target ACH of 1.0



1

2

3	

221.9 (1.01 ACH)

200.6 (0.99 ACH)

190.2 (1.01 ACH)	

Not Used

104.9 (0.99 ACH)

77.0 (1.01 ACH)	

82.8 (1.00 ACH)

73.4 (0.99 ACH)

74.9 (1.01 ACH)	

79.9 (1.00 ACH)

Not Tested

74.2 (1.00 ACH)

a Actual air exchange rates during each test are given in parentheses.

Other products used in the house were assessed for formaldehyde emission
characteristics. For the paint used on the wallboard and interior doors
at the study house, the manufacturer’s data indicated that the
wallboard paint contained 0.001% formaldehyde by weight and the door
paint contained 0.007% formaldehyde. A carpet/padding sample was sent to
EPA’s Air and Energy Engineering Research Laboratory (AEERL) in
Research Triangle Park for small-chamber testing. Results of the chamber
tests, conducted at 0.5-0.55 ACH, 74 ( F and 55-60 % RH, indicated an
emission rate of 1.4 µg/m2-h for the carpet and the same rate for the
padding (after subtracting the chamber background) at an elapsed time of
24 hours following insertion in the chamber, or an emission rate of 2.8
µg/m2-h for these two constituents combined. The combined emission rate
is one to two orders of magnitude lower than that for UF-bonded wood
products.

Small-chamber tests were conducted to evaluate sink characteristics of
the carpet and padding.  Two tests were conducted under identical air
exchange rates of 1.5 ACH, but with different formaldehyde
concentrations in the input stream because the rate of adsorption to the
sink was suspected to be a function of the air concentration. In both
tests, a 7.1-inch square piece of carpet and padding was placed in a
tight-fitting aluminum pan with edges equal to the depth of the carpet
and padding. The pan containing the carpet and pad was then placed
horizontally on the chamber floor. The chamber was sourced with a
formaldehyde stream of 0.2 ppm during the first test and 0.12 ppm during
the second. Comparison of the theoretical (concentrations that would be
expected in the absence of sinks) and measured concentrations indicated
that formaldehyde was adsorbed by the carpet, and the trend in the data
appeared to reflect a first-order removal process. 

Chamber tests for gypsum wallboard were conducted to evaluate its sink
characteristics. As with the carpet tests, these two tests were
conducted under identical conditions with the exception of the
formaldehyde feed-stream concentration, to evaluate the effect of
concentration on the rate of mass transfer to and from the wallboard.
For each test, two 10.5-inch square pieces of painted wallboard were
securely fastened back-to-back and edge sealed with sodium silicate,
leaving a total of 220.5 square inches of exposed wallboard surface
area. The wallboard was then placed in the center of chamber in a
vertical position, supported by stainless steel wire. The chamber was
sourced with a formaldehyde stream of 0.2 ppm during the first test and
0.12 ppm for the second, at 1.5 ACH. The resulting data from these tests
indicated that the sink effect was more pronounced for wallboard than
for the carpet/padding ensemble. The concentration data were fit to a
Langmuir isotherm model to estimate adsorption and desorption rate
constants for the two types of sinks.

Tests to evaluate the behavior of underlayment both with and without a
carpet/padding barrier were conducted under identical chamber conditions
using a tight-fitting aluminum pan with edges equal to the depth of the
underlayment, carpet and padding. In the first test, a 7.1-inch square
piece of underlayment was placed in the pan with nothing above it. In
the second test, a 7.1-inch square piece of underlayment was placed in
the pan with an equally sized piece of padding and carpet placed above
it. For both tests, the pan was then placed horizontally on the chamber
floor. The resulting data indicated that the underlayment was releasing
formaldehyde.  While it was tempting to conclude from the results that
the carpet and padding had no effect on the emission of formaldehyde,
the carpet/padding sink test described above indicated otherwise. 

A.4	STUDY CONCLUSIONS

The following conclusions, primarily related to the objectives given in
Section A.1, were reached as a result of the pilot study:

Some logistical difficulties were encountered that prevented the study
from being completed as outlined in the QAPP. Among the greatest
difficulties were locating a study house and maintaining humidity levels
in the house within the prescribed range. Humidity was not a primary
experimental variable for the study, but was considered to be an
important covariate that should be tightly controlled. Logistical
aspects such as acquiring, storing and accessing the materials
associated with house loading went remarkably smoothly, due in part to
the high level of cooperation from various industry representatives and
their close working relationship with the study team.  

Little difficulty was encountered in controlling most experimental
variables at the study house. The primary variables – product loading
rate, emission rate, temperature, and air exchange – all were
generally kept within their respective target ranges. Some emissions
decay was evident for wood products stored for successively later
loadings.  

It was demonstrated that different test results could be obtained across
different experimental conditions and that sufficient precision in
measurement response could be obtained. Differences across the
experimental conditions (medium versus high loading) could be
distinguished, although these differences were partially dampened by the
substantial sink effect of the gypsum wallboard. Appropriate precision
was obtained largely because of the consistently high resolution,
accuracy and precision for the sampling method (solid adsorbent
cartridge) with a 24-hour sampling duration.

The study data were marginally sufficient for estimating the extent of
variability of results and the variation with changes in experimental
conditions. The limited number of loadings, coupled with the unique
circumstances of the first loading (little mass in the indoor sinks at
the outset), made it difficult to properly estimate the variance
component associated with repeats of the same loading configuration.
However, it was demonstrated that the variance across repeats of the
high loading, for a specific sampling location at the same elapsed time
relative to loading of the house, was only marginally higher than the
variance for duplicate formaldehyde samples.

Formaldehyde carryover between successive tests at the conventional
house, due to inherent sinks, was not trivial. The sink effect, due
largely to adsorption of formaldehyde by painted gypsum wallboard, was
not eliminated by the prescribed “airing out” of the house. The
monitoring results, however, did suggest that baseline formaldehyde
values were reasonably reproducible once the house had been loaded with
UF-bonded wood products.

The ability to control and vary the air exchange rate of the
conventional house using a heat recovery ventilator (HRV) was
demonstrated. This ability was demonstrated for two of the three target
air exchange rates (0.2 and 0.5 ACH) planned for the full study; the
third target (1.2 ACH) could not be addressed because the HRV lacked
sufficient capacity. Achieving the lower target of 0.2 ACH was made
possible by sealing procedures that substantially reduced the air
leakage of the house and, thus, the sensitivity of its air exchange rate
to weather conditions.



Appendix B

Model Sensitivity

B.1 	ANALYSIS BY MATTHEWS

Prior work by Matthews and colleagues at Oak Ridge National Laboratory
during the mid-1980s forms the primary basis for FIAM-pwp, as noted in
Section 1 of the main document. The researchers generated a series of
reports documenting their work for the U.S Consumer Product Safety
Commission. One of these reports included a sensitivity analysis for
formaldehyde concentration and emission-rate models; the concentration
model was very similar to that used for FIAM-pwp, including temperature
and humidity adjustments. The measure of sensitivity was the fractional
change in model output per unit change in each coefficient; for example,
a sensitivity of 0.1 would mean that a 10% change in the model is
expected with a 100% change in the coefficient.

The results of the sensitivity analysis depended on both the values of
the model coefficients determined for a given pressed-wood product data
set and the environmental conditions that are substituted into the
model. Results of the analysis indicated that the concentration model
was most sensitive to values for the coefficients representing the
temperature and relative humidity (RH) dependence, regardless of product
type. The highest value for the sensitivity measure, for MDF, was 0.45
for temperature; similar, but somewhat lower, sensitivity values were
found for particleboard and paneling. The highest sensitivity value for
RH was 0.40, for paneling.

B.2 	ANALYSIS BY VERSAR

Versar performed sensitivity analysis on a prior, mathematically
identical, version of FIAM-pwp to guide the choice of factors to be
varied and the relative number of variations for each as part of an
earlier set of model runs. This analysis was conducted for two cases
available in the prior version of FIAM-pwp – a conventional home
(i.e., the conventionally built single-family detached home used in the
EPA formaldehyde pilot study; see Appendix A) and a manufactured home.
The conventional home, with a volume of 471 cubic meters, is
conceptualized as having two zones whereas the manufactured home is a
single-zone structure with a volume of 222 cubic meters. 

As summarized in Table B-1 for the conventional home, ten factors (input
parameters) were systematically varied for the sensitivity analysis. The
background concentration and average concentration encountered when an
individual is away from home both were set to zero for this analysis, so
that modeled concentrations and exposures would reflect only the impact
of PWPs in the house. The source and sink loading areas in the table
represent one of the actual loading conditions that were used in the
pilot study. 

Sensitivity was examined by varying each input parameter – one at a
time – by 20 percent in each direction (i.e., 20 % lower and 20 %
higher) from the base condition shown in Table B-1. Five model outputs
were examined for sensitivity to the ten factors: (1) initial
concentration in zone 1 (upstairs); (2) concentration 24 months later in
zone 1; (3) average daily concentration (ADC); (4) average daily dose
(ADD); and (5) % of time the concentration is ≥ 0.01 ppm (10 ppb).
Values of these outputs for the base condition are listed in Table B-2.

Table B-1. Input Values for Sensitivity Analysis – Base Condition for
Conventional Home

Input Parameter(s)	Base Value(s)

Air Exchange Rate (ACH)	0.59

Emissions Half Life (years)	2.92

Temperature (°C)	23

Relative Humidity (%)	50

Source Loading Areasa (m2) 	46, 59, 35, 4

Sink Loading Areasa (m2)	94, 420

House Age at Move-in (years)	5

Years Lived in House	5

Hours in House on Weekday, Weekend Day	15, 18

Inhalation Rate (m3/day)	13.3

a Sources are underlayment, cabinets, interior doors, and countertop;
sinks are carpet/padding and painted wallboard.

Table B-2. Model Results for Base Condition for the Conventional Home

Initial Conc in Zone 1, 

mg/m3	Conc 24 Months Later in Zone 1, mg/m3	Average Daily Concentration,
mg/m3	Average Daily Dose, 

mg/kg-day	% of Time Concentration  ≥ 0.01 ppm

0.0514	0.0320	0.0061	0.0011	13.7



Changes in each model output associated with the change in each input,
in each direction, were averaged and then expressed as a percent change
from the result for the base condition. The sensitivity results are
summarized in Figure B-1 as the percent change for each model output
associated with a 20 % change in each input. Certain inputs, by
definition, do not affect certain model outputs. For example, the
emissions half life has no effect on the initial concentration and the
inhalation rate affects only the average daily dose. Based on this
analysis, the most sensitive input parameters, in order of relative
sensitivity (most sensitive first), are indoor temperature, emissions
half life, house age, and indoor humidity.

Figure B-1.  Relative Model Sensitivity for the Conventional Home.

For the manufactured home (Table B-3), the same ten factors were
systematically varied for the sensitivity analysis. As with the
conventional home, the background concentration and average
concentration encountered when an individual is away from home both were
set to zero for the analysis. A broader set of PWPs, with generally
higher loading rates, was used for this case. As a result, the modeled
initial concentration was about three times as high as the conventional
home case (Table B-4). Due to the higher concentration, a higher
threshold (0.02 ppm) was used to examine the relative frequency of
higher concentrations. As with the conventional home, the most sensitive
input parameters for this case are indoor temperature, emissions half
life, house age, and indoor humidity (Figure B-2).

Table B-3. Input Values for Sensitivity Analysis – Base Condition for
Manufactured Home

Input Parameter(s)	Base Value(s)

Air Exchange Rate (ACH)	0.50

Emissions Half Life (years)	2.92

Temperature (°C)	23

Relative Humidity (%)	50

Source Loading Areasa (m2) 	95, 53, 26, 3, 43, 36

Sink Loading Areasa (m2)	72, 158

House Age at Move-in (years)	5

Years Lived in House	5

Hours in House on Weekday, Weekend Day	15, 18

Inhalation Rate (m3/day)	13.3

a Sources are PB underlayment, paneling, interior doors, countertop
(particleboard),  cabinets, and closet shelving (PB/MDF); sinks are
carpet/padding and wallboard.

.

Table B-4. Model Results for Base Condition for the Manufactured Home

Initial Conc in Zone 1, 

mg/m3	Conc 24 Months Later in Zone 1, mg/m3	Average Daily Concentration,
mg/m3	Average Daily Dose, 

Figure B-2.  Relative Model Sensitivity for the Manufactured Home.



Appendix C

Basis for Emission Classes

As described in Sections 2.3.1 and 2.3.3 of the main document, default
values for FIAM emission-rate parameters have been developed for four
emission classes: Baseline; CARB1; CARB2; and NAF. Further details on
the basis for these default values are described below, under the
headings of Baseline Emissions Class (Section C.1) and Reduced Emissions
Classes (Section C.2; includes CARB1, CARB2, and NAF). This appendix has
been extracted/adapted from the draft formaldehyde exposure assessment
report cited in Appendix B (see Section B.2).

C.1	BASELINE EMISSIONS CLASS

Between June and October 2010, EPA administered a survey to domestic
manufacturers of composite wood panels; the survey asked about both
current and planned emission levels. Because many CWP manufacturers are
in the midst of modifying their production technology and raw materials
to achieve compliance with Phase 2 of the CARB Air Toxics Control
Measure (ATCM), using current emission levels to estimate emission
levels in 2013 would not accurately reflect the 2013 baseline. Instead,
baseline emission levels were estimated using responses to a survey
question about emission levels that mills planned to achieve within the
next three years. These levels were weighted by production levels that
mills reported in the survey to estimate average emission levels by
certification standard. The survey data also were used to estimate the
share of total production volume represented by products meeting each of
the emission standards. The survey results are reported in Table C-1.
For comparison purposes, the CARB emission limits are shown in Table
C-2. 

Table C-1. Baseline Emission Levels (2013) For Domestically Produced
CWPs, 

by Product Type and Emissions Certification Standard

Emissions Standard	Product Type

	HWPW	MDF	PB

Weighted Average Emissions (ppm)

None	0.058	---	0.058

CARB 1 a	---	---	0.090

CARB 2	0.032	0.082	0.057

CARB ULEF b

0.024	0.042

CARB NAF c	0.013	0.035	0.013

Share of Production

None	0.4%	---	0.2%

CARB 1	---	---	1.8%

CARB 2	60.2%	86.1%	88.5%

CARB ULEF

4.6%	2.8%

CARB NAF	39.4%	9.3%	6.7%

a Some mills producing products currently certified under CARB Phase 1
(CARB 1) reported that they would continue this product line after the
CARB Phase 2 (CARB 2) effective date.

b Hardwood plywood mills meeting CARB 2 and CARB Ultra Low Emission
Formaldehyde (ULEF) were grouped together in the CARB 2 category.
Several HWPW mills appeared to be meeting CARB ULEF standards but did
not report this in their response. For example, they listed the
certification standard as “CARB 2” but the resin type as “ULEF”
and/or reported maximum emissions that satisfy ULEF requirements. Due to
difficulties in distinguishing between CARB 2 and CARB ULEF product
lines and the fact that emission levels appeared similar among mills
reporting either CARB 2 or CARB ULEF certification standards, these
mills were grouped together.

c NAF = No Added Formaldehyde.

Table C-2. Formaldehyde Emissions Standards in the CARB Air Toxic
Control Measure for Formaldehyde Emissions from Compressed Wood
Productsa (ppm)

Emissions Standard	HWPW	MDF	Particleboard

CARB 1	0.08	0.21	0.18

CARB 2	0.05	0.11b	0.09

CARB ULEF –       less frequent testing	0.05	90% of samples < 0.06

All samples < 0.09c	90% of samples < 0.05

All samples < 0.08

CARB ULEF – exemption from third party certification	90% of samples <
0.04

All samples < 0.05	90% of samples < 0.04

All samples < 0.06	90% of samples < 0.04

All samples < 0.06

CARB NAF	90% of samples < 0.04

All samples < 0.05	90% of samples < 0.04

All samples < 0.06	90% of samples < 0.04

All samples < 0.06

a The Formaldehyde Standards for Composite Wood Products Act (FSCWPA)
emission limits are the same as the CARB2 ATCM limits.

b The CARB 2 emission limit for thin MDF is 0.13 ppm.

c The CARB ULEF limits for thin MDF are 90% of samples < 0.08 ppm and
all samples < 0.11 ppm.

Among imported products, baseline emission levels were estimated
separately for products from Canada versus those from the rest of the
world. The known composite wood manufacturers in Canada were compared to
the CARB list of certified mills, and most Canadian mills were found to
be CARB-certified. As a result, Canadian mills were assumed to achieve
the same baseline emissions as U.S. mills (whose emissions are shown
above in Table C-1).

The baseline emission levels for non-Canadian imported CWPs were
estimated using data from a 2009 presentation by Professional Service
Industries, a CARB Third Party Certifier (TPC). Although PSI is located
in the U.S., the mills that it certifies are located mainly in Asia. The
mills certified by PSI are located in countries that are major sources
of pressed wood products imported into the United States; thus, the data
from the PSI presentation were assumed to represent emission levels from
all non-Canadian CWPs imported into the U.S. 

The PSI presentation included histograms of observed formaldehyde
emission levels for hardwood plywood, MDF, and particleboard. The data
included emission levels for 387, 198, and 193 samples of hardwood
plywood, MDF, and particleboard, respectively, and for all three product
types combined. After imputing the data underlying these graphs, the
average emission values reported in the presentation were successfully
replicated. This information was used to estimate the number of samples
at different emission levels.

The emission levels reported in the PSI presentation ranged from near
zero to levels exceeding the CARB 2 emission standards by a substantial
margin. The emission levels observed in these data were from a period
after CARB 1 took effect but before CARB 2 did. Because many of the
mills that met the CARB 1 standards at the time when samples were taken
may subsequently achieve CARB 2 levels, baseline emissions were
predicted by assuming that products with emission levels between the
CARB 1 and CARB 2 levels in the PSI data would achieve CARB 2 levels by
2013. Emission levels of 0.06 and 0.05 ppm were used for CARB 2 baseline
emissions for MDF and particleboard, respectively, because these levels
represent the midpoints of the highest emission intervals in the PSI
data that meet the CARB 2 limits. None of the intervals in the PSI data
for HWPW reflected the CARB 1 emission limits (the CARB 1 limit of 0.08
ppm is spanned by the interval of 0.075 to 0.125 ppm in the PSI data and
the CARB 2 limit of 0.05 ppm is spanned by the interval of 0.0375 to
0.075 ppm). Therefore, no adjustments were made to hardwood plywood for
the shift from CARB 1 to CARB 2. After making the MDF and PB
adjustments, the average emissions by product type were calculated (see
Table C-3).

Although a number of samples in the PSI data had emission levels close
to zero, it was not possible to determine whether the products qualify
for certification as ultra low-emitting formaldehyde (ULEF) or no added
formaldehyde (NAF); the PSI presentation did not identify the resin type
used for each sample. Also, as indicated in a footnote to C-2 above, NAF
and ULEF have emissions requirements that must be met by 90% of the
samples from a mill as well a ceiling value that must be met by all
samples. This distribution of emissions could not be determined from the
PSI presentation; thus, baseline emission levels were not estimated
separately for non-Canadian imports of ULEF and NAF products. Such
products are included in the baseline emission estimates for the CARB-2
category.

Table C-3 also includes the share of imports for each product type and
emissions standard. This share was estimated by assuming that the
fraction of samples for each category in the PSI presentation is
equivalent to the share of imports for that category across all
non-Canadian composite wood products imported into the U.S.

Table C-3. Baseline Emission Levels for Non-Canadian Imported Composite
Wood Products, by Product Type and Emissions Standard

Emissions Standard	Hardwood Plywood	MDF	Particleboard

Average Emissions (ppm)

None	0.163	0.502	0.384

CARB 2	0.015	0.047	0.036

Share of Imports

None	26%	11%	38%

CARB 2	74%	89%	62%



The estimates presented above in Table C-3 rely on several critical
assumptions that are a source of uncertainty in this analysis, the most
important being as follows:

The sample is representative of imported non-Canadian composite wood
products.

It is not known whether the sample is representative of products
destined for the United States. The destination of these products could
be other countries or their country of origin. On the one hand, U.S. end
users may be more likely to insist on lower-emitting products. On the
other hand, U.S. end users not shipping to California face less
stringent formaldehyde emissions standards than many other potential
destinations. (Although many end users prefer to handle a single type of
product, rather than one product that meets the California requirements
and another type for the other 49 states). The fact that the samples
were being tested for formaldehyde emissions by a CARB TPC suggests that
many of them were produced by mills whose products ultimately end up in
the U.S.  

The distribution of emission levels will not change for mills that were
producing products with emission levels that currently exceed the CARB 1
standards.

If some of these mills decide to obtain CARB 2 certification, it is not
known whether these will be the higher or lower emitting mills on
average. Some additional mills (aside from those below the CARB 1 limits
but above the CARB 2 limits) may subsequently decide to lower their
emission levels.  

Conversely, the PSI data may under-represent imported products with very
high emission levels. Mills making products likely to exceed the CARB
levels may not have had their products tested because they intended to
sell them in the other 49 states.

The percentage of samples with current emissions below CARB 1 levels is
a reasonable estimate of the percentage of mills that will be achieving
CARB 2 levels in the baseline scenario:

The PSI presentation was given after CARB 1 was in effect but before
CARB 2 was. It is not known whether more overseas producers ultimately
will meet the Phase 2 emission limits than initially met the Phase 1
limits.

  

For the purposes of TSCA Title VI, laminated products are a subset of
hardwood plywood. The statute defines a laminated product as one made by
affixing a wood veneer to a particleboard, medium-density fiberboard, or
veneer-core platform. The statutory definition goes on to state that
laminated products are component parts used in the construction or
assembly of a finished good, and that a laminated product is produced by
the manufacturer or fabricator of the finished good in which the product
is incorporated. EPA is given the authority to modify the statutory
definition of laminated product through rulemaking. EPA also is directed
to use all available and relevant information to determine whether the
definition of hardwood plywood should exempt engineered veneer or any
laminated product.  

Note that baseline emissions data are not available for hardwood plywood
defined as laminated products. The CARB 1 standard for hardwood plywood
is used to represent the average baseline emission level for laminated
products made in the U.S. and Canada. The CARB 1 level represents the
average emission level that CARB found prior to promulgating the ATCM.
It reflects a mix of products using different resin types, including
both UF and NAF resins. This level is considered to be a reasonable
proxy for domestically produced laminated products. These products are
thought to generally be made with UF or PVA resin, so their emissions
may be similar to those from stock hardwood panel products before the
ATCM went into effect. 

For laminated products made outside the U.S. and Canada, average
baseline emissions were assumed to be the same as the levels for
imported hardwood plywood that does not meet an emission standard
(estimated from the PSI presentation). Emissions data prior to the CARB
ATCM are not available for non-Canadian imported hardwood plywood.
Because the PSI presentation was given after CARB 1 was in effect, it
presumably reflects a reduction in emissions compared to pre-ATCM
levels. Therefore, if baseline laminated product emissions are similar
to pre-ATCM emissions, using the average of all non-Canadian imported
hardwood plywood would underestimate emission levels for laminated
products. Using the estimated average emissions from hardwood plywood
that does not meet an emissions standard reflects the possibility that
many of these imported laminated products are made with UF resins.

Table C-4 lists the emission levels used for the baseline 2013 scenario
for CWPs, including laminated products; the values listed reflect the
assumptions and choices described above. The estimated share of U.S.
consumption also is listed in the table for each product type and
associated country of origin.

Table C-4. Baseline Emission Levels for Composite Wood Products by
Product Type, Country of Origin, and Emission Standard

Emissions Standard	Hardwood Plywood	MDF	Particleboard	Laminated Products

	U.S./Can.	Non-U.S./Can.	U.S./Can.	Non-U.S./Can.	U.S./Can.	Non-U.S./Can.
U.S./Can.	Non-U.S./Can.

Average Emissions (ppm)

None	0.058	0.163	---	0.502	0.058	0.384	0.058	0.163

CARB 1	---	---	---	---	0.090	---	---	---

CARB 2	0.032	0.015	0.082	0.047	0.057	0.036	---	---

CARB ULEF

	0.024

0.042



	CARB NAF	0.013

0.035

0.013



	Share of U.S. Consumption

None	0.1%	16.6%	---	0.8%	0.2%	1.1%	36.0%	64.0%

CARB 1	---	---	---	---	1.7%	---	---	---

CARB 2	19.7%	47.4%	79.2%	7.2%	85.0%	1.9%	---	---

CARB ULEF

	4.2%

3.0%



	CARB NAF	16.2%

8.6%

7.1%



	

C.2	REDUCED EMISSIONS CLASSES

Seven analytical options were evaluated in the EPA exposure assessment,
as listed and described briefly in Table C-5. The first four options –
CARB1, CARB2/Statutory Option, CARB2/Statutory Option including
laminated products at NAF, and NAF – assume that progressively higher
percentages of CWPs meet stricter emission standards. The last three
scenarios are similar to CARB1, CARB2/Statutory Option, and NAF, with
the single exception that laminated products are not regulated as HWPW
under TSCA. 

Table C-5. Alternative Analytical Options Evaluated in the Exposure
Assessment

Analytical Option	Description

CARB1, Including Laminated Productsa	All CWPs, including laminated
productsb, are at emission levels that meet CARB Phase 1 standards1.
Products that meet CARB Phase 1 standards at baseline are assumed to
remain at that emission level.

CARB2 / Statutory Option, Including Laminated Productsa,c	All CWPs,
including laminated products, are at emission levels that meet CARB
Phase 2 standards. Products that meet CARB Phase 2 standards at baseline
are assumed to remain at that emission level.

CARB2 / Statutory Option, with Laminated Products at NAF	As above,
except that laminated products are assumed to meet NAFd standards.

NAF, Including Laminated Productsa	All CWPs, including laminated
products, are at emission levels that meet NAF standards. Products that
meet NAF standards at baseline are assumed to remain at that emission
level.

CARB1, Excluding Laminated Products	Like CARB1 above, except that
laminated products are not regulated as HWPW and, thus, are assumed
remain at baseline emission levels.

CARB2/Statutory Option,

Excluding Laminated Products	Like CARB2 above, except that laminated
products are not regulated as HWPW and, thus, are assumed remain at
baseline emission levels.

NAF, Excluding Laminated Products	Like NAF above, except that laminated
products are not regulated and, thus, are assumed remain at baseline
emission levels.

a Included among the emission classes in FIAM-pwp

b 30% of all HWPW products are assumed to be laminated products.

b The CARB emission standards for this scenario are equivalent to those
specified in the FSCWPA.

d NAF = no added formaldehyde.

In the economic analysis presented to the Small Business Advocacy Review
(SBAR) Panel for the formaldehyde rule, EPA estimated that HWPW
production (excluding engineered wood flooring) was 1,749 thousand
square feet (on a 3/8-inch basis) in 2006, and that laminated product
production (excluding engineered wood flooring) was about 408 million
square feet on a 5/8-inch basis, equivalent to about 680 million square
feet on a 3/8-inch basis.  Thus, based on these estimates, laminated
products account for about 28% of combined HWPW production. The Hardwood
Plywood and Veneer Association has indicated that laminated products
represent about 30% of engineered wood flooring production. Therefore,
this exposure analysis assumes that laminated products represent 30% of
the volume of HWPW.

Table C-6 lists emission levels and market shares for the same set of
CWPs under the assumption that CARB1 emission limits would be in effect
for all CWPs, including laminated products. In compiling this table, it
was first assumed that emission levels of all products meeting CARB1 or
stricter (CARB2/ULEF/NAF) standards at baseline would not change. Next,
it was assumed that any products listed as not meeting any emission
standard at baseline, but whose emission levels would comply with CARB1,
also would remain at their baseline emission levels. For example, some
U.S.-produced HWPW and PB products are not certified as meeting an
emission standard at baseline but nonetheless have emission levels that
comply with CARB Phase 1 emission limits; thus, their emission levels
are assumed to remain at the baseline level under this
scenario/analytical option. As noted in a table footnote, for the CARB1
analytical option laminated products were assumed to be at the same
emission levels as HWPW if regulated and at baseline emission levels
(see Table C-4) if not regulated.

Table C-6. Emission Levels for CARB1 Analytical Option by CWP Type,
Origin, and Emissions Standard

Emissions Standard	Hardwood Plywooda	MDF	Particleboard

	U.S.	Non-U.S.	U.S.	Non-U.S.	U.S.	Non-U.S.

Average “Emissions” (ppm)

None	---	---	---	---	---	---

CARB1 (previously NOT CARB1)	0.058b	0.058	---	0.14c	0.058b	0.090

Already at CARB1	---	---	---	---	0.090	---

CARB2	0.032	0.015	0.082	0.047	0.057	0.036

ULEF	---	---	0.024	---	0.042	---

NAF	0.013	---	0.035	-	0.013	-

Share of U.S. Consumption

None	---	---	---	---	---	---

CARB1 (previously NOT CARB1)	0.1%	16.6%	---	0.8%	0.2%	1.1%

Already at CARB1	---	---	---	---	1.7%	---

CARB2	19.7%	47.4%	79.2%	7.2%	85.0%	1.9%

ULEF	---	---	4.2%	---	3.0%	---

NAF	16.2%	---	8.6%	---	7.1%	---

a For the CARB1 scenario that includes laminated products, those
products are assumed to have the same emission levels as hardwood
plywood; for the CARB1 scenario that excludes laminated products, those
products are assumed to have baseline emission levels as listed in Table
C-4.

b Products in these cells have baseline emission levels that comply with
CARB1, even though producers reported that they were not intentionally
meeting any particular standard. Because the baseline levels already
comply with CARB1, the emission levels for these products are assumed
not to change and, thus, are the same as listed in Table C-4.

c Products in this cell are not CARB1-compliant at baseline; thus, their
emission levels are assumed to change for the CARB1 scenario. The
assumed emission level is 67 % of the level that would exactly comply
with CARB1; as noted in the text, industry generally sets target
emission levels below the standard to ensure that it is being met, given
variability across/within production lots as well as measurement
uncertainty.

For the CARB1 scenario it was further assumed that products listed as
not meeting any emission standard at baseline, and whose emission levels
exceed the CARB Phase 1 limit, would attain emission levels equal to
those of corresponding U.S. products listed as not meeting any emission
standard but nonetheless complying with CARB1 at baseline. For example,
non-U.S. HWPW not complying with CARB1 at baseline, with an average
emission level of 0.163 ppm, was assumed to match the level (0.058 ppm)
of U.S.-produced HWPW complying with CARB1 at baseline, and its market
share (16.6%) was shifted from “None” to “CARB1” in the table. 

Lastly, it was assumed that products not meeting any emission standard
at baseline, and for which there are no comparable products complying
with CARB1 at baseline, would obtain an emission level equal to 67% of
the CARB1 limit – the CWP industry generally sets targets for emission
levels below the standard to ensure that it is being met, considering
the variability across/within production lots as well as measurement
uncertainty. For example, non-U.S. MDF, with an emission level at
baseline (0.502 ppm) that is not CARB1-compliant, was assumed to change
to a level equal to 67 percent of the CARB1 standard (0.67 x 0.21 ppm =
0.14 ppm).

Table C-7 lists emission levels and market shares for the analytical
option where CARB2/TSCA Title VI statutory emission limits would be in
effect. As with the CARB1 scenario, it was assumed that emission levels
of all products already meeting CARB2 or stricter standards (ULEF/NAF)
would not change. Next, it was assumed that products not meeting CARB2
at baseline (including those meeting the less strict CARB1 standard)
would obtain emission levels equal to U.S. products of the same CWP type
that comply with CARB2 at baseline. For example, both U.S.-produced HWPW
(baseline level of 0.058 ppm) and non-U.S. HWPW (baseline level of 0.163
ppm) were assumed to shift to a CARB2-compliant emission level of 0.032
ppm. For this analytical option, laminated products were assumed to be
at the same emission levels as HWPW if regulated and at baseline
emission levels (see Table C-4) if not regulated.

Table C-7. Emission Levels for CARB2 Analytical Option by CWP Type,
Origin, and Emissions Standard

Emissions Standard	Hardwood Plywooda	MDF	Particleboard

	U.S.	Non-U.S.	U.S.	Non-U.S.	U.S.	Non-U.S.

Average “Emissions” (ppm)

None	---	---	---	---	---	---

CARB1	---	---	---	---	---	---

CARB2 (previously NOT CARB2)	0.032	0.032	-	0.082	0.057	0.057

Already at CARB2	0.032	0.015	0.082	0.047	0.057	0.036

CARB2	---	---	0.024	---	0.042	---

ULEF	0.013	---	0.035	---	0.013	---

NAF	---	---	---	---	---	---

Share of U.S. Consumption

None	---	---	---	---	---	---

CARB1	---	---	---	---	---	---

CARB2 (previously NOT CARB2)	0.1%	16.6%	---	0.8%	1.9%	1.1%

Already at CARB2	19.7%	47.4%	79.2%	7.2%	85.0%	1.9%

CARB2	---	---	4.2%	---	3.0%	---

ULEF	16.2%	---	8.6%	---	7.1%	---

NAF	---	---	---	---	---	---

a For the CARB2 scenario that includes laminated products, those
products are assumed to have the same emission levels as hardwood
plywood; for the CARB2 scenario that excludes laminated products, those
products are assumed to have baseline emission levels as listed in C-4.

Table C-8 lists emission levels and estimated percentages under the
analytical option whereby NAF emission limits would be in effect. First,
it was assumed that emission levels of all products meeting the NAF
standard at baseline would not change. Next, it was assumed that
products not meeting NAF at baseline (including those meeting less
strict standards) would attain the emission levels for U.S. products of
the same CWP type that are NAF-certified at baseline. For example,
U.S.-produced HWPW accounts for 36 percent of all HWPW production. A
subset of the U.S. HWPW production, accounting for 19.8 percent, does
not comply with NAF at baseline but attains a NAF-compliant emission
level of 0.013 ppm under the NAF analytical option; the remaining 16.2
percent, already NAF-compliant, likewise has an emission level of 0.013
ppm. Lastly, it was assumed that MDF products meeting ULEF at baseline,
but with average emission levels lower than those for NAF-compliant MDF
products at baseline, would not change. As with the analytical options
described above (CARB1 and CARB2), for the NAF analytical option
laminated products were assumed to be at the same emission levels as
HWPW meeting the NAF standard if regulated and at baseline emission
levels (see Table C-4) if not regulated.

Table C-8. Emission Levels for NAF Scenario by CWP Type, Origin, and
Emissions Standard

Emissions Standard	Hardwood Plywooda	MDF	Particleboard

	U.S.	Non-U.S.	U.S.	Non-U.S.	U.S.	Non-U.S.

Average “Emissions” (ppm)

None	---	---	---	---	---	---

CARB P1	---	---	---	---	---	---

CARB P2	---	---	---	---	---	-

NAF (previously not NAF with non-NAF emissions)	0.013	0.013	0.035	0.035
0.013	0.013

NAF (previously not NAF with NAF emissions)	---	---	0.024	---	---	---

Already at NAF	0.013	---	0.035	---	0.013	---

Share of U.S. Consumption

None	---	---	---	---	---	---

CARB P1	---	---	---	---	---	---

CARB P2	---	---	---	---	---	---

NAF (previously not NAF with non-NAF emissions)	19.8%	64.0%	79.2%	8.0%
89.9%	3%

NAF (previously not NAF with NAF emissions)	---	---	4.2%	---	---	---

Already at NAF	16.2%	---	8.6%	---	7.1%	---

a For the NAF scenario that includes laminated products, those products
are assumed to have the same emission levels as hardwood plywood; for
the NAF scenario that excludes laminated products, those products are
assumed to have baseline emission levels as listed in Table C-4.

Table C-9 lists emission levels and market shares under the analytical
option whereby CARB2 emission limits would be in effect, but with the
added assumption that all laminated products (assumed to account for 30
percent of all HWPW production) are NAF-compliant (e.g., using NAF cores
laminated with NAF resins) at an emission level of 0.013 ppm. Note that
the estimated percentages for the U.S. and non-U.S. HWPW columns have
been adjusted in the table such that all HWPW production, including
laminates, sums to 100 percent.

Table C-9. Emission Levels for CARB2 Scenario with Laminates at NAF by
CWP Type, Origin, and Emissions Standard

Emissions Standard	Hardwood Plywood	MDF	Particleboard	Laminated
Productsa

	U.S.	Non-U.S.	U.S.	Non-U.S.	U.S.	Non-U.S.	U.S.	Non-U.S.

Average “Emissions” (ppm)

None	---	---	---	---	---	---	---	---

CARB1	---	---	---	---	---	---	---	---

CARB2 (previously NOT CARB2)	0.032	0.032	---	0.082	0.057	0.057	---	---

Already at CARB2	0.032	0.015	0.082	0.047	0.057	0.036	---	---

CARB2	---	---	0.024	---	0.042	---	---	---

ULEF	0.013	---	0.035	---	0.013	---	---	---

NAF	---	---	---	---	---	---	0.013	0.013

Share of U.S. Consumption

None	---	---	---	---	---	---	---	---

CARB1	---	---	---	---	---	---	---	---

CARB2 (previously NOT CARB2)	0.1%	11.6%	---	0.8%	1.9%	1.1%	---	---

Already at CARB2	13.8%	33.2%	79.2%	7.2%	85.0%	1.9%	---	---

CARB2	---	---	4.2%	---	3.0%	---	---	---

ULEF	11.3%	---	8.6%	---	7.1%	---	---	---

NAF	---	---	---	---	---	---	10.8%	19.2%

	a Assumed to account for 30% of all HWPW production.

C.3	SLOPES AND INTERCEPTS FOR EMISSION RATES

Appendix D lists emission rates for various types of PWPs that were
derived from sets of chamber emission tests conducted during the 1980s
and up through the mid-1990s. A distinguishing feature of these chamber
tests, although somewhat dated, is that they were conducted at multiple
air exchange rates and/or product loading ratios, enabling estimation of
model parameters – slope and intercept – that reflect the dependence
of the emission rate on the indoor formaldehyde concentration. The slope
is a mass transfer coefficient that reflects the “backpressure”
effect of the indoor formaldehyde concentration on the emission rate,
whereas the intercept reflects the hypothetical emission rate when the
indoor formaldehyde concentration is zero.

The following equation, identical to Eqn. 3-4 in Section 3.2 of the main
document, was used as a basis for determining the intercept (b)
associated with products meeting certain emission standards:

[CH2O]SS = [CH2O]out + (CH2OER( * Area)/Q	(Eqn. C-1)

where:

[CH2O]SS	=	steady-state formaldehyde concentration inside the
compartment (mg/m3)

[CH2O]out	=	steady-state formaldehyde concentration outside the
compartment (mg/m3)

CH2OER’	=	the emission rate of a solid formaldehyde source inside the
compartment (mg/h)

Area	=	the surface area of the formaldehyde source (m2)

Q	=	airflow rate into and out of the compartment (m3/hr)

The emission rate of a solid source is:

CH2OER( = -m * [CH2O]V + b		(Eqn. C-2)

where:

m	=	the mass transfer coefficient (m/hr)

[CH2O]V	=	the CH2O concentration in the vapor phase (mg/m3)

b	=	a constant; the emission rate at zero CH2O concentration in the air
(mg/m2-hr).

Assuming that [CH2O]out is zero and that [CH2O]V  = [CH2O]SS , and
substituting Eqn. C-2 into Eqn. C-1, we can solve for b as follows:

	b = [CH2O]SS * (1 + m * Area / Q) * (Q / Area)	(Eqn. C-3)

ASTM E 1333, Standard Test Method for Determining Formaldehyde
Concentrations in Air and Emission Rates from Wood Products Using a
Large Chamber, is a commonly used method for demonstrating compliance
with formaldehyde emission standards. The method prescribes an air
exchange rate for testing of 0.5 hr-1 (i.e., 0.5 ACH) and product
loading ratios of 0.43 m2/m3for PB, 0.95 m2/m3for HWPW, and 0.26
m2/m3for MDF. Using these rates along with an assumed slope (m) for a
given PWP type, the corresponding intercept can be determined for a
hypothetical source that just meets a given emissions standard. 

For example, the CARB2 emission standard for MDF is 0.11 ppm or 0.135
mg/m3. An arbitrary chamber volume of 100 m3, together with a value of
50 m3/hr for Q (corresponding to an air exchange rate of 0.5/hr), was
used in the calculation. The volume assumed does not matter, as the
exposed surface area scales to the volume via the loading ratio – with
a volume of 100 m3 the corresponding surface area is 26 m2. A slope of
1.06 m/hr was assumed; this value is the average slope from about 30
tests of different MDF specimens, as listed in Appendix D. 

Substituting the assumed values given above in Eqn. 4-2, the intercept
(b) that corresponds to the CARB2 emission standard for MDF is
calculated as follows:

b = 0.135 mg/m3 * (1 + 1.06 m/hr * 26m2 / 50 m3) * (50 m3 / 26 m2)

= 0.40 mg/m2-hr

Table C-10 lists the calculated intercepts associated with various
emission levels for HWPW, MDF and PB, as listed previously in Tables C-6
through C-9, using the calculation method described above and an assumed
slope for each CWP type as indicated in the table.

Table C-10. Calculated Intercepts for Various Emission Levels for Each
CWP Type 

Emission Standard –  Origin	Emission Levela	Assumed Slope	Calculated
Intercept

	ppm	mg/m3	m/hr	mg/m2/hr

Hardwood Plywood

None – U.S.	0.058	0.072	0.27	0.057

None – non-U.S.	0.163	0.201	0.27	0.160

CARB1 – U.S. & non-U.S.	0.058	0.072	0.27	0.057

CARB2 – U.S.	0.032	0.040	0.27	0.031

CARB2 – non-U.S.	0.015	0.019	0.27	0.019

NAF – U.S.	0.013	0.016	0.27	0.013

Medium Density Fiberboard

None – U.S.	---	---	1.06	---

None – non-U.S.	0.502	0.620	1.06	1.849

CARB1 – U.S.	0.140	0.173	1.06	0.516

CARB2 – U.S.	0.082	0.101	1.06	0.302

CARB2 – non-U.S.	0.047	0.058	1.06	0.173

ULEF – U.S.	0.024	0.030	1.06	0.088

NAF – U.S.	0.035	0.043	1.06	0.129

Particleboard

None – U.S.	0.058	0.072	0.7	0.133

None – non-U.S.	0.384	0.474	0.7	0.883

CARB1 – U.S. 	0.058	0.072	0.7	0.133

CARB1 – non-U.S.	0.090	0.111	0.7	0.207

CARB2 – U.S.	0.057	0.070	0.7	0.131

CARB2 – non-U.S.	0.036	0.044	0.7	0.083

NAF – U.S. & non-U.S.	0.013	0.016	0.7	0.030

a mg/m3 formaldehyde = ppm formaldehyde * 1.23

Providing standard-specific slopes and intercepts for each CWP type
would prove cumbersome, exceeding the number of input rows available for
sources in the FIAM model. It was determined experimentally that using a
composite intercept together with the corresponding total exposed
surface area for a given CWP type yielded identical modeling results to
that from the more tedious approach of using values for the individual
components meeting different emission standards. The composite intercept
was calculated as a weighted average of the individual intercepts, using
the market share as the weight; the resultant composite intercepts are
shown for the baseline scenario in Table C-11 for each CWP type, along
with standard-specific intercepts. 

Table C-11. Composite Intercepts by CWP Type – Baseline Scenario

Emission Standard – Origin	Slope	Intercept	Market Share

Hardwood Plywood

None – U.S.	0.27	0.057	0.1%

None – non-U.S.	0.27	0.160	16.6%

CARB2 – U.S.	0.27	0.031	19.7%

CARB2 –  non-U.S.	0.27	0.015	47.4%

NAF – U.S.	0.27	0.013	16.2%

Composite	0.27	0.042	100 %

Medium Density Fiberboard

None – non-U.S.	1.06	1.849	0.8%

CARB2 – U.S.	1.06	0.302	79.2%

CARB2 non-US	1.06	0.173	7.2%

ULEF – U.S.	1.06	0.088	4.2%

NAF – U.S.	1.06	0.129	8.6%

Composite	1.06	0.281	100 %

Particleboard

None – U.S.	0.70	0.133	0.2%

None – non-U.S.	0.70	0.883	1.1%

CARB1 – U.S.	0.70	0.207	1.7%

CARB2 – U.S.	0.70	0.131	85.0%

CARB2 –  non-U.S.	0.70	0.083	1.9%

ULEF – U.S.	0.70	0.097	3.0%

NAF – U.S.	0.70	0.030	7.1%

Composite	0.70	0.131	100 %



Composite intercepts for baseline and other emission scenarios –
CARB1, CARB2 and NAF – are shown by CWP type in Table C-12, without
the underlying details. The most notable difference across scenarios is
that the NAF intercepts are substantially lower than for the other three
scenarios, for each CWP type.

Table C-12. Composite Intercepts by CWP Type and Emissions Scenario

CWP Type	Composite Intercept

	Baseline	CARB1	CARB2	NAF

Hardwood Plywood	0.042	0.025	0.021	0.013

Medium Density Fiberboard	0.281	0.271	0.269	0.127

Particleboard	0.131	0.124	0.122	0.030



The modeling also accounted for some PWP types that would not be subject
to emission standards because they are specifically exempt from TSCA
Title VI. Underlayment in homes typically is either tongue-and-groove
softwood plywood (SWPW) or oriented strand board (OSB). The emission
rates from earlier chamber tests on SWPW are thought to be applicable to
either type; for example, a study by Hodgson et al. reported virtually
the same intercept (0.029) for OSB as did the earlier tests for SWPW
(0.030). 

Interior doors were represented using estimated emission rates from
chamber testing of 6-panel white doors that were used in the EPA pilot
study; this is the only known study in which such doors have been tested
at multiple air exchange/loading rates, enabling calculation of both a
slope and an intercept. The emission rates for these doors also were
used to represent those for coated (but not laminated) CWP materials,
assuming a similar barrier effect.  

Because the interiors of such coated materials presumably would be in
compliance with any rules governing CWP emissions, the intercept for
interior doors and coated CWPs was varied across emission scenarios. It
was observed that, at baseline, the intercept for these products was
close to halfway between the intercepts for HWPW and PB. Thus, the
average of those two intercepts for other emission scenarios – CARB1,
CARB2 and NAF – was used for doors and coated CWPs; the average
intercepts were 0.074 for CARB1, 0.071 for CARB2, and 0.0215 for NAF. 
Emission factors for underlayment, interior doors, and coated CWPs are
listed in Table C-13.

Table C-13. Emission Rates for Underlayment and Coated CWPs – Baseline
Scenario

PWP Type/Class	Proxy Used for Emission Rate 	Slope	Intercept

Underlayment (OSB)	Softwood Plywood

(PF Resin)	0.61	0.030

Interior Doors & Coated CWPsa	6-panel Doors Used

 in EPA Pilot Study 	0.52	0.082

a CWPs could be coated, for example, with primer/paint, vinyl film, or
thermally fused paper.



Appendix D

Historical Testing of PWPs to Derive Slope and Intercept Values

D.1	BACKGROUND

During the 1980s and up through the mid-1990s, a number of organizations
conducted sets of chamber emission tests at multiple air exchange rates
and/or product loading ratios, enabling the estimation of model
parameters – slope and intercept – that reflect the dependence of
the emission rate on the indoor formaldehyde concentration. The
following source types were tested:

Particleboard underlayment

Mobile home decking

Industrial particleboard

Medium density fiberboard

Hardwood plywood paneling (with a print face)

Hardwood plywood paneling (with a paper face)

Hardwood plywood paneling (with a wood veneer face)

Hardwood plywood paneling (unspecified face material)

Softwood plywood (phenol-formaldehyde resin)

Hardboard (phenol-formaldehyde resin)

Kitchen cabinets

Interior doors. 

D.2	DETAILED CHAMBER TESTING RESULTS

Table D-1 lists, by product type, products for which multiple chamber
tests were conducted in the 1980s-1990s to enable estimation of emission
rate model parameters (slope and intercept). The table also provides
references for the data and additional information, as available, on
chamber testing conditions (air exchange rate and loading ratio). The R2
values listed in the right-hand column of the table, which indicate the
degree of correspondence between calculated and predicted emission
rates, are based on regression analysis of calculated emission rates
against measured chamber concentrations for the various products tested,
from which the slope and intercept were estimated (see Section D.4 for
an example of this type of analysis). Not all investigators reported
these values; values are listed in cases where they were reported.

The abbreviated names assigned to individual available sources in the
table reflect the organization (e.g., EPA, ORNL) or individual that
performed the testing from which modeling parameters were estimated.
Several Available Sources – EPA-PBU for Particleboard Underlayment,
EPA-HWP for Hardwood Plywood Paneling (Unspecified), EPA-CABINET for
Kitchen Cabinets, and EPA-DOOR for Interior Doors – are based on
chamber testing performed as part of the EPA pilot study on formaldehyde
(see Appendix A).

	Table D-1.  Calculated Slopes and Intercepts for Matthews Model from
Chamber Tests Conducted at 23 (C, 50 % RH or at 25 (C, 50 % RHa

	(Listed Cases are for Conditions of 23 (C and 50 % RH)

Product Type	

Product Codeb	

No. of Data Pointsc	

N/L Ranged	

Measured Concentration Range (ppm)	

Matthew’s Emission Modele

ER = -m (conc) + b





	

Slope (m/hr)	

Intcpt. (mg/m2/hr)	

R2 Valuef



Particleboard Underlayment	

ORNL-PCB #1

ORNL-PCB #2

ORNL-PCB #3

ORNL-PBU 1 #4

ORNL-PBU 3 #3

ORNL-PBU 3 #6

GP-2

GP-4

GP-5

NBS-USM2-2B

NBS-USM2-3B

NBS-USM5-1A

NBS-U-5,8,9,12,18

EPA-PBU	

3

4

4

4

4

3	

0.13 to 0.63

0.32 to 5.70

0.10 to 0.95

0.16 to 9.18

0.44 to 6.67

0.60 to 2.37	

0.077 to 0.162

0.057 to 0.354

0.084 to 0.158

0.084 to 0.215

0.058 to 0.442

0.062 to 0.118	

0.32

0.70

0.88

1.72

0.70

0.93

0.84

1.57

0.76

0.36

0.38

0.62

0.60

1.25	

0.09

0.44

0.19

0.46

0.51

0.47

0.40

0.37

0.58

0.18

0.27

0.33

0.18

0.28	

0.92

0.98

0.92

0.77

0.96

0.52

0.81

0.98

0.42

0.76



Mobile Home Decking	

LEH-B

GP-MHD	

	

	

	

1.35

0.77	

0.66

0.39	





Industrial Particleboard	

ORNL-PBI 3#2

LEH-C

LEH-D

LEH-H

LEH-I

LEH-J

LEH-K

LEH-L

LEH-M

LEH-N

LEH-O

LEH-P

LEH-Q	

4

8

9

6

6

6

6

6

6

6

6

6

6	

0.32 to 3.31

0.76 to 6.10

0.31 to 6.10

0.59 to 3.82

0.59 to 3.82

0.59 to 3.82

0.59 to 3.82

0.59 to 3.82

0.59 to 3.82

0.59 to 3.82

0.59 to 3.82

0.59 to 3.82

0.59 to 3.82	

0.096 to 0.457	

0.47

1.00

0.54

1.02

0.87

1.30

0.96

1.07

0.40

1.32

0.58

1.58

0.98	

0.42

1.24

0.36

0.45

0.60

0.27

0.32

0.53

0.14

0.24

0.22

0.37

0.39	

0.88



Industrial Particleboard (continued)	

GP-1A

GP-3A

GP-6A

GP-7A

GP-8A

GP-9A

GP-11A

NBS-USM7-1B

NBS-USM2-1A

NBS-USM6-1B

NBS-USM6-2A

NBS-USM6-3B

NBS-USM4-1B

GP-1N

GP-2N

GP-3N

GP-4N

GP-5N

GP-6N

GP-7N	

	

	

	

0.41

0.63

0.46

0.38

0.63

0.62

0.55

0.45

0.39

0.80

0.33

0.64

0.28

0.70

0.72

0.81

0.55

0.50

0.46

0.58	

0.12

0.33

0.34

0.24

0.33

0.35

0.29

0.12

0.30

0.18

0.18

0.29

0.20

0.53

0.53

0.35

0.27

0.37

0.34

0.48	

0.98

0.67

0.98

0.32

0.94

0.98



Medium Density Fiberboard	

ORNL-MDF 1 #5

ORNL-MDF 3 #5

ORNL-MDF 1 #4

ORNL-MDF 2 #4

ORNL-MDF-U

LEH-E

LEH-F

LEH-G

GP-1

GP-2

GP-3

GP-4

GP-5

GP-6	

4

4

9

9

15

	

1.47 to 12.5

1.33 to 12.5

0.31 to 6.10

0.31 to 6.10

0.15 to 7.09	

0.170 to 0.952

0.160 to 0.936	

0.89

1.00

1.25

1.58

0.91

0.48

0.80

0.80

4.98

1.17

2.00

2.22

3.87

0.78	

2.78

2.49

1.62

1.09

4.24

0.32

0.79

0.83

1.27

0.77

0.84

1.33

0.80

0.50	

0.98

0.81



Medium Density Fiberboard (continued)	

GP-7

GP-8

GP-9

GP-10

GP-11

GP-12

GP-13

GP-14

GP-15

NBS-LMDF-1C

NBS-LMDF-2B

NBS-MDF 1

NBS-MDF 2

NBS-MDF 3

NBS-MDF 4

NBS-MDF 5	

	

	

	

0.73

0.43

0.69

1.83

0.81

1.45

0.90

1.30

0.77

0.88

0.71

1.50

1.16

1.09

0.76

0.65

	

0.58

0.46

0.32

0.92

0.58

1.31

0.63

0.75

0.39

0.64

0.73

1.73

1.14

1.31

1.29

1.43	





Hardwood Plywood Paneling (Print)	

ORNL-PNPR 2 #3

ORNL-PNPR 3 #3

ORNL-PNPR 2 #5

ORNL-PNPR 3 #1

ORNL-PAN #2	

4

4

4	

0.31 to 6.57

0.44 to 3.34

0.32 to 1.90	

0.085 to 0.719

0.029 to 0.127

0.021 to 0.055	

0.46

0.37

0.40

0.85

0.48

	

0.67

0.13

0.25

0.26

0.05	

0.86

0.94

0.77



Hardwood Plywood Paneling (Paper)	

ORNL-PNP 2 #4

ORNL-PNP 3 #1,2g

ORNL-PNP 3 #1,2h	

4

4

3	

0.31 to 6.57

0.20 to 1.11

0.17 to 1.11	

0.037 to 0.426

0.063 to 0.225

0.129 to 0.402	

0.27

0.10

0.27

	

0.27

0.09

0.22	

0.61

0.69

0.98



Hardwood Plywood Paneling (Veneer)	

ORNL-PND 1 #1,2g

ORNL-PND 1 #1,2h

ORNL-PND 3 #5

ORNL-PAN #1	

4

4

3

4	

0.15 to 0.35

0.12 to 1.06

0.32 to 1.11

0.14 to 0.95	

0.105 to 0.194

0.055 to 0.248

0.062 to 0.137

0.072 to 0.162	

0.19

0.11

0.34

0.45	

0.08

0.08

0.11

0.13	

0.36

0.81

0.90

0.85



Hardwood Plywood Paneling (Unknown)	

NBS-P6-1

NBS-P8-1

NBS-P10-1

NBS-P14-1

NBS-P17-1

NBS-P21-1

EPA-HWP

	

8

5

11

8

5	

0.04 to 1.91

0.03 to 1.14

0.04 to 4.86

0.18 to 5.20

0.08 to 7.36

0.26 to 1.04	

0.022 to 0.253

0.027 to 0.239

0.045 to 0.370

0.014 to 0.098

0.020 to 0.224

0.066 to 0.106	

0.12

0.12

0.45

0.67

0.52

0.22

0.96

	

0.05

0.04

0.30

0.08

0.15

0.03

0.16

	

0.83

0.96

0.71

0.66

0.72

0.18

0.98



Softwood Plywood 

(PF Resin)	

ORNL-PFPLY #1	

5	

0.23 to 1.05	

0.017 to 0.029	

0.61

	

0.03	

--



Hardboard

(e.g., "Masonite")	

ORNL-HBD 1	

7	

--	

0.02 to 0.095	

1.39

	

0.17	

0.98



Kitchen Cabinets	

EPA-CABIN	

3	

0.39 to 2.29	

0.023 to 0.075	

0.49	

0.08	

0.94



Interior Doors	

EPA-DOOR	

3	

0.58 to 2.44	

0.022 to 0.061	

0.52	

0.08	

0.86



a	The data listed in this table are based on the results of emission
rate tests performed at Oak Ridge National Laboratory (ORNL) for the
Consumer Product Safety Commission (CPSC), at the National Bureau of
Standards (NBS) for EPA and CPSC, at Georgia Pacific Corp. (GP), and at
Weyerhauser Co. by Dr. W. F. Lehmann. The data listed in the table for
individual sources were obtained from or are based on the results
reported in Progress Reports No. I, II, XIV, XV, XVI, and XVIII
submitted to CPSC by ORNL; NBS Report #NBSIR 85-3255 to CPSC; a 1987 NBS
report to EPA (Grot et al. 1987); comments submitted by Georgia Pacific
Corp. to EPA (Howlett 1988); Lehmann (1987) (also personal communication
between B. Lehmann  of Weyerhauser Co. and G. Schweer of EPA on June 30,
1987); and Koontz et al. 1996.

The ORNL reports listed the Matthew’s model parameters for the
individual sources and presented summary test data for the tested N/L
conditions. The Matthews model parameters listed in the table are as
listed in the ORNL reports. Lehmann (1987) and Howlett (1988) reported
HBF model parameters for the tested boards; Matthews model parameters
for these products were calculated from the HBF model parameters. The
NBS report listed Matthews model parameters.

b	Product codes are reported as listed in the ORNL Progress Reports
(except for interior plywood, hardboard, and MDF-U) for the ORNL
products. The products reported in Lehmann (1987) are denoted by "LEH"
followed by the letter code used in his report. The products reported in
the NBS reports are denoted by "NBS" followed by the code used in the
reports. The products tested by Georgia Pacific Corp. are denoted by
"GP" followed by a digit representing the order presented in the
report’s tables. The products tested for the EPA Pilot Study (Koontz
et al. 1996) are denoted by "EPA."

c	Refers to the number of distinct N/L experiments conducted.

d	N/L is the ratio of the experimental air exchange rate (N), in air
changes per hour, to the product loading in the chamber (L), in m2 of
product surface area per m3 of chamber volume.

e	ER is the emission rate of the product in units of mg of formaldehyde
per m2 of product surface area per hour; the intercept has the same
units. The slope units are m/hr.

f	Indicator of the degree of correspondence between calculated emission
rates and rates predicted by the regression equation.

g	Only decorative side exposed.

h	Both sides exposed.

REFERENCES CITED IN TABLE D-1:

ANSI 1999. American National Standard: Particleboard. ANSI 208.1-1999,
approved February 8, 1999.

ANSI 2002. American National Standard: Particleboard. ANSI 208.2-2002,
approved May 13, 2002.

R.A Grot, S. Silberstein, and K. Ishiguro. 1985. Validation of Models
for Predicting Formaldehyde Concentrations in Residences Due to Pressed
Wood Products. U.S. Dept. of Commerce, National Bureau of Standards.
Report #NBSIR 85-3255.

R.A. Grot, S. Nabinger, and S. Silberstein. 1987. Formaldehyde from
Low-emitting Pressed Wood Products and the Effectiveness of Various
Remedial Measures for Reducing Formaldehyde Emissions. U.S. Dept. of
Commerce, National Bureau of Standards.

C.T. Howlett. 1988. Data provided to EPA on January 14, 1988, and on
June 6, 1988 by C.T. Howlett (Georgia Pacific Corp.).

M.D. Koontz, H.E. Rector, D.R. Cade, C.R. Wilkes, and L.L. Niang. 1996.
Residential Indoor Air Formaldehyde Testing Program: Pilot Study. Report
No. IE-2814, prepared by GEOMET Technologies, Inc. for the U.S.
Environmental Protection Agency, Washington, DC, under EPA Contract No.
68-D3-0013.

W.F. Lehmann. 1987. Effect of Ventilation and Loading Rates in Large
Chamber Testing of Formaldehyde Emissions from Composite Panels. Forest
Products Journal 37(4):31-37.

D.3	AVERAGE VALUES FOR SLOPES AND INTERCEPTS

Average slope and intercept values were determined from the data
displayed in Table D-1. All data shown in the table were used for each
of the product types summarized in Table D-2, except three cases for MDF
that were deemed outliers because they had either a substantially higher
slope (Product Codes GP-1 and GP-5) or a substantially higher intercept
(Product Code ORNL-MDF-U) than other cases within that product group.
Although, as noted previously, the data on which these averages are
based are somewhat dated, the slopes in the table are useful for
estimating hypothetical intercept values for products assumed to meet
certain emission standards. Illustrative calculations in this regard are
provided in Appendix C (see Section C.1) and in Section 2.3.3 of the
main document.

Table D-2.  Calculated Average Slopes and Intercepts for Eight Product
Types

Product Type	Average Slope	Average Intercept

Particleboard Underlayment	0.83	0.34

Mobile Home Decking	1.06	0.53

Industrial Particleboard	0.70	0.35

Medium Density Fiberboarda	1.06	1.02

Hardwood Plywood Paneling (Print	0.51	0.27

Hardwood Plywood Paneling (Paper)	0.21	0.19

Hardwood Plywood Paneling (Veneer)	0.27	0.10

Hardwood Plywood Paneling (Unknown)	0.44	0.12

a Two outliers excluded from the calculation; see text.

D.4	ILLUSTRATIVE REGRESSION ANALYSIS ON CHAMBER DATA SET

The data set for this example originated from the EPA pilot study
described in Appendix A. Among the product types subjected to chamber
testing was a kitchen cabinet ensemble. Test conditions and results for
kitchen cabinets prior to the first house loading are shown in Table
D-3. The chamber tests were conducted at three air exchange rates –
0.17 ACH, 0.51 CH, and 1.0 ACH – to provide a basis for estimating the
relationship between formaldehyde concentration and emission rate. The
chamber tests were conducted at a target temperature of 77 °F (25 °C)
and a target relative humidity (RH) of 50 percent. The measured
“steady-state” chamber concentrations shown in the table are based
on adjustment to standard conditions of 77 °F and 50 % RH, per ASTM
Standard E-1333. The chamber volume was 1080 ft3 (30.6 m3) and the
product loading ratio was 0.13 ft2/ft3 (0.43 m2/m3).

Table D-3.  Chamber Conditions and Test Results for Kitchen Cabinets –
EPA Pilot Study

Air Exchange Rate, ACH	Temperature, °F	RH, 

%	Chamber Concentration, mg/m3	Calculated Emission Rate, mg/m2-hr

1.00	75.0	50	0.0353	0.0821

0.51	77.5	48	0.0645	0.0765

0.17	77.5	51	0.1143	0.0452



As shown in the table, the increase in the chamber concentration was
less than proportional to the decrease in air exchange rate. For
example, if there were no dependence of the emission rate on the air
concentration (i.e., if there were no “backpressure” effect), then
one would expect a six-fold decrease in the air exchange rate (from 1.0
to 0.17 ACH) to result in a six-fold increase in the concentration (from
0.035 to 0.212 mg/m3); however, the measured chamber concentration at
0.17 ACH (0.114 mg/m3) was about half the expected value, indicative of
a backpressure effect.

The emission rates shown in the table were calculated under a
steady-state assumption as follows:

	ER = (Css * AER * V) / Area	(D-1)

		Where:

		ER	= 	Emission rate, mg/m2-hr

		Css	= 	Steady-state concentration, mg/m3

		AER	= 	Air exchange rate, 1/hr

		V 	= 	Chamber volume, m3

		Area	= 	Exposed surface area, m2 = loading ratio (m2/m3) * chamber
volume (m3)

The volume term appears in both the numerator and denominator; thus,
Eqn. D-1 can be simplified to: 

	ER = (Css * AER) / L	(D-2)

		Where:

		L 	= 	Loading ratio, m2/m3

The calculated emission rate is then regressed against the chamber
concentration, as follows: 

	ER = -m * Css + b	(D-3)

		Where:

		m	= 	Regression slope (reflecting the backpressure effect)

		b	= 	Regression intercept

The resultant regression estimates from this procedure are 0.49 for the
slope and 0.10 for the intercept. The regression intercept obtained here
is slightly different from that shown in Table D-1 (0.08), but there is
uncertainty as to the specific data points that were used by the
originating researchers. A comparison of the predicted and calculated
emission rates (Table D-4) indicates a relatively high degree of
correspondence, as reflected in an R2 value of 0.94 for the regression.

Table D-4.  Predicted vs. Calculated Emission Rates for Kitchen Cabinets
– EPA Pilot Study

Air Exchange Rate, ACH	Temperature, °F	RH, 

%	Predicted Emission Rate, mg/m2-hr	Calculated Emission Rate, mg/m2-hr

1.00	75.0	50	0.0861	0.0821

0.51	77.5	48	0.0719	0.0765

0.17	77.5	51	0.0475	0.0452



Appendix E

 

Mass balance equation for zone 1 (well-mixed assumption):

 	(1)

where:

  (Matthew’s implementation) 	(2)

         m3/hr 	(3)

          mg/hr	(4)

m = mass transfer coefficient, m/hr

b = emission rate when the air concentration is zero, mg/(m2 hr)

Q = flow rates, m3/hr

C = concentrations, mg/m3

Area = emission area, m2

t = time, hr

V = volume, m3

Mass balance equation for zone 2 (well-mixed assumption):

 	(5)

where:

 	(6)

   	(7)

 	(8)

Assume flows are constant and set dC/dt = 0 (steady state condition):

  	(9)

And:

 	(10)

Solving for C1:

 	(11)

Solving for C2:

 	(12)

Substitute equation 12 into equation 11:

 	(13)

Rearranging and combining terms:

 

 

 

 

 	(14)

Similarly:

 	(15)

 

 In this report, the term “composite wood product” (CWP) refers only
to hardwood plywood, medium density fiberboard, or particleboard whereas
“pressed wood product” (PWP) refers to a broader set of wood
materials including, for example, softwood plywood and oriented strand
board.

 T. Matthews, T. Reed, B. Tromberg, C. Daffron, and A. Hawthorne. 1983.
Formaldehyde Emissions from Combustion Sources and Solid Formaldehyde
Resin Containing Products: Potential Impact on Indoor Formaldehyde
Concentration and Possible Corrective Measures. Proceedings of ASHRAE
Symposium Management of Atmospheres in Tightly Enclosed Spaces, Santa
Barbara, CA. 

 T. Matthews, A. Hawthorne, and C Thompson. 1987. Formaldehyde Sorption
and Desorption Characteristics of Gypsum Wallboard. Environmental
Science & Technology 21: 629-634.

 M. Koontz, H. Rector, D. Cade, C. Wilkes, and L. Niang. 1996.
Residential Indoor Air Formaldehyde Testing Program: Pilot Study. Report
No. IE-2814, prepared by GEOMET Technologies, Inc. for the USEPA Office
of Pollution Prevention and Toxics under EPA Contract No. 68-D3-0013,
Washington, DC.

 CDC. 2008/2010. Final Report on Formaldehyde Levels in FEMA-Supplied
Travel Trailers, Park Models, and Mobile Homes. Centers for Disease
Control and Prevention. Available at:   HYPERLINK
"http://www.cdc.gov/nceh/ehhe/trailerstudy/assessment.htm#final" 
http://www.cdc.gov/nceh/ehhe/trailerstudy/assessment.htm#final . July
2008 (amended December 2010).

 DHS. 2009. FEMA Response to Formaldehyde in Trailers. OIG-09-83.
Department of Homeland Security, Office of Inspector General,
Washington, DC. Available at: 

  HYPERLINK
"http://www.dhs.gov/xoig/assets/mgmtrpts/OIGr_09-83_Jun09.pdf" 
http://www.dhs.gov/xoig/assets/mgmtrpts/OIGr_09-83_Jun09.pdf .

 See   HYPERLINK
"http://www.msnbc.msn.com/id/23168160/ns/us_news-life/t/cdc-tests-confir
m-fema-trailers-are-toxic/" 
http://www.msnbc.msn.com/id/23168160/ns/us_news-life/t/cdc-tests-confirm
-fema-trailers-are-toxic/ .

 See   HYPERLINK "http://www.arb.ca.gov/toxics/compwood/compwood.htm" 
http://www.arb.ca.gov/toxics/compwood/compwood.htm   for recent activity
related to this ATCM.

 Available at  HYPERLINK
"http://www.arb.ca.gov/toxics/compwood/compwood.htm"
http://www.arb.ca.gov/toxics/compwood/compwood.htm .

 USEPA. 2012. Formaldehyde from Composite Wood Products: Exposure
Assessment. Draft Final Report. U.S. Environmental Protection Agency,
Office of Pollution Prevention and Toxics, Exposure Assessment Branch,

1200 Pennsylvania Avenue NW, Washington, DC 20460.  July 2012.

 Formaldehyde Indoor Air Model – Pressed Wood Products: Model
Documentation. Prepared  by Versar, Inc. for   EPA/OPPT by under
Contract No. EP-W-04,035, Work Assignment No. 4-2. August 2009.

 Peer Review Results for the Formaldehyde Indoor Air Model – for
Pressed Wood Products. Prepared by Eastern Research Group, Inc for
EPA/OPPT under Contract No. EP-W-05-014. October 2009.

 See   HYPERLINK "http://www.epa.gov/oppt/exposure/pubs/gems.htm" 
http://www.epa.gov/oppt/exposure/pubs/gems.htm .

 Centers for Disease Control and Prevention. 2008/2010. Final Report on
Formaldehyde Levels in FEMA-Supplied Travel Trailers, Park Models, and
Mobile Homes. July 2008 (amended December 2010).

Available at:   HYPERLINK
"http://www.cdc.gov/nceh/ehhe/trailerstudy/assessment.htm#final" 
http://www.cdc.gov/nceh/ehhe/trailerstudy/assessment.htm#final . 

 Department of Homeland Security. 2009. FEMA Response to Formaldehyde in
Trailers. OIG-09-83. Office of Inspector General, Washington, DC.
Available at: 

  HYPERLINK
"http://www.dhs.gov/xoig/assets/mgmtrpts/OIGr_09-83_Jun09.pdf" 
http://www.dhs.gov/xoig/assets/mgmtrpts/OIGr_09-83_Jun09.pdf .

 See, for example,   HYPERLINK
"http://www.eia.doe.gov/emeu/recs/contents.html" 
http://www.eia.doe.gov/emeu/recs/contents.html .

 Available at   HYPERLINK
"http://rredc.nrel.gov/solar/old_data/nsrdb/1991-2005/tmy3/" 
http://rredc.nrel.gov/solar/old_data/nsrdb/1991-2005/tmy3/ .

  According to the Residential Energy Consumption Survey, 84 percent of
U.S. housing units had air-conditioning       equipment as of 2005 (see 
 HYPERLINK
"http://205.254.135.24/emeu/recs/recs2005/hc2005_tables/detailed_tables2
005.html" 
http://205.254.135.24/emeu/recs/recs2005/hc2005_tables/detailed_tables20
05.html ).

  M.D. Koontz and H.E. Rector. 1995. Estimation of Distributions for
Residential Air Exchange Rates. Report prepared for EPA/OPPT under
Contract No. 68-D9-0166, Work Assignment No. 3-19. 

 Burmaster D.M. Murray and D.E. Burmaster. 1995. Residential Air
Exchange Rates in the United States: Empirical and Estimated Parametric
Distribution by Season and Climatic Region. Risk Analysis 15: 459-465.

 F. J. Offermann, J. Robertson, D. Springer, S. Brennan, and T. Woo.
2008. Window Usage, Ventilation, and Formaldehyde Concentrations in New
California Homes: Summer Field Sessions. ASHRAE IAQ 2007 Conference,
Baltimore, MD.

 R. Maddalena, M. Russell, D.P. Sullivan, and M.G. Apte. 2008. Aldehyde
and Other Volatile Organic Chemical Emissions in Four FEMA Temporary
Housing Units – Final Report. LBNL-254E, Ernest Orlando Lawrence
Berkeley National Laboratory, Environmental Energy Technologies
Division, Berkeley, CA. Available at:

  HYPERLINK
"http://www.cdc.gov/nceh/ehhe/trailerstudy/pdfs/LBNL-254E.pdf" 
http://www.cdc.gov/nceh/ehhe/trailerstudy/pdfs/LBNL-254E.pdf .

 A. Persily, A. Musser, and S.J. Emmerich. 2010. Modeled Infiltration
Rate Distributions for U.S. Housing. Indoor Air 20: 473-485.

 A. Berge, B. Mellagaard, P. Hanetho, and E. Ormstad. 1980. Formaldehyde
Release from Particleboard – Evaluation of a Mathematical Model. Holz
Als Roh-und Werkstoff 38: 252-255. 

 G. Myers. 1985. The Effects of Temperature and Humidity on Formaldehyde
Emission from UF-bonded Boards: A Literature Critique. Forest Products
Journal 1985: 35:20-31. 

 See, for example, M.A. Jayjock. 1994. Back Pressure Modeling of Indoor
Air Concentration from Volatilizing Sources, Am. Ind. Hyg. Assoc. J. 55
(3): 230-235.

 ASTM. 2010. Standard Test Method for Determining Formaldehyde
Concentrations in Air and Emission Rates from Wood Products Using a
Large Chamber. Designation E1333-10, American Society for Testing and
Materials.

 See, for example,   HYPERLINK
"http://www.epa.gov/oswer/riskassessment/sghandbook/chemicals.htm" 
http://www.epa.gov/oswer/riskassessment/sghandbook/chemicals.htm .

    HYPERLINK
"http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=209866" 
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=209866 

 USEPA. 2011. Draft Final. Assessment of Occupational Exposure to
Formaldehyde from Composite Wood Products. U.S. Environmental Protection
Agency, Office of Pollution Prevention and Toxics, Chemical Engineering
Branch, 1200 Pennsylvania Avenue NW, Washington, DC 20460. 6 October
2011.

 The most recent DOE/EIA Commercial Building Energy Consumption Survey
(CBECS) indicated that about 5 % of office buildings and 8 % of
education buildings were of recent construction (i.e., 3 years old or
less); see   HYPERLINK
"http://www.eia.gov/emeu/cbecs/cbecs2003/detailed_tables_2003/detailed_t
ables_2003.html" 
http://www.eia.gov/emeu/cbecs/cbecs2003/detailed_tables_2003/detailed_ta
bles_2003.html , Table B-8.

 CARB. 2004. Environmental Health Conditions in California’s Portable
Classrooms. Report to the California Legislature by the California Air
Resources Board (CARB) and the California Department of Health Services,
November 2004. Available at:   HYPERLINK
"http://www.arb.ca.gov/research/apr/reports/l3006.pdf" 
http://www.arb.ca.gov/research/apr/reports/l3006.pdf .

 F. J. Offermann, J. Robertson, D. Springer, S. Brennan, and T. Woo.
2008. Window Usage, Ventilation, and Formaldehyde Concentrations in New
California Homes: Summer Field Sessions. ASHRAE IAQ 2007 Conference,
Baltimore, MD.

 A.T Hodgson, A.F. Rudd, D. Beal, and S. Chandra. 2000. Volatile Organic
Compound Concentrations and Emission Rates in New Manufactured and
Site-Built Houses. Indoor Air 10: 178-192.

   HYPERLINK "http://www.epa.gov/iaq/base/" 
http://www.epa.gov/iaq/base/ 

 USEPA. 1996. Sources and Factors Affecting Indoor Emissions from
Engineered Wood Products: Summary and Evaluation of Current Literature.
EPA-600/R-96-067, NTIS PB96-183876, June 1996. Project Summary is
available at:   HYPERLINK "http://nepis.epa.gov/Adobe/PDF/P1000I7X.PDF" 
http://nepis.epa.gov/Adobe/PDF/P1000I7X.PDF .

 T. Matthews, T. Reed, B. Tromberg, C. Daffron, and A. Hawthorne. 1983.
Formaldehyde Emissions from Combustion Sources and Solid Formaldehyde
Resin Containing Products:  Potential Impact on Indoor Formaldehyde
Concentration and Possible Corrective Measures. In Proceedings of ASHRAE
Symposium “Management of Atmospheres in Tightly Enclosed Spaces,”
Santa Barbara, CA. 

 T. Matthews, A. Hawthorne, C. Daffron, T. Reed, and M. Corey. 1983b.
Formaldehyde Release from Wood Products. In Proceedings of the 17th
International Particleboard Symposium, Washington State University,
Pullman, WA.

 T. Matthews, A. Hawthorne, and C. Thompson.  1987. Formaldehyde
Sorption and Desorption Characteristics of Gypsum Wallboard. 
Environmental Science & Technology 21: 629-634.

 J. Pickrel, L. Griffis, B. Mokier, G. Kanapilly, and C. Hobbs. 1984.
Formaldehyde Release Rate Coefficients from Selected Consumer Products.
Environmental Science & Technology 18: 682-686.

 T. Godish and B. Kanyar. 1985. Formaldehyde Source Interaction Studies.
Forest Products Journal 35: 13-17.

 G Myers. 1984. Effect of Ventilation Rate and Board Loading on
Formaldehyde Concentration: A Critical Review of rhe Literature. Forest
Products Journal 34: 59-68.

 J Hoetjer. 1978. Introduction to a Theoretical Model for the Splitting
of Formaldehyde from a Composite Board. Methanol Chemie Netherland, June
8.

 A Berge, B Milligaard, P Haneto, and E Ormstad. 190. Formaldehyde
release from Particleboard – Evaluation of a Mathematical Model. Holz
als Roh-Und Werkstaff 38:251-255.

 S Fujii, T Suzecki, and S Koyagaehiro. 1973. Study on Liberated
Formaldehyde as Renewal for JIS Particleboard. Kenjal Shiken Joho 9:
10-14.

 Brown, S.K. 1999. Chamber Assessment of Formaldehyde and VOC Emissions
from Wood-Based Panels. Indoor Air  9: 209–215.

 L. Mølhave, L. Dueholm, and L.K. Jenson. 1995. Assessment of Exposures
and Health Risks Related to Formaldehyde Emissions from Furniture: A
Case Study. Indoor Air 5: 104-119.

 The use of changing values for the exponential decay can be viewed as a
simplified or constrained version of the double-exponential model; the
full, or unconstrained, model would use both exponentials in parallel
rather than in series. The current configuration of FIAM-pwp does not
allow use of the full double-exponential model, due in part to
uncertainty regarding appropriate parameter values to assign to the rate
constants for faster and slower decline. Faster and slower exponentials
can be applied sequentially, as done here, with FIAM-pwp by making a
series of two related model runs, as illustrated in Section 2.7.5.  

 Versar. 1988. Formaldehyde Exposure in Residential Settings: Sources,
Levels, and Effectiveness of Control Options. Update (September 30,
1988) of 1986 report by Versar, Inc. for the USEPA Office of Toxic
Substances, prepared under Contact No. 68-02-3968.

 T. Matthews, T. Reed, B. Tromberg., K. Fung, C. Thompson, J. Simpson,
and A. Hawthorne. 1985. Modeling and Testing of Formaldehyde Emission
Characteristics of Pressed-Wood Products: Report XVIII to the U.S.
Consumer Product Safety Commission. ORNL/TM-9867, Oak Ridge National
Laboratory, Oak Ridge, TN.

 Zinn, T.W., Cline, D. and Lehmann, W.F. 1990. Long-term Study of
Formaldehyde Emission Decay from Particleboard. Forest Products Journal
40(6): 15-18.

 California Air Resources Board (CARB). 2007. Initial Statement of
Reasons for Proposed Rulemaking, Proposed Airborne Toxic Control Measure
to Reduce Formaldehyde Emissions from Composite Wood Products. Available
at:   HYPERLINK
"http://www.arb.ca.gov/regact/2007/compwood07/fro-final.pdf" 
http://www.arb.ca.gov/regact/2007/compwood07/fro-final.pdf ..

 W. Groah. 2005. Decay or the Decrease in Formaldehyde Concentrations or
Emissions over Time and UF-bonded Wood Panel Products. Report prepared
for the Composite Panel Association, Formaldehyde Council Inc., Hardwood
Plywood and Veneer Association, and the Kitchen Cabinet Manufacturers
Association.

 W. Groah and G. Gramp. 1988. An Estimate of Home Occupant Exposures to
Formaldehyde Gas from Plywood and Reconstituted Wood Products Bonded
with Formaldehyde Based Adhesives. CPA/HPVA 1988 Exposure and Assessment
Package Report, Hardwood Plywood and Veneer Association, Reston, VA.
Available at:

  HYPERLINK
"http://www.ecobind.com/research/Decay_or_Decrease_in_Emissions.pdf" 
http://www.ecobind.com/research/Decay_or_Decrease_in_Emissions.pdf .

 ASTM. 1996a. Standard Test Method for Determining Formaldehyde
Concentrations in Air and Emission Rates from Wood Products Using a
Large Chamber. Designation E 1333-96 (Reapproved 2010), American Society
for Testing and Materials, Philadelphia, PA.

 A. Berge, B. Mellagaard, P. Hanetho, and E. Ormstad. 1980. Formaldehyde
Release from Particleboard – Evaluation of a Mathematical Model. Holz
Als Roh-und Werkstoff 38: 252-255.

  G. Myers. 1985. The Effects of Temperature and Humidity on
Formaldehyde Emission from UF- bonded Boards: A Literature Critique.
Forest Products Journal 1985: 35:20-31. 

 S. Silberstein, R. Grot, K. Ishiguro, and J. Mulligan. 1988. Validation
of Models for Predicting Formaldehyde Concentrations in Residences due
to Pressed-Wood Products. JAPCA 38: 1403-1411.

 References cited here are listed below Table 4-13; portions of this
section have been extracted from the following source:

USEPA. 2012. Formaldehyde from Composite Wood Products: Exposure
Assessment. Draft Final Report. U.S. Environmental Protection Agency,
Office of Pollution Prevention and Toxics, Exposure Assessment Branch,

1200 Pennsylvania Avenue NW, Washington, DC 20460.  July 2012.

 Portions of this appendix have been extracted or adapted from the
following sources:

D. Hare, R. Margosian, W. Groah, S. Abel, G. Schweer, and M. Koontz.
1996. Evaluating the Contribution of UF-bonded Building Materials to
Indoor Formaldehyde Levels in a Newly Constructed House. In Proceedings
of the 30th International Particleboard/Composite Materials Symposium,
Washington State University, Pullman, WA, pp. 93-108. Available at:  
HYPERLINK
"http://www.ecobind.com/research/Evaluating_the_Contribution.pdf" 
http://www.ecobind.com/research/Evaluating_the_Contribution.pdf .

M. Koontz, H. Rector, D. Cade, C. Wilkes, and L. Niang. 1996.
Residential Indoor Air Formaldehyde Testing Program: Pilot Study. Report
No. IE-2814, prepared by GEOMET Technologies, Inc. for the USEPA Office
of Pollution Prevention and Toxics under EPA Contract No. 68-D3-0013.

 USEPA. 1990. Compendium of Methods for the Determination of Air
Pollutants in Indoor Air.  Report No. EPA/600/4-90/010, Atmospheric
Research and Assessment Laboratory, USEPA, Research Triangle Park, NC. 

 ASTM. 1996. Standard Test Method for Determining Formaldehyde
Concentrations in Air from Wood Products Using a Small-Scale Chamber.
Designation D 6007-02 (Reapproved 2008), American Society for Testing
and Materials, Philadelphia, PA.

 T. Matthews, T. Reed, B. Tromberg., K. Fung, C. Thompson, J. Simpson,
and A. Hawthorne. 1985. Modeling and Testing of Formaldehyde Emission
Characteristics of Pressed-Wood Products: Report XVIII to the U.S.
Consumer Product Safety Commission. ORNL/TM-9867, Oak Ridge National
Laboratory, Oak Ridge, TN.

 Extracted/adapted from Formaldehyde from Composite Wood Products:
Exposure Assessment. Draft Final Report. U.S. Environmental Protection
Agency, Office of Pollution Prevention and Toxics, Exposure Assessment
Branch,1200 Pennsylvania Avenue NW, Washington, DC 20460.  July 2012.

  The model output no longer includes ADD values.

 PSI. 2009. Professional Services Industries, Inc. (PSI). “The TPC
Perspective,” available at   HYPERLINK
"http://university.ahfa.us/documents/fmhyde309_schutfort.pdf" 
http://university.ahfa.us/documents/fmhyde309_schutfort.pdf .

 According to CARB’s list of certified mills, as of November 12, 2010,
PSI was the TPC for 95 mills in 9 countries.  The countries, and the
number of mills certified by PSI in each, are as follows:  Argentina
(1); Chile (1); China (55); Ecuador (2); Indonesia (2); Malaysia (26);
the Philippines (1); Taiwan (1); and Thailand (6).

 The emission levels in the graphs seemed to be labeled according to the
interval midpoints (e.g., a reported value of 0.05 ppm represented an
interval of 0.0375 to 0.075 ppm). The maximum values reported in the
presentation for MDF and particleboard were used for the maximum
intervals. The median value was used for the median interval for MDF.
The data in the hardwood plywood graphic were consistent with an average
value of 0.07 ppm rather than the 0.05 ppm value that could be
calculated from the summary statistics shown in the presentation for all
three product types combined. For hardwood plywood, the three highest
emission levels (all of which exceeded 1.0 ppm) seemed to be driving
this discrepancy; thus, they were dropped from the analysis. After
dropping these outliers the imputed average emissions matched the 0.05
ppm value that could be calculated from the other summary statistics
shown in the presentation.

 Resin formulations that meet the NAF definition are those that do not
contain any added formaldehyde in their formulation.

 Emission levels for U.S. CARB2-compliant products (as opposed to
non-U.S. products) were chosen because the data set from which average
emission levels were determined was larger and, thus, considered more
reliable.

 A.T Hodgson, A.F. Rudd, D. Beal, and S. Chandra. 2000. Volatile Organic
Compound Concentrations and Emission Rates in New Manufactured and
Site-Built Houses. Indoor Air 10: 178-192.

 The degree to which the historical data in Table D-1 reflects
current/recent domestic or imported production is not known.

  PAGE   \* MERGEFORMAT  1 

  PAGE   \* MERGEFORMAT  67 

A-  PAGE   \* MERGEFORMAT  10 

B-  PAGE   \* MERGEFORMAT  4 

C-  PAGE   \* MERGEFORMAT  15 

D-  PAGE   \* MERGEFORMAT  8 

	

E-  PAGE   \* MERGEFORMAT  3 

Notice

 h

hEx

␃ഁ׆Āቈ愁Ĥ摧硅

(

Š

Ý

B

¢

	

"

#

$

%

&

'

(

)

*

F

G

H

I

L

M

h

i

j

„

…

,…

†

‡

ˆ

‰

Š

‹

Œ

¨

©

ª

«

®

¯

»

¼

½

×

Ø

Ù

Ú

Û

Ü

Ý

Þ

ß

û

ü

ý

þ

	

	



 

 

!

;

<

=

?

@

A

B

C

D

`

a

b

c

i

j



€

›

œ

Ÿ

 

¡

¢

£

¤

À

Á

-Á

Â

Ã

É

Ê

Þ

ß

à

ú

û

ü

þ

ÿ



y

y

y

 h²$

y

y

y

y

y

y

y

y

y

y

y

y

 h

h

 h

 h

h

h

h

h

h

h

h

h

h

h

h

h

h

  hŒ,

h

h

 h

h

h

h

h

h

h

h

h

h

h

h

h

h

h

h

h

h

  hŒ,

h

 hŒ,

h

h

 hŒ,

h

h

 h

h

gdy

a$gdy

a$gdy

gdy

°

gdy

gdy

gdy

°

gdy

a$gdy

gdy

 h(

  h(

a$gdy

a$gdy

 hŒ,

  hŒ,

愀Ĥ摧୹ò؀

␃愃̤摧挲§

̤̀☊଀ๆ愀̤摧挲§

 hÅ

  hÅ

 hÅ

 hd

  hd

h]

H*

愀Ĥ摧どi

愀Ĥ摧どi

愀Ĥ摧どi

愀Ĥ摧どi

愀Ĥ摧どi

晆؏

摧ၡ®

晆ಜ

h

hÑ

hÑ

h

h

h

h

h

h

h

h

h

h

h

h

h

gd

gd

h

 h

h]

h

h

h

K

L

Š

í

î

ꐓx摧ㅆé

␃愃̤摧孋É

̤̀☊଀ፆ᐀碤愀̤摧孋É

 

$

(

?

@

x

y

€

ƒ

‡

•

¼

È

ì

í

î

ï

h

h

h

h

h

h

h

h

h

h

h

h

h

h

h

h

h

h

h

h

愀Ĥ摧৽g

 h

h

h

h

h

h

h

h

h

h

圀Մ怀಄愀Ĥ摧た£

ž

 hJ

hJ

hJ

hJ

hJ

hJ

hJ

 hJ

hJ

 hJ

hJ

hJ

 hE

kd

h˜

h˜

kd

  h˜

 h˜

 h˜

  h˜

  h˜

 h˜

  h˜

 h˜

Æ

Æ

Æ

Æ

Æ

`úØ	š

„š

]„Ð^„š

Æ

`úØ	š

„–

]„Ð^„–

jX

Æ

Æ

Æ

`ú@

Æ

Æ

`ú@

Æ

Æ

Æ

@ˆþÿ

Æ

Æ

Æ

Æ

$

$

 hw	

ᔐ२顈ᘀ땨鵟䀀ﺈÿ̙䉪ᙌᔀ睨ഉᘀ땨鵟䔀嗿Ĉ̫繪䤁
੓Ĉ栕ॷ

 

E

F

´

Þ

␃ᐃ愀̤摧䠉ഀF

±

´

µ

·

¸

Þ

ê

ó

6

7

;

<

_

q

¤

©

Õ

ñ

ô

␃ഃÆ

Æ

Æ

h(f

h(f

愀Ĥ摧珵t

愀Ĥ摧珵t

摧珵t

摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

摧珵t

摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

摧珵t

摧珵t

摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

摧珵t

摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

摧珵t

 hž

摧珵t

摧珵t

愀Ĥ摧珵t

愀Ĥ摧珵t

̤̀☊଀ⵆ฀䒄崁䒄愁̤摧⓬L

愀Ĥ摧玎¹

§

愀Ĥ摧玎¹

§

愀Ĥ摧玎¹

§

§

 h§f

 h§f

$

@

 h§f

 h§f

 h§f

 hõ9

@

@

@

$

@

$

@

@

@

$

@

@

@

  h˜

¾

;

¾

Ë

愀Ĥ摧玎¹

 hÅ

  h(

  h(

  h(

愀Ĥ摧玎¹

  h(

  h(

d

d

Æ

`úØ	š

„–

]„Ð^„–

Æ

`úØ	š

„–

]„Ð^„–

Æ

`úØ	@

„@

]„Ð^„@

ùñæñæñæñæØÏñæñæñæñæÅ¾æñæñæñæñæ±ª¦ªŸª
¦ª¦ª¦ª˜‹˜ª„¦„¦„¦ h?

 hò

 h?

혈\䀄닿舌⤚耣爆

耀퀆

鐇Ǔ혈\䀄닿舌⤚耣爆

耀퀆

̤̀☊଀ņ㄀$␷㠀$⑈愀̤摧㴙g
̤̀☊଀ņ᐀㲤㄀$␷㠀$⑈愀̤摧㴙gᘀ

  h

°ð h

  h

愀Ĥ摧玎¹

摧玎¹

摧玎¹

摧玎¹

摧玎¹

摧玎¹

摧玎¹

 h

$

@

„

`„

' h 

	

	

'

/

X

e

’

Ÿ

 

¡

¡

¢

®

ê

ì

í

	

B

C

E

F

G

H

!H

µ

»

¼

Å

Ñ

Ò

Ö

×

␃ฃ뮄ༀᄀ᪄ᓿ㲤崀뮄帀怀᪄懿̤摧⓬L଀

  hˆ

 hˆ

 hˆ

  hl

„

`„

„

`„

„

`„

ð

ñ

ñ

ñ

 ñ

"ñ

#ñ

(ñ

1ñ

2ñ

>ñ

?ñ

@ñ

Nñ

Oñ

Pñ

Uñ

^ñ

_ñ

dñ

mñ

nñ

tñ

}ñ

~ñ

ƒñ

Œñ

ñ

’ñ

›ñ

´ñ

µñ

Èñ

Éñ

Êñ

Ëñ

Íñ

Îñ

äñ

åñ

es. Persons seeking information on specific regulatory requirements
should consult 40 CFR Part 767, and the preamble for the regulatory
action.

Mention of the names of specific companies, organizations, or entities
does not constitute an endorsement by EPA.

Upper Portion

Middle Portion

Lower Portion

Display Area

Selection Area

Upper Portion

Lower Portion

Middle Portion

Upper Portion

Lower Portion

Fast Decay

Slow Decay

 

 

 

Q12

Q21

Q01

Q02

Q10

Q20

Zone 1

V1

S1

Zone 2

V2

S2

