Performance Concerns with Wood-Frame Attics

PERFORMANCE CONCERNS
WITH WOOD-FRAME ATTICS
MARCUS DELL, PENG; GRAHAM FINCH, PENG; BRIAN HUBBS, PENG;
AND ARIEL LEVY, PE
RDH BUILDING SCIENCES INC.
308 S.W. First Ave., #300, Portland, OR 97204
Phone: 503-243-6222 • Fax: 503-243-5052 • E-mail: alevy@rdhbe.com
S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1 L E V Y • 9 1
9 2 • L E V Y S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1
ABSTRACT
As the quantity of insulation used in wood-frame attics has increased to reduce energy
consumption, we have observed a disturbing increase in moisture damage and mold growth
on the framing components within those attics. Proposed building codes in some jurisdictions
intend to increase minimum insulation values further in the next few years. Code ventilation
requirements for attics in North America have not changed in nearly 60 years.
However, insufficient ventilation is unlikely to be the sole cause of the observed problems.
This research utilizes case studies and computer model simulation to review the contributing
factors to the observed attic moisture problems. We also explore parallels between
moisture-related problems in wall assemblies that the Pacific Northwest has experienced
over the last several decades.
This paper will discuss the impact of the following:
• Reduced conductive heat loss through the insulation
• Uncontrolled air leakage, both through the building enclosure and through mechanical
systems
• Discharging exhaust vents through soffits
• Venting rainscreen cavities into the attic space
• Fireplaces and ice damming
In conclusion, this paper discusses the above factors as they relate to steps that should
be considered when developing remedial solutions. It will also discuss alternate insulating
strategies and how those methods may address the currently observed problems.
SPEAKER
ARIEL LEVY, PE — RDH BUILDING SCIENCES INC., PORTLAND, OR
ARIEL LEVY is a principal with RDH Building Sciences and manages its Portland,
Oregon, office. RDH focuses on the integration of building science, engineering, architecture,
construction management, and risk assessment services for new and existing building
enclosures. Levy has spent his career designing and investigating building enclosure systems
in most climate zones of North America. He is regularly asked to provide expert testimony
on building enclosure performance problems and is recognized by both plaintiff and
defendant parties as an objective consultant. As such, he is often asked to perform independent
third-party evaluations and to perform directly as a mediator’s technical consultant.
INTRODUCTION AND
BACKGROUND
In the early 1990s, the construction
industry identified a widespread trend of
water ingress and resultant deterioration in
wood-framed buildings. Although the problems
had existed for some time, there was a
significant recognition from design and construction
professionals alike that the problems
were systemic. Some of the wellknown
causes of increased moisture accumulation
and subsequent deterioration in
wall assemblies are the following:
• An increased need/desire for higher
levels of insulation
• An increased need/desire for more
airtight wall assemblies
• A reduction in the resilience of construction
materials to moisture
• A decrease in the vapor permeability
of sheathing materials, due in part
to the glues (e.g., plywood and OSB)
In the Pacific Northwest, this realization
resulted in a comprehensive 1996 study
funded by the Canada Mortgage and
Housing Commission (CHMC), Survey of
Building Envelope Failures in Coastal
Climate of British Columbia (Study). The
study focused on exterior wall assemblies
and identified that generational shifts in
design, materials, and energy use result in
a shift of the wetting and drying balance of
the building enclosure. In many cases,
water penetration into the exterior walls is
no longer able to dry before deteriorating
wood components.
The study, along with other research,
kindled significant changes in design and
construction practices of exterior walls.
Many of these design and construction
practice changes are documented in another
CMHC study (Guide), called the Best
Practice Guide: Wood-Frame Envelopes in
the Coastal Climate of British Columbia
(CMHC 1999). The guide laid the foundation
for industry acceptance in British Columbia
of rainscreen and drained-wall cavity construction
methods, which have been used to
reduce the risk of wood-frame deterioration
since the late 1990s. The performance of
these rainscreen wall-cladding assemblies
in similar climates is now well documented
by the industry and the results available in
published papers, including “The Performance
of Rainscreen Walls in British
Columbia” (Finch 2007).
RDH has investigated attics in woodframe
buildings for decades and, more
recently, has identified a problematic level
of premature deterioration of the wood
sheathing, growth of mold, and other related
moisture problems. We find parallels
between the previously described problems
in wall assemblies in the 1990s and the
problems more
recently observed
in attics. The
attics discussed
in the following
case studies are
associated with
multiunit residential
woodframe
buildings,
typically two or
three stories tall.
Figures 1 and 2
show buildings
indicative of the
type discussed
herein.
This research was primarily
focused in the coastal area of
British Columbia, which is a temperate
rainforest. However, the
weather in the greater Vancouver
area is similar to much of climate
zone 4C (as identified in the current
editions of the International
Energy Conservation Code and ASHRAE
90.1). As might be expected, the water
ingress problems and resultant deterioration
identified in coastal British Columbia
in the early 1990s also occur in other geographic
locations, including other parts of
Canada and the United States, as well as in
much of Europe, New Zealand, Australia,
and other countries with similar building
practices and climates.
In North America, the basics of attic
design (from a building science perspective)
have not changed significantly in decades.
PERFORMANCE CONCERNS
WITH WOOD-FRAME ATTICS
Figure 1 – Typical multiunit
three-story wood-frame
building with attic problems.
Figure 2 – Typical two-story
wood-frame multiunit
building with attic problems.
Figure 3 – Typical attic configuration.
S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1 L E V Y • 9 3
Figure 3 illustrates typical components of a
sloped-roof assembly with vented attic. The
ventilation is typically provided at the soffits
and near the ridge. The upper venting is
either provided through button-style vents
or through vented ridge shingles. All of the
case studies discussed in this paper have
this same basic roof and attic configuration
(framing varies).
As in the aforementioned wall discussion,
the quantity of insulation used in
wood-frame attics has increased over time
in an effort to reduce energy consumption
and to improve occupant comfort.
The current standard in many
parts of Canada and northern
portions of the U.S. is around R-
40. Also similar to the wall discussion,
sheathings have shifted
away from loosely fitted planks to
heavily glued plywood and OSB
products. Proposed building
codes in some jurisdictions intend
to continue increasing
the insulation
requirements
of
ceilings and
attics to accommodate
the
growing need
for energy consumption
reduction.
Ho w e v e r ,
although there
has been a
shift in insulation requirements, building
codes in most North American jurisdictions
for ventilation of attics have not changed, in
some cases, for as many as 60 years. The
ventilation requirement for attics in most
North American jurisdictions is 1/150,
which can be further reduced to 1/300, in
many cases, with the addition of a ceiling
vapor barrier. Most jurisdictions require
that this ventilation space be split between
the upper and lower halves of the space.
The following case studies evaluate the
cause of moisture-related damage and fungal
growth in wood-framed attics.
The last case study discusses
unintended heat loss into the
attic. In each case, a
number of other factors
may have contributed to
the observed problems.
For brevity, these other
factors are not discussed
here.
CASE STUDY 1
Case Study 1 (Figures 4 and 5) is an 82-
unit townhouse development located in
Squamish, BC (about one hour’s drive north
of Vancouver). The development consists of
20 multiunit townhouse buildings that are
accessed by common laneways. The buildings
were constructed in 2002 and are all of
similar design and construction. We investigated
the project in the fall of 2007.
The primary visual symptom of the
problem was extensive mold on the underside
of the plywood sheathing, as well as
moist sheathing. During our investigation,
we identified the following:
• There are approximately 12 in of R-
40 blown fiberglass insulation in the
attics.
• There was no identifiable water leakage
through the roofing that may
have contributed to the observed
leakage.
• Ventilation is provided through ridge
vents (Figure 5), full-length perforated
soffit vents (Figure 6), and gable
end-wall vents (Figure 7). Our calculations
confirm that the quantity of
venting exceeded building code
requirements. We also confirmed
that the venting was adequately distributed.
• Mold growth (Figure 8) was more
extensive near the bathroom
exhaust fans and ducting from the
clothes dryers. Exhaust vents are
ducted to the eaves and pointed
down toward the perforated soffits
but are not hard-connected. We
used tracer smoke injected into the
bathroom fans to trace airflow pat-
Figure 4 – General
view of
Case Study 1.
Figure 5 –
General view of
shingle-clad
roofs with
ridge venting.
Figure 6 – Typical
perforated soffit vents.
Figure 8 – Mold on sheathing. Note the
generous use of vent baffles.
Figure 7 – Typical
gable end vent.
9 4 • L E V Y S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1
terns. During this testing,
the discharge from
the ducts short-circuited
back into the attic, up
through soffit vent baffles,
as opposed to exiting
out through the soffit-
mounted grills. As
built, the exhaust ducts
were terminated several
inches short of the vent
grills (Figures 9 and
10). We verified that the
vent grills and ducts
were not plugged and
that surrounding soffits
are fully perforated.
However, the buoyancy
of the discharge air and
the natural convective airflow
through the attic caused the tracer
smoke to rise back up into the attic
through the baffles. This flow path
mimics the observed mold patterns
on the plywood sheathing (Figure
11).
• Although south-facing roofs receive
more solar heating and, as such, are
often less subject to moisture accumulation
and mold growth, the location
of the bathrooms and the
exhaust systems were more dominant
factors for mold growth than
the roof orientation (i.e., solar heating
was not sufficient to balance the
increased moisture load).
• Ridge vent capacity is reduced
because they contain a filter fabric
(Figures 12 and 13), presumably
used as a bug screen. Our calculations
confirmed that the airflow values
published in the product literature
did not allow for the restriction
caused by the filter fabric. However,
we estimate that the overall
attic ventilation maintained
the building code
requirements (we did not perform
testing to confirm the airflow
through the filter fabric).
• We also used smoke testing to confirm
that some air leakage was
occurring around the attic access
hatch and through poorly sealed
dryer ducts.
• The extent of the mold growth varies
from unit to unit, possibly as a
result of changes in construction,
but more likely as a result of occupant
behavior.
The Venting Process
In a typical attic, ventilation
occurs between the soffit and
ridge roof vents. Under normal
wintertime conditions, effects of
buoyancy will draw the air out of
the ridge vent and pull air into the
attic through the soffit vents
(although local wind
pressures can counteract
this natural flow).
When insulation is added to the conventional
ceiling (attic floor), heat loss from the
living space below is reduced, which in turn
reduces the capacity of the attic air to
absorb moisture (cool air has less capacity
to hold moisture than warm air). Reducing
the moisture absorption capacity of the air
in the attic reduces the rate of drying and
therefore may increase the risk of mold
growth (provided local surface temperatures
remain adequate for growth). Also, reducing
the heat flow into the attic reduces the
quantity of natural ventilation that occurs
Figure 9 – Typical grill at bathroom
exhaust vent.
Figure 10 – Duct not continuous to grill.
Figure 11 – Mold is worse inside of the baffle.
Figure 13 –
Effectiveness of
ridge vent
significantly
restricted by
geotextile fabric.
Figure 12 – Ridge vent.
S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1 L E V Y • 9 5
(air changes per hour) because the thermal
buoyancy of the air is reduced.
Attic Humidity
Attic humidity problems will most often
occur in the winter months, when warm,
moist air from the interior leaks into the
attic space and cannot be removed fast
enough by ventilation. If the leakage rate is
high enough, the relative humidity (RH) will
continue to increase, eventually leading to
moisture accumulation and condensation
on the cold roof sheathing and truss surfaces.
In Case Study 1, airflow was significantly
restricted though the ridge vents by the
integral filter fabric bug screen, and the
quantity of moisture entering the attic was
excessive because of the short-circuiting of
the bathroom exhausts at the soffits.
Consequently, the warm, moist air was not
removed quickly enough, resulting in an
accumulation of water vapor and increased
RH within the attic space. This high RH led
to moisture accumulation and condensation
on the cold roof surfaces.
The moisture sources are summarized
in Figure 14.
Rehabilitation
The rehabilitation process was twofold:
• Increase ventilation capacity on the
high side of the attic.
• Reduce moisture load.
The aforementioned were
addressed through remedial
repairs that included the following:
• All of the joints in the
bathroom ducts were fully
sealed with foil-face tape.
The bathroom exhaust ducts were
extended down and through the perforated
soffits. New hoods were
added to the exterior ends of the
ducts, which were designed and
installed to drain condensation that
formed in the outer portions of the
duct.
• All of the joints in the dryer exhaust
ducts were fully sealed, and an airtight
seal was obtained at the ductto-
sheathing interface. The exterior
of the exhaust ducts was insulated.
• Smoke testing was performed on
randomly selected bathroom and
dryer exhausts to confirm continuity
and airtightness.
• Other air-sealing work was also performed
to minimize air from within
the occupied space from entering
the attic (i.e., attic hatch seals).
• Additional button-style vents were
added. The final ventilation area
exceeded code requirements.
CASE STUDY 2
Case Study 2 consists of 110 residential
townhouse units within 16 multiunit buildings,
which are accessed by common pathways
and/or by street access. The complex
is located in New Westminster, a Vancouver,
BC, suburb. The buildings were originally
constructed in two phases (1999 and 2004),
and all are of similar design and construction.
We performed our investigation in
summer 2008. Similar to Case Study 1, the
primary visual symptom of a problem was
Figure 14 – Summary of attic moisture sources and airflow paths.
Figure 15 – Button vents near
ridges and hips.
Figure 17 – Gooseneck vents at
dryer duct termination.
Figure 16 – Poorly installed baffles
at perforated soffits.
9 6 • L E V Y S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1
mold growth on the underside of the
sheathing, which existed in the majority of
the attics.
Field review and testing determined the
following:
• The attics are insulated with
approximately 12 in of R40 fiberglass
insulation.
• The attics are ventilated through the
use of button vents near the ridges
(Figure 15) and perforated soffits
with baffles (Figure 16). We confirmed
that the button vents were
properly installed and functioning
as intended.
• The clothes dryers are
located in the upper level of
the units, and the ducts for
the dryers extend vertically
through the ceiling (attic
floor) and exit at gooseneck
vents (Figure 17) through
the roof sheathing. Ducts
are reasonably well sealed
along their length and are
insulated with fiberglass
insulation with a polyethylene
wrap.
• The exterior ends of the
ducts are sealed to the
underside of the sheathing
with aluminum foil tape
(Figure 18). Extensive mold
around the duct terminations
indicates that air
leakage is occurring. We
used tracer smoke to verify
this air leakage path.
• Despite the clear correlation
between the dryer duct
terminations and the deterioration
patterns, mold
growth is also present at
other locations away from
these terminations (Figure 19), specifically above a rainwater leader discharge
(Figures 20 and 21). Figure 22 illustrates mold growth on sheathing at the top ventilation
zone of the rainscreen cavity that is adjacent to the rainwater leader discharge
point.
Rainscreen Cavity Ventilation
In Figure 21, the rainscreen wallcladding
assembly (behind the blue vertical
arrows) incorporates ventilation and a
drainage cavity. The top of the cavity, how-
Figure 18 – Dryer duct
termination at underside of
sheathing.
Figure 19 – Localized mold on
underside of sheathing near
soffit at northern slope of roof
shown in Figure 20.
Figure 20 – The rainwater leader at this
location discharges onto a lower gable roof
projection near its ridge. The water runs
down the roof parallel to the wall until it is
directed away from the wall at the roof valley
approximately 2 ft from the rainwater leader.
Figure 21 – The rainscreen wallcladding
assembly at this location
incorporates a ventilation and
drainage cavity.
Figure 22 – Photograph taken inside
of soffit directly above lower roof
shown in Figure 21. Note that the
rainscreen cavity terminates above
the level of the soffit material.
S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1 L E V Y • 9 7
ever, extends over the soffit above and vents
into the attic space (Figure 22). The condition
is exacerbated by the rainwater leader
at the base of the rainscreen cavity, which
provides additional moisture within the
ventilation air space. In Figure 22, mold is
visible directly above the rainwater leader
location.
Attic Ventilation
Ventilation wass somewhat restricted by
poor installation of the foam vent baffles at
the soffits. The restricted ventilation was
unable to adequately dry the moisture from
poorly terminated dryer ducts and the
moisture that entered the attics through
ventilation of the rainscreen cavity above
the soffit level.
Rehabilitation
Remedial repairs included the following:
• Premanufactured duct receivers
were installed at the duct-to-sheathing
interface. The duct receivers
have flanges for tie-in to the sheathing
to minimize the risk of air leakage.
The duct receivers include a
solid duct extension that was inserted
into the gooseneck vent on the
exterior of the roof. The joint
between the duct receiver and the
gooseneck vent was sealed with
polyurethane foam. The perforated
soffits surrounding the duct discharge
points were replaced with
solid soffit material, which will
reduce the potential for the flow of
exhaust air back up into the attic.
• All of the dryer ducts were insulated
after air sealing.
• All damaged baffles were replaced
and additional baffles added.
Additional button vents were also
added in select locations where the
vent distribution was not adequate.
• Where a few originally installed
dryer exhausts vented through the
soffits, they were rerouted to extend
vertically up through the roof
sheathing. The advantage of vertical
ducts is that the natural buoyancy
of the air assists with discharge from
the building.
• Extensions were added to the rainwater
leaders so that the discharge
was carried continuously into the
gutter on the lower roof. This will
extend the life of the shingles on the
lower roof, as well as reduce the
moisture load in the attic.
There was some consideration of
modification to the rainscreen cavity
ventilation design. However, for budgetary
purposes, relation of the
downspouts and increased ventilation
were more practical and were adequate
to resolve the issue.
CASE STUDY 3
The problems studied in Case
Study 3 are similar to those in Case
Studies 1 and 2. Specifically, the primary
moisture source was air leakage
from bathroom exhausts, clothes dryer
exhausts, and from the occupied space
into the attic.
This case study illustrates that the orientation
of the roofs to solar radiation (and
its subsequent heating and drying capacity)
is not singularly sufficient to prevent mold
growth on the underside of the sheathing
when moisture loads are high.
Case Study 3 is a townhouse development
located in Burnaby, BC, a suburb of
Vancouver. The development consists of 31
detached three-story townhouse buildings
with five to seven units per building, for a
total of 177 units (Figures 23 and 24). The
first phase of the development was occupied
four years prior to our review.
The roof shingles are a dark color; and
the roofs are generally exposed to solar
radiation, as they lack shade by trees or
other buildings. The attics are vented
through button vents near the ridge and
through perforated soffits at the perimeter.
Based on our calculations, the vent area
and distribution comply with building code
requirements.
The photographs in Figures 25 and 26
were taken within the same attic. As illustrated,
the extent and distribution of the
mold growth is similar on both the
north- and south-facing aspects of the
attic. Although shingle heating on the
south-facing attics likely increases the
drying capacity of the sheathing, it is
not adequate to mitigate mold growth.
Further, it is likely that redistribution of
moisture within the attic during periods
of cloud cover or at night maintain the
ability for even mold distribution.
The problems were addressed
through remedial repairs similar to
those for Case Studies 2 and 3.
Figure 23 – General view of
Case Study 3 complex.
Figure 25 – Fungal growth on
south-facing roof sheathing
throughout the attic.
Figure 26 – Fungal growth on
north-facing roof sheathing
throughout the attic.
9 8 • L E V Y S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1
Figure 24 – General view of
Cast Study 3 complex.
CASE STUDY 4
Case Study 4 differs from the previous
case studies in that the problem was associated
with ice damming and water penetration
(Figures 27 and 28), as opposed to mold
growth or air leakage. Case Study 4 is located
in Panorama, BC, a ski town located in
the interior of the province. The complex
was constructed in 2005. The owners
reported problems the first winter.
Field review and testing determined the
following:
• The units have electric baseboard
heaters and propane fireplaces.
Each unit is individually metered for
electrical use, while the propane for
the fireplace (and the cost) is shared
equally amongst all of the units. The
property manager reported that
most occupants use the fireplaces
extensively for heating as well as for
ambiance.
• The roofs and attics have a complicated
configuration (Figure 29)
because of the unit layout and an
architectural need for fire separation
within the attics. This complex
configuration increases the difficulty
of obtaining uniformly distributed
ventilation.
• The architectural plans called for the use of cupola vents at intervals
along the roof ridges, but continuous ridge venting was installed at the
time of construction (Figure 30) in lieu of cupolas. Ridge vents become
buried by snow during the winter season.
• All of the soffits contained perforated vents (Figure 31).
• Calculations confirmed that the vent areas (during the summer
months) conformed to building
code requirements. Observations
confirmed that the vents were adequately
disbursed.
• Several of the buildings have gable
end vents (Figure 32). Even though
the vents are large, their effectiveness
is limited because of the fire
separations within the attics.
• The vents for the fireplaces pass
Figure 27 – Heavy ice damming
and icicle formation.
Figure 28 – Heavy icicle
formation.
Figure 30 – Typical
continuous ridge vent.
Figure 29 – Typical roof
configuration above area of
ice damming and icicle
formation. Circle marks gas
fireplace exhaust vent.
Figure 31 – Continuous
soffit venting.
Figure 32 – Gable end vents were not effective
because of attic compartmentalization for fire
separation.
S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1 L E V Y • 9 9
through the attics. The
fireplace vents are uninsulated
flexible ducting.
The temperature of the
fireplace vent ducts was
measured above 128°F (53°C) within
one hour of operation (Figure 33).
A combination of conditions contributed
to the ice damming on this project, but the
most significant were the high attic temperatures
caused by heat from the fireplace
ducts. Other contributors include:
• The thermal resistance of the snow
(approximately R1 per inch) contained
the heat within the attics,
increasing the melting rate of the
snow adjacent to the shingles.
• The ineffectiveness of the ridge vents
once the vents were buried in snow.
• The complex configuration of the
roofs, which results in a large roof
area draining down onto a short section
of eave, as well as creating the
difficulty of adequate ventilation.
Remediation
Interestingly, most of the adjacent
buildings have raised cupola vents on the
roof ridges that likely remain elevated above
the winter snow level (Figure 34). To remediate
this building, we added insulation to
the fireplace ducts and improved attic ventilation
below the roof surface. The remedial
repairs did not eliminate the icicles but
reduced the problem to a tolerable
level (Figures 35 and 36).
ANALYSIS BEYOND THE
CASE STUDIES
Hygrothermal Assessment –
The Impacts of Improved
Energy Efficiency on Attic
Moisture Tolerance
We performed a series of
hygrothermal simulations using
the WUFI 5.1 Pro (WUFI)
computer model to
demonstrate the performance of increasing
insulation levels, air leakage, and attic ventilation
on the moisture sensitivity of ventilated
wood-frame attic assemblies. The
model runs used Vancouver, BC, weather
data. The WUFI model takes into account
material properties, measured exterior climatic
data, and indoor climatic data to simulate
heat and moisture transfer through
assemblies. WUFI is one of the most
advanced commercially available hygrothermal
moisture programs in use today, and
its accuracy at prediction of wood-frame
assemblies has been demonstrated in
numerous field and laboratory studies
(www.wufi.de), including those performed
by our firm (Finch 2007).
For this assessment, a model of a conventional
ventilated wood-frame attic
assembly was constructed within WUFI,
and three-year simulations were performed
to evaluate year-over-year trends. The
model is shown schematically in Figure 37
to demonstrate the changing variables,
which we independently assessed.
We know that independently assessing
variables is somewhat unrealistic because
many of the variables are interrelated.
Figure 33 – Fireplace exhaust
vents reached 53.4ºC (128.12ºF)
within one hour of fireplace
operation.
Figure 35 – After remedial work.
Figure 37 – WUFI model schematic.
1 0 0 • L E V Y S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1
Figure 36 – After remedial work.
Figure 34 – Raised cupola
vents on adjacent building.
However, this approach allows us to assess
the impact of each variable separately. To
evaluate the results, the moisture content
of the plywood sheathing, along with the
relative humidity and temperature at the
sheathing surface and within the attic
space, are evaluated for each case. Typical
daily or seasonal results are provided, along
with a count of the number of hours where
certain predetermined thresholds are met.
Impact of Insulation R-value
The first analysis demonstrates the
impact of increasing insulation R-value on
the conditions within an attic. The model
assumes an average ventilation rate of four
air changes per hour (ACH), no incidental
air leakage from the house into the attic,
and varying insulation depth from R-0 (no
insulation) to R-100 (superinsulated) in
increments of R-20.
Figures 38 and 39 show the sheathing
surface temperature and surface RH for the
attic, respectively, occurring during a typical
winter week. A typical, indoor-air dewpoint
temperature is also plotted against
the surface temperature to assess the
potential for indoor air leakage to result in
condensation on the sheathing surface. The
majority of wintertime hours are below the
indoor air dew point, which indicates the
potential for condensation to occur where
air leakage is present (i.e., at leaking penetrations).
If bathroom fan exhaust was considered
with warm, moist air due to showers,
it would be warmer and at higher relative
humidity than typical interior air. This
warm, moist air would have relatively high
dew point temperatures, closer to 20°C,
such that condensation would be almost
sure to occur from this sort of air leakage,
as attic temperatures are well below its dew
point.
As insulation R-value increases, the
temperature of the sheathing decreases and
the surface RH increases, since less heat
energy is lost through the ceiling to warm
the attic. The number of hours in a typical
year where the RH is above 80% at the interior
of the plywood surface and within the
open attic airspace is shown in Figure 40.
The numbers of hours above 80% RH are
representative of risk of mold growth when
temperatures are favorable (i.e., between
0°C and 40°C, or 32°F and 104°F). The
resulting sheathing moisture contents are
shown in Figure 41, and the number of
hours exceeding 15% MC (as there is no
occurrence that exceeded 20% under normal
conditions) is shown in Figure 42.
Figure 38 – Sheathing temperature; impact of insulation R-value.
Figure 40 – Annual hours of
RH >80% at plywood
sheathing and center of
attic; impact of insulation Rvalue.
Figure 39 – RH at plywood sheathing; impact of insulation R-value.
Figure 41 – Typical year annual plywood moisture content, impact of insulation
R-value.
Figure 42 –
Annual hours of
plywood
sheathing above
15% MC; impact
of insulation Rvalue.
S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1 L E V Y • 1 0 1
Analysis Comments
In each of the cases, the simulation
results demonstrate that greater insulation
levels reduce heat loss into the attic, resulting
in colder sheathing temperatures and
increased RH levels. The largest reduction in
temperature and increase in RH occurs
between an uninsulated attic and R-10.
Changes are observed in insulation increments
up to about R-40, and conditions
between R-60 and R-100 are not much cooler/
humid than those at R-40. Note that all of
these simulations use the same air change
rate (4 ACH). In reality, the number of the
ACH is influenced by convective airflow
caused by thermal buoyancy of the air in the
attic as the air is warmed. Reducing the heat
flow into the attic will reduce the ACH, which
will in turn increase the attic RH.
Impact of Insulation R-value With Air
Leakage
The second analysis demonstrates the
impact of increasing insulation R-value on
the conditions within an attic where there is
air leakage from the interior into the attic
(either through the ceiling or leaking duct
work). The model assumes an average attic
ventilation rate of 4 ACH, a small amount of
air leakage from the house into the attic,
and varying insulation depth from R-0 (no
insulation) to R-100 (superinsulated) in
increments of R-20. The amount of air leakage
is set within the WUFI model and is
based on a leakage rate that results in wetting
of the sheathing up to 25% MC in the
baseline model. Increasing insulation with
this same air leakage rate will therefore
reduce the sheathing temperature, resulting
in increased condensation events.
Figures 43 and 44 show the sheathing’s
surface RH and MC, respectively, which
occurs during the winter months (November
through April). Figures 45 and 46 show the
number of hours per year that the sheathing
exceeds 80% RH and 20% MC, respectively.
Analysis Comments
Similar to the first analysis, as the insulation
level increases, the time spent at elevated
RH and MC levels increases at the
plywood sheathing. Because the sheathing
is maintained at a temperature well below
the indoor dew point for much of the winter,
the risk for air leakage condensation is
high. The results, however, do not show a
significant increase in moisture levels when
insulation levels are above R-40. The
results highlight the sensitivity of attics to
air leakage and the need to address ceiling
and ductwork air tightness.
Impact of Attic Ventilation Rate
The third analysis demonstrates the
impact of the average attic ventilation rate
on the conditions within an attic. The model
parameters are as follows:
• The average ventilation rate is varied
from ≈0 ACH to 10 ACH (unventilated
to very well ventilated).
• There is no air leakage from the
house into the attic.
• There is a fixed insulation R-value of
R-10, representative of an older,
poorly insulated attic.
Figure 43 – RH at plywood sheathing; impact of insulation R-value with constant
air leakage.
Figure 44 – MC of plywood sheathing; impact of insulation R-value with constant
air leakage into attic.
Figure 45 –
Annual hours of
RH at plywood
sheathing and
center of attic;
impact of insulation
R-value with
constant air
leakage into attic.
Figure 46 –
Annual hours
of plywood
sheathing above
20% MC; impact
of insulation
R-value with
air leak.
1 0 2 • L E V Y S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1
Hourly ventilation rates from this average
are subsequently calculated within
WUFI to adjust for diurnal cycles and
impacts of solar radiation, etc. Based on
industry research and measurements, average
attic ventilation rates of between 1 and
4 ACH are expected in wood-frame attics
(Walker 1997). The impact of attic ventilation
rate on the attic air temperature and
RH for a typical winter week is shown in
Figure 47 and Figure 48. For comparison,
Figure 49 plots the moisture content of the
plywood sheathing as a result of the varying
ventilation rate for the same time period.
Three interesting findings become
apparent in the analysis of varying the ventilation
rate in the coastal climate of British
Columbia:
• Higher ventilation rates (above 2
ACH) result in cooler temperatures
within the attic and of the sheathing,
as the outdoor air is colder than
the attic. The attic air is warmed by
heat loss through the insulation
from below and by a reduction in the
amount of air that is exchanged with
colder outdoor air during warmer
seasons. An attic with no ventilation
(≈0 ACH) results in the warmest
temperatures.
• No ventilation (≈0 ACH) results in
warm attic temperatures, but higher
RH levels. The highest RH levels,
however, come from the higher ventilation
rates (above 4 ACH) because
in this climate, the outdoor air is
more humid than the air within the
attic. At around 4 ACH, the RH levels
are the same whether the attic is
ventilated or not. Less ventilation
results in lower RH levels to a point
of a minimum RH near 0.1 to 0.05
ACH. Below 0.05 ACH (e.g., 0.01
ACH), the RH within the attic
increases, demonstrating a lower
inflection point and potentially, a
theoretical optimal ventilation rate
for this attic. This is highlighted in
Figure 50, which shows the number
of hours per year exceeding 80% RH
at the sheathing in the middle of the
attic for different ventilation rates.
• The highest sheathing moisture content
results from no ventilation or
very low ventilation rates, closely
followed by the highest ventilation
rate (because the outside air is
depositing more moisture than the
ventilation is removing). This is
highlighted in Figure 51, which
Figure 47 – Attic air temperature; impact of attic ventilation rate.
Figure 48 – Attic RH; impact of attic ventilation rate.
Figure 51 –
Annual hours the
plywood sheathing
exceeds 15%
moisture content;
impact of attic
ventilation rate.
Figure 50 –
Annual hours of
RH at plywood
sheathing and
open attic
airspace; impact
of attic ventilation
rate.
Figure 49 – Plywood sheathing moisture content; impact of attic ventilation rate.
S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1 L E V Y • 1 0 3
shows the number of hours per year
the plywood sheathing moisture
content exceeds 15% for different
ventilation rates.
Analysis Comments
The analysis demonstrates the impact of
ventilation on the sheathing moisture levels
and RH within an attic. Too little ventilation
is detrimental and results in high RH levels
within the attic and at the sheathing.
However, too much ventilation can also
cause elevated RH levels and sheathing
moisture contents. For this analysis, there
appears to be an optimal ventilation rate of
less than 1 ACH to a minimum of 0.1 ACH
for attics within this climate. Further
research is needed to evaluate how this
would be achieved or constructed in practice
(e.g., vent area and distribution, etc.)
and the impacts of air leakage.
Impact of Attic Ventilation Rate With
Air Leakage
The final analysis demonstrates the
impact of the average attic ventilation rate
on the conditions within an attic. The model
parameters are these:
• The model varies the ventilation rate
from 0.5 ACH to 4 ACH (low to high
ventilation).
• Air leakage from the house into the
attic is included.
• There is a fixed insulation R-value of
R-10, representative of an older,
poorly insulated attic.
The same-size air leak is introduced in
each case, calibrated from the previous
example. The impact of attic ventilation rate
on the RH at the plywood surface and MC of
the plywood for a typical winter season is
shown in Figure 52 and Figure 53, compared
to a baseline attic with no air leakage.
The total hours per year that the plywood
sheathing surface RH exceeds 80% is
shown in Figure 54. An attic insulation Rvalue
of R-40 is also compared to show the
increasing sensitivity with increasing Rvalue
in the same figure.
Analysis Comments
These results demonstrate the positive
impact that attic ventilation has on removing
excess moisture from an attic where air
leakage is able to occur. While less ventilation
appears to be more beneficial under
normal conditions and reduces the potential
for mold growth on the sheathing,
where air leakage can occur, the need for
greater ventilation is shown. However, it
should be noted that even very high ventilation
rates cannot sufficiently remove
enough moisture so that elevated RH levels
and moisture contents can be reduced. The
importance of controlling air leakage is
demonstrated, as well as the need for airtight
ceilings and airtight and properly terminated
ducts.
GENERAL CONCLUSIONS AND
COMMENTARY
Based on the site observations documented
in the case studies and the results
of computer modeling, we draw the conclusions
listed below. However, we must add
that the number of variables and their
interaction is complex, and further research
is required before final conclusions can be
drawn. Also, it should be noted that these
comments are for the climate experienced
in the Pacific Northwest; significantly different
climate zones could result in different
conclusions.
• There is an increased risk of moisture
damage in vented roof assemblies
with higher insulation levels
due to cooler surface temperatures
that are created on exterior components.
Moisture problems will likely
become worse as insulation levels
increase with upcoming building
code changes unless other changes
are made. Specifically within British
Columbia, there is a mandate
toward net-zero building enclosures
by 2020. Anticipated changes to the
BC Building Code in 2012 will
increase attic insulation levels to
greater than R-50.
• Increasing the quantity of insulation
in attics increases the significance of
air leakage from the occupied por-
Figure 52 – RH at plywood sheathing; impact of ventilation rate with air leakage.
Figure 53 – MC of plywood sheathing; impact of ventilation rate with air leakage.
Figure 54 –
Annual hours of
80% RH at plywood
sheathing;
impact of attic
ventilation rate
with air leakage
(R-10 and R-40).
1 0 4 • L E V Y S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1
tions of the building and, more significantly,
air leakage from mechanical
equipment such as bathroom
and clothes dryer exhausts.
• It is critical that all mechanical ventilation
is continuous to the outside
of the attic, and there is no opportunity
for warm, moist air to short-circuit
back into the attic. As with the
case studies, exhaust ducts should
not discharge into areas with perforated
soffits, and exhaust ducts
must be well air-sealed into vents
that penetrate the roof sheathing.
Even relatively small quantities of
air leakage from mechanical systems
can have a negative impact on
the performance of attics.
• If air leakage into the attic from
occupied space and mechanical systems
is controlled, there is minimal
negative effect of increasing the Rvalue
above R-40.
• If there is no air leakage from occupied
space or from mechanical
equipment to warm the attic air
temperature, increased ventilation
may actually lead to increased moisture
content in the sheathing. This
occurs because the exterior air that
is brought into the attic deposits
moisture that may condense on cold
surfaces, specifically during nightsky
radiant cooling.
• Care must be taken to ensure that
the cavity from rainscreen wall
assemblies is not vented
into attics, particularly
if there is a large
moisture source, such
as a rainwater leader,
at the base of the rainscreen
cavity.
Additional Thoughts
One option for reducing the risk of air
leakage from occupied spaces into attics is
the use of spray-in-place low-density foam
insulation (½ pcf) as a base layer prior to
installation of blown insulation above
(Figure 55 through 57) in a flash-and-fill
attic assembly. In this approach, the polyethylene
is not relied on for airtightness,
nor is the electrical or HVAC contractor
required to ensure continuity of airtightness.
The spray foam is applied after all
ceiling work is complete and provides continuity
of airtightness but still requires the
use of a vapor barrier because low-density
spray-in-place foam is relatively vaporopen.
In addition to making the ceiling airtight,
mechanical exhaust ducts should run
vertically through the attic (i.e., up through
the roof sheathing), and it is imperative that
all joints/connections are fully sealed and
the full length of the duct insulated.
The impact of improved ceiling airtightness
at current R-40 insulation levels is
Figure 55 – Flash-and-fill attic assembly (photo by Murray Frank,
2011).
Figure 56 – Low-density spray-foam
application (photo by Murray Frank,
2011).
Figure 57 – Blown insulation over spray foam.
Figure 58 – Plywood moisture content in a typical year.
S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1 L E V Y • 1 0 5
compared to current practice (as based on
the case studies, where air leakage into the
attic occurs at penetrations and ducts, etc.)
and to an historically air-leaky older attic
with minimal insulation (R-10) using the
WUFI model. Figure 58 plots the moisture
content of the sheathing for a typical year,
and Figure 59 plots the RH at the underside
of the sheathing. The graphics show the relative
difference in the order of magnitude of
MC and RH as a result of changing insulation
and the importance of addressing air
leakage.
Although not addressed in this paper,
another consideration is the use of warmattic
technology. This option should consider
the need for ventilation between the shingles
and the roof sheathing (plywood).
REFERENCES
Graham Finch, The Performance of
Rainscreen Walls in British
Columbia, University of Waterloo,
Ontario, Canada. 2007.
Morrison Hershfield Ltd., “Survey of
Building Envelope Failures in the
Coastal Climate of British
Columbia,” Canadian Mortgage and
Housing Corporation, Burnaby, BC.,
1996.
RDH Building Engineering Ltd., “Wood
Frame Envelopes in the Coastal
Climate of British Columbia,”
Canadian Mortgage and Housing
Corporation, Vancouver, BC, 1999.
I.S. Walker, “Attic Ventilation and
Moisture,” Canada Mortgage and
Housing Corporation, Ottawa, 1997.
Figure 59 – RH at interior surface of sheathing in a typical year.
1 0 6 • L E V Y S Y M P O S I U M O N B U I L D I N G E N V E L O P E T E C H N O L O G Y • OC T O B E R 2 0 1 1