Wind Design – Not Just Another SPRI Wind/ES Presentation

Wind Design – Not Just Another SPRI Wind/ES Presentation

Mike Ennis, RRC, CDT
SPRI
Reynoldsburg, Ohio
Bob LeClare
W.P. Hickman
Asheville, North Carolina
ABSTRACT
Roofing systems are one of the most commonly damaged portions of the building
envelope during high wind events. For this reason, additional emphasis has been
placed in the International Building Code on designing low-slope commercial roofing
systems to reduce damage during high wind events.
Reducing damage to roof systems during high wind events requires a full system
approach, including proper design of the field, perimeter, corner, and edge metal of
the roofing system. This presentation will summarize:
1. Lessons learned from post hurricane inspections conducted by the Roofing
Industry Committee on Weather Issues (RICOWI).
2. SPRI standards that can be used to design the roofing system to be in com¬
pliance with the requirements of the International Building Code, including:
• SPRI WD-1 – Wind Design Standard for Low Slope Roofing Systems
• ANSI/ SPRI RP-4 – Wind Design Standard for Ballasted Single-Ply Roofing
Systems
• ANSI/SPRI ES-1 – American National Standard Wind Design Standard for
Edge Systems Used with Low Slope Roofing Systems
SPEAKERS
Mike Ennis is a graduate of Ohio State University with a degree in environmental sci¬
ences. He worked for The Dow Chemical Company in various research and product
development capacities for 30 years and was the North American Application
Technology Leader for commercial products in Dow’s Building Solutions business
unit. Mike is past president of SPRI, Inc., the association representing flexible sheet
roofing membrane manufacturers and component suppliers, and is currently its
technical director. Mr. Ennis is a member of RCI and is a Registered Roof Consultant
and is a member of the Construction Specifications Institute and has received his
Construction Documents Technologist certification. He is also a member of the
International Concrete Repair Institute (ICRI), The American Architectural
Manufacturers Association (AAMA), the Sealant Waterproofing and Restoration
Institute (SWRI), the National Roofing Contractors Association (NRCA), and the
Western Construction Consultants Association (WESTCON).
Contact Information: Phone – 614-501-8909; E-mail – m.ennis@mac.com
Bob LeClare is vice president of sales for the WP Hickman Company, a position he
has held since joining the company in 2002. The WP Hickman Company has been in
business for over 60 years, developed the first pre-manufactured roof edge system,
has patented over 25 products, and is considered an industry leader in the metal roof
edge market. Bob has a bachelor’s degree from Purdue University and has 25 years
of experience in the architectural metals industry. He has experience in multiple
areas of the industry, including engineering, fabrication, installation, and sales of
architectural metals and commercial roofing products. Bob chairs SPRI’s ES-1 task
force, is a member of RCI and CSI and has received his CDT certification.
Contact Information: Phone – 828-274-4000; E-mail – bob@wph.com
Ennis and LeClare – 86 Proceedings of the RCI 23rd International Convention

Wind Design – Not Just Another SPRI Wind/ES Presentation

INTRODUCTION
Combined insurance losses
from Hurricanes Katrina and Rita
exceeding $60 billion, by far the
most expensive natural disasters
on record for the U.S. (Greenberg
Traurig Alert, October 2005).
Factory Mutual Global reports
that roofing claims related to wind
damage to commercial roofing
systems account for $210 million/
year in insurance losses,
10% of the total losses. (FM
Global, “Protecting Roofing Sys¬
tems Against Windstorm Dam¬
age.”) Proper design and installa¬
tion of roofing systems will help
mitigate these losses and the
associated losses from business
disruption.
The International Building
Code (IBC) contains requirements
for roofing systems that must be
met to resist wind uplift forces.
Investigations conducted after
high wind events have determined
that in many instances, the roof
system was not installed in accor¬
dance with the applicable code,
whether due to deficient design or
installation. Problems were also
noted with a lack of maintenance.
If the requirements of the code
had been met, losses due to high
wind events could have been dra¬
matically reduced.
This paper discusses stan¬
dards and recommendations de¬
veloped by SPRI, the trade associ¬
ation representing the single-ply
roofing industry, that meet the
requirements of the International
Building Code for wind resistance
of roofs.
BALLASTED SINGLE-PLY
ROOF SYSTEMS
Ballasted single-ply roofing
systems have been in use since
the 1970s. Designing these sys¬
tems to resist wind load forces
was one of the initial concerns
with ballasted systems. To
address this concern, the roofing
industry conducted wind tunnel
and full-scale mockup testing of
ballasted systems. In addition to
the testing, field observations
were performed by manufacturers
and consultants to develop addi¬
tional data on the wind perfor¬
mance of these systems. Based on
this information, a Standard
Design Guide was developed.
Much of the information that
was used as the basis for the
development of the Standard
Design Guide was presented at
the Second International Sympo¬
sium on Roofing Technology held
in 1985.
The following papers were pre¬
sented at that symposium regard¬
ing ballasted roof systems. The
information from this work is the
basis for much of the information
presented in this paper:
• “Wind Design Guide for
Ballasted Roofing Sys¬
tems,” Richard J. Gillen¬
water
• “Wind Tunnel Tests on
Loose-Laid Roofing Sys¬
tems for Flat Roofs,” R.J.
Kind
• “Stone Ballast Design
Criteria on Loose-Laid
Single-Ply Ballasted Roofs
for Wind Speed, Size and
Weight,” Thomas E. Pha¬
len Jr.
• “A Study of the Behavior of
Loose-Laid Ballasted
Single-Ply Roofing Sys¬
tems Subjected to Violent
Winds,” Kenneth G.
Schneider Jr.
The IBC requires that ballast¬
ed single-ply roof systems be
installed in accordance with
ANSI/ SPRI Standard RP-4, “Wind
Design Standard For Ballasted
Single-ply Roofing Systems.”
ANSI/ SPRI RP-4 was first includ¬
ed in the regional building codes
(BOCA, SBCCI, IBCO) starting in
1986. It is imperative that stoneballasted
roof systems be de¬
signed in accordance with this
standard design guide to assure
that stone blow-off will not occur.
Ballasted single-ply roof systems
have performed very well during
high wind events when designed
in accordance with the ANSI/ SPRI
standard, and even in some
instances when they have not.
As an example, Figure 1 shows
a stone-ballasted EPDM roof in
Mississippi City, Mississippi, after
Hurricane Katrina that did not
appear to meet the requirements
of ANSI/SPRI RP-4. The estimated
wind speeds at this location were
120 to 130 mph. As can be seen,
the roof system performed very
well, with just a few areas of local¬
ized stone scouring as shown in
Figure 2. This is due to high wind
loads in this area from wind flow¬
ing around roof-top equipment. In
this case, the wind came from the
backside of the equipment and
caused wind scouring on the
opposite or near side in this
photo. Methods of addressing
stone scour are addressed in the
Proceedings of the RCI 23rd International Convention Ennis and LeClare – 87
Figure 2 – Example of stone scour that can
occur in areas of localized high wind loads
(photo courtesy of RICOWI Inc.).
Figure 1 – Typical roof area at this location
(photo courtesy of RICOWI Inc.).
. The investigators on this roof did
not find ballast stone that had
blown off the roof.
Even though in this example
stone blow-off was not observed,
SPRI’s position is that all ballast¬
ed roof systems should be
installed in strict adherence to the
code-mandated ANSI/SPRI RP-4
standard. Not doing so can lead to
stone blow-off, providing debris
that has been observed to cause
collateral damage to surrounding
buildings and vehicles (see FEMA
549).
Section 1504.4 of the 2007
Supplement to the International
Building Code (the most current
version of the code) states,
“Ballasted low-slope (roof slope
<2:12) single-ply roof system cov¬
erings installed in accordance
with Section 1507.12 and
1507.13 shall be designed in
accordance with Section 1504.8
and ANSI/SPRI RP-4. This
Section provides a direct reference
to the ANSI/SPRI Standard in the
International Building Code.”
Section 1504.8 describes re¬
quirements for aggregate surfac¬
ing materials, while the
ANSI/SPRI standard describes
ballasting requirements for large
stones.
The standard contains five
sections that will be discussed in
detail in the portion of this report
that details how to use the stan¬
dard:
General design considerations
This section contains defini¬
tions and information that applies
to designing the roof system to
resist wind loads for any type of
system.
System requirements
This section contains require¬
ments for single-ply membranes
to be used in ballasted systems,
along with requirements for the
various types of ballast that can
be used.
Design options
This section describes the
methods for installing Systems 1,
2, and 3 as called out in the
design tables. The design provi¬
sions become more resistant to
wind loads as the number in¬
creases. This section also de¬
scribes how Protected Membrane
Roof systems should be ballasted.
Design provisions
This section describes how to
handle special considerations for
example large openings in the wall
and eaves and overhangs.
Design tables
This section contains tables
that allow the user to determine
which system design will be
required, based on the design
wind speed, building height, para¬
pet height, and exposure catego¬
ry.
It is important to note that
this is a very conservative stan¬
dard. The following conservative
approaches were taken in devel¬
oping the requirements included
in this standard:
• The ballast design tables
have been developed so
that the ballast will not
blow off the roof at the de¬
sign wind speed. There
has been some concern
expressed with the grada¬
tion that occurs within a
specified stone type in
ASTM D448, Standard
Classification for Sizes of
Aggregate for Road and
Bridge Construction,
which is the standard ref¬
erenced for ballast stone
size in the RP-4 standard.
For example, ASTM D448
Type 4 stone is nominal
1-1/2 inches in diameter;
however, it can range from
greater than 3/8 inch to
less than 2 inches in dia-
Ennis and LeCIare – 88 Proceedings of the RCI 23rd International Convention
meter. It has been the ex¬
perience of SPRI manufac¬
turers that have investi¬
gated the performance of
these systems that the
smaller stones migrate to
the bottom and are not
available to become windborne
debris.
• This standard is based on
having no deliberately in¬
stalled air retarders for all
systems with 10 Ibs/sq ft
or more of ballast weight.
This was done because it
is recognized that the
weight of stone or other
ballast may not always be
adequate to resist uplift
loads that result from
some internal or other
under membrane pres¬
sures. Therefore, the
worse-case scenario was
considered in the design of
this standard.
• For lighter weight systems,
air retarders are required,
but this standard assumes
the air retarder is imper¬
fect. The standard in
eludes a discussion on
where air retarders may be
required.
A few examples of restrictions
placed on the use of ballasted roof
systems included in the ANSI/
SPRI standard are:
• When the maximum build
ing height exceeds 150
feet, the roof design shall
be based on an expert’s
design method and ap¬
proved by the authority
having jurisdiction.
• When the maximum wind
speed exceeds 140 miles
per hour, the roof design
shall be based on an ex¬
pert’s design method and
approved by the authority
having jurisdiction.
• In areas designated as
windborne debris regions,
ballast designs using
stone ballast shall use a
minimum nominal stone
diameter of 2-1/2 inches.
• In hurricane-prone re¬
gions, buildings exceeding
60 feet in height shall not
use stone ballast in the
corners and perimeters
unless the parapet height
exceeds 36 inches.
In the above restrictions, the
use of expert design is required if
the building height is above 150
feet or if the design wind speed is
above 140 mph. In doing the
expert design analysis, the key
factor is determining the antici¬
pated wind speed at the roof sur¬
face. This will be significantly
impacted by the parapet height.
Once this is determined, an excel¬
lent reference is NRC Report
Number 15544. Design of Rooftops
Against Gravel Blow-Off This
report provides an analysis of
wind tunnel testing that was con¬
ducted to evaluate the critical
speeds at which stone ballast
would begin to move. Data from
this report can be compared to
the rooftop wind speed and be
used as the basis of a rational
design.
HOW TO USE THE ANSI/ SPRI
RP-4 STANDARD
When considering the use of a
ballasted single-ply roof system,
the designer must first verify that
the roof structure and deck will
support the ballast load in combi¬
nation with all other design loads.
A licensed architect or an engi¬
neer should make this determina¬
tion.
Once the structure has been
determined to be adequate, the
following variables must be deter¬
mined in order to identify the
proper way to ballast the system.
Wind Speed
The wind speed used in the
standard is the Basic Wind Speed
as provided in the ANSI/ASCE 7-
2005 standard or the local au¬
thority having jurisdiction when
local values exceed ASCE 7-2005.
This is the 3-second gust speed at
33 ft (10 m) above the ground in
Exposure C. The intensifying
effects of abrupt or unique topo¬
graphical features need to be
accounted for in the design (See
ASCE-7). Both the Commentary of
the standard and sections within
the standard address how this
should be accomplished.
Building Height
The building height is mea¬
sured from ground level to the
roof system surface at the roof
edge. If multiple roof levels are
present, each one must be de¬
signed separately.
Edge Condition
If a gravel stop is used at the
building perimeter, the top edge of
the flashing must be at least 2 in
above the top surface of the mem¬
brane and higher than the top of
the ballast.
If the edge of the roof uses a
parapet, the height of the parapet
is the distance from the top of the
roof system membrane to the top
of the parapet for conventional
ballasted systems (roof deck,
loose-laid insulation, loose-laid
membrane, ballast). For Protected
Membrane Systems (roof deck,
loose-laid or adhered membrane,
loose-laid insulation, fabric, and
ballast), the parapet height is the
distance from the top of the insu¬
lation to the top of the parapet.
If the edge of the roof consists
of a parapet of variable height,
special conditions may influence
the measurement of parapet
height. The standard defines how
to calculate parapet height in
these situations.
Building exposure
The terrain surrounding the
building will influence the degree
of exposure of the building to the
wind. The building is classified as
Proceedings of the RCI 23rd International Convention Ennis and LeClare – 89
either protected or unprotected.
Protected exposures
Surface Roughness B: Ur¬
ban and suburban areas,
wooded areas, or other ter¬
rain with numerous closely
spaced obstructions having
the size of single-family
dwellings or larger. Use of
this exposure category shall
be limited to those areas for
which terrain representative
of Exposure B prevails in the
upwind direction for a dis¬
tance of at least 2,600 ft
(800 m) or 20 times the
height of the building,
whichever is greater.
Unprotected exposures
Surface Roughness C:
Open terrain with scattered
obstructions having heights
generally less than 30 ft (9.1
m). This category includes
flat open country, grass¬
lands and all water surfaces
in hurricane-prone regions.
Exposure C shall apply for
all cases where exposures B
or D do not apply.
Surface Roughness D: Flat,
unobstructed areas and
water surfaces outside hur¬
ricane-prone regions. This
category includes smooth
mud flats, salt flats, and
unbroken ice. Use of this
exposure category shall be
limited to those areas for
which terrain representative
of Exposure D prevails in
the upwind direction for a
distance of at least 5,000 ft
(1524 m) or 20 times the
height of the building,
whichever is greater.
In most instances, the unpro¬
tected exposure will be used.
DETERMINING BALLAST
DESIGN
You now have the basic infor¬
mation necessary to determine
BLDG. SYSTEM 1 SYSTEM 2 SYSTEM 3
Ht. Ft Exp.C Exp. B Exp. C Exp. B Exp. C Exp. B
0-15 100 105 115 115 130 140
>15-30 100 105 110 115 130 140
>30-45 90 100 100 115 130 140
>45-60 NO NO 95 115 120 140
>60-75 NO NO 90 110 120 120
>75-90 NO NO NO NO NO NO
>90-105 NO NO NO NO NO NO
>105-120 NO NO NO NO NO NO
>120-135 NO NO NO NO NO NO
>135-150 NO NO NO NO NO NO
Table 1-A – From 2-in high gravel stop to less than 6-in high
parapet maximum allowable wind speed (mph).
BLDG. SYSTEM 1 SYSTEM 2 SYSTEM 3
Ht. Ft Exp.C Exp. B Exp. C Exp. B Exp. C Exp. B
0-15 100 105 115 115 140 140
>15-30 100 105 110 115 140 140
>30-45 90 105 105 115 140 140
>45-60 NO 90 95 115 130 140
>60-75 NO 90 90 110 120 130
>75-90 NO NO 90 110 110 120
>90-105 NO NO 90 100 110 110
>105-120 NO NO 85 100 100 110
>120-135 NO NO NO 100 100 110
>135-150 NO NO NO 95 100 110
TABLE 1-C – For parapet heights from 12 to less than 18
inches maximum wind speed (mph).
BLDG. SYSTEM 1 SYSTEM 2 SYSTEM 3
Ht. Ft Exp.C Exp. B Exp. C Exp. B Exp. C Exp. B
0-15 110 110 120 120 140 140
>15-30 110 110 120 120 140 140
>30-45 100 110 120 120 140 140
>45-60 95 110 105 120 140 140
>60-75 90 100 100 120 140 140
>75-90 90 100 100 120 140 140
>90-105 90 90 100 110 130 140
>105-120 85 90 100 110 130 140
>120-135 85 90 100 110 130 140
>135-150 NO 85 100 110 130 140
TABLE 1-F – For parapet heights from 36 to less than 72
inches maximum wind speed (mph).
Ennis and LeClarc – 90 Proceedings of the RCI 23rd International Convention
the required ballast design. Table
1 provides an example of a ballast
design table.
The Design Tables cover bal¬
last designs for various parapet
heights. This example contains
designs for some of the heights.
The standard actually contains
designs for the following parapet
heights:
Part A: From 2-in gravel
stop to less than 6-in-high
parapet
Part B: From 6-in gravel
stop to less than 12 -in-high
parapet
Part C: From 22-in gravel
stop to less than 18-in-high
parapet
Part D: From 18-in gravel
stop to less than 24-in-high
parapet
Part E: 24-in gravel stop to
less than 36-in-high parapet
Part F: 36-in gravel stop to
less than 72-in-high parapet
Part G: 72-in gravel stop
and above
The Design Tables also refer¬
ences System 1, System 2, and
System 3. These are references to
different ballasting schemes. The
resistance to wind loads increases
as the System number increases.
The designs are as follows:
System 1
System 1 requires that the
roof covering be ballasted with
nominal 1-1/2-inch smooth,
river-bottom stone of ballast gra¬
dation size #4, or alternatively,
#3, #24, #2, or #1 as specified in
ASTM D-448, “Standard Sizes of
Coarse Aggregate” spread at a
minimum rate of 1,000 pounds
per 100 square feet; standard
concrete pavers (minimum 18
psf); or interlocking, beveled, dow¬
eled, or contoured fit lightweight
concrete pavers (minimum 10
psf).
System 2
System 2 requires that the
field of the roof be ballasted in the
same manner as was used in
System 1 and the perimeters and
corners are ballasted with nomi¬
nal 2-1/2-inch smooth, river-bot¬
tom stone of ballast gradation size
#2, or alternatively #1, as speci¬
fied in ASTM D 448, “Standard
Sizes of Coarse Aggregate” spread
at a minimum rate of 1,300
pounds per 100 square feet; con¬
crete pavers (minimum 22 psf); or
approved interlocking, beveled,
doweled or contoured fit; light¬
weight concrete pavers (minimum
10 psf) when documented or
demonstrated as equivalent.
The perimeter is defined as
the rectangular roof section paral¬
lel to the roof edge and connecting
the corner areas with a width
measurement equal to 40% of the
building height, but no less than
8.5 feet.
The corner is defined as the
space between intersecting walls
forming an angle greater than 45
degrees but less than 135
degrees. The corner area is
defined as the roof section with
sides equal to 40% of the building
height. The minimum length of a
side is 8.5 feet.
Unlike ASCE-7, which allows
either 40% of the building height
or 10% of the building width,
whichever is less, to determine
the perimeter and corner areas,
RP-4 requires the use of 40% of
the building height. For tall build¬
ings, this results in very large
perimeter and corner areas.
System 3
System 3 is the most stringent
design and is required in areas
where the design wind speed
exceeds 120 mph. For this design,
the field of the roof is ballasted in
the same manner as the perime¬
ters and corners are in the
System 2 design. The corner and
perimeter areas must use either a
mechanically attached or adhered
system that is designed to with¬
stand the uplift force in accor¬
dance with ANSI/ASCE 7-2005 or
the local building code. No loose
stone can be used in these areas.
If a protective covering is re¬
quired, then a fully adhered mem¬
brane system must be used and
covered with minimum 22 psf
pavers or other material approved
by the authority having jurisdic¬
tion.
At the junction of the looselaid
roof membrane with the
adhered or mechanically attached
membrane areas, a mechanical
termination providing a minimum
100 pounds per linear foot hold¬
ing power must be provided.
Practical example of determin¬
ing the ballast requirements
Example building:
Building height – 60 ft
Edge condition – gravel stop
Design Wind Speed – 90 mph
Exposure – unprotected (C)
To determine the ballast
requirements for this building,
look in Table 1, Part A, from 2-
inch high gravel stops to less than
6-inch high parapets. Find the
appropriate building height in
Column 1. Look across the row
until the required design wind
speed is located. Once it is locat¬
ed, make sure it matches the
appropriate exposure categoiy. In
this case, Ballast System 1 cannot
be used; you must use Ballast
System 2. The maximum building
height for these conditions that
would allow for the use of Ballast
System 1 is 45 ft.
The importance of parapet
height can also be seen. Ballast
System 1 could not be used on
this building unless there were a
parapet that was at least 36 inch¬
es high (Table 1, Part F).
The importance of correctly
identifying the exposure category
Proceedings of the RCI 23rd International Convention Ennis and LeClare – 91
can also be observed. In Table 1,
Part C, Ballast System 1 could be
used if the parapet height was
increased to 12 inches and the
exposure category was changed to
protected (B). If the exposure cat¬
egory remained as unprotected,
System 1 could not be used.
Conditions that impact ballast¬
ing requirements
There are a number of condi¬
tions that will influence the
required ballast loading. These
conditions and the action that
should be taken are summarized
in Table 2.
The RP-4 standard is available
free of charge from the SPRI web¬
site, www.spri.org.
MECHANICALLY ATTACHED
AND ADHERED SINGLE-PLY
ROOF SYSTEMS
Both mechanically attached
and adhered single-ply roofing
assemblies have performed well in
high-wind events (see Figures 3
and 4).
Figure 3 shows a mechanically
attached roof after exposure to
Hurricane Katrina. The building is
located in Bay St. Louis, Missis¬
sippi, and was exposed to wind
speeds of 120 to 130 mph. Figure
4 shows an adhered single-ply
roof assembly after exposure to
Hurricane Ivan. The building is
located in Escambia County,
Florida, and was exposed to wind
speeds of 110 to 120 mph.
RICOWI, Inc. has conducted
field investigations on three hurri¬
canes: Charley, Ivan, and Katrina.
Reports from these investigations
are available on the RICOWI Web
site, www.ricowi.com. These in¬
vestigations have found that ad¬
hered and mechanically attached
single-ply membrane systems can
be installed to perform well in
high wind events. However, in
some instances, unsatisfactory
performance was observed. In
these situations, the unsatisfacto-
1. – The RP-4 Standard provides definitions for each of these conditions.
2. – Importance Factor
Condition1 Action
Large openings in a wall Roof area above the opening must be
designed as a corner area of the re¬
spective System 2 or System 3 designs.
For System 1 designs, use the corner
area specifications of a System 2 design.
Positive pressure in
building between 0.5 and
1 inch of water
Increase the roof-top wind speed by 20
mph from the basic wind speed from
the wind map.
Rooftop projections (See
Figure 2 to see potential
issues with rooftop
projections.
The roof area that extends four feet out
from the base of such projections shall
have the same design as the corner
area of the roof.
Overhangs, eaves, and
canopies – pervious decks
The design of the entire overhang, eave,
or canopy area shall be upgraded to
the corner design of the next level sys¬
tem for wind resistance over the
applicable design. System 3 is still
designed to System 3.
Overhangs, eaves and
canopies – impervious
decks
Eaves and overhangs are designed as a
perimeter of the applicable design.
Canopies are designed as a corner sec¬
tion of the applicable design.
Exposure D Increase the roof-top wind speed by 20
mph from the basic wind speed from
the wind map.
Importance factor2 For buildings fitting category III or IV
(high importance), increase the roof¬
top wind speed by 20 mph from the
basic wind speed from the wind map.
Category I: Buildings that represent a low hazard to human life.
Category II: Buildings not covered by categories I, III or IV.
Category III: Buildings that represent a substantial hazard to human life.
Category IV: Buildings that are considered essential facilities.
ry performance was related to
deficiencies in either the design or
installation of the system; either
could have been the cause. In
addition to design or installation
issues, puncturing of the mem¬
brane was noted as a problem in
high-wind events. The punctures
were caused by flying debris, or in
some instances, rooftop equip¬
ment coming loose and rolling
across the surface of the roof.
Section 1504.3 of the 2007
supplement to the International
Building Code requires that roofs
be designed to resist wind loads
as determined by Chapter 6 of
ASCE-7. Once the appropriate
wind loads have been determined,
Section 1504.3.1 of the code
requires that an assembly tested
to resist the determined load be
used. Test results from a code¬
approved testing laboratory and
tested in accordance with ap-
Ennis and LeClare – 92 Proceedings of the RCI 23rd International Convention
design cor¬
ner and per¬
imeter loads.
If the first
layer is ad¬
hered, then
the adhesive
bead spac¬
ing is de¬
creased. The
enhance¬
ments are
based on the
known hold¬
ing power of
the mechan-
Figure 3 – Mechanically fastened thermoplastic
single-ply membrane after Hurricane Katrina.
(Photo courtesy of RICOWI Inc.)
proved methods may be used to
demonstrate compliance with the
code. Approved methods are FM
4450, FM 4470, UL 580 or FM
1897.
ical fastener
or adhesive.
SPRI has
proposed a
code change for the 2007/2008
IBC Code cycle to include this new
standard in the IBC.
the Quick Reference tables pro¬
vided in this Standard Practice or
by calculating these values follow¬
ing the requirements of the cur¬
rent version of the ASCE 7
Standard, Minimum Design Loads
for Buildings and Other Struc¬
tures. The Quick Reference Tables
are based on ASCE 7-05 and can
only be used if a particular build¬
ing meets the criteria identified in
the standard.
Second Part – Select an
appropriate roofing system as¬
sembly by comparing the tested
wind uplift resistance capacity to
the calculated design loads. It is
strongly recommended that a
safety factor be applied to the
tested wind uplift resistance be¬
fore comparison to the design
pressures. A 2 to 1 safety factor is
commonly used. The standard
refers to the “factored load.” The
Some states have developed
requirements that must be met
when designing roof assemblies.
For example, the Florida has a
code that has been specifically
developed for high-velocity hurri¬
cane zones.
SPRI has recently developed a
national consensus standard,
ANSI/ SPRI WD-1, “Wind Design
Standard Practice for Roofing
Assemblies.” This wind design
standard allows the user to deter¬
mine wind loads through a series
of easy-to-read tables that have
been calculated using Chapter 6
of ASCE-7, thus meeting the
requirements of the International
Building Code. This standard
practice also provides installation
guidelines to enhance the attach¬
ment of the roofing system at the
perimeters and corners where
wind loads are higher. The pre¬
scriptive enhancements vary
based on the attachments method
used for the first layer. If the first
layer is mechanically attached,
then the standard provides a
method to calculate the increase
in fasteners required to resist the
factored load is:
Factored Tested Load
Capacity = tested uplift capacity
(Lt) / safety factor, psf
HOW TO USE ANSI/SPRI
STANDARD WD-1
ANSI/SPRI WD-1 consists of
three primary sections.
General Design Considerations
and Definitions
Wind loads are higher at the
perimeter and corner areas of the
The information in this sec¬
tion is consistent with the same
type of information provided in
ANSI/SPRI RProofing
assembly. For this reason,
enhanced attachment is necessaiy
in these areas. The standard
provides a method to extrapolate
4 and was cov¬
ered earlier in
this report and
will not be re¬
peated here.
Two-part
Methodology
First Part
– Calculate the
wind uplift de¬
sign loads for
the field, peri¬
meter, and cor¬
ner areas of a
building.
This is ac¬
complished by Figure 4 – Adhered single-ply membrane after
either using Hurricane Ivan. (Photo courtesy of RICOWI Inc.)
Proceedings of the RCI 2 3rd Intentational Convention Ennis and LeClare – 93
the field of roof rating to the
perimeter and corner regions,
assuming certain conditions are
met. These conditions and meth¬
ods for various types of attach¬
ment are:
ADHERED SYSTEM
ASSEMBLIES
The adhered roofing system
assembly extrapolation method is
only applicable when all of the fol¬
lowing criteria are met:
1. The adhered roofing sys¬
tem assembly utilizes
either mechanical fasten¬
ers or ribbons/beads of an
adhesive for insulation
attachment, and
2. The tested wind uplift load
capacity of the proposed
adhered roofing system
assembly was determined
utilizing a test chamber of
sufficient size to allow
side-by-side positioning of
a minimum of three fullsize
insulation/ cover
board/ substrate boards/
panels on the test frame,
and
3. The calculated field design
load does not exceed 53
psf.
Extrapolation for adhered
roofing system assemblies is not
possible when the insulation
layer(s) is (are) attached using a
100% coverage rate of an adhe¬
sive.
Mechanically Attached
Insulation
For insulation attached with
mechanical fasteners, determine
the increased number of fasteners
per insulation board (Fn) needed
to meet the calculated design
load(s) using the following equa¬
tion: Fn = (Ft x Ld) / Lt
Where:
Fn is the number of fasten¬
ers per board needed to
meet the design load.
Ft is the number of fasten¬
ers per board used to
achieve the tested load
capacity.
Ld is the calculated design
load for the perimeter or cor¬
ner area of a roof, psf.
Lt is the factored tested load
capacity, psf.
Ribbon/Bead Adhesive-
Attached Insulation
For insulation attached with
ribbons/beads of adhesive, deter¬
mine the reduced ribbon/bead
spacing (Rn) needed to meet the
calculated design load(s) using
the following equation:
Rn = Rt /(Ld / Lt)
Where:
R n is the ribbon/bead spac¬
ing needed to meet the
design load, inches.
Rt is the ribbon/bead spac¬
ing used to achieve the test¬
ed load capacity, inches.
Ld is the calculated design
load for the perimeter or cor¬
ner area of a roof, psf.
Lt is the factored tested load
capacity, psf.
Note: When ribbon/bead-attached
insulation is applied di¬
rectly to a fluted steel deck, the
ribbon/bead spacing will be dic¬
tated by the center-to-center
spacing of the top (high) flutes of
the steel deck. The extrapolated
ribbon/bead spacing must be
rounded down (when necessary)
to coincide with a top (high) flute.
If the extrapolated ribbon/bead
spacing is less than the center-tocenter
spacing of the top (high)
flutes of a steel deck, ribbon/bead
attachment of the insulation in
that area is not acceptable.
The Fn and Rn equations can
only be used to increase the num¬
ber of fasteners or decrease the
spacing of ribbons/beads of adhe¬
sive needed in the corner and
perimeter areas. These equations
cannot be used to extrapolate
backwards and reduce the num¬
ber of fasteners or increase the
spacing of ribbons/beads of adhe¬
sive used in the field of the roof.
EXTRAPOLATION METHOD –
MECHANICALLY FASTENED
SYSTEM ASSEMBLIES
The mechanically fastened
roofing system assembly extrapo¬
lation method is only applicable
when the following criteria are
met:
1. The tested wind uplift load
capacity of the proposed
linearly-attached (rows),
mechanically fastened
roofing system assembly
was determined utilizing a
test chamber of sufficient
size to allow positioning of
a minimum of three
attachment rows on the
test frame. The minimum
frame width shall be 8
feet.
2. The tested wind uplift load
capacity of the proposed
spot-attached, mechani¬
cally fastened roofing sys¬
tem assembly was deter¬
mined utilizing a test
chamber of sufficient size
to allow positioning of a
minimum of nine attach¬
ment locations on the test
frame. The minimum
frame width shall be 8
feet.
For mechanically fastened
system assemblies, first deter¬
mine the influence area per fas¬
tener for the tested assembly (IAt)
by multiplying the row spacing by
the fastener spacing (along the
row). For spot-attached systems,
multiply the distance between the
attachment locations in each
Ennis and LeClare – 94 Proceedings of the RCI 23rd International Convention
direction (2 ft x 2 ft, 2 ft x 3 ft,
etc.). This gives the number of
square feet of membrane held in
place by one fastener. Next, calcu¬
late the influence area needed to
meet the design load using the fol¬
lowing equation:
IAn = (Lt x IAt) / Ld
Where:
IAn is the area of membrane
needed to be held in place by
one fastener to meet the
design load, ft2.
IAt is the area of membrane
held in place by one fastener
for the tested assembly, ft2.
Ld is the calculated design
load for the perimeter or cor¬
ner area of a roof, psf.
L* is the factored tested load
capacity, psf.
The fastener row spacing or
the spot attachment grid spacing
of the assembly being evaluated
must be reduced so the ft2 of
membrane held in place by each
fastener does not exceed IAn. Use
the same fastener spacing (along
the row) as was tested.
For mechanically fastened
system assemblies with linear
(row) attachment, only the spac¬
ing between fastener rows can be
reduced to meet IAn. This extrap¬
olation method cannot be used to
reduce the spacing between fas¬
teners along the row (12 inches to
6 inches, for example) in place of
reducing the spacing between fas¬
tener rows. This extrapolation
method also cannot be used to
extrapolate backwards and in¬
crease the spacing between fas¬
teners along the row (12 inches to
18 inches, for example) or in¬
crease the spacing between fas¬
tener rows (8 feet to 10 feet, for
example) .
Quick Reference Tables
The Quick Reference Tables
have been developed using the
ASCE 7-05 Standard (Minimum
Design Loads For Buildings And
Other Structures). These tables
are applicable to buildings in
exposure categories B, C and D
when all of the following criteria
are met:
• The building is not situat¬
ed on a hill, ridge, or
escarpment.
• The building is Category
II. 1.
• The building is enclosed.
• The roof slope does not
exceed 1.5 inches per foot
(7 degrees).
PRACTICAL EXAMPLE FOR
USE OF THIS STANDARD
Example Building Criteria
A 40-ft high warehouse build¬
ing located outside of Pittsburgh,
PA, has a plan dimension of 200 ft
by 400 ft. The building has metal
roof deck with flutes spaced 6
inches on center. The walls have
no large openings. The roof slope
is 1/2 in per ft. The architect/
designer has selected a 2.0 safety
factor to be used for this project.
Task
Design a system that uses an
adhered membrane over insula¬
tion with mechanical fasteners.
First Part: Calculate the wind
uplift design loads for the field,
perimeter, and corner areas of the
building that will be used for all
three examples.
Step 1: Determine If the Quick
Reference Tables Contained in
This Document Can be Used:
• Building is Category II
(or Table 1-1 of ASCE 7-
05)
• Building is not situated on
a hill.
• Building is enclosed (from
Section 6.2 of ASCE 7-05).
• Building is in 90 mph
wind zone (from figure
6-1 of ASCE 7-05).
• Roof slope is <= 70 (1.5
in/ft).
All the conditions are met so
the Quick Reference Tables can
be used. If this were not the case,
the design loads would need to be
calculated in accordance with the
current ASCE 7 Standard. The
equations used to calculate the
design loads are referenced con¬
tained in the Standard.
Step 2: Determine Design
Loads Using the Quick
Reference Tables
Refer to Table 3.
Field Design Load = -25.5
Perimeter Design Load = -42.8
Corner Design Load = -64.4
The negative sign merely indi¬
cates that the uplift load is out¬
ward (away from the building).
The negative sign will be ignored
for calculation purposes.
Second Part: Select an appro¬
priate roofing system assembly by
comparing the tested wind uplift
resistance capacity of that assem¬
bly to the design loads.
System 1 – Adhered Roofing
System Assembly Selection Exam¬
ple for Mechanically Fastened In¬
sulation
The adhered membrane roof¬
ing system assembly being con¬
sidered for this building was test¬
ed on a 12 ft x 24 ft test chamber
to a maximum wind uplift resis¬
tance capacity of 90 psf (Lt) using
16 fasteners (Ft) per board. Apply
1 – When a building is classified as Category I, the Quick Reference Tables are usable if the field, perimeter, and corner design loads
are multiplied by 0.85. Likewise, when a building is classified as Category III or IV, the Quick Reference Tables are usable if the field,
perimeter, and corner design loads are multiplied by 1.15.
Proceedings of the RCI 23rd International Convention Ennis and LeClare – 95
the 2.0 safety factor to the 90 psf
tested value to determine the
Factored Tested Load Capacity:
Factored Tested Load Capacity =
Lt / 2.0 = 90 psf / 2.0 = 45 psf
The factored tested load capa¬
city (45 psf) exceeds the design
loads for both the field (25.5 psf)
and perimeter (42.8 psf) areas of
the roof but not the corner area
(64.4 psf). Consequently, the astested
assembly is acceptable for
use in the field and perimeter
areas. To determine if extrapola¬
tion is acceptable for the corner
areas, check the extrapolation
requirements of the Extrapolation
Method – Adhered System Assem¬
blies. Since all the extrapolation
method requirements are satis¬
fied, extrapolation is acceptable.
To determine the number of
fasteners (Fn) needed per insula¬
tion board for the corner areas of
the roof, use the equation:
Fn = (Ft x Ld) / Lt.
Where:
Fn is the number of fasten¬
ers per board needed to
meet the corner design
load.
F^. is the number of fasten¬
ers per board used to
achieve the tested load
capacity.
Ld is the calculated design
load for the corner area of
the roof, psf.
Lt is the factored testee
load capacity.
Corner Area
Fn = (16 fasteners x 64.4 psf) /
45 psf = 23 fasteners per board
The final design for this
assembly scenario is to use 16
fasteners per insulation board in
the field and perimeter areas and
23 fasteners per board in the cor-
Building
Height, ft.
Field Design
Load, psf
Perimeter
Design Load, psf
Corner Design
Load, psf
0 – 15 -20.8 -34.8 -52.4
20 -22.1 -37.0 -55.7
25 -23.0 -38.6 -58.1
30 -24.0 -40.2 -60.5
40 -25.5 -42.8 -64.4
50 -26.7 -44.7 -67.3
60 -27.6 -46.3 -69.7
70 -38.4 -60.3 -82.1
80 -39.7 -62.2 -84.8
90 -40.6 -63.7 -86.9
100 -41.2 -64.7 -88.2
120 -43.0 -67.5 -91.9
140 -44.6 -69.9 -95.3
160 -45.5 -71.4 -97.3
180 -46.9 -73.7 -100.4
200 -47.9 -75.1 -102.4
250 -50.1 -78.6 -107.1
300 -52.1 -81.8 -111.5
350 -53.7 -84.3 -114.9
400 -55.3 -86.8 -118.3
450 -56.7 -89.0 -121.3
500 -58.0 -91.0 -124.0
Table 3 – Building Category II, Exposure C – 90 mph peak gust
wind zone.
ner areas. The extra seven fasten¬
ers added to the corner areas
shall be evenly distributed around
the tested fastener layout pattern.
Fastening pattern examples for
insulation boards are included in
Appendix B of the Standard.
The WD-1 Standard is avail¬
able free of charge from the SPRI
Web site www.spri.org.
EDGE METAL
ATTACHMENT
In the RICOWI hurricane
investigations, the most common
source of low-slope roof system
damage was failure of the edge
metal system, resulting in expo¬
sure of the edge of the roofing sys¬
tem allowing for the membrane
and insulation to be peeled off the
roof. Figures 5 and 6 are pictures
of a modified bitumen roof system
in Pass Christian, MS, after
Hurricane Katrina. The roof was
exposed to wind speeds of 120 to
130 mph. It appears that the edge
metal system was lost and the
membrane then peeled off the
insulation.
The FEMA Mitigation Assess¬
ment Team also investigated this
roof (see FEMA 549). Its report
also concluded that the edge sys¬
tem was lost resulting in progres¬
sive roof failure. The FEMA team
determined that the edge system
was lost due to inadequate
attachment of the wood nailer.
To address the need for more
robust edge-metal attachment,
and the need for a standard pro¬
cedure for measuring the strength
of various attachment methods,
SPRI developed ES-1, “Wind De¬
sign Standard for Edge Systems
Used with Low-Slope Roofing
Systems.” Section 1504.5 of the
2007 supplement to the Inter¬
national Building Code requires
that the resistance of edge metal
systems be tested in accordance
Ennis and LeCIare – 96 Proceedings of the RCI 23rd International Convention
with this standard.
At the time of this writing,
ANSI/SPRI ES-1 was being routed
through the ANSI canvassing
process. The standard was updat¬
ed to combine ES-1 and Factory
Mutual Standard 4435. This will
result in the following changes:
• Title changes to SPRI ES-1
and FM 4435, and, pre
sumably, to ANSI/SPRI/
FM 4435 – ES-1.
• A 2.0 safety factor has
been added into the design
calculation.
• Tables have been added
listing the Field Design
Pressures for given build¬
ing heights and wind
speeds for each exposure
factor, thus reducing the
amount of calculations
required.
• The RE-1 test for Depend
ently Terminated Systems
will now be performed to
failure (previously, this
was a pass /fail at 100
Ib/ft). A table listing the
Membrane Tension Design
Load, based upon the field
design pressure and the
membrane fastener spac¬
ing, has been included.
• The angle of pull for the
RE-1 test has changed
from 45 degrees to 25
degrees.
• Nailer attachment has
been included and two ad¬
ditional tests have been
added:
• RE-4 to test fastener
pull out of substrate
• RE-5 to test fastener
pull through nailer
• Test loads shall increase
in 15 psf increments (pre¬
viously was 10 psf incre¬
ments).
HOW TO USE ANSI/SPRI ES-1
The ES-1 standard addresses
copings and horizontal roof edges
but does not
address gut¬
ters. It focuses
primarily on
design for wind
resistance;
however, it also
addresses cor¬
rosion and fas¬
cia thicknesses
that provide sat¬
isfactory flat¬
ness.
The stan¬
dard consists of
three test pro¬
cedures:
RE-1: Test for
Figure 5 – Edge metal attachment lost (photo
courtesy of RICOWI Inc.).
Roof Edge Ter¬
mination of
Ballasted or
Mechanically
Attached
Roofing Mem¬
brane Systems
This test is
designed to de¬
termine the
force required
to allow the
membrane to
come free of the
edge termina¬
tion, or for the
termination to
come free of its
Figure 6 – Membrane peeled off the insulation
(photo courtesy of RICOWI Inc.).
mount. It is re¬
quired for systems for which the
edge termination is expected to
secure the membrane. Figure 7
shows the set-up of the test appa¬
ratus.
A minimum 12-inch-wide
mock-up of the edge device sys¬
tem is evaluated. The jaws of the
test unit are clamped to the mem¬
brane and the load is applied until
either the membrane comes free
of the membrane termination or
the termination comes free of its
vide a minimum-load resistance
of 100 Ibs/ft.
For mechanically attached
assemblies, the required mini¬
mum-load resistance is a function
of the distance between the firstrow
membrane fasteners and the
perimeter edge, or the first row of
membrane fasteners parallel to
the edge in the corner regions,
with the requirement being
whichever load is greater. The fol¬
lowing equations are used:
mount. F = (D) (P) / 2
Performance Criteria
For ballasted systems, the
edge device assembly must pro-
F corner “ 1 ■ 5(D corner)(P)/2
Proceedings of the RCI 23rd International Convention Ennis and LeClare • 97
Where:
F = minimum load
resistance
D = distance between first
row of membrane fas¬
teners and roof edge (ft)
P = design pressure (psf)
Typically, the fastening rate is
increased in the corner, so the
equation for outside the corner
areas would be used.
RE-2 Pull-Off for Edge
Flashings Where Exposed
Horizontal Component is 4
Inches or Less
This test is designed to mea¬
sure the fascia blow-off resis¬
tance. The test specimen for this
evaluation is full size in width and
all other dimensions with a length
equal to the average length de¬
signed for use on the project, with
a minimum of 8 ft. If the mini¬
mum length designed for the pro¬
ject is less than 8 ft, then the
longest design length must be
used.
The load is applied incremen¬
tally at a point no greater than 12
inches to the centerline of vertical
face of the edge flashing (see
Figure 8). The load
is held for a mini¬
mum of 60 seconds
after stabilization
and then removed
until the specimen
stabilizes. The next
incremental load is
then applied. This
continues until
there is a loss of
attachment of any
component of the
roof edge system or
deformation that
would result in
loss of weather
protection of the
edge.
Fascia Blow-Off Test Set Schematic
(Force at Failure x Face Area =
Blowoff Resistance)
Test results
The maximum
Figure 8 – RE-2 Test set-up.
load (outward force)
is converted to
pressure using the following for¬
mula:
Pressure = Outward Force/ Face
Height x Face Length
Where:
2
Pressure is measured in Ibs/ft .
Force is measured in lbs force.
Face height is measured in ft.
Performance requirements
The test results must exceed
the design outward wind pres¬
sures for the building.
RE-3 Pull-Off Test for Copings
– Where Exposed Horizontal
Flange Depth Exceeds 4 Inches
This test is designed to determine
the force necessary to pull the
copings off the substrate. The
specimens for this test must be
full size in width and all other
dimensions using the same mate¬
rials, details, and methods of
construction and anchoring
devices as used on the actual
building. The specimen length is
the average length designed for
field use on the project and a
minimum of 8 ft unless the
longest length designed for the
project is less than 8 ft. In this
case, the longest design length
should be used.
The load is applied to the top
of the coping and to the face of the
coping simultaneously in the ratio
of (Face height x Horizontal Gcp)
to (Top Width x Vertical Gcp). (Gcp
is the gust coefficient described in
ASCE 7.) See Figure 9. Loads are
applied incrementally on centers
no greater than 12 inches to the
top of the coping and to one of the
faces of the coping at the same
time. Both the face and back leg
are tested, using separate speci¬
mens for each evaluation. The
load is held for 60 seconds and
then released. The next incremen¬
tal load is then applied. This
process continues until there is a
loss of attachment of any compo¬
nent of the coping system or
deformation that would result in
loss of weather protection at the
edge.
Test results
The outward and upward
maximum force at failure is
recorded. These forces are con¬
verted to pressure using the fol¬
lowing formulas:
Ennis and LeClare – 98 Proceedings of the RCI 23rd International Convention
Figure 9 – RE-3 test set-up.
damage in 79 per¬
cent of 145 cases
of insurance losses
experienced after
high-wind events.
The peel-failure
phenomenon can
occur in two basic
ways. In the first
case, the edge ter¬
mination fails and
becomes a sail to
catch wind, allow¬
ing the wind to
progressively peel
the membrane
from the perimeter
edge inward toward
shows a mechanically attached
EPDM membrane that uses a type
of peel-stop design. This roof is
located in Bay St. Louis, MS, and
even though it was exposed to
wind speeds of 120 to 130 mph
during Hurricane Katrina, there
was no damage to the attachment
of the membrane.
SPRI members recommend
the inclusion of a peel-stop device
in high velocity hurricane zones,
or when the designer is concerned
about the possibility of high wind
events as a common-sense design
enhancement.
CONCLUSION
Outward pressure = Outward
force/face height x face length
Upward pressure = Upward
force/ coping width x coping
length
Where:
2 Pressure is measured in lbs/ ft .
Force is measured in lbs force.
Face length is the test sample
length in feet.
Face height is in feet.
Face refers to back leg or front
leg.
Performance requirements
The test results for the coping
design must exceed the design
upward and outward wind pres¬
sures on both the front and back
leg tests.
ANSI/ SPRI ES-1 is available
free of change from the SPRI web¬
site at www.spri.org.
PEEL STOP
RECOMMENDATION
As noted earlier in this paper,
one of the findings from the
RICOWI hurricane investigations
is that a common mode of roof
system damage is edge failure. In
fact, Factory Mutual cites edge
the center of the
roof.
In the second failure mode,
the membrane remains attached
at the edge but it separated from
the substrate and balloons. This
ballooning action creates in¬
creased peel forces around the
edges of the balloon and causes
the ballooned area to progressive¬
ly expand. In either case, the peel
action will continue until stopped
by a physical feature or roof sys¬
tem enhancement.
The basic concept of a peel¬
stop design (also referred to as a
storm strip, hurricane strip, or
Proper design of roofing sys¬
tems to resist anticipated wind
loads is a key component of a sus¬
tainable roofing system. Inves¬
tigations conducted by RICOWI
after hurricane events show that
roof systems that were designed
and installed in a manner that
met manufacturer and building
code requirements performed well
in high wind events.
SPRI offers the following code¬
complaint test standards free of
charge on its Web site,
www.spri.org. These can be used
by the roof design professional to
design roof systems that will
hurricane bar)
is to install a
termination
device approxi¬
mately 12 in¬
ches away
from the roof
edge or para¬
pet wall around
the en¬
tire roof peri¬
meter. The de¬
vice is at¬
tached to the
structural deck
with mechan¬
ical anchors
spaced 6 inch¬
es on center
(see Figure 10).
This Figure
Figure 10 – This perimeter fastening is designed
for high wind resistance. (Photo courtesy of
RICOWI Inc.)
Proceedings oj the RCI 23rd International Convention Ennis and LeClare – 99
resist significant damage in highwind
events.
• ANSI/SPRI WD-1 – Wind
Design Standard for Low-
Slope Roofing Systems
• ANSI/SPRI RP-4 – Wind
Design Standard for Bal¬
lasted Single-Ply Roofing
Systems
• ANSI/SPRI ES-1 – Amer¬
ican National Standard
Wind Design Standard for
Edge Systems Used with
Low-Slope Roofing Sys¬
tems
REFERENCES
ANSI/SPRI WD-1 – Wind De¬
sign Standard for Low-
Slope Roofing Systems,
2007.
ANSI/SPRI RP-4 – Wind De¬
sign Standard for Bal¬
lasted Single-Ply Roofing
Systems, 2002.
ANSI/SPRI ES-1 – American
National Standard Wind
Design Standard for Edge
Systems Used with Low-
Slope Roofing Systems,
2003.
Baskaran, A and Smith, T.L.,
editors, A Guide for the
Wind Design of Mech¬
anically Attached Flexible
Membrane Roofs, National
Research Council of Can¬
ada, Institute for Research
in Construction, 2005.
FEMA 489, Mitigation Assess¬
ment Team Report: Hurri¬
cane Ivan in Alabama and
Florida, Observations,
Recommendations and
Technical Guidance, Aug¬
ust 2005.
FEMA 488, Mitigation Assess¬
ment Team Report: Hur¬
ricane Charley in Florida,
Observations, Recommen¬
dations and Technical
Guidance, April 2005.
FEMA 549, Mitigation Assess¬
ment Team Report: Hur¬
ricane Katrina in the Gulf
Coast, Building Perfor¬
mance Observations, Rec¬
ommendations and Tech¬
nical Guidance, July 2006.
FM Global, “Protecting Roof¬
ing Systems Against Wind¬
storm Damage,” 2003.
Gillenwater, Richard J., Wind
Design Guide for Ballasted
Roofing Systems,” Proceed¬
ings of the Second Inter¬
national Symposium on
Roofing Technology, 1985.
Greenberg Traurig Alert,
October 2005.
Kind, R.J., “Wind Tunnel
Tests on Loose-laid
Roofing Systems for Flat
Roofs,” Proceedings of the
Second International Sym¬
posium on Roofing Tech¬
nology, 1985.
Kind, R.J., J.R. McDonald,
Thomas L. Smith, “Hurri¬
cane Hugo: Evaluation of
Wind Performance and
Wind Design Guidelines
for Aggregate Ballasted
Single-Ply Membrane Roof
Systems,” Proceedings of
the VII International
Roofing and Waterproofing
Congress, Madrid, 1992,
pp. 598.
Kind, R.J., and R.L. Wardlaw,
“Design of Rooftops Against
Gravel Blow-Off,”
National Research Council
of Canada Report Number
15544, 1976.
McDonald, J.R., and Thomas
L. Smith, “Performance of
Roofing Systems in Hur¬
ricane Hugo,” Institute for
Disaster Research, Texas
Tech University, August
1990.
Phalen, Thomas E. Jr., “Stone
Ballast Design Criteria on
Loose-Laid, Single-Ply Bal¬
lasted Roofs for Wind
Speed, Size and Weight,”
Proceedings of the Second
International Symposium
on Roofing Technology,
1985.
RICOWI Inc., Hurricane Katri¬
na Investigation Report,
September 2007.
RICOWI Inc., Hurricanes
Charley and Ivan Investi¬
gation Report, April 2006.
Schneider, Kenneth G. Jr., “A
Study of the Behavior of
Loose-laid Ballasted Sin¬
gle-ply Roofing Systems
Subjected to Violent
Winds,” Proceedings of the
Second International Sym¬
posium on Roofing Tech¬
nology, 1985.
Smith, Thomas L., “Hurricane
Hugo’s Effects on Edge
Flashings,” International
Journal of Roofing Tech¬
nology, 1990, pp. 65.
Smith, Thomas L., “Improving
Wind Performance of
Mechanically Attached
Single-Ply Membrane Roof
Systems: Lessons from
Hurricane Andrew,” Pro¬
ceedings of the IX Congress
of the International Water¬
proofing Association, Am¬
sterdam, April 1995, pp.
321.
SPRI, “Enhancing Single-Ply
Wind Performance,” June
2007.
Ennis and LeClare – 100 Proceedings of the RCI 23rd International Convention