Testing Built-up and Modified Bitumen Roofs for Hail Damage

24 • IIBEC Interface July 2022
There are numerous testing laboratories
across the country that test built-up and
modified bitumen roof (bituminous roof)
samples for evidence of hail damage. The
problem is that there is no specific ASTM
International standard test method for performing
these tests. As a result, testing labs
are using a variety of test methods for testing
bituminous roof samples for hail damage.
Roof Technical Services Inc. (Rooftech) has
been testing bituminous roof samples for evidence
of hail damage since the 1980s. Most
of the testing from different laboratories that
we have reviewed applied adapted test protocols
from ASTM D2829, Standard Practice for
Sampling and Analysis of Existing Built-Up Roof
Systems,1 and/or ASTM D3746, Standard Test
Method for Impact Resistance of Bituminous
Roofing Systems.2 Additionally, there are acceptance
criteria that are sometimes referenced
in the laboratory reports included in Factory
Mutual (FM) Class Number 4470, Single-Ply,
Polymer-Modified Bitumen Sheet, Built-up Roof
(BUR) and Liquid Applied Roof Assemblies for
Use in Class 1 and Noncombustible Roof Deck
Construction3,4 (Appendix F: Susceptibility
to Hail Damage Test Standard); ANSI
FM 4473, Impact Resistance Testing of Rigid
Roofing Materials by Impacting with Freezer
Ice Balls5; and Underwriters Laboratory (UL)
2218, Standard for Safety—Impact Resistance of
Prepared Roof Covering Materials.6
This paper has two broad objectives. The
first is to provide an understanding of the
test methods in ASTM D2829 and ASTM
D3746 and how they relate to testing bituminous
roofs for hail-caused impact damage, as
well as the acceptance criteria in UL 2218,
FM 4470, and FM 4473. The second is to
provide a method for evaluating test samples
for evidence of hail damage based on laboratory
testing. To accomplish these objectives,
we performed ASTM D3746 tests on insulated
and noninsulated bituminous roofs. The purpose
of this testing was to simulate hail-caused
impact damage on these samples, to document
the resulting hail-caused impact damage, and
to provide a comparative visual standard to
evaluate bituminous test samples for hail damage.
This is an ongoing project, which will be
expanded to more varieties of bituminous roof
systems.
ASTM D3746
ASTM D37462 is a test protocol used to
assess the hail resistance of bituminous roofing,
such as built-up and modified bitumen roofing.
This test procedure utilizes a free-falling steel
missile to replicate the impact energy of a 2 in.
(50 mm) hailstone. This test method provides a
standard protocol for analyzing the bituminous
roofing for resistance to hail-caused impact
damage to bituminous roofs.
ASTM D3746 Section 10.7, “Damage
Assessment” establishes a test protocol for
evaluating the roof for impact damage as follows.
Section 10.7 also establishes a standard
for desaturating felts for evaluation. The process
uses a solvent bath to remove the bituminous
material from the reinforcement.
Reinforcements typically include fiberglass
felts; polyester mats; combination fiberglass
and polyester; and, in the case of older roofs,
organic or asbestos felts. The desaturation process
makes it easier to evaluate the felts for
evidence of hail-caused damage.
This paper was originally presented at the 2022 IIBEC International Convention and Trade Show.
July 2022 IIBEC Interface • 25
Section 10.7 Damage Assessment
10.7.1 Remove any slag or gravel surfacing
from the specimen carefully with a
hot scraper, such as a putty knife.
10.7.2 Record the extent of obvious
damage to the membrane, such as dents
or fractures, by photograph or sketch and
written description.
10.7.3 Cut the four Impact Areas from
the specimen using a hot knife. Staple
the felts in each area together and extract
the bitumen by immersing in warm 1,1,1
trichloroethane in a fume hood. Do not
heat the trichloroethane to boiling. (For
tarred felt and pitch membranes, use
xylene in place of trichloroethane.)
ASTM D3746 Section 10.8, “Rating of
Impact Damage,” establishes a protocol for rating
the impact damage as follows. There is only
a protocol for rating the samples based upon
the evidence of dents and cracks or splits. There
is nothing in the protocol that establishes a pass
or fail rating, so the interpretation is left to the
reader of the report.
Section 10.8 Rating of Impact Damage
10.8.1 Rate the impact damage which
occurs in each ply in each of the four
quadrants by assigning the number
which most accurately describes the
impact damage, as follows:
0 = no damage;
2 = dents, indentations only;
4 = any cracks or splits
10.8.2 After assigning the numbers to all
plies within each quadrant, add up all
the numbers and divide by four times
the number of plies to obtain an average
for the membrane. (Note: No passing or
failing criteria are provided.)
ASTM D2829
ASTM D28291 is a test for the analysis of
existing built-up roofs to determine whether
the roof sample contained the appropriate
number of plies, the appropriate amount of
asphalt or coal tar pitch (bitumen), an appropriate
flood coat, and an appropriate gravel surfacing,
and whether there are excessive installation
voids in the interply. The following section
from ASTM D2829 describes the test. There is
nothing in the scope of this test method that
deals with hail-caused impact damage. It is
a test protocol for “determining approximate
quantities of the various components.”
1. Scope
1.1 This practice is a guide for
removing test specimens from existing
built-up roofing systems in the field and
for determining the “approximate” quantities
of the components of that specimen
(Note 1). Components determined may
be: 1.1.1 Insulation components when
they are part of the roof membrane
system, 1.1.2 Plies of roofing felt, 1.1.3
Interply layers of bituminous material,
1.1.4 Top coating, and 1.1.5 Surfacing.
NOTE 1—This procedure is for the
investigation of existing roofs and is not
intended for new construction inspection.
1.2 This practice is applicable to both
914-mm (36-in.) and 1000-mm (39⅜-
in.) wide felt rolls.
1.3 The values stated in SI (metric) units
are to be regarded as standard.
1.4 This standard does not purport to
address all of the safety concerns, if any,
associated with its use. It is the responsibility
of the user of this standard to
establish appropriate safety and health
practices and determine the applicability
of regulatory limitations prior to use.
For specific precautionary information,
see 6.3.2.1.
The test protocol includes methods for
extracting samples from the field and for
delaminating the felts, both of which are commonly
used in testing bituminous roofs for
hail-caused impact damage.
Section 8, “Report,” describes the reporting
protocol as follows. Again, the report is
based on analyzing the roof components and
is not related to testing for hail-caused impact
damage.
8. Report
8.1 Describe the built-up roof, including
the type and class of bituminous material,
type of surfacing, type of insulation,
type of roof decking, and the type and
number of felts or roofing sheets.
8.2 Fully identify the origin and roof
location of each specimen.
8.3 Report the mass per unit area of
surfacing, average interply bituminous
material, top coating bituminous material,
total applied bituminous material,
and the total specimen (minus insulation).
See Table 3 for summary of results
and conversion to conventional units of
measurement.
8.4 Diagram the felt lapping to show the
number of plies and the lap relationship,
if determined (6.8).
FM 4470
The requirements for hail damage resistance
are included in FM 4470 Section 4.4,
There are numerous testing laboratories across
the country that test built-up and modified
bitumen roof (bituminous roof) samples for
evidence of hail damage. The problem is that
there is no specific ASTM International standard
test method for performing these tests.
26 • IIBEC Interface July 2022
“Hail Damage Resistance Test.”3 The test is
included in Appendix F, “Susceptibility to Hail
Damage Test Standard,” which was included
in the older standard4 and is only referred to as
“Susceptibility to Hail Damage Test Standard”
in the newer standard.
The FM test is similar to ASTM D37462
in that the test method includes dropping steel
balls to simulate hail. The acceptance criteria are
included in FM 4420 Section 4.4.1, “Conditions
for Hail Damage Resistance,” as follows.
4.4.1 Conditions for Hail Damage
Resistance
Both unconditioned (unweathered) and
conditioned (weathered) samples of roof
cover are inspected for damage. Neither
the roof cover nor the field seam (if present)
shall show any signs of cracking
or splitting. The field seam shall not
show any signs of cracking, splitting,
separation, or rupture when examined
closely under 10X magnification. Under
adhered conditions, minor separations of
the roof cover from the substrate (directly
under the Impact Areas) is acceptable
for monolithic decks only (i.e., structural
concrete or gypsum) or lightweight insulating
concrete insulation.
ANSI FM 4473
The ANSI FM 44735 standard utilizes ice
balls to simulate hail impact. The ice balls are
propelled at a velocity that simulates the kinetic
energy established for the various sizes of hail.
Ice ball testing is generally more representative
of actual hail impact than steel balls. Steel balls
are a simple way of simulating impact energy—
for example, drop a 1 lb (0.5 kg) steel ball
1 ft (0.305 m) and one gets 1 ft-lb (0.14 kg-m)
of impact energy. A good example is clay and
concrete tile roofing. Steel balls that generate
the same impact energy as hail will break tile,
while ice balls with the same impact energy
will not, as ice shatters upon impact and steel
does not. The difference is related to the difference
in momentum between ice balls and steel
balls. The acceptance criteria are included in
Section 4, “Pass/Fail Criteria,” as follows:
4.1.1 The test specimen shall show no
evidence of visible cracking or breakage
or any damage such as splits, punctures,
fractures, disengagement of lap elements
or exposure of materials not so intended.
4.1.2 When a test specimen fails to meet
the acceptance criteria for a tested classification,
two consecutive test specimens
must successfully meet the acceptance criteria
to qualify for the given classification
UL 2218
UL 22186 is similar to ASTM D37462 in
that steel balls are dropped to simulate hail
impact. UL 2218 utilizes 1.25-in. (32-mm), 1.5-
in.- (38-mm), 1.75-in. (44-mm), and 2.00-in.
(50-mm)-diameter steel balls that are dropped
from 12.0 ft (3.7 m), 15.0 ft (4.6 m), 17 ft
(5.2 m), and 20.0 ft (6.1 m), respectively, to simulate
the impact energy of 1.25-in-., 1.50-in.-,
1.75-in.-, and 2.00-in.-diameter hail. The
acceptance criteria are included in Section 7,
“Acceptance Criteria,” as follows:
7.1 The prepared roof covering material
is to be examined after being subjected
to the test procedure described in Section
6. The prepared roof covering material
exposed surface, back surface and underneath
layers shall show not evidence of
tearing, fracturing, splitting, rupture,
crazing or other evidence of opening
through any prepared roof covering layer.
7.2 For asphalt shingles, a visible crack
of the asphalt on the back of the shingles
shall be determined to be a failure.
7.3 For wood, tile, concrete, fiber-cement,
plastic and metal roof coverings, a surface
crack shall not be determined to be a
failure. A crack that extends through the
cross-section of the roof covering material
layer shall be determined to be a
failure.
7.4 Cosmetic damage in and of itself
shall not be determined to be a failure.
Cosmetic damage such as denting, damage
not extending through the cross-sectional
area of a roof covering material
layer, crack of any paint finish, etc. shall
not be determined to be a failure.
TESTING PROTOCOL
We utilize a modified version of ASTM
D37462 for analyzing hail-caused impact damage
and the methodology for delamination and
desaturating the felts to evaluate the samples to
determine whether there is damage. In general,
the protocol includes the following:
1. Visually examining the top and bottom
of the samples for evidence of impact
damage to the surface of the roof.
This would include evidence of spatter
marks, denting, displaced granules
or gravel, and evidence of crushed or
cracked bitumen.
2. Delaminating the samples in general
accordance with ASTM D28291 and
visually examining the interply bitumen
for evidence of denting, crushed
interply, and/or fracturing of the
reinforcement.
3. Desaturating the samples in general
accordance with ASTM D37462 and
visually examining the desaturated
reinforcements for evidence of denting
and fracturing of the reinforcement.
It should be noted that some
labs do not desaturate the samples and
rely on the examination of the delaminated
plies for evidence of fracturing.
However, desaturation is part of the
ASTM D3746 protocol and provides
for a more reliable assessment of the
reinforcement.
4. Examining the samples under various
magnifications, including 10-power
magnification, at each step of the
testing in general accordance with
Susceptibility to Hail Damage Test
Standard, Section 4.4.1.7
The following is an excerpt from a typical
report describing our testing protocol:
Each of the six mineral granule surfaced
modified bitumen roof membrane
samples was logged, visually inspected
under various magnifications and
photographed top and bottom. The roof
membrane samples were then delaminated,
inspected, and photographed.
The roof membrane samples were
desaturated and evaluated in general
accordance with ASTM D3746, Impact
Resistance Analysis of Bituminous
Roofing Systems. Each individual ply
was photographed top and bottom and
visually inspected. Plies were examined
under microscope at various magnifications.
Any anomalies detected were
photographed and recorded.
Our intent is to visually document each
step of the testing to provide transparency in
our reporting. Each step of the testing protocol
is photographed, including photographs of the
front and back of the sample upon arrival, the
individual plies after delamination, and the
individual plies after desaturation, along with
magnified views of points of interest. A diagram
of the sample configuration in general accordance
with ASTM D28291 is also provided.
July 2022 IIBEC Interface • 27
ANALYSIS AND
INTERPRETATION OF THE DATA
The next step is to analyze and interpret
the data, which can be subjective. As noted earlier,
the general guidelines included in ASTM
D37462 are as follows:
Rate the impact damage which occurs
in each ply in each of the four quadrants
by assigning the number which most
accurately describes the impact damage,
as follows:
0 = no damage;
2 = dents, indentations only;
4 = any cracks or splits
We do not provide a rating analysis as
described. Our reports document evidence of
granule or gravel loss, denting of the surface
and/or the reinforcement, cracks or splits in
the reinforcement, and crushed bitumen at the
point or points of interest. ASTM D3746 provides
an unambiguous description of impact
damage. Are the desaturated felts dented or
fractured? No dents or fractures receive a rating
of 0; dents receive a 2, and fractures receive a 4.
The test method does not consider the surfacing
or the interply bitumen for damage. The
issues of what constitutes damage may become
subjective, depending on the various definitions
of damage—for example, the definition of
damage included in an insurance policy.
It is important to consider that ASTM
D3746 is designed to test and rate a bituminous
roof sample for resistance to hail. Typically, the
test samples used are from new construction
and often prepared for the purpose of testing. In
the case of these samples, the area of impact is
known, and the impacted area can be analyzed
and compared with the nonimpacted areas.
This is not the case of test samples taken from
existing roofs.
Existing bituminous roofs have been subjected
to construction traffic, maintenance traffic,
and, in many cases, years of weathering.
Bituminous roofs are typically installed with
heat (hot asphalt and torches) and are susceptible
to foot and general construction traffic
during the installation of the roof, particularly
when the roof is hot. Anomalies from installation
traffic are common to virtually all bituminous
roofs, and these anomalies are often
confused with impact damage from hail.
The intent of our ASTM D37462 testing
was to provide clear visual examples of impact
damage to bituminous roofs—that is, what hailcaused
impact damage looks like on bituminous
roofs.
ASTM D3746 TESTING
We performed ASTM D37462 testing on
aged aggregate-surfaced fiberglass asphalt
built-up roof samples and on new and aged
granule-surfaced styrene butadiene styrene
(SBS) modified bitumen roof samples.
The built-up roofing samples were tested
on insulated (relatively soft) and noninsulated
(firm) substrates. The new granulesurfaced
modified bitumen samples were tested
on a firm substrate of 0.5 in. (13 mm) gypsum
cover board and an insulated (relatively soft)
substrate. The aged granule-surfaced modified
bitumen samples were tested on insulated and
noninsulated substrates. ASTM D3746 uses
a missile with a 2.0 in. (50 mm) diameter to
replicate the impact energy of 2.0 in. hail. Many
bituminous roof systems are resistant to 2.0 in.
hail. The fundamental purpose of our testing
was to replicate hail-caused damage. Therefore,
we modified the missile size to replicate 2.5
in. (64 mm) hail and to develop 57.48 ft-lb
of impact energy as defined by the National
Bureau of Standards.3
Piping on roofs constantly moves, which can result in roof
damage. Wood or rubber blocks used as pipe supports don’t
allow pipe movement. The solution? MAPA engineered
rooftop pipe supports. They help prevent roof abrasion and
add years to the life of a roof.
www.mapaproducts.com
Innovative rooftop supports since 1998
Severe damage to roof
and pipe due to the use
of wood blocks.
PIPE
PLACED HERE
PROTECTS
ROOFS.
28 • IIBEC Interface July 2022
The modified missile was 2.5 in. (64 mm) in diameter and 6 in.
(152 mm) long, and weighed 7.67 lb (3.48 kg). A 24 × 24 in. (610 × 610
mm) testing table was constructed using two-by-fours placed on edge
and bolted together in accordance with ASTM D3746. The test sample
was placed on the testing table, and the missile was dropped from
87⅞ in. (2232 mm) onto the approximate center of each of the four
quadrants. Figure 1 shows a test sample positioned on the platform,
and Fig. 2 shows how a missile was dropped onto the approximate
center of each of the four quadrants.
Testing was performed on test samples taken from aged aggregate-
surfaced asphalt fiberglass built-up samples and from new and
aged granule-surfaced SBS modified bitumen samples. Each sample
was examined and photographed in accordance with the our protocol
described previously.
TEST RESULTS FROM AGGREGATESURFACED
BUILT-UP ROOF SAMPLES
Evaluation of Surfacing at Impact Area
The test results related to surface damage were consistent. The
impact from a 2.5 in. (64 mm) missile resulted in surface damage to
the samples on both the insulated and noninsulated substrates. The
results of the laboratory impact damage to the surface were compared
with test samples of roofs that had been damaged by hail as
well as examples included in Haag Engineering’s Built-up Roofing: A
Pictorial Guide.8 The test sample before the missile drop in shown in
Fig. 3. The test sample after the missile drop and after the loose gravel
was removed is shown in Fig. 4. The red
arrows identify the area of impact in Fig. 4.
The areas of impact were very similar on
all tests. There was a general displacement
of the imbedded aggregate and exposure of
the asphalt flood coat. The aggregate at the
point of impact was crushed at some of the
impacts (Fig. 5). In no case did the impact
result in aggregate being driven into the
sample, affecting the felts below. The crushed
aggregate was likely the result of using a steel
missile rather than an ice ball for testing, as
hail typically shatters on impact and steel
does not.
There was localized crushing of the
asphalt flood coat at the point of impact, as
shown in Fig. 6.
Figure 1. A test sample positioned on the platform.
Figure 4. The test sample after the missile drop
and after the loose gravel was removed.
Figure 3. The test sample before the missile
drop.
Figure 2. A missile was dropped onto the
approximate center of each of the four quadrants.
July 2022 IIBEC Interface • 29
The area of impact in our simulated testing
is also consistent with roof samples that we
have tested from roofs damaged by hail. Figure
7 shows a test sample taken from a roof that
was damaged by hail. The meteorological and
physical evidence indicated that the hail was in
the 2.5 in. (64 mm) and larger range.
The surface damage occurring at each of
the impact areas from our testing was consistent
in appearance, was consistent with surface
damage from actual hail observed in the
field and in the laboratory, and was consistent
with the photographs included in Built-up
Roofing: A Pictorial Guide.8 Our conclusion is
that these illustrations of surface damage to
aggregate-surfaced built-up roofs are representative
of actual hail damage and can be used for
comparative analysis.
Evaluation of Interply Bitumen—
Insulated and Noninsulated Samples
There is no protocol for the evaluation of
the interply bitumen in ASTM D3746.2 The
rating system protocol for evaluation of damage
is limited to the desaturated felts. Therefore,
crushed or disturbed interply is not a factor in
the evaluation process of ASTM D3746. In the
case of the noninsulated samples, there was no
denting or cracking of the desaturated felts;
therefore, based on ASTM D3746, this roof
would have been rated as having no hail-caused
impact damage.
The sample was desaturated and the felts
were evaluated using the rating protocol in
ASTM D3746. Early on, the vast majority of our
testing focused on ASTM D28291 and ASTM
D3617.9 Both standards were used to determine
whether the roof system was installed in accordance
with industry standards. These testing
standards consisted of weighing and measuring
the components of the roof and comparing
the results to a recommendation or guideline,
either a manufacturer’s
or the
National Roofing
C o n t r a c t o r s
Association’s. These
tests were used for
quality assurance or
forensic purposes.
Part of the protocol
included the delaminating
the felts to
determine the lapply
configuration. In addition, we examined the
interply for evidence of voids or dry spots.
We began testing roof samples for evidence
of hail damage in the 1980s. Our standard test
protocol for evaluating hail damage utilized
ASTM D2829 testing standards for removal of
the samples, for preparation of the samples, and
for delamination of the felts. The delaminated
felts were then desaturated and evaluated using
the ASTM D3746 protocol for assessment. On
Figure 5. The test sample after the missile drop
showing crushed gravel at point of impact.
Figure 6. Localized crushing of the asphalt
flood coating at the point of impact is
shown in this 10-power photograph.
Figure 7. A test sample taken from a roof that was damaged by hail.
30 • IIBEC Interface July 2022
rare occasions, in the case of very large hail,
we observed crushed interply asphalt between
the plies and began including these observations
in our reports. It should be noted that
crushed interply was never observed in coal tar
pitch samples, which is likely explained by the
self-healing properties of coal tar pitch.
Observations regarding crushed or disturbed
interply
bitumen
were included
in our study.
Our study was
based on evaluating
damage
from 2.5 in. (64 mm) missiles representing
very large hail. Crushed interply bitumen was
observed at each area of impact on the insulated
samples. Crushed interply was observed at
some of the impact areas on the noninsulated
samples. Figure 8 shows the back side of the
top ply of felt, which was partial lap-ply. The
sample was only impacted by test drops 1 and 2.
There was no evidence of crushed
interply at this level within the
sample.
Figure 9 shows a
close up of impact 2
with no evidence of
crushed or disturbed
interply.
Figure 10 shows
the impact area at
10-power with no
evidence of crushed
asphalt.
There was evidence of crushed interply in
the lower layers of felt. Figure 11 shows impact
area 4 on the back of the first full ply.
Figure 12 shows impact area 4 at 10-power
and the crushed interply asphalt at the point of
impact.
The crushed interply in our simulated testing
was also consistent with roof samples we
have tested from roofs damaged by hail. One
example is a roof designed by us that was
approximately 15 years old at the time of the
event. The hail was in the 2.5 in. (64 mm) and
larger range.
There was evidence of crushed interply
in all of the areas of impact on the insulated
samples. There was evidence of crushed
Figure 9. Close-up of impact 2 with no
evidence of crushed or disturbed interply.
Figure 8. The back side of the top ply
of felt, which was partial lap-ply.
Figure 10. Impact area at
10-power with no evidence
of crushed asphalt.
Figure 11. Impact area 4 on
the back of the first full ply.
Figure 12. Impact area 4 at
10-power and the crushed interply
asphalt at the point of impact.
July 2022 IIBEC Interface • 31
interply in some of the
areas of impact on the
noninsulated samples.
In general, the softer
the substrate (typically
insulated), the more
susceptible the roof is to
hail damage. Crushing
of the interply may also
be a function of the
thickness of the asphalt.
The crushed interply
bitumen occurring at
each of the impact areas
from our testing was
consistent in appearance
and was also consistent
with crushed
interply bitumen from
actual hail observed in
the field and laboratory.
Evaluation of Desaturated Felts—
Insulated and Noninsulated Samples
In ASTM D3746,2 the protocol for evaluating
hail-caused impact damage to desaturated
felts is clearly defined as dents or cracks or
splits (fractures). The desaturated felts were
examined for evidence of damage at each of the
four impact areas. The results of the laboratory
impact damage to the surface were compared
with our test samples of roofs that had been
damaged by hail and to examples included
in Haag Engineering’s Built-up Roofing: A
Pictorial Guide.8
Our study showed that there were fractures
in the felts at each of the impact areas on the
insulated samples, but there were no fractures
in the felts at each of the impact areas on the
noninsulated samples. There were no dents or
indentations in the felts in any of the impact
areas. The denting criterion is probably a holdover
from testing on organic or asbestos felts, as
it has been our experience that hail impact does
not typically result in dents in fiberglass felts.
Dents and other anomalies in the top layer of
felt commonly occur as a result of construction
traffic during installation, particularly from
loose aggregate stepped on during construction
before the flood coat has been applied.
Our test results confirm the importance of
the substrate in the hail resistance of bituminous
roofs. In general, the softer the substrate (typically
insulated) is, the more susceptible the roof
is to hail damage. Impact from very large hail
can cause localized deflection in the membrane
at the point of impact. Figure 13 illustrates how
the softer substrate allows for more deflection
in the membrane, resulting in tension in the
bottom layers of the felts.
Figure 14 shows the desaturated bottom
ply of felt in the noninsulated (firm substrate)
sample. The tension is greatest in the bottom
ply, so the bottom ply is the most likely ply to
fracture. The tension in the felt was limited by
the firm substrate, and there was no fracturing.
Also, there was no denting in the fiberglass felts
on the insulated and noninsulated samples.
There were fractures at all impact areas
in the insulated sample. Figure 15 shows the
bottom full ply with impact fractures identified
with the red arrows. Figure 16 shows a magnified
view of the fracture in the felt at impact
area 3.
The fractures at the areas of impact in our
simulated testing are also consistent with roof
samples we have tested from roofs damaged by
hail. Figure 17 shows fractured felts in a test
sample from a roof that was damaged by hail.
This was a roof that was designed by us and
was approximately 15 years at the time of the
hail event. The hail was in the 2.5 in. (64 mm)
and larger range.
The area of impact in our simulated testing
is also consistent with the hail-caused impacts
illustrated in reference 8.
Figure 13. The softer substrate allows for
more deflection in the membrane, resulting
in tension in the bottom layers of the felts.
Figure: Adapted from Haag Engineering’s
Built-up Roofing: A Pictorial Guide.
Figure 14. Desaturated bottom ply of felt in the
noninsulated (firm substrate) sample.
Figure 15. Bottom full ply with impact
fractures identified by red arrows.
32 • IIBEC Interface July 2022
There was no evidence of fracturing in the
desaturated felts in the noninsulated sample.
The was evidence of fractured felts at all of the
areas of impact on the insulated samples. There
was no evidence of denting in the insulated
or noninsulated samples. The fractured felts
occurring at the impact areas from our testing
were consistent with fractures occurring as a
result of actual hail and consistent with published
literature.8
TEST RESULTS FROM NEW GRANULESURFACED
SBS MODIFIED BITUMEN
ROOF SAMPLES
Evaluation of Surfacing at Impact Area
The test results related to surface damage to
the granule-surfaced modified bitumen
samples varied. There was very
little evidence of displaced granules
on the new samples on insulated or
noninsulated substrates. It should be
noted that the noninsulated samples
were supported by 0.50 in. (13 mm)
gypsum cover board, which was not
as firm as the wood test table but
is considered to be a firm
substrate. The insulated
samples were supported by
0.75 in. (19 mm) perlite,
which is considered to be
a soft substrate. The samples
were positioned and
impacted as described in
the protocol for the aggregate-covered
built-up roofs.
Figure 18
shows the new
modified bitumen
sample on noninsulated
substrate
after the sample
was impacted by the
four missile drops.
Figure 19
shows a closer view
of impact area 1.
There is a slight difference
in the color
of the granules.
Figure 20 shows a 10-power view of impact
area 1. There are no discernible displaced granules.
The change in color is the result of localized
crushing of the granules at the point of
impact.
Figure 21 shows a 10-power view of a typical
impact area on the insulated sample. There
are no discernible displaced granules. There is
no evidence of localized crushed granules on
the insulated samples.
The area of impacts on the new modified
bitumen exhibited no displaced granules. The
crushed granules were likely the result of the
steel missile and were unlikely to occur on
simulated ice balls or natural hail. We have
observed similar results on newer installations
that were impacted by large hail. The adhe-
Figure 16. Magnified view of the fracture in the felt at
impact area 3.
Figure 19. A closer view
of impact area 1. There
is a slight difference in
the color of the granules.
Figure 17. Fractured felts in a test sample from a roof that
was damaged by hail.
Figure 20. A 10-power view of impact
area 1. There are no discernible
displaced granules. The change in
color is the result of localized crushing
of the granules at the point of impact.
Figure 18. The new
modified bitumen
sample on noninsulated
substrate after the
sample was impacted by
the four missile drops.
July 2022 IIBEC Interface • 33
sion of the granules on new SBS modified bitumen
roofs is generally very good, and limited
or no granule loss may occur on newer installations
of modified bitumen roofs. The crushed
granules may weather away over time and result
in an area of localized granule loss, but this was
not verified.
Evaluation of Interply Bitumen—Insulated
and Noninsulated Samples
As stated previously, there is no protocol
for the evaluation of the interply bitumen in
ASTM D3746.2 However, the interply bitumen
was examined on the modified bitumen samples.
Crushed interply bitumen was observed
on the insulated sample but not on the noninsulated
sample. The lack of crushed interply is
possibly explained by the quantity of interply
asphalt. It has been our experience that thicker
applications of interply asphalt are more prone
to crushed interply. Figure 22 shows an example
of crushed interply on the insulated sample.
There was evidence of
crushed interply in all of the
areas of impact on the insulated
samples. There was no
evidence of crushed interply
in the areas of impact on
the noninsulated samples. In
general, the softer the substrate
(typically insulated) is,
the more susceptible the roof
is to hail damage. Crushing
of the interply may be also
a function of the thickness
of the asphalt. The crushed
interply bitumen occurring
at each of the impact areas
from our testing was consistent
with crushed interply
bitumen from actual hail observed in the field
and laboratory.
Evaluation of Reinforcement—Insulated
and Noninsulated Samples
The modified bitumen membrane had a
dual-carrier mat with a combination of polyester
and fiberglass reinforcement. Figure 23
shows the dual-carrier mat with no fractures
or denting.
The second ply was fiberglass asphalt felt.
There were fractures in the fiberglass felt at all
four missile drops on both the insulated and
noninsulated samples. It should be noted that
there was a depression at the area of impact in
the gypsum cover board,
which resulted in more
tension in the bottom ply
than the built-up samples
on the wood testing
table. Figure 24 shows a
fracture in the bottom ply
(fiberglass felt) on the noninsulated sample. The
fractures on the insulated samples were more
pronounced than those on the noninsulated
samples. The fractured felts occurring at the
impact areas from our testing were consistent
with fractures occurring as a result of actual
hail, and consistent with published literature.8
TEST RESULTS FROM AGED GRANULESURFACED
SBS MODIFIED BITUMEN
ROOF SAMPLE
Evaluation of Surfacing at Impact Area—
Noninsulated Substrate
The aged modified bitumen sample
was only tested over a noninsulated (wood)
Figure 21. A 10-power view of a typical impact area on the
insulated sample. There are no discernible displaced granules.
There is also no evidence of localized crushed granules. Figure 22. Example of crushed interply on the insulated sample.
Figure 24. Fracture in the bottom ply
(fiberglass felt) on the noninsulated sample.
Figure 23. Dual-carrier mat with no fractures or denting.
34 • IIBEC Interface July 2022
substrate. The exact age of the sample is unknown but is believed to be at least 10 years
old. The test results related to surface damage to the granule-surfaced modified bitumen
samples were similar to results for the noninsulated new modified bitumen sample, with
the exception that there was some granule displacement at the point of impact on the aged
sample. Figure 25 shows the aged modified bitumen roof on the noninsulated substrate
after the sample was impacted by the four missile drops.
Figure 26 shows a close-up of impact area 3. There are crushed granules similar
to the crushed granules on the new modified bitumen sample on the firm substrate
as well as some granule
displacement not evident
in the sample of the new
roof. Figure 27 shows a
10-power view of impact
area 3.
Figure 28 shows a
close up of impact area
4. There are crushed
and displaced granules.
Exposed reinforcement is
also visible.
Figure 29 shows
impact area 4 at 10-power.
The area of impacts on the
aged modified bitumen
exhibited crushed granules
and some displaced
granules. It is possible that
additional granule loss would occur over time if the sample were exposed to
normal weathering.
We have observed localized granule loss at impacts from actual hail
on aged modified bitumen roofs. Figure 30 shows an example of localized
granule loss on an aged modified bitumen roof. The pattern of granule loss
was consistent with the random distribution of the large hail that fell and
matched the pattern of the larger hail-caused impacts on the air-conditioning
units and larger spatter marks.
Figure 31 shows a closer view of the hail-caused granule loss on the
same project.
Figure 29. Impact area 4 at 10-power.
Figure 28. Close-up of impact area 4.
Figure 25. Aged modified bitumen
roof on the noninsulated substrate
after the sample was impacted by
the four missile drops.
Figure 26. Close-up
of impact area 3.
Figure 27. A 10-power
view of impact area 3.
July 2022 IIBEC Interface • 35
Figure 32 shows the impact at 10-power.
There are visible fractures in the surface at
the point of impact. Also, note that the surface
of the exposed modified bitumen is relatively
smooth. There is no evidence of shrinkage
cracking or oxidized bitumen typically seen on
older modified bitumen roofs.
Figure 33 shows a test sample from a modified
bitumen roof with an area of localized
granule loss at the point of interest noted. This
of type localized granule loss is also often confused
with localized
granule loss
from hail impact.
In some cases, a
close examination of the area of granule loss
will exhibit signs of weathering, as shown in
Fig. 33, indicating that the granule loss occurred
before the hail event.
Figure 34 shows a closer view of the point
of interest. There is evidence of shrinkage
cracking in the surface. There is localized granule
loss in areas where the shrinkage
cracking has converged. The point
of interest is the largest
area of localized granule
loss visible on the
sample.
Figure 35 shows
the point of interest
at 10-power. The oxidized
modified bitumen
and shrinkage
cracking in the modified
bitumen surfacing
are visible. There
is no evidence of impact damage to the surface,
and there were no fractures of crushed
interply below the point of impact. There was
no evidence of any other damage to the sample
consistent with hail-caused impact. This type
of localized granule loss is often confused with
granule loss from hail-caused impact.
Another type of localized granule loss is
related to contaminants such as bird droppings.
Figure 32. Impact at 10-power.
Figure 30. An example of localized granule
loss on an aged modified bitumen roof.
Figure 31. A closer view of the hail-caused
granule loss on the same project as Fig. 30.
Figure 34. A closer view of
the point of interest.
Figure 33. Test sample from a modified
bitumen roof with an area of localized
granule loss at the point of interest noted.
36 • IIBEC Interface July 2022
This type of localized area of granule loss tends to be in specific areas where birds
congregate, as shown in Fig. 36. This type of localized granule loss is often confused
with granule loss from hail-caused impact. Figure 37 shows a closer view of localized
granule loss occurring as a result of contaminants from bird droppings. Figure 38
shows the progression of granule loss occurring as a result of the bird droppings.
Remnants of the bird droppings are visible. The granules typically continue to come
off over time. Figure 39 shows the progression of granule loss from small semicircular
areas of granule loss to complete circular granule loss.
The surface damage occurring as a result of our impact testing of the granule-surfaced
modified bitumen ranged from no displaced or crushed granules on the new
modified bitumen on insulated (soft) substrate to crushed granules on firm substrates
with no discernible granule loss. This is consistent with field observations on relatively
new modified bitumen roofs that were impacted by large hail.
The aged modified bitumen sample tested on a firm substrate resulted in crushed
granules and some displaced granules. The displaced granule loss
observed on the aged sample tested was not as pronounced as observations
of granule loss on aged modified bitumen roofs impacted by
large hail. The age and degree of surface deterioration are likely contributing
factors, as well as the angle of strike from actual hail.
Evaluation of Interply Bitumen—Noninsulated Samples
The interply bitumen was examined on the modified bitumen
samples. Crushed interply bitumen was observed on the sample at all
four areas of impact. Figure 40 shows an example of crushed interply
on the insulated sample. There was evidence of crushed interply in all
of the areas of impact on the samples. The crushed interply bitumen
occurring at each of the impact areas from our testing was consistent
in appearance and was also consistent with crushed interply bitumen
from actual hail observed in the field and
laboratory.
Evaluation of Reinforcement—
Noninsulated Substrate
The membrane was a fiberglassreinforced
modified bitumen installed
over two plies of glass felt. There was
no denting or fracturing of the modified
bitumen reinforcement. There were fractures
in the bottom ply at all four impact
Figure 35. Point of interest at 10-power.
Figure 36. Localized areas of
granule loss tend to be in specific
areas where birds congregate.
Figure 37. A closer view
of localized granule loss
occurring as a result
of contaminants from
bird droppings.
Figure 38. Progression of
granule loss occurring as a
result of the bird droppings.
Figure 39. Progression of granule loss
from small semicircular areas of granule
loss to complete circular granule loss.
July 2022 IIBEC Interface • 37
areas. Figure 41 shows the fractured bottom ply of
the membrane. The fractures were small and difficult
to see without desaturating the felts.
The testing results of the aged modified bitumen
sample were consistent with the testing of the
new modified bitumen samples, with the exception
that there was discernible granule displacement on
the aged modified bitumen sample.
DISCUSSION
The use of highly magnified photographs of
samples and particularly of desaturated felts and
reinforcements to demonstrate damage has become
increasingly prevalent. By way of demonstration, we
performed desaturation on a new roll of Type IV
fiberglass felt. A 12 × 12 in. (305 × 305 mm) sample
was removed and placed on a light table, illustrating
the normal holes in the new felt, as shown in Fig. 42.
Figure 43 shows the portion of the felt in the red box (test area) of Fig. 42 at a
magnification showing the normal holes in a typical fiberglass felt.
Figure 44 shows the desaturated test area. The individual fibers can be
seen. We have seen an increasing number of this type of photograph, which is
represented as evidence of hail-caused damage to bituminous roofs. Holes like
the one circled in red are often represented as being the result of hail impact.
Anomalies resulting from construction or maintenance damage are also often
represented as evidence of hail-caused damage. It is important to distinguish
normal surface anomalies and normal holes in felts from actual hail-caused
impact damage.
CONCLUSION
The use of the assessment protocol in ASTM D37462 is an appropriate method
for evaluating impact damage to the reinforcements of bituminous roofs.
The assessment protocol in ASTM D3746 does not address impact to the surfacing of the
sample or to the
interply bitumen.
The results of our
testing utilizing
ASTM D3746
impact testing
protocol to simulate
hail-caused
impact damage
provide a graphic
Figure 40. An example of
crushed interply on the
insulated sample.
Figure 41. Fractured
bottom ply of the
membrane.
Figure 44.
Desaturated
test area.
Figure 42. A 12 × 12 in. sample was removed and
placed on a light table illustrating the normal holes
in the new felt. Note: 1 in. = 25.4 mm.
Figure 43. When magnified, the portion of the felt in
the red box (test area) of Fig. 42 shows the normal
holes in a typical fiberglass felt.
38 • IIBEC Interface July 2022
example of what hail-caused damage looks like.
The results of our testing were consistent with
observations and testing of bituminous roofs that
have been damaged by hail, are consistent with
the research performed by Haag Engineering,8
and are consistent with testing reports by others
that we have reviewed over the years.
Hail-caused impact damage has a specific
signature, as demonstrated by this testing,
and the results of this testing provide a graphic
comparative standard for hail-caused damage
to bituminous roofs. The results of this testing
provide a way to distinguish actual hail-caused
damage to bituminous roofs from normal anomalies
common on bituminous roofs, including
construction traffic, maintenance traffic, and
contaminants. Most bituminous roofs are resistant
to 1.5 in. (38 mm) hail, and many are resistant
to 2.0 in. (50 mm) or larger hail, so it takes
large hail to damage this type of roof. We are
continuing to research different types of roofing
to provide standards for assessing hail-caused
damage on various combinations of roofing.
There are a variety of testing standards that
provide protocols for addressing hail-caused
impact. Unfortunately, there is no specific
ASTM test method for evaluating hail damage
to existing roof systems. The lack of a specific
standard has led to confusion in the industry
and the use of widely varying test methods for
analyzing hail-caused impact damage. A new
ASTM test standard that specifically addresses
testing for hail-caused impact damage to existing
roofs would help eliminate the confusion
in the industry and provide more consistent
testing and analysis.
REFERENCES
1. ASTM International. 2019. Standard
Practice for Sampling and Analysis
of Existing Built-Up Roof Systems.
ASTM D2829M-07(2019)e1. West
Conshohocken, PA: ASTM International.
2. ASTM International. 2015, Standard
Test Method for Impact Resistance
of Bituminous Roofing Systems.
ASTM D3746M-85 (2015)e1.
West Conshohocken, PA: ASTM
International.
3. Factory Mutual Insurance Company
(FMIC). 2012. Single-Ply, Polymer-
Modified Bitumen Sheet, Built-Up Roof
(BUR) and Liquid Applied Roof Assemblies
for Use in Class 1 and Noncombustible
Roof Deck Construction. FM Class
Number 4470. Johnson, RI: FMIC.
4. FMIC. 2002. Single-Ply, Polymer-
Modified Bitumen Sheet, Built-Up
Roof (BUR) and Liquid Applied
Roof Assemblies for Use in Class
1 and Noncombustible Roof Deck
Construction. FM Class Number 4470.
Johnson, RI: FMIC.
5. FMIC. 2005. Impact Resistance Testing
of Rigid Roofing Materials by Impacting
with Freezer Ice Balls. FM Class Number
4473. Johnson, RI: FMIC.
6. Underwriters Laboratories (UL). 2020.
Standard for Safety—Impact Resistance
of Prepared Roof Covering Materials.
UL 2218. Northbrook, IL: UL.
7. FM Approvals. 2010. Approval Standard
for Class 1 Roof Covers Appendix 7:
Susceptibility to Hail Damage Test
Standard. Norwood, MA: FM Approvals.
8. Haag Engineering Co. 2004. Built-up
Roofing: A Pictorial Guide.
9. ASTM International. 2017. Standard
Practice for Sampling and Analysis
of Built-Up Roof Systems During
Application. ASTM D3617/D3617M-17.
West Conshohocken, PA: ASTM.
Stephen L. Patterson,
RRC, PE, has been in
the roofing industry
for almost 50 years.
He founded Roof
Technical Services
Inc. (ROOFTECH)
in 1983 and has been
an active consulting
engineer and roof
consultant ever since.
ROOFTECH has
provided laboratory
testing, including testing
for hail damage, since the late 1980s. Patterson
has been technical director/director of engineering
for two roofing manufacturers and managed a
roof contracting company for four years.
Stephen L. Patterson,
RRC, PE
Publish in IIBEC Interface
INTRODUCTION
In evaluating building enclosure
problems, the author has encountered
many newly constructed, wood-framed,
low-slope roofs and exterior balconies
and decks that exhibit excessive/sustained
ponding of water (Figure 1). These
conditions can lead to interior water
damage through premature deterioration
of roof coverings and/or excessive
deflection of roof framing members. The
ponding (and associated creep of the
framing) can be so significant that it
may ultimately lead to failure of the roof
framing.
The purpose of this article is to provide
insight into the most likely causes
of these problematic ponding conditions
as they relate to commonly accepted
design and construction methods.
36 • IIBEC IntErfaCE OCtOBEr 2019
Figure 1 – Excessive ponding water
on a roof.
Figure 2 – Ponding typically occurs prior to reaching discharge points.
INTRODUCTION
The concept of building for resilience
has been increasingly adopted by various
organizations over the past five years.
Organizations use different definitions or
phrases to describe resilience and the hazards
that are included in resilient design.
These definitions from six sources are compared
and a single definition incorporating
these is developed.
RESILIENCE AS DEFINED BY SELECT
ORGANIZATIONS
Industry Statement
Twenty-one organizations, including the
U.S. Green Building Council (USGBC), the
American Society of Heating, Refrigerating,
and Air-Conditioning Engineers (ASHRAE),
the American Institute of Architects (AIA),
the American Society of Civil Engineers
(ASCE), the Building Owners and Managers
Association (BOMA), and the National
Institute of Building Sciences (NIBS) issued
an industry statement on resilience[1] that
stated (the bold or red text is theirs):
Representing more than 750,000
professionals, America’s design and
construction industry is one of the
largest sectors of this nation’s economy,
generating over $1 trillion in
GDP. We are responsible for the
design, construction, and operation
of the buildings, homes, transportation
systems, landscapes, and public
spaces that enrich our lives and
sustain America’s global leadership.
We recognize that natural and
manmade hazards pose an increasing
threat to the safety of the public
and the vitality of our nation. Aging
infrastructure and disasters result
in unacceptable losses of life and
property, straining our nation’s ability
to respond in a timely and efficient
manner. We further recognize
that contemporary planning, building
materials, and design, construction,
and operational techniques can
make our communities more resilient
to these threats.
Drawing upon the work of the
National Research Council, we define
resilience as the ability to prepare
8 • IIBEC IntErfaCE SEptEmBEr 2019
This article is reprinted with permission
from Advances in Civil Engineering
Materials, Vol. 7, No. 1, 2018, copyright
ASTM International, 100 Harbor Drive,
West Conshohocken, PA 19429
www.astm.org.
IIBEC Interface journal is seeking submissions for the following issues. Optimum article size is 2000 to
3000 words, containing five to ten high-resolution graphics. Articles may serve commercial interests
but should not promote specific products. Articles on subjects that do not fit any given theme may be
submitted at any time.
Submit articles or questions to Executive Editor Christian Hamaker at 800-828-1902
or chamaker@iibec.org.
ISSUE SUBJECT SUBMISSION DEADLINE
November 2022 Design of Cavity Walls July 15, 2022
December 2022 Technology August 15, 2022
January 2023 The Building Enclosure September 15, 2022
February 2023 Decks and Balconies October 15, 2022
March 2023 Energy Issues November 15, 2022
SIPL