Design Guide 40

Rain Loads
and
Ponding
Design Guide 40

Rain Loads
and
Ponding
Mark D. Denavit, PE, PhD
James M. Fisher, PE, PhD, Dist. M.ASCE

American Institute of Steel Construction

 © AISC 2024

by

American Institute of Steel Construction

All rights reserved. This book or any part thereof must not be reproduced
in any form without the written permission of the publisher.
The AISC logo is a registered trademark of AISC.

The information presented in this publication has been prepared following recognized principles of design
and construction. While it is believed to be accurate, this information should not be used or relied upon
for any specific application without competent professional examination and verification of its accuracy,
suitability and applicability by a licensed engineer or architect. The publication of this information is not a
representation or warranty on the part of the American Institute of Steel Construction, its officers, agents,
employees or committee members, or of any other person named herein, that this information is suitable
for any general or particular use, or of freedom from infringement of any patent or patents. All represen-
tations or warranties, express or implied, other than as stated above, are specifically disclaimed. Anyone
making use of the information presented in this publication assumes all liability arising from such use.

Caution must be exercised when relying upon standards and guidelines developed by other bodies and
incorporated by reference herein since such material may be modified or amended from time to time sub-
sequent to the printing of this edition. The American Institute of Steel Construction bears no responsibility
for such material other than to refer to it and incorporate it by reference at the time of the initial publication
of this edition.

Printed in the United States of America

Authors
Mark D. Denavit, Ph.D., P.E., is an associate professor at the University of Tennessee, Knoxville. He obtained his doctoral
degree from the University of Illinois at Urbana-Champaign and previously worked as a design engineer at Stanley D. Lindsey
and Associates, Ltd. in Atlanta, Georgia. He is a member of the AISC Committee on Specifications, Task Committee 5 on Com-
posite Design, and Task Committee 7 on Evaluation and Repair; a member of the ASCE/SEI 7-28 Snow and Rain Loads Sub-
committee; a professional engineer in the state of Georgia; and recipient of the AISC Terry Peshia Early Career Faculty Award.

James M. Fisher, Ph.D., P.E., Dist. M.ASCE, received a B.S. degree in civil engineering from the University of Wisconsin and
his M.S. and Ph.D. degrees in structural engineering from the University of Illinois at Urbana-Champaign. Dr. Fisher’s structural
expertise has resulted in authoring and co-authoring 16 books and numerous technical publications. He has lectured for AISC
throughout the United States and Mexico.
Dr. Fisher served as the chair of the AISC Committee on Specifications from 2003 to 2010. He continues to be active on several
AISC Specification Task Committees. He is an emeritus member of the AISI Cold-Formed Specification Committee. In addition,
Dr. Fisher serves as the consulting engineer to the Steel Joist Institute.
In 1984, Dr. Fisher received the T.R. Higgins Lectureship Award presented by AISC, and in 2000, he received the AISC Lifetime
Achievement Award. Both awards recognized his exceptional contribution to the steel construction industry. He received AISC’s
highest award, the J. Lloyd Kimbrough Award, in 2006. In 2020, Dr. Fisher received the AISC Steel Conference Speaker award.

Acknowledgments
The authors thank the American Institute of Steel Construction for funding the development of this document and for assistance
in its preparation. The authors thank Paxton Lifsey, master’s student at the University of Tennessee, Knoxville, for assistance
preparing figures. The authors thank the following reviewers of this document, including members of the American Institute of
Steel Construction and the Steel Joist Institute, for their constructive comments.

Josh Buckholt Sean Homem J.R. Mujagic

Michael Desch Larry Kruth Chris Raebel
Cindi Duncan Margaret Matthew Joe Voigt
Arvind Goverdhan

Preface
This AISC/SJI Design Guide is based on the 2022 edition of ASCE/SEI 7, Minimum Design Loads and Associated Criteria for
Buildings and Other Structures, the 2022 edition of the AISC Specification for Structural Steel Buildings, and the 2020 edition of
the SJI Standard Specifications for K-Series, LH-Series, and DLH-Series Open Web Steel Joists, and for Joist Girders. It provides
guidance for the design of roof systems to avoid or resist ponding instability and water accumulation. The motivation for this
Design Guide is recent changes to design standards to require consideration of ponding in the determination of rain loads and
not as a separate check. The Design Guide highlights specific areas of design for the benefit of architects, building inspectors,
building officials, designers, engineers, students, and others.

iii
iv
Table of Contents
CHAPTER 1 INTRODUCTION . . . . . . . . . . . . . . . . . . 1 CHAPTER 4 SJI ROOF BAY
ANALYSIS TOOL . . . . . . . . . . . . . . . . . . . . . . . 21
1.1 BEHAVIOR OF ROOFS UNDER
RAIN LOADS . . . . . . . . . . . . . . . . . . . . . . . . 1 4.1 INPUT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.2 CURRENT CODE PROVISIONS AND 4.2 ANALYSIS PROCEDURE . . . . . . . . . . . . . . 24
REQUIREMENTS . . . . . . . . . . . . . . . . . . . . . 3 4.3 OUTPUT . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.2.1 International Building Code . . . . . . . . . . 7 4.4 DESIGN PROCEDURE . . . . . . . . . . . . . . . . 25
1.2.2 International Plumbing Code . . . . . . . . . 7
1.2.3 Material-Specific Standards . . . . . . . . . . 7 CHAPTER 5 DESIGN EXAMPLES . . . . . . . . . . . . . 27
1.2.4 Provisions from FM Global . . . . . . . . . . 7
5.1 EXAMPLE 1—EXISTING DEAD
1.2.5 Ponding and Snow Loads . . . . . . . . . . . 8
FLAT ROOF . . . . . . . . . . . . . . . . . . . . . . . . 27
1.3 OBSOLETE METHODS OF DESIGN
5.2 EXAMPLE 2—OPEN WEB STEEL JOIST
FOR PONDING . . . . . . . . . . . . . . . . . . . . . . 8
ROOF DESIGN . . . . . . . . . . . . . . . . . . . . . . 30
5.3 EXAMPLE 3—STRUCTURAL STEEL
CHAPTER 2 RECOMMENDED METHOD OF
ROOF DESIGN . . . . . . . . . . . . . . . . . . . . . . 33
DESIGN FOR PONDING . . . . . . . . . . . . . . . . . 11
5.4 EXAMPLE 4—BAY WITH LOW SLOPE AND
2.1 DESIGN BASIS . . . . . . . . . . . . . . . . . . . . . 11 SECONDARY MEMBERS PERPENDICULAR
2.2 CALCULATION OF REQUIRED TO A FREE-DRAINING EDGE . . . . . . . . . . . 42
STRENGTH . . . . . . . . . . . . . . . . . . . . . . . . 11 5.5 EXAMPLE 5—BAY WITH LOW SLOPE AND
2.3 CALCULATION OF RAIN LOAD SECONDARY MEMBERS PARALLEL TO A
INCLUDING PONDING EFFECTS . . . . . . . . 11 FREE-DRAINING EDGE . . . . . . . . . . . . . . . 45
2.4 CALCULATION OF SNOW LOAD
INCLUDING PONDING EFFECTS . . . . . . . . 12 CHAPTER 6 CONCLUDING NOTES . . . . . . . . . . . 53
2.5 CALCULATION OF AVAILABLE
STRENGTH . . . . . . . . . . . . . . . . . . . . . . . . 12
SYMBOLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.6 BAYS WITH LOW SLOPE . . . . . . . . . . . . . . 13
2.7 SERVICEABILITY . . . . . . . . . . . . . . . . . . . 14
2.8 PRACTICAL CONSIDERATIONS . . . . . . . . . 14 GLOSSARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.9 RECOMMENDATIONS WHEN DRAIN
SIZES ARE NOT YET KNOWN . . . . . . . . . . 16 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
CHAPTER 3 METHODS OF ANALYSIS
FOR PONDING . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.1 CLOSED-FORM SOLUTIONS . . . . . . . . . . . 17
3.2 AMPLIFIED FIRST-ORDER ANALYSIS . . . . 17
3.3 NEGATIVE SPRING STIFFNESS . . . . . . . . . 18
3.4 ITERATIVE ANALYSIS . . . . . . . . . . . . . . . . 18

v
Chapter 1
Introduction
The 2022 editions of both the AISC Specification for Struc- to design for ponding of roofs constructed with structural
tural Steel Buildings, ANSI/AISC 360-22 (AISC, 2022), steel, open web steel joists, and joist girders.
hereafter referred to as the AISC Specification, and ASCE/
SEI 7, Minimum Design Loads and Associated Criteria for 1.1 BEHAVIOR OF ROOFS UNDER RAIN LOADS
Buildings and Other Structures (ASCE, 2022), hereafter
Ponding is a process where water, gravitating to low points in
referred to as ASCE/SEI 7, include significant changes rel-
the deflected surface of a roof system, causes progressively
evant to design for ponding. Most importantly, the appendix
increasing deflection and load. When sufficient strength and
included in previous editions of the AISC Specification that
stiffness are provided, the result of ponding is equilibrium
described methods of ponding analysis has been removed,
(Figure 1-1), albeit with greater loads in a different distribu-
and the definition of rain load in ASCE/SEI 7 has been
tion than would be expected based on the undeformed roof.
revised to be based on the deflected shape of the roof. The
When either the strength or stiffness is insufficient, the result
new provisions require a more realistic evaluation of the
of ponding is collapse (Figure 1-2).
effects of ponding that is consistent with other aspects of
Ponding is a potential problem in northern as well as
structural design, but one that may be unfamiliar to design
southern climates. Water can come from rain, snowmelt, or
professionals. The goal of this Design Guide is to help design
both. Heat from within a building can thaw areas of snow on
professionals properly and efficiently use the new provisions
a roof to produce melt water. The effect can be exacerbated

Fig. 1-1. Accumulation of water on a roof due to blocked drains (photo courtesy of Wiss, Janney, Elstner Associates, Inc.).

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 1

with repetitive freeze-thaw cycles. Furthermore, primary roof and the deflected surface of the roof, resulting in added
drains may become blocked, and water may accumulate load. The total load on the beam is:
faster than it can be discharged by secondary drains.
Because impounded water conforms to the deflected w ( x ) = wi + γ S y ( x ) (1-1)
surface of the roof, the load from the water is not uniform
across the roof system, which may pose further challenges where
for systems that are designed for uniform dead, live, or snow S = beam spacing, in.
loads as is common for steel joist roofs. w(x) = total load on beam, kips/in.
To explore the behavior of roofs subjected to rain loads in y(x) = total deflection along span, in.
more detail, consider a flat roof consisting of equally spaced γ = u nit weight of water, 62.4 lb/ft3 = 3.61 × 10 -5
structural steel beams or open web steel joists—hereafter kips/in.3
collectively referred to as steel beams—spanning between
two walls. Assuming the deck does not deflect between the Assuming small displacements, the governing differential
steel beams, the behavior of this roof can be captured in a 2D equation is formed noting that loading is proportional to the
analysis of a simply supported beam. The beam has length, fourth derivative of the deflection.
L, flexural stiffness, EI, and uniformly distributed load, wi,
as shown in Figure 1-3(a). The uniformly distributed load EIy′′′′ ( x ) = wi + γ Sy ( x ) (1-2)
consists of some combination of dead load and the weight of
water above the undeflected surface of the roof. Neglecting Note that Equation 1-2 does not include shear deforma-
ponding, the maximum moment in the beam is wi L2/ 8, and tions, but the additional flexibility from shear deformations
the maximum deflection is 5wi L4/ 384EI. However, water can be included in an approximate manner by reducing the
will fill the volume between the undeflected surface of the moment of inertia. Silver (2010) derived the solution to this
differential equation as:

Fig. 1-2. Partial roof collapse due to ponding (photo courtesy of Wiss, Janney, Elstner Associates, Inc.).

2 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

 ⎧ ⎡ ⎤ ⎡ ⎤ ⎫ γ SL4
⎪ cos ⎢ 4 γ S ⎛ L − x⎞ ⎥ cosh ⎢ 4 γ S ⎛ L − x⎞ ⎥ ⎪ C=
wi ⎪ EI ⎝ 2 ⎠ EI ⎝ 2 ⎠ ⎪ π 4 EI
⎣ ⎦ ⎣ ⎦
y (x) = ⎨ + − 2⎬ 3 4
( 62.4 lb/ft3 )⎛ 1,000 lb ⎞ ⎛ 12 in.⎞ (10 ft ) ⎛ 1 ft ⎞ ⎡⎢( 40 ft ) ⎛ 1 ft
2γ S ⎪ ⎡ γS ⎛ L ⎞ ⎤ ⎡ γS ⎛ L ⎞ ⎤ 1 kip 1 ft 12 in. 12 in.⎞ ⎤
⎪
⎪ cos ⎢ EI ⎝ 2 ⎠ ⎥
4 cosh ⎢ 4 ⎥ ⎪ ⎝ ⎠⎝ ⎠ ⎝ ⎠⎣ ⎝ ⎠ ⎥⎦
⎩ ⎣ ⎦ ⎣ EI ⎝ 2 ⎠ ⎦ ⎭ =
π ( 29,000 ksi ) (199 in. )
4 4

(1-3)
= 0.409

An important observation from Equation 1-3 is that the (1-4)

deflections approach infinity as the value of the flexibility
coefficient, C, given by Equation 1-4, approaches unity. For this case with 10 lb/ft2 dead load and 2 in. of water above
the undeflected surface of the roof, the deflection of the beam
γ SL4 is shown in Figure 1-3(b) and the bending moment is shown
C= (1-4) in Figure 1-3(c). This case, with a span-to-depth ratio of 35, is
π 4 EI 
likely too flexible for any practical situation. However, the flex-
In other words, C = 1 is the limit of elastic ponding stabil- ibility coefficient is well below the limit of elastic ponding sta-
ity for this problem. However, even when C < 1, the applied bility. The more pressing design consideration is the increased
load, deflections, bending moments, etc., will be greater than required strengths.
if ponding is not considered. For example, if the roof beams
are W14×22 spanning 40 ft and spaced at 10 ft, then C = 1.2 CURRENT CODE PROVISIONS AND
0.409. REQUIREMENTS
Given the multidisciplinary nature of roof drainage, rain loads,
and ponding, many standards include provisions that are relevant
to the design and assessment of roof systems. The provisions

wi

(a) Structure

0 100
Including ponding effect
No ponding effect
Bending moment (kip-in.)

-1 80
Deflection (in.)

-2 60

-3 40

-4 20
Including ponding effect
No ponding effect
-5 0
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
Position along length of beam (ft)        Position along length of beam (ft)

(b) Deflection (c) Bending moment

Fig. 1-3. Ponding on an example flat, simply supported W14× 22 beam (C = 0.409).

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 3

in ASCE/SEI 7 are particularly relevant for structural engi- of the means of discharge and could include a system with
neers and provide a good starting point for discussion. drains and piping, scuppers, or a free draining edge.
ASCE/SEI 7, Section 8.2, specifies rain loads as the Figure 1-4 shows a roof drain (i.e., the primary drainage
weight of water supported by the roof in a condition where system) and a scupper (i.e., the secondary drainage system).
vulnerable drainage systems are blocked, rainwater has The vertical separation between the inlet of the drain and
accumulated to the next point of discharge, and rainwater is the bottom of the scupper (i.e., the invert elevation) is large
being discharged from the roof at a rate computed based on enough that the scupper can act as the SDSL for this roof.
a specific storm intensity. One reason for the vertical separation is so that activation of
Vulnerable drainage systems are defined as (1) the pri- the SDSL—for example, as shown in Figure 1-5—serves as
mary drainage system, (2) secondary drainage systems with a warning that the primary drainage system is blocked and in
an inlet that is vertically separated from the inlet to the pri- need of maintenance.
mary drainage system by less than 2 in., (3) secondary drain- Note that the provisions of ASCE/SEI 7 do not prohibit
age systems that share drain lines with the primary drainage the installation or use of secondary drainage systems that
system, and (4) secondary drainage systems with controlled would be classified as vulnerable. The provisions of ASCE/
flow roof drains. The drainage system through which water SEI 7 only require that such systems are assumed to be
is discharged when all vulnerable drainage systems are blocked when determining rain load. Poor design team coor-
blocked is called the secondary drainage system for struc- dination and use of secondary drainage systems that are clas-
tural loading (SDSL). The term SDSL is used regardless sified as vulnerable per ASCE/SEI 7 may result in very large
rain loads. In an extreme case where no secondary drainage

Fig. 1-4. Roof drain and adjacent scupper.

4 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

system is provided or where the secondary drainage system the static head requires information on the elevation of the
is vulnerable, the next point of discharge may be flow over SDSL with respect to the roof surface and the roof slope,
a parapet. Note that consideration of extreme cases such both of which are typically not designed by the structural
as this are not required per the National Building Code of engineer and may not be known at the time of structural
Canada, where rain load is limited by the one-day rainfall design. Recommendations for design when the drainage
(NRCC, 2020). system is not yet known are presented in Section 2.9 of this
Rain load is computed as the weight of water that will Design Guide. Camber can be considered when determining
be present on the roof in the design condition. The equa- the static head, and it is often beneficial to do so.
tion provided in ASCE/SEI 7 for computing the rain load The hydraulic head, dh, is the depth of water on the unde-
(ASCE/SEI 7, Equation 8.2-1, or Equation 1-5 herein) is the flected roof above the inlet of the SDSL required to achieve
unit weight of water multiplied by the total depth of water, the design flow. The design flow rate, Q, is computed as
which is categorized into three components: the static head, a rainfall intensity multiplied by a tributary area (ASCE/
ds, the hydraulic head, dh, and the ponding head, dp. SEI 7, Equation C8.2-1, or Equation 1-6 herein).

R = 5.2 ( ds + d h + dp) (1-5) Q = 0.0104 Ai (1-6)

Each component of the depth of water can vary over the roof where
surface, as can be seen in Figure 1-6. As a result, rain load is A = tributary area, ft2
typically not a uniform load.
Q = flow rate, gal/min
The static head, ds, is the depth of water from the unde-
flected roof surface up to the inlet of the SDSL. Determining i = rainfall intensity, in./hr

Fig. 1-5. Water flowing from scuppers, likely due to a blocked primary drainage system
(photo courtesy of Wiss, Janney, Elstner Associates, Inc.).

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 5

 The rainfall intensity is that which corresponds to the where
15-min duration storm with the return period based on the Q = flow rate, gal/min
risk category of the structure. Structures in risk categories Ws = scupper width, in.
I and II use a design storm return period of 100 yr, struc- cd = coefficient of discharge = 0.6
tures in risk category III use a design storm return period of
dh = hydraulic head, in.
200 yr, and structures in risk category IV use a design storm
return period of 500 yr. The Hydrometeorological Design g = acceleration of gravity = 386 in./s2
Studies Center at the National Weather Service, an agency The ponding head, dp, is the depth of water due to deflec-
of the U.S. government and part of the National Oceanic and tions of the roof subjected to unfactored rain load and unfac-
Atmospheric Administration (NOAA), maintains the Pre- tored dead load. It is determined using structural analysis.
cipitation Frequency Data Server at https://hdsc.nws.noaa A recommended procedure for determining ponding head is
.gov/pfds (NOAA, 2023). This server provides rainfall inten- presented in Chapter 2 of this Design Guide.
sity data based on geographic location for various durations The inclusion of the ponding head when computing rain
and recurrence intervals, including those needed for deter- load was a change made in the 2022 edition of ASCE/SEI 7.
mination of rain loads per ASCE/SEI 7. Note that the data on In previous editions of ASCE/SEI 7, the rain load was based
this server are generally provided in inches of rainfall for the only on the static head and the hydraulic head. In other
specified duration and need to be converted to units of inches words, the rain load was based on the weight of water on
per hour by dividing the expected rainfall by the duration. the undeflected roof. Neglecting the ponding head underes-
As described in the ASCE/SEI 7 Commentary, the tribu- timated rain loads and required a separate check for pond-
tary area for a drain should include the appropriate roof area ing instability. Inclusion of the ponding head eliminates the
plus one-half the wall area that diverts rainwater onto the need for a separate check for ponding instability in bays that
roof. accumulate water. An analysis by Denavit and Scott (2021)
Once the flow rate for the drain is computed, the hydrau- confirmed that use of the modified rain load results in a suf-
lic head, dh, is determined based on hydraulic test data or ficient level of reliability.
hydraulic calculations. The ASCE/SEI 7 Commentary Bays without accumulated water may still be susceptible
in­cludes several tables that relate the flow rate and hydraulic to ponding if the roof slope is small enough that low points
head. The tables for drains and rectangular scuppers were form in the deflected shape of the roof, allowing water to
reproduced from FM Global Property Loss Prevention Data accumulate. ASCE/SEI 7, Section 8.3, requires that bays
Sheet 1-54, Roof Loads and Drainage (FM Global, 2021). with low slope be investigated and provided with sufficient
The values in the table for open-topped scuppers can be stiffness and strength to preclude ponding instability. A rec-
replicated with high accuracy using Equation 1-7 from fluid ommended procedure for evaluating bays with low slope for
mechanics for flow through a rectangular weir (Finnemore ponding is presented in Section 2.6 of this Design Guide.
and Franzini, 2001). ASCE publishes a guide to the rain load provisions of
q ASCE/SEI 7 (O’Rourke and Lewis, 2020) that provides
⎛ 3.85Q ⎞ additional background information and many helpful exam-
dh = ⎜ ⎟ (1-7)
⎝ qcd Ws 2g ⎠  ples. However, note that the edition of the ASCE guide pub-
lished in 2020 is based on the 2016 edition of ASCE/SEI 7

Secondary drainage
(scupper) dh = 0
ds = 0
Undeflected
dh shape of roof

ds
dp

Deflected
Primary drainage
Wall shape of Wall
Blocked roof

Fig. 1-6. Schematic illustrating the three components of rain load.

6 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

(ASCE, 2016), which has significant differences from the the rainfall intensity for sizing secondary drains per Sec-
2022 edition of ASCE/SEI 7. tion 1108.3 of the International Plumbing Code is differ-
ent than that defined in ASCE/SEI 7 for the determination
1.2.1 International Building Code of the hydraulic head. Communication with the plumbing
engineer to determine the most efficient solution overall may
The International Building Code (ICC, 2021a), hereafter
be required if the drainage design results in large rain loads.
referred to as IBC, includes additional provisions relevant to
roof drainage, rain loads, and ponding.
1.2.3 Material-Specific Standards
IBC Section 1502 addresses roof drainage. Key provi-
sions in this section include (1) the design and installation of The AISC Specification includes requirements for design for
roof drainage systems must comply with Chapter 11 of the ponding in Section B3.10:
International Plumbing Code (ICC, 2021b), (2) secondary
The roof system shall be investigated through structural
(emergency overflow) drains or scuppers are required when
analysis to ensure stability and strength under ponding
roof drains are required, and (3) scuppers must be at least
conditions unless the roof surface is configured to pre-
4 in. wide. Additionally, the dependance of rain loads on the
vent the accumulation of water.
roof drainage system is emphasized through multiple refer-
Ponding stability and strength analysis shall consider
ences to IBC Section 1611.
the effect of the deflections of the roof’s structural fram-
IBC Section 1611 addresses rain loads. The provisions for
ing under all applicable loads present at the onset of
rain load in IBC Section 1611.1 are similar to those in the
ponding and the subsequent accumulation of rainwater
2016 edition of ASCE/SEI 7 but do not include the pond-
and snowmelt.
ing head. Accordingly, IBC Section 1611.2 requires a sepa-
The nominal strength and resistance or safety factors
rate evaluation for ponding instability. IBC Section 1611.1
for the applicable limit states are specified in Chapters
also does not include greater design storm return periods
D through K.
for structures assigned to risk categories III and IV or the
requirement that vulnerable secondary drainage systems be Prior to the 2022 edition, the AISC Specification included,
assumed to be blocked. Note, however, that determining rain in Appendix 2, methods of evaluating stability and strength
loads in accordance with ASCE/SEI 7, Chapter 8, satisfies of a roof system under ponding conditions. These methods
the requirements of IBC Section 1611.1, because the rain were widely referenced in other standards but are no lon-
loads are greater, and IBC Section 1611.2, because inclu- ger included in the AISC Specification. In fact, given the
sion of ponding head in rain loads is a rational approach for requirement that available strength be determined using
ensuring that the roof possesses adequate stiffness to pre- Chapters D through K, the methods of design for ponding
clude progressive deflection and adequate strength to resist presented in Appendix 2 of prior editions of the AISC Speci-
the additional ponding load. fication are no longer applicable. Section 1.3 of this Design
IBC Section 1507 addresses requirements for roof cov- Guide includes further discussion of the methods of design
erings, including minimum slope for different types of for ponding that were included in previous editions of the
construction. For example, IBC Section 1507.10.1 states AISC Specification.
“Built-up roofs shall have a design slope of not less than AISI S100 (AISI, 2016) has a provision that mirrors that
4 unit vertical in 12 units horizontal (2-percent slope) for of the AISC Specification.
drainage, except for coal-tar built-up roofs that shall have a SJI Standard Specifications for K-Series, LH-Series, and
design slope of not less than 8 unit vertical in 12 units hori- DLH-Series Open Web Steel Joists, and for Joist Girders,
zontal (1-percent slope).” The minimum slope of 4 in. per ANSI/SJI 100-2020 (SJI, 2020), hereafter referred to as the
foot is applicable to other types of roof construction as well, SJI Specifications, Section 5.11, states “The ponding inves-
including modified bitumen roofs, single-ply membrane tigation shall be performed by the specifying professional.”
roofs, sprayed polyurethane foam roofs, and liquid-applied Like the provisions in the AISC Specification and AISI S100,
roofing. this provision highlights the importance of ponding, but also
places the responsibility for ensuring the roof is specified to
1.2.2 International Plumbing Code have sufficient stiffness and strength on the specifying pro-
fessional, and not the joist manufacturer.
Structural engineers generally do not need to be familiar with
the provisions in the International Plumbing Code. How-
1.2.4 Provisions from FM Global
ever, provisions in Chapter 11 of the International Plumbing
Code govern the design and installation of roof drains and FM Global is a mutual insurance company that provides
may be helpful to approximate the number and size of the insurance to a significant percentage of the commercial and
drains on a roof for determination of rain loads before the industrial building market in the United States. FM Global
plumbing engineer has performed the design. Importantly, Property Loss Prevention Data Sheet 1-54, Roof Loads

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 7

and Drainage (FM Global, 2021) does not apply to all new Two methods were presented; both evaluated individual
construction but contains additional requirements for many susceptible bays. The first of the two methods, “simplified
buildings. The additional requirements include: design for ponding,” was based only on the stiffness of the
• Minimum rain loads must be based on 6 in. depth of water roof bay. In effect, it stated that if the stiffness of the bay
at the low point of the roof. was sufficiently high, then the increased load due to ponding
could be neglected. The second of the two checks, “improved
• The inlet elevation of overflow drains and the invert eleva-
design for ponding,” considered both stiffness and aspects
tion of overflow scuppers must be at least 2 in. above the
of the strength of the roof bay. In effect, it evaluated the
low point of the roof surface.
increased flexural demands on the primary and secondary
• Roofs and their drainage inlets must be inspected after members and compared those demands to the yield moment
roof construction, prior to the start of the rainy or tropi- with a factor of safety of 1.25. A susceptible bay was consid-
cal cyclone seasons, following storms, and at least every ered adequate if either of the two methods indicated that the
3 months. Obstructions or accumulations of foreign mat- bay was adequate.
ter must be cleared as frequently as necessary. Underlying these methods was a closed-form solution for
ponding analysis of an idealized interior roof bay. There-
1.2.5 Ponding and Snow Loads fore, the methods were only strictly applicable under the
assumptions used to developed that closed-form solution.
The codes and standards also have provisions related to
The assumptions include:
ponding and snow loads. The weight of snow places signifi-
cant demands on roofs and often controls roof design. Snow • The roof is perfectly flat (i.e., no slope, members are not
combined with water, either rain or snowmelt, places even cambered, and tapered insulation is not considered).
greater demands and introduces the potential for ponding. • The bay under consideration is rectangular.
Snow slows but does not stop the flow of water on the roof, • The adjacent bays are identical to that of the bay under
resulting in a challenging condition to analyze and one that consideration.
is not fully understood. As before, the provisions of ASCE/
SEI 7 provide a good starting point for discussion. • The primary members are of equal stiffness.
ASCE/SEI 7, Section 7.10, specifies a rain-on-snow sur- • The secondary members are uniformly spaced and of
charge load of 8 lb/ft2 that applies when the ground snow equal stiffness.
load is less than, or equal to, the minimum snow load for • All members are simply supported.
low-slope roofs, but not zero, and when the roof slope (in
degrees) is less than the horizontal distance from eave to • Axial loads are not present in any member.
ridge (in feet) divided by 50. This surcharge is based on • Water covers the entire bay.
analysis of flow though snowpack on rigid roofs (O’Rourke
and Downey, 2001). These idealized conditions seldom exist in real structures.
ASCE/SEI 7, Section 7.11, specifies that susceptible bays Additionally, potentially important aspects of design are
must be designed to preclude ponding instability. It further neglected in their use.
notes that roof deflections caused by full snow loads shall be The assumption of a perfectly flat roof is particularly prob-
evaluated when determining the likelihood of ponding insta- lematic given that a minimum slope is required by the IBC
bility. IBC Section 1608.3 echoes this requirement. (ICC, 2021a) and steel joists and joist girders are typically
constructed with camber. SJI Technical Digest 3 (Fisher
1.3 OBSOLETE METHODS OF DESIGN and Denavit, 2018) lists several adaptations to the Appen-
FOR PONDING dix 2 methodologies that expand their range of applicability
including to low-sloped roofs. However, these adaptations
Prior to the 2022 editions of ASCE/SEI 7 and the AISC are based largely on engineering judgment and lack rigorous
Specification, rain loads were computed based on the unde- justification.
flected roof, and a separate ponding analysis was performed, The evaluation of available strength is inconsistent and
typically in accordance with the methods presented in incomplete in the Appendix 2 methods. For example, roofs
Appendix 2 of previous editions of the AISC Specification that satisfy the requirements of the simplified method can
(e.g., AISC, 2016). These methods were based on the work still see load amplifications due to ponding of up to 40%
of Marino (1966). At the time of development, the methods (Denavit, 2019). Even when strength is considered in the
were a significant advancement because they considered the improved method, only the limit state of flexural yielding is
flexibility of both the primary and secondary members of a considered, and the strength is not consistent with the provi-
roof framing system. sions of AISC Specification Chapter F. No check is made for
moment and shear strength that vary along the length (as is

8 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

common in joists and joist girders), nor are the strengths of loads, additional potential limit states arise that are not nec-
the connections checked. essarily captured by the moment and shear strength enve-
The Appendix 2 method was originally developed for lopes. For joists, a region of high applied distributed load
roof systems with solid web steel beams and girders where can cause bending failure between the panel points. For joist
a check of the maximum moment is sufficient to assess girders, high applied panel point loads can cause failure of
strength. Open web steel joists and joist girders are typi- web verticals.
cally designed for uniform loads. Accordingly, their shear The improved method in Appendix 2 is only intended for
and moment strength can vary along their length, requiring use with ASD load combinations. No provisions are pro-
strength to be checked along the entire length. This is espe- vided for use with LRFD; however, Carter and Zuo (1999)
cially important because (1) the maximum moment experi- provide guidance for ponding calculations in LRFD.
enced under ponding conditions may not occur at mid-span, While the Appendix 2 method was pioneering when first
where the available moment strength is at a maximum, and developed in the 1960s, it has seen little modification in the
(2) shear reversals near mid-span can occur under ponding decades since despite advances and changes in nearly every
conditions causing web members that are designed for ten- other aspect of structural design. The changes made in the
sion to be subjected to compression. 2022 editions of ASCE/SEI 7 and the AISC Specification
Furthermore, when joists and joist girders that have been to design for ponding by direct computation of load bring
designed for uniform loads are subjected to nonuniform much needed consistency in requirements.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 9

10 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40
Chapter 2
Recommended Method of Design for Ponding
Noting the limitations of the traditional methods of design approach for investigating these bays is described in Sec-
for ponding, a more general method of design for ponding tion 2.6 of this Design Guide.
that is compliant with current standards is presented in this Bays that do not have accumulated water and do not have
chapter. Methods of analysis that can be used with the rec- low slope will not be subjected to rain loads nor will they be
ommended method of design are described in Chapter 3. susceptible to ponding instability. No further investigation is
An implementation of the method of design within the SJI needed for these bays.
Roof Bay Analysis Tool (SJI, 2022) is described in Chap-
ter 4. Design examples using the method are presented in 2.2 CALCULATION OF REQUIRED STRENGTH
Chapter 5.
In the recommended method of design for ponding, the
effects of ponding are captured within the rain load, R, and
2.1 DESIGN BASIS
the snow load, S. Once R and S including ponding effects are
Rain loads and ponding are only applicable to portions of computed, these loads can be used in basic load combina-
roofs that can accumulate water. Therefore, the first step is to tions as defined in ASCE/SEI 7 with no further consideration
determine the level to which water will rise on the roof. Per of ponding. Alternatively, because linear elastic analysis will
the requirements of ASCE/SEI 7, Section 8.2, this will be typically be appropriate for roof systems, once R and S have
the hydraulic head above the inlet to the SDSL. The next been determined, load effects from the ponding analysis to
step is to identify which portions of the roof will accumulate determine R or S may be combined with other load effects
water if water rises to this level. Often it is appropriate to via superposition.
assess a roof one bay at a time.
For bays that accumulate water in the design condition, 2.3 CALCULATION OF RAIN LOAD INCLUDING
the rain and snow loads will be computed considering the PONDING EFFECTS
deformed shape of the roof system. Once computed, the
The rain load, R, including ponding effects should be deter-
rain and snow loads are used in standard strength load com-
mined in accordance with ASCE/SEI 7, Chapter 8, using
binations to determine required strengths, which are then
Equation 1-5:
compared to available strengths per AISC Specification Sec-
tion B3, SJI Specifications Section 4.2.1 or 4.2.2, etc. If the
R = 5.2 (d s + d h + dp) (1-5)
available strength of each structural component equals or
exceeds the required strength, no further ponding checks are The static head, ds, the hydraulic head, dh, and the pond-
required. Otherwise, the roof design needs to be revised. The ing head, dp, can all vary across a roof system. By contrast,
following modifications can be implemented: the water level in the design condition (i.e., hydraulic head
• Stiffen the roof system to reduce the additional load above the SDSL) does not vary and provides a convenient
caused by ponding. definition of rain load. The selection of drain types, loca-
• Strengthen the roof system so that it can support the addi- tions, and sizes is generally not the responsibility of the
tional load caused by ponding. structural engineer; however, the structural engineer must
either evaluate or be provided with the characteristics of the
• Adjust the shape of the roof surface to reduce the addi-
SDSL in order to determine the rain load.
tional loads caused by ponding (roof slope, camber,
It is important for the design professional to have a physi-
tapered insulation, sloping fill, etc.).
cal understanding of the loading condition assumed in the
• Increase drainage capacity (e.g., add drains near columns). design, including how water flows, accumulates, and is
• A combination of the above. eventually discharged from the roof. Outcrops of building
lines along the low side of a sloped roof can lead to potential
For bays that do not accumulate water in the design con- problems if the flow of water is not controlled (see ASCE/
dition (i.e., their elevation is greater than the design water SEI 7, Figure C8.2-1).
level), it is still possible for water to accumulate if the slope The ponding head should be determined from an elastic
of the bay is low enough that deflections cause low points analysis that considers ponding nonlinearities, specifically
in the roof surface. Identified in ASCE/SEI 7, Section 8.3, that the depth of water is dependent on the deflection of the
are cases that require further investigation. A recommended roof. In general, geometric nonlinearity (e.g., P-Δ effects)

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 11

need not be captured. Several alternative methods of analysis load (i.e., the hydraulic head above the inlet to the SDSL).
are presented in Chapter 3 of this Design Guide. The analy- This condition is based on a situation where the primary
sis should consider all relevant member and component drainage system is blocked. Noting the reduced probability
deformations. For open web steel joists and joist girders, it of a roof system experiencing both full snow loads and a
is customary to account for the effect of shear deformations blocked drain simultaneously, a reduced snow load is used.
by dividing the moment of inertia by a factor of 1.15. For Investigations performed by the authors have indicated that
structural steel members, it is recommended to apply a stiff- this condition is generally conservative with respect to the
ness reduction factor of 0.8 to account for partial yielding of methods presented in Appendix 2 of previous editions of
the cross section. the AISC Specification for flat roofs and roughly equiva-
If a convergent solution cannot be obtained from the pond- lent to results from the direct analysis method for ponding
ing analysis, this indicates that the roof is not stiff enough to described in SJI Technical Digest 3 (Fisher and Denavit,
avoid ponding instability. The roof design must be revised to 2018) for sloped roofs. The design professional may con-
increase stiffness. sider alternative conditions.
The ponding analysis should be performed with unfac- Both the impounded water load and snow load represent a
tored dead load and unfactored rain load. For the D + R physical quantity of material above the roof surface. In situ-
load combination in ASD, the load effects from the pond- ations where these physical quantities overlap, as shown in
ing analysis can be used directly as required strengths. For Figure 2-1, the load can be reduced to account for the fact
other load combinations and LRFD, the rain load, R, is the that the unit weight of a mix of water and snow is less than
weight of water determined from ponding analysis. Alterna- the unit weight of water plus the unit weight of snow. It is
tively, the load effects due to rain load, R, can be isolated by recommended that the unit weight of the region where water
subtracting out load effects due to dead load, D, from D + and snow overlap be taken as the unit weight of water. The
R. The resulting isolated load effects due to rain load can be unit weight (density) of snow is defined by ASCE/SEI 7,
combined with other load effects via superposition. Equation 7.7-1. It is conservative to neglect the overlap and
consider the water and snow loads independently.
2.4 CALCULATION OF SNOW LOAD
INCLUDING PONDING EFFECTS 2.5 CALCULATION OF AVAILABLE STRENGTH
Provisions for evaluating ponding in the presence of snow Available strengths for structural steel members and con-
are less specific than for ponding without snow. Ponding in nections, open web steel joists, joist girders, and deck are
the presence of snow is a complex physical phenomenon determined in accordance with material standards, such as
that is currently not well understood. The recommendations the AISC Specification, SJI Specifications, or the SDI Stan-
presented here are a logical extension of the provisions for dard for Steel Deck, ANSI/SDI SD-2022 (SDI, 2022), just
ponding without snow. as for any other loading condition. Note that the increased
Like ponding without snow, where the effects of ponding allowable stress (i.e., 0.8Fy) used in the obsolete Appendix 2
were included in the rain load, R, the effects of ponding in method is not applicable.
the presence of snow will be included in the snow load, S, For joists and joist girders, moment and shear need to be
as part of the recommended design methodology. The snow checked along the entire length, and the applied loads need
load, S, including ponding effects should be considered as an to be checked to prevent local overstress. Steel joists and
additional case of the snow load to be evaluated along with joist girders are typically designed for uniform loads. Thus,
other cases determined in accordance with ASCE/SEI 7, generally, the available moment and shear strength vary
Chapter 7. along the length. KCS-Series joists are an exception where
Stated in ASCE/SEI 7, Section 7.11, “Susceptible bays the available shear and moment strength is constant along
shall be designed to preclude ponding instability. Roof the length of the joist.
deflections caused by full snow loads shall be evaluated Typical moment and shear envelopes for joists are shown
when determining the likelihood of ponding instability.” in Figure 2-2. The moment envelope is a parabolic shape
This provision was written when the methods presented in with a maximum value of wL2/ 8, where w is the total uni-
Appendix 2 of previous editions of the AISC Specification formly distributed load-carrying capacity given in the SJI
(e.g., AISC, 2016) were among the only options for ponding Load Tables and L is the design length of the joist. The shear
analysis. It is recommended that full snow loads be used to envelope has a maximum value of wL/ 2 and follows a para-
determine if low points where water can accumulate form on bolic equation. A minimum value of 25% of the maximum
the roof. However, for strength design, it is recommended end reaction is given based on the requirements of SJI Speci-
that 75% of the controlling uniform design snow load be fications Section 4.4.2 for shear in the design direction.
used with a surplus load equal to the additional weight of Depending on the roof slope, there may be conditions
water rising to a level equal to that computed for the rain where ponding could create an unbalanced loading condition

12 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

with a much heavier ponding load toward one end of the select a larger joist, specify a larger joist girder panel point
joist. When this occurs, there is potential for stress reversals load, or take no action if the applied loads are determined by
in the joist webs near midspan. All web members inherently analysis or judgement to not cause overstress.
have some stress reversal capacity, but for unbalanced pond-
ing conditions such as this, to ensure the joist web members 2.6 BAYS WITH LOW SLOPE
have adequate stress reversal capacity, it is recommended to
The surest way to avoid a ponding collapse is to construct
specify the joists be designed for a minimum shear equal
a roof with sufficient slope and free drainage so that water
to 12.5% of the maximum end reaction. The SJI Roof Bay
never accumulates. If the slope is small on the undeflected
Analysis Tool (described in Chapter 4 of this Design Guide)
roof, low points where water can accumulate can form when
assumes a minimum vertical shear reversal strength of
loads are applied.
12.5% of the end reaction.
ASCE/SEI 7, Section 8.3, states:
Typical moment and shear envelopes for a joist girder
with eight joist spaces are shown in Figure 2-3. The moment Free-draining bays and internal bays not subjected to
envelope is a multilinear shape with a maximum value in the accumulated rain load required by Section 8.2 shall
center based on the panel point loads. The shear envelope is a have adequate strength and stiffness to preclude pro-
multilinear stair-step shape with a minimum value of 25% of gressive deflection (i.e., instability) and resist poten-
the maximum end reaction based on the requirements of SJI tial ponding rain loads where either of the following
Specifications Section 4.4.2. For shear reversals, the strength conditions are met: 1. The roof slope is less than 4 in:
is 25% of the shear envelope in the opposite direction based per foot (1.19 degrees), or 2. The bay is adjacent to a
on the requirements of SJI Specifications Section 4.4.2.2. free-draining edge with secondary members parallel to
Due to the nonuniform nature of ponding loads, the applied the free-draining edge and the roof slope is less than β
load on a portion of the length of a joist or joist girder can where β = ( Ls S + π ) 20 (in. per foot).
occasionally exceed the allowable load, even when the shear
Having a roof slope less than the limits prescribed in
and bending moment diagrams fall within their respective
ASCE/SEI 7, Section 8.3, does not mean that low points
design envelopes. In such instances, there is the potential
will form. The limits are intended to identify bays requiring
for overstress due to top chord bending between the panel
further investigation. The first step of the investigation is to
points for joists or compression failure of web verticals for
determine if low points will indeed form on the deflected
joist girders. In these cases, the specifying professional may

el
Lev
Sno
w nly
wO
Water Level Sno
Water
Only now
nd S oof
a te r a
p e of R
W Sha
f l e cted
e
Und

(a) Undeformed

l
ve
Le ly
Sn
o w On
o w
Water Level Sn
Water Only
Snow
r and Roo
f
Wate ap e of
d Sh
ecte
Defl

(b) Deformed

Fig. 2-1. Schematic of water and snow loading.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 13

roof surface. This can be achieved by performing a struc- 2.7 SERVICEABILITY
tural analysis of the roof system under dead and full snow
Ponding, as defined in ASCE/SEI 7, is a strength design con-
load. Noting that water load is not included in this step, the
sideration. The method of design presented in this chapter is
analysis does not need to consider ponding nonlinearity. The
intended to prevent collapse and does not guarantee eventual
analysis should consider deflections of primary and second-
drainage of the water from the roof. Lengthy exposure to
ary members, roof slope, camber, and roofing materials of
water can be detrimental to a roof assembly and increases
nonuniform thickness. If the analysis indicates that positive
the chance of water penetrating the roof membrane. To
slope to the free-draining edge or adjacent bay exists, even
ensure eventual drainage, a positive slope to the drain or
under load, then it has been shown that ponding will not
free-draining edge must be provided throughout the roof
occur, and no further investigation is necessary.
system. Deflections should be considered when evaluating
If the analysis indicates that a low point forms in the
the slope as described in Section 2.6 of this Design Guide. A
interior of the bay (i.e., there is not positive slope to the
tapered insulation system or insulating fill can also be used
free-draining edge or adjacent bay), then ponding will occur,
to help provide adequate slope to drain (Wilen, 2012). Addi-
and further analysis is required. If low points form, a pond-
tionally, maintenance of drains and scuppers to keep debris
ing analysis should be performed as described in this chap-
from accumulating should be routine.
ter. The water level for this analysis should be selected as
a nominal height (e.g., 2 in.) above the low point on the
2.8 PRACTICAL CONSIDERATIONS
perimeter of the bay.
The preceding discussion assumes that the roof is free The recommended method of design for ponding presented
of ice dams that would inhibit the flow of water to the in this chapter is a rational approach to address the strength
free-draining edge. Alternative methods are necessary when of roofs supporting water from rain or snowmelt. However,
evaluating roofs where ice dams may form. it requires the design professional to have a physical under-
standing of the assumed loading condition, including how

wL2/8

(a) Moment envelope

(b) Shear envelope

Fig. 2-2. Strength envelopes for a steel joist.

14 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

water flows, accumulates, and is eventually discharged from The approximate gross moment of inertia of a joist girder
the roof. The method also implies that aspects of roof design can be determined using Equation 2-2 for ASD or Equa-
not typically associated with strength, such as deflections, tion 2-3 for LRFD.
roof slopes, camber, and the locations and sizes of roof
drains, can strongly influence the safety and reliability of the I jg = 0.027NPc Ld (ASD) (2-2)
roof system. Therefore, the design professional should treat
these aspects as they would any other aspect that influences I jg = 0.018NPc Ld (LRFD) (2-3)
strength.
The approximate gross moment of inertia of a K, LH, where
or DLH-Series joist can be determined using Equation 2-1 Ijg = approximate gross moment of inertia (not adjusted
which can be found in the introduction to the Standard Load for shear deformations), in.4
Tables (SJI, 2020). L = design length of joist girder (span − 3 ft), ft
N = number of joist spaces
I j = 26.8 × 10 − 6 wL 360 L3 (2-1)
Pc = specified panel point load, kips
d = depth of the joist girder, in.
where
Ij = approximate gross moment of inertia (not Open web steel joists and joist girders can be expected
adjusted for shear deformations), in.4 to have approximately 15% more deformation than a solid
L = design length of joist (span − 3 ft), ft web member due to greater shear deformation. However,
wL/360 = nominal live load which will produce an approxi- many of the equations and methods described in this Design
mate deflection of 1/360 of the span (red figure Guide were developed neglecting shear deformations (i.e.,
in the Load Table), lb/ft Bernoulli-Euler beam theory). In these situations, a reduced

Mmax

(a) Moment envelope

(b) Shear envelope

Fig. 2-3. Strength envelopes for a joist girder (example with eight joist spaces).

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 15

value, the effective moment of inertia computed by dividing in AISC Design Guide 36, Design Considerations for Cam-
Ij or Ijg by 1.15, is used in lieu of the gross moment of inertia. ber (Kloiber and Burmeister, 2020).
For special joists and joist girders, the moment of inertia
can be specified along with the loads. In such cases, it is 2.9 RECOMMENDATIONS WHEN DRAIN SIZES
in the specifying professional’s best interest for the strength ARE NOT YET KNOWN
and stiffness requirements to be roughly consistent. While
Roof slopes and drainage systems must be known to accu-
exact chord size will not be known when specifying the joist,
rately determine rain loads. However, these may not yet be
they can be approximated by assuming the effective depth is
known at the time of initial structural design. To proceed
equal to 95% of the overall depth and noting that steel with
with the structural design of the roof, rain loads may be
yield strength of Fy = 50 ksi is typically used for joists and
approximated. Assumptions based on interpretation of the
joist girders.
requirements in the International Plumbing Code or previ-
When designing roofs with low slopes, parallel chord
ous designs of similar buildings can help guide approxima-
joists with end supports at different elevations are more eco-
tions. Alternatively, rough assumptions such as the inlet of
nomical than providing pitch into the joist top chords. The
the secondary drainage system being located 2 in. above the
web system of a nonparallel chord joist and the joist, as a
low point of the roof surface and hydraulic head of 4 in. may
whole, is more expensive to manufacture.
also provide sufficient accuracy for initial structural design.
Camber can reduce ponding loads. Open web steel joists
In any case, assumptions made for initial structural design
and joist girders are provided with camber by default. SJI
must be verified through the request for information (RFI)
Specifications Table 4.6-1 lists the approximate camber as
process prior to issuing drawings for construction. Small
a function of top chord length. A camber different than the
changes in the size of the outlet of a drain, for example, can
default can be specified to the joist manufacturer. Design
lead to large changes in rain load and an unsafe design.
considerations for cambering structural steel are presented

16 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

Chapter 3
Methods of Analysis for Ponding
The recommended method of design for ponding presented 1
in Chapter 2 of this Design Guide requires structural analyses Bp = (3-1)
1 − 1.15Cp − Cs 
that consider ponding nonlinearity. Tracking the movement
of water on the roof and resulting fluid-structure interaction
where
would be complex and is not required for design. Static anal-
Cp = f lexibility coefficient for primary members, Equa-
ysis in which the applied load is dependent on the deflec-
tion 3-2 or 3-4
tion is sufficient. Other simplifications are also appropriate.
Given the relatively low slope of roofs where ponding is a Cs = f lexibility coefficient for secondary members, Equa-
consideration, it is generally acceptable to treat the rain load tion 3-3 or 3-5
as a vertical load rather than more accurately as one that acts The flexibility coefficients, Cp and Cs, can be computed
normal to the roof surface. Given that roof members typi- using consistent units as:
cally have low axial loads and do not have the restraint nec-
essary to develop catenary action, it is generally acceptable γ Ls L4p
to not consider geometric nonlinearities (e.g., P-Δ effects). Cp = (3-2)
Lastly, it is acceptable to use elastic analyses with no explicit π 4 EIp 
tracking of material nonlinearities (e.g., steel yielding). If γSL4s
necessary, the effects of partial yielding of the cross section Cs = (3-3)
π 4 EIs 
can be captured with a stiffness reduction factor.
Four methods of analysis for ponding are discussed in this
where
chapter, each of which captures the nonlinear effect arising
E = modulus of elasticity of steel
from the applied loads being dependent on the deflected
shape of the roof. The methods vary in the range of appli- Ip = moment of inertia of primary members
cability and ease of use. The engineer may select whichever Is = moment of inertia of secondary members
method of analysis is most appropriate for the situation. Lp = length of primary members
Ls = length of secondary members
3.1 CLOSED-FORM SOLUTIONS S = spacing of secondary members
The methods of design for ponding presented in the 2016 γ = unit weight of water
edition of AISC Specification Appendix 2 (AISC, 2016)
were based on closed-form solutions for deflections and The flexibility coefficients, Cp and Cs, can also be computed
bending moments in an idealized bay. While the methods in standard units as:
of design no longer meet the requirements of ASCE/SEI 7,
the underlying closed-form solutions remain accurate for the 32Ls L4p
Cp = (3-4)
cases to which they apply. Other closed-form solutions have 10 7 I p 
also been developed. Silver (2010) presents several solutions
32SL4s
for single beams, as well as beam and girder systems. While Cs = (3-5)
each of these solutions is limited to specific simple cases, 10 7 I s 
use of closed-form solutions to perform ponding analysis is
computationally efficient. where
Ip = moment of inertia of primary members, in.4
3.2 AMPLIFIED FIRST-ORDER ANALYSIS Is = moment of inertia of secondary members, in.4
Amplified first-order analysis is perhaps the simplest method Lp = length of primary members, ft
of analysis for ponding. This method, developed by Denavit Ls = length of secondary members, ft
(2019), is similar in nature to the B1/B2 method for approxi- S = spacing of secondary members, ft
mate second-order analysis of frames described in AISC
Specification Appendix 8. Results from a first-order analysis The amplification factor, Bp, is based on the closed-form
(i.e., one in which the loads are based on the undeformed solution developed by Marino (1966) for his idealized rect-
shape of the roof) are amplified by the factor Bp. angular bay. Calibration of the amplification factor targeted

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 17

the maximum moment in the secondary members, so it mar- from one iteration to the next will either increase or decrease
ginally underestimates the maximum loads at the center of as successive iterations are performed. With increasing
the bay where the deflections are greatest and overestimates increments of rain load, the deflections and internal forces
the forces in the primary members and the total water load will diverge to infinity, indicating instability. With decreas-
in the bay. ing increments of rain load, the deflections and internal
forces will converge to equilibrium. Once convergence is
3.3 NEGATIVE SPRING STIFFNESS achieved within a suitable tolerance, the iterations can stop.
Iterative analysis is most general because it can be easily
It is also possible to perform a ponding analysis using neg-
applied to cases where only a portion of the roof is below
ative stiffness springs as described by Baber and Rigsbee
free water level and can incorporate consideration of snow
(2010). This method is amenable to general finite element
loads. This method extends naturally to analyses that include
packages but becomes difficult to implement when only a
material and geometric nonlinearity.
portion of the roof is below the water level (as is common).
Determining loads from the deflected shape of the roof
is an important step in the iterative analysis. The process is
3.4 ITERATIVE ANALYSIS
relatively straightforward for 2D analyses but more complex
Iterative analysis is the most general method of analysis for for 3D analyses. Colombi (2006) presents a method for com-
ponding. In this method, a series of analyses are performed puting rain loads on a deflected roof. In this method, the roof
with rain loads updated in each iteration based on the deflec- is discretized into a grid of cells as shown in Figure 3-1. For
tions from the previous iteration. The change in rain load each cell, the depth of water at each vertex is identified, then

(a) Isolation of a single cell

(b) Transformation to point loads

Fig. 3-1. Illustration of the calculation of rain loads.

18 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

the volume of water and point loads at each vertex are com- Bay Analysis Tool for rectangular bays constructed with
puted based on the depths and the geometry of the cell. open web steel joists or steel beams and joist girders or steel
The repetitive computation and application of loads girders. The PyPonding Python package includes tools for
required for this method demand automation to be per- performing iterative ponding analyses, including functions
formed efficiently. However, once automated, the process is for computing water volume and load (Denavit et al., 2022).
efficient. Iterative analysis is implemented in the SJI Roof

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 19

20 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40
Chapter 4
SJI Roof Bay Analysis Tool
The SJI Roof Bay Analysis Tool is a Microsoft Excel spread- the recommended method of design presented in this Design
sheet that can be used as a design aid for estimating purposes Guide.
and the selection of roof framing. It also has the capability Within the tool, general bay design (not considering the
to perform the recommended method of design for ponding effects of ponding) is performed in the “Joists & Joist Gird-
described in Chapter 2 of this Design Guide. It is applica- ers,” “Joists & WF Girders,” or “Beams & WF Girders”
ble only to rectangular bays with equally spaced secondary sheets, depending on selected member types (wide flange
members, but it explicitly accounts for roof slope (as defined is abbreviated as WF in the SJI Roof Bay Analysis Tool).
by the elevation of the four columns), camber, and rigid Ponding analyses are performed in the “Ponding Analy-
perimeter supports. The tool is available on the Steel Joist sis” sheet regardless of selected member type. Notes and
Institute website: https://steeljoist.org (SJI, 2022). instructions regarding the ponding analysis are given in the
This chapter describes version 5.1 of the tool. Previous “Ponding Instructions” sheet and analysis results formatted
versions of the tool implemented the direct analysis method for printing are given in the “Ponding Load Results” sheet.
for ponding as described in SJI Technical Digest 3 (Fisher A screenshot of the general information sheet is shown in
and Denavit, 2018). The direct analysis method for pond- Figure 4-1.
ing is a predecessor to the provisions of ASCE/SEI 7 and

General Information for Roof Framing Analysis Tool

Steel Joist Institute

140 W. Evans St. Suite 203
Florence, SC 29501
Phone: (843) 407-4091
www.steeljoist.org

Purpose:
This workbook provides a design aid for estimating purposes for the selection of roof framing.

Input:
Input values are in boxed cells with highlighted light yellow background.
The user must be familiar with the analysis and design procedures of the SJI and SDI Specifications.

ANSI/SJI 100 - 2020

STANDARD SPECIFICATION FOR K-SERIES, LH-SERIES, AND DLH-SERIES OPEN WEB STEEL JOISTS
AND FOR JOIST GIRDERS

4. Design Basis
4.2.1 Design Using Load and Resistance Factor Design (LRFD)
Joists and Joist Girders shall have their components so proportioned that the required stresses,
fu, shall not exceed ФFn
4.2.2 Design Using Allowable Strength Design (ASD)
Joists and Joist Girders shall have their components so proportioned that the required stresses,
fa, shall not exceed Fn / Ω

Notes:
1. Minimum design loads and load combinations are in accordance with ASCE 7-16 or ASCE 7-22
2. Bay sizes are limited to Joist and Joist Girder Load Tables

Fig. 4-1. SJI Roof Bay Analysis Tool.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 21

4.1 INPUT • Water level relative to zero datum. The primary input
defining the load from the impounded water is the water
The “Ponding Analysis” sheet references the other sheets
level. This level is calculated by the user outside of the
for key general input data such as bay dimensions, selected
tool, typically as the hydraulic head above the inlet of
member sizes, and loads (Figure 4-2). Additional data spe-
SDSL. The spreadsheet does not distinguish between
cific to the ponding analysis is defined in the “Ponding Anal-
rainwater and snowmelt. The datum is arbitrary, but the
ysis” sheet (Figure 4-3). This additional data includes:

General Input (Defined in Joists & Joist Girders spreadsheet)

Design Methodology ASD
Minimum Design Loads ASCE 7-16
Primary Member Type Joist Girder
Secondary Member Type Joist
Joist Span 50.00 ft
Joist Girder Span 42.00 ft
Joist Size 30K11
Joist Girder Size 48G6N15.1K
Number of Joist Spaces 6
Dead Load Supported by Deck 7.14 psf
Dead Load Supported by Joists 14.17 psf
Joist Girder Self Weight 44.00 lb/ft
Snow Load 28.00 psf

z = 0.000 in. z = 21.000 in.

TOP PRIMARY MEMBER

DECK
SPAN
SECONDARY MEMBER 1

SECONDARY MEMBER 2

SECONDARY MEMBER 3

SECONDARY MEMBER 4

SECONDARY MEMBER 5

SECONDARY MEMBER 6

BOTTOM PRIMARY MEMBER

z = 0.000 in. z = 21.000 in.

MIRRORED BAYS
(OPTIONAL)

Fig. 4-2. General input.

22 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

 same datum must be used for both the water level and top computed in accordance with the 2022 edition of ASCE/
of roof elevations. SEI 7 (ks is the concurrent snow load factor, S is the snow
• Compute load on deformed roof. The load must be com- load, γs is the snow density, and 0.7 is the ASD load factor
puted on the deformed roof for the analysis to capture the for snow in the 2022 edition of ASCE/SEI 7). When com-
nonlinear effect of ponding. However, an option is pro- puting ponding loads, the spreadsheet accounts for the
vided to compute the load on the undeformed roof for possibility of snow and water occupying the same physi-
special investigations or evaluation of rain loads from pre- cal space. In this case, the unit weight of snow-plus-water
vious standards (e.g., the 2016 edition of ASCE/SEI 7). is taken as that of just water. To override this behavior and
conservatively allow snow and impounded water to over-
• Snow density, γ s. The snow density (unit weight) can be lap, a value of zero for snow density may be inputted.
determined using ASCE/SEI 7, Equation 7.7-1, as a func-
tion of the ground snow load, pg. The snow density is used • Concurrent snow load factor, ks. The concurrent snow
to compute the height of snow in the bay. The height of load factor is applied to the snow load when performing
snow is computed as ks S/ γs when minimum design loads the ponding analysis. A value of zero excludes snow from
are computed in accordance with the 2016 edition of the ponding analysis and is appropriate for determining
ASCE/SEI 7 or 0.7ks S/ γs when minimum design loads are rain loads. Values greater than zero are appropriate for

Ponding Specific Input

Water level relative to zero datum: 4.00 in.
Compute load on deformed roof: Y (Y or N)
Snow density, γs: 15.30 lb/ft3
Concurrent snow load factor, k s: 0.75 ---
Top of roof elevation:
Top Left 0.000 in.
Top Right 21.000 in.
Bottom Left 0.000 in.
Bottom Right 21.000 in.
Camber:
Joist 1.042 in.
override: in.
Top Joist Girder 0.735 in.
override: in.
Bottom Joist Girder 0.735 in.
override: in.
Bay is mirrored:
Left N (Y or N)
Right Y (Y or N)
Top Y (Y or N)
Bottom Y (Y or N)
Joist support is wall:
Top N (Y or N)
Bottom N (Y or N)
Joist is rigid:
Joist 1 (Leftmost) N (Y or N)
Joist 7 (Rightmost) N (Y or N)
Effective moment of inertia:
Joist 541.8 in.4
override: in.4
Joist Girder 4,254 in.4
override: in.4

Fig. 4-3. Ponding specific input.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 23

 assessing ponding in the presence of snow. A value of the method described by Colombi (2006) and illustrated
0.75 is recommended in Chapter 2 of this Design Guide. in Figure 3-1. The total load from the impounded water
• Top of roof elevation. The elevations are used to define from each iteration is displayed in the spreadsheet as an
the roof slope. The datum for defining the elevations is indication of the speed of convergence.
arbitrary, but the same datum must be used for both the • Perform structural analysis. The primary and secondary
top of roof elevations and the water level. The spreadsheet members are analyzed separately; each is analyzed as a
does not check for minimum slope required by IBC, nor 2D simply supported beam. The secondary members are
does it account for crickets, sloping fill, or other non- subjected to the computed loads, and primary members are
uniform thickness of roofing materials. subjected to the reactions from the secondary members.
• Camber. For structural steel members, the default cam- While the applied loads are nonlinear, no other sources of
ber is zero. For open web steel joists and joist girders, the nonlinearity are considered (i.e., a first-order elastic anal-
default camber is based on the SJI Specifications and the ysis is performed). To account for adjacent bays, loads on
span of the member. Default values can be overridden by the primary members and edge secondary members are
the user. Manufacturing tolerances for camber need not doubled, as applicable according to the input.
be considered when using the SJI Roof Bay Analysis Tool • Check for convergence. After the structural analysis is
because other means of providing a margin of safety (e.g., performed, convergence is checked by comparing the
load factors) are employed. computed point loads to those computed in the previous
• Bay is mirrored. Mirroring of the bay accounts for addi- iteration.
tional loads from outside the bay on edge members. When If the analysis does not converge, that is typically indica-
selected as “yes,” the loads on the edge members are dou- tive of ponding instability, and the roof system must be made
bled in the analysis. stiffer. For ASD, once convergence is obtained, the moments,
• Beam/joist support is wall. When selected as “yes,” the shears, strength ratios, and other results are computed and
beams or joists on that edge are assumed to be supported displayed. For LRFD, once convergence is obtained, one
by an infinitely stiff wall. additional analysis is performed with factored load, then the
• Beam/joist member is rigid. When selected as “yes,” the moments, shears, strength ratios, and other results are com-
identified beam or joist is assumed to be infinitely stiff and puted and displayed.
without camber.
4.3 OUTPUT
• Effective moment of inertia. The effective moment of iner-
tia of structural steel members is taken as 0.8 multiplied Results from the analysis are presented in data tables.
by the gross moment of inertia to account for partial yield- The water depth output table presents the maximum water
ing of the cross section. The effective moment of inertia of depth in the bay and along each edge of the bay. This output
joists and joist girders is automatically calculated and the can be helpful for interior bays with low slope where the
1.15 factor to account for shear deformations is automati- water level is defined with respect to the deflected shape of
cally included. the roof as just enough for the water to flow to the adjacent
bay.
4.2 ANALYSIS PROCEDURE The deck output table presents stiffness and strength
checks for the deck. Deck deflections are not included in
The ponding analysis is performed when the user presses the the ponding analyses; therefore, their stiffness must be
button labeled “RUN ANALYSIS,” initiating a macro. The large enough that potential amplification in load due to their
analysis is iterative; the following loop is performed until deflections is negligible. The deck moment of inertia is com-
convergence is obtained: pared to the limit given by Equation 4-1, which is the same
• Compute loads. Loads due to the combination of dead, limit that applied in the methods of previous editions of
snow, and impounded water are computed based on the AISC Specification Appendix 2 (e.g., AISC, 2016).
current deflected shape of the roof. To compute the loads,
the bay is broken into a grid of cells: 20 cells are used Id ≥ 25 × 10 − 6 S 4 (4-1)
along the length of the secondary members, and a number
of cells equal to the number of secondary member spaces where
is used along the length of the primary members. For each Id = m
 oment of inertia of the steel deck supported on sec-
cell, the height of snow and water are computed at each ondary members, in.4/ft
of the four corners of the cell, and the physical volume of
S = s pacing of secondary members, ft
water and snow is converted to four point loads following

24 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

 The secondary member (i.e., beam or joist) output table 3. Determine the hydraulic head based on the flow rate, Q,
displays the maximum shear and moment for each beam or and the characteristics of the secondary drainage system
joist as well as a strength check. The total applied loads in for structural loading.
the ponded configuration are nonuniform as depth of water a. The flow rate can be determined using ASCE/SEI 7,
varies with the deflection of the bay. The shear and moment Equation C8.2-1.
strength ratios are computed as the maximum ratio between
required strength and available strength along the length of b. The hydraulic head can be determined based on
the member. The equivalent loads in shear and moment are ASCE/SEI 7, Tables C8.2-1, C8.2-3, and C8.2-5.
uniform loads that would produce the same strength ratio as c. O’Rourke and Lewis (2020) provide additional
the nonuniform ponding load. guidance.
To better visualize the structural behavior of the bay and 4. Determine the water level as the hydraulic head above
to perform checks against local overstressing of the joists, the inlet to the secondary drainage system for structural
three tables of applied loads are generated. The first table loading.
displays the unfactored ponded water load. The second table
displays the total distributed load on the secondary mem- 5. Input the required data into the “Ponding Analysis”
bers. The third table displays the total distributed load on the sheet. Data from the “Joist & Joist Girders,” “Joists &
joist in excess of the allowable load for joists. The third table WF Girders,” and “Beams & WF Girders” sheets is auto-
is not active when the secondary members are beams. matically transferred to the “Ponding Analysis” sheet.
The primary member (i.e., wide-flange girder or joist 6. Run the ponding analysis.
girder) output table displays reactions from the secondary a. If stability is achieved, the analysis is complete.
members, point loads, maximum moment and shear, as Ponding stability is achieved when the strength ratio
well as a strength check. The strength ratios are computed for all members is less than or equal to 1.0. If local
through a point-by-point comparison as the ratio of required load effects are indicated (loads highlighted with
strength to available strength. red font) determine based on analysis or judgement
whether the indicated applied loads will cause a
4.4 DESIGN PROCEDURE localized overstress. If the loads are determined to
The following describes the design procedure using the SJI cause overstress, rerun the analysis with larger joists,
Roof Bay Analysis Tool: beams, or girders.
1. Conduct a preliminary design using the “Joist & Joist b. If stability is not achieved, return to the “Joist & Joist
Girders,” “Joists & WF Girders,” or “Beams & WF Gird- Girders,” “Joists & WF Girders,” or “Beams & WF
ers” sheets. Girders” sheet and select stiffer or stronger primary
2. Identify the secondary drainage system for structural or secondary members by inputting larger loads in
loading (SDSL). See the requirements in ASCE/SEI 7, the “optional increased load data” section. Rerun the
Section 8.2. ponding analysis until stability is achieved.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 25

26 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40
Chapter 5
Design Examples
This chapter presents several examples that illustrate the use of the recommended method of design for ponding and the SJI Roof
Bay Analysis Tool.

5.1 EXAMPLE 1—EXISTING DEAD FLAT ROOF

Given:
Evaluate the strength of the roof of an existing industrial building located in Tampa, Florida, and classified as Risk Category II to
determine if the existing members need to be reinforced as part of an alteration to the building. The primary members are ASTM
A36/A36M (ASTM, 2019) W24×68 girders that span 50 ft. The secondary members are 24J8 open web steel joists spaced at
6 ft-3 in. that span 38 ft. The joists are Fy = 36 ksi steel. Material and geometric properties of the steel girders are from the AISC
Steel Construction Manual (AISC, 2023). The roof is constructed without slope and without secondary drainage. The dead load
is 15 lb/ft2, including an allowance for the self-weight of the steel framing, and the roof live load is 20 lb/ft2.

Solution:
From AISC Manual Table 2-4, the material properties are as follows:
Girders
ASTM A36/A36M
Fy = 36 ksi
From AISC Manual Table 1-1, the geometric properties are as follows:
Girders
W24×68
Ix = 1,830 in.4
Zx = 177 in.3
From SJI Technical Digest #3, Table B.1 (Fisher and Denavit, 2018), the joist geometric properties are as follows:
Joists
Is = 258.5 in.4
ASCE/SEI 7, Section 8.2, requires that the primary drainage system is assumed to be blocked. The next point of overflow which
is not required to be blocked is the SDSL. For the existing building, which was not constructed with secondary drains, the next
point of discharge would be over the parapet. Because this would result is very large rain loads, scuppers or a secondary drainage
system should be installed with the alteration.
The design storm return period for Risk Category II is 100 years (ASCE/SEI 7, Table 8.2-1). The rainfall intensity, i, for a 15 min
duration storm with a 100 yr return period in Tampa, Florida, is 7.55 in./hr as determined from the ASCE Hazard Tool (ASCE,
2023).
i = 7.55 in./hr

The ground snow load is 3 lb/ft2 as determined from the ASCE Hazard Tool and will be neglected in this investigation.
Assume that secondary drains are installed at every column; therefore, the tributary area of each secondary drain is equal to the
area of the bay.
A = ( 50 ft ) ( 38 ft )
= 1,900 ft 2

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 27

The flow rate for each drain is determined using Equation 1-6 (ASCE/SEI 7, Equation C8.2-1) as:
Q = 0.0104 Ai (1-6)
= 0.0104 (1,900 ft 2 ) ( 7.55 in./hr )
= 149 gal/min 

Per ASCE/SEI 7, Table C8.2-1, a roof drain with an 8-in.-diameter overflow dam, 4 in. drain outlet, and 2 in. drain bowl depth can
accommodate that flow rate with a hydraulic head of dh = 1.5 in. Select this drain size. The inlet of the secondary drain should be
located 2 in. above the inlet to the primary drainage system so that the secondary drain is not required to be blocked per ASCE/
SEI 7, Section 8.2. Because the inlet to the primary drainage system is at the roof surface, the static head is ds = 2 in.
The ponding head is determined by ponding analysis using the amplified first-order analysis method. The flexibility coefficients
for the primary and secondary members are computed from Equations 3-4 and 3-5 as:
32Ls L4p
Cp = (3-4)
10 7 Ip
32 ( 38 ft ) ( 50 ft )4
=
10 7 ( 0.8 ) (1,830 in.4 )
= 0.519 
32SL4s
Cs = (3-5)
10 7 I s
32 ( 6.25 ft ) ( 38 ft )4
=
10 7 ( 258.5 in.4 )
= 0.161 

where Ip is the major-axis moment of inertia for the girder, Ix, with the recommended 0.8 stiffness reduction applied, and Is, the
effective moment of inertia of the joist, is determined from SJI Technical Digest 3, Table B.1 (Fisher and Denavit, 2018).
The ponding amplification factor is computed using Equation 3-1 as:
1
Bp = (3-1)
1 − 1.15Cp − Cs
1
=
1 − 1.15 ( 0.519 ) − ( 0.161)
= 4.13 

When the roof is subjected to unfactored dead load and unfactored rain load, the uniform load excluding the ponding head is:
D + 5.2 (ds + d h) = (15 lb/ft 2 ) + 5.2 ( 2 in. + 1.5 in.)
= 33.2 lb/ft 2

Amplifying this load by Bp, gives the approximate uniform load including the ponding head.
Bp ( 33.2 lb/ft 2 ) = 4.13 ( 33.2 lb/ft 2 )
= 137 lb/ft 2

28 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

Subtracting out the dead load, gives the rain load.
R = 137 lb/ft 2 − 15 lb/ft 2
= 122 lb/ft 2

From ASCE/SEI 7, Chapter 2, determine the factored uniform load. The rain load is greater than the roof live load, so use R in
the load combinations.

LRFD ASD

1.2D + 1.6R = 1.2 (15 lb/ft 2 ) + 1.6 (122 lb/ft 2 ) D + R = 15 lb/ft 2 + 122 lb/ft 2
= 213 lb/ft 2 = 137 lb/ft 2

Determine the uniform load on the joist.

LRFD ASD

wu = ( 213 lb/ft 2 ) ( 6.25 ft ) wa = (137 lb/ft 2 ) ( 6.25 ft )

= 1,330 lb/ft = 856 lb/ft

From previous editions of the SJI Load Tables (SJI, 2003), the allowable total safe load on a 24J8 spanning 38 ft is 249 lb/ft
(LRFD strength is not given). Therefore, the joist is not strong enough and will need to be reinforced.
Approximating the loading on the girder as uniform, the maximum moment is:

LRFD ASD

wu = ( 213 lb/ft 2 ) ( 38 ft ) wa = (137 lb/ft 2 ) ( 38 ft )

= 8,090 lb/ft = 5,210 lb/ft
wu L p2 wa Lp2
Mu = Ma =
8 8
(8,090 lb/ft )( 50 ft )2 ( 5,210 lb/ft )( 50 ft )2
= =
8 8
1 kip ⎞ ⎛ 12 in.⎞ ⎛ 1 kip ⎞ ⎛ 12 in. ⎞
= ( 2,530,000 lb-ft ) ⎛ = (1,630,000 lb-ft )
⎝ 1,000 lb ⎠ ⎝ 1 ft ⎠ ⎝ 1,000 lb ⎠ ⎝ 1 ft ⎠
= 30,400 kip-in. = 19,600 kip-in.

From AISC Specification Section F2.1, the plastic bending moment, representing the upper bound of the girder’s nominal flexural
strength, is:
Mp = Fy Z x (Spec. Eq. F2-1)
= ( 36 ksi ) (177 in. ) 3

= 6,370 kip-in. 

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 29

 LRFD ASD
ϕMp = 0.90 ( 6,370 kip-in.) Mp 6,370 kip-in.
=
= 5,730 kip-in. Ω 1.67
ϕMn ≤ ϕM p < Mu n.g. = 3,810 kip-in.
Mn M p
≤ < Ma n.g.
Ω Ω

Therefore, the girder is also not strong enough and will need to be reinforced. Reinforcement that stiffens the joists and girders
will reduce load amplification due to ponding, thus reducing the required strengths.
The calculation of required strengths can also be performed with the SJI Roof Bay Analysis Tool. For this case, the effective
moment of inertia of the joist must be entered manually using the override option because 24J8 joists are not currently produced
and not available in the tool. The required strengths calculated with the SJI Roof Bay Analysis Tool are lower and more exact than
those obtained using the approximate method, but they are still far greater than the available strengths computed in this example.
The difference in required strengths is due to inaccuracies in the amplification factor for very flexible roofs (e.g., Bp ≥ 2). While
required strengths computed using the SJI Roof Bay Analysis Tool can be used for this example, available strengths and strength
checks from the tool cannot be used because 24J8 joists are not available in the tool, and the tool assumes Fy = 50 ksi for wide-
flange members.
Note that this example is based on one presented by Marino (1966), where the member sizes were found to be adequate for
ponding. There are several differences between the calculations by Marino and those presented here. Marino did not explicitly
compute rain loads, but rather assumed one-quarter of the live load (5 lb/ft2) was on the roof at the outset of ponding, which is
less than the combined static head and hydraulic head (18.2 lb/ft2) now required by ASCE/SEI 7. Marino employed a factor of
safety of 1.25 against flexural yielding, which is less conservative than implied by current standards. Marino did not reduce the
stiffness when determining flexibility coefficients. Lastly, use of the amplification factor, Bp, is more approximate and somewhat
more conservative than the original equations presented by Marino, especially when Bp ≥ 2.

5.2 EXAMPLE 2—OPEN WEB STEEL JOIST ROOF DESIGN

Given:
Design the framing for the roof shown in Figure 5-1. Use joist girders as the primary members and open web steel joists as the
secondary members. The joists and joist girders are Fy = 50 ksi steel. The structure is located in Memphis, Tennessee, and clas-
sified as Risk Category II. The depth of the girders is limited to 36 in. for architectural reasons. The spacing of the secondary
members is 5 ft.
Dead load on the roof consists of a uniform 18 lb/ft2 (which includes an allowance for the weight of the secondary members)
and self-weight of the primary members, which is approximated as an additional unform 1 lb/ft2. The live load on the roof is
20 lb/ft2 and is reducible. The controlling uniform snow load is 24.1 lb/ft2 (which includes the rain-on-snow surcharge) for a
ground snow load of 23 lb/ft2 in accordance with the provisions of ASCE/SEI 7.
The primary roof drains are located at the interior columns. The secondary roof drainage system consists of two 24-in.-wide
channel scuppers located at opposite ends of the structure. The roof is sloped down to the interior columns at 4 in./ft. The top
of deck at the interior columns is at an elevation of −12 in. The thickness of the roofing and insulation is 2 in. The top of roof at
the interior columns and inlet to the primary drains are at an elevation of −10 in. The invert elevation of the scuppers is −8 in.

Solution:
The first step is to conduct a preliminary design using the “Joist & Joist Girders” sheet of the SJI Roof Bay Analysis Tool. The
tool assumes a yield strength of Fy = 50 ksi, which is appropriate for this design. Enter relevant data from the problem state-
ment. The maximum joist girder depth of 36 in. needs to be entered, otherwise the tool will identify a deeper joist girder as more
efficient. The option to minimize X-bridging was selected to reduce total cost. 24K7 joists and 36G8N9.4F joist girders were
selected. LRFD was used for this example, results using ASD are similar.

30 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

With the preliminary design complete, rain load and ponding will be evaluated next using the “Ponding Analysis” sheet. When
the “Joists & Joist Girders” case is selected, general input will be automatically copied from the preliminary design (Figure 5-2).

Overflow scupper
invert elevation -8"
40'-0"

Primary drain
inlet elevation -10"
C

Top of roof elevation +0" Top of roof elevation +0"

along gridline 1
40'-0"

along gridline 3

Primary drain
inlet elevation -10"
B
40'-0"

Overflow scupper
invert elevation -8"

A
40'-0" 40'-0"

1 2 3

Fig. 5-1. Roof framing plan for Examples 2 and 3.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 31

Next, the ponding specific input needs to be defined, starting with the water level. The primary drains adjacent to the interior
columns are required to be blocked per ASCE/SEI 7, Section 8.2. The scuppers at either end of the building are not required to
be assumed to be blocked and, therefore, will be the secondary drainage system for structural loading (SDSL). There are two
overflow scuppers; each has a tributary area of half the total area of the roof.
A = (60 ft ) (80 ft )
= 4,800 ft 2

The rainfall intensity, i, for a 15 min duration storm with a 100 yr return period in Memphis is 6.09 in./hr as determined from the
ASCE Hazard Tool.
i = 6.09 in./hr

The flow rate, Q, for each scupper is determined using Equation 1-6 (ASCE/SEI 7, Equation C8.2-1) as:
Q = 0.0104 Ai (1-6)
= 0.0104 ( 4,800 ft 2 ) ( 6.09 in./hr )
= 304 gal/min 

The hydraulic head, dh, associated with this flow rate and a 24-in.-wide channel scupper can be determined from Equation 1-7 as:
q
⎛ 3.85Q ⎞
dh = ⎜ (1-7)
⎝ qcdWs 2g ⎟⎠
q
⎡ 3.85 ( 304 gal/min ) ⎤
=⎢ ⎥
⎢⎣ q ( 0.6 ) ( 24 in.) 2 ( 386 in./s2 ) ⎥⎦
= 2.68 in. 

Alternatively, the hydraulic head can be determined by linear interpolation from ASCE/SEI 7, Table C8.2-3. The invert elevation
of the scuppers is −8 in.; therefore, the water level for the design condition is:
− 8 in. + d h = − 8 in. + 2.68 in.
= − 5.32 in.

General Input (Defined in Joists & Joist Girders spreadsheet)

Design Methodology LRFD
Minimum Design Loads ASCE 7-22
Primary Member Type Joist Girder
Secondary Member Type Joist
Joist Span 40.00 ft
Joist Girder Span 40.00 ft
Joist Size 24K 7
Joist Girder Size 36G8N9.4F
Number of Joist Spaces 8
Dead Load Supported by Deck 10.00 psf
Dead Load Supported by Joists 18.00 psf
Joist Girder Self Weight 40.00 lb/ft
Snow Load 24.10 psf

Fig. 5-2. General input for Example 2.

32 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

The load should be computed on the deformed roof in accordance with the recommended method of design for ponding. The
snow density, γs, can be computed using ASCE/SEI 7, Equation 7.7-1.
γ s = 0.13pg + 14 ≤ 30 lb/ft 3
= 0.13 ( 23 lb/ft 2 ) + 14 ≤ 30 lb/ft 3
= 17.0 lb/ft 3 ≤ 30 lb/ft 3
= 17.0 lb/ft 3

The concurrent snow load factor is taken as 0.75 in accordance with the recommended method of design for ponding.
The preliminary design applies to all bays of the structure. For ponding, the bays between grid lines B and C will control because
the water level is the same, and the roof level is higher for the other bays leaving more room for water. Performing the design
for one of the bays between grid lines B and C and applying the results to all bays in the structure will be conservative. The tool
requires the top of roof elevation for each corner of the bay to define the slope. Nearly identical bays are located adjacent to the
left, right, and bottom edges of the bay under investigation; therefore, those edges are defined to be mirrored and not rigid. The
top edge is the wall along grid line 3, which is assumed to be rigid and not mirrored. Note that the designations of left, right, top,
and bottom in the tool are rotated 90° from what is shown in Figure 5-1.
The default values of camber and effective moment of inertia are appropriate. A summary of the ponding specific input is shown
in Figure 5-3.
The analysis is performed by clicking the “Run Analysis” button. The joist output table (Figure 5-4) indicates that the joists are
sufficient for the ponding condition. The joist girder output table (Figure 5-5), on the contrary, indicates that the joist girder does
not have sufficient strength for the ponding condition. Thus, the design needs to be revised.
The results of the ponding analysis show that the panel point loads are as large as 11.96 kips (Figure 5-5). Increasing the joist
girder panel point load to 11.96 kips would be more than sufficient because the increase in strength will also increase stiffness
(thus reducing the load amplification due to ponding). After iteration, a panel point load of 11.7 kips is found to be sufficient.
The increased panel point load can be defined on the optional increased load data table on the “Joist & Joist Girders” sheet
(Figure 5-6).
The joist girder output table for the revised design (Figure 5-7) indicates that the strength of the joist girders is sufficient. The
other tables also indicate a safe design, and thus the ponding analysis is complete with 24K7 joists and 36G8N11.7F joist girders.

5.3 EXAMPLE 3—STRUCTURAL STEEL ROOF DESIGN

Given:
Design the framing for the roof shown in Figure 5-1. Use ASTM A992/A992M (ASTM, 2022) wide-flange girders as the primary
members and ASTM A992/A992M wide-flange beams as the secondary members. Project and loading details are identical to
those for Example 2.

Solution:
From AISC Manual Table 2-4, the material properties are as follows:
All members
ASTM A992/A992M
Fy = 50 ksi
As with Example 2, the first step is to conduct a preliminary design that does not consider rain loads or ponding. This preliminary
design can be conducted using the “Beams & WF Girders” sheet of the SJI Roof Bay Analysis Tool. The tool assumes a yield
strength of Fy = 50 ksi, which is appropriate for this design. Enter relevant data from the problem statement. Unlike with joists
and joist girders, the SJI Roof Bay Analysis Tool does not automatically select structural steel members, but rather provides a list
of least weight shapes that satisfy requirements for each family of shapes. From the provided list, W12×19 beams and W24×55
girders are selected. ASD was used for this example, results using LRFD are similar.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 33

 Ponding Specific Input
Water level relative to zero datum: -5.32 in.
Compute load on deformed roof: Y (Y or N)
Snow density, γs: 17.00 lb/ft3
Concurrent snow load factor, ks: 0.75 ---
Top of roof elevation:
Top Left 0.000 in.
Top Right 0.000 in.
Bottom Left -10.000 in.
Bottom Right -10.000 in.
Camber:
Joist 0.667 in.
override: in.
Top Joist Girder 0.000 in.
override: in.
Bottom Joist Girder 0.667 in.
override: in.
Bay is mirrored:
Left Y (Y or N)
Right Y (Y or N)
Top N (Y or N)
Bottom Y (Y or N)
Joist support is wall:
Top Y (Y or N)
Bottom N (Y or N)
Joist is rigid:
Joist 1 (Leftmost) N (Y or N)
Joist 9 (Rightmost) N (Y or N)
Effective moment of inertia:
Joist 215.0 in.4
override: in.4
Joist Girder 1,681 in.4
override: in.4

Fig. 5-3. Ponding specific input for Example 2.

34 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

Joist Output

Joist Max Shear Equiv. Load Shear Max Moment Equiv. Load Moment Strength
Number kips lb/ft Strength Ratio kip-ft lb/ft Strength Ratio Check
1 5.41 270.3 0.712 46.93 266.6 0.702 OKAY
2 5.53 276.4 0.728 47.79 272.5 0.718 OKAY
3 5.70 284.9 0.751 48.99 280.7 0.740 OKAY
4 5.82 290.8 0.766 49.85 286.5 0.755 OKAY
5 5.86 292.9 0.772 50.16 288.6 0.760 OKAY
6 5.82 290.8 0.766 49.85 286.5 0.755 OKAY
7 5.70 284.9 0.751 48.99 280.7 0.740 OKAY
8 5.53 276.4 0.728 47.79 272.5 0.718 OKAY
9 5.41 270.3 0.712 46.93 266.6 0.702 OKAY

NOTES: 1. Loads and load effects correspond to LRFD load combinations.

2. Shear strength ratio computed assuming shear capacity equal to 12.5% of the end reaction for
shear reversals, see Note 15 on the Ponding Instructions spreadsheet.

Fig. 5-4. Joist output for Example 2.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 35

Joist Girder Output Joist Girder size can be revised by inputting larger loads in the ''optional increased load
data'' section of the Joists & Joist Girders spreadsheet
Top Joist Girder Bottom Joist Girder
Joist Joist React. Panel Point Load Joist React. Panel Point Load
Number kips kips kips kips
2 4.36 5.53 11.30
3 4.42 5.70 11.63
4 4.47 5.82 11.87
5 4.48 5.86 11.96
6 4.47 5.82 11.87
7 4.42 5.70 11.63
8 4.36 5.53 11.30

Max Shear (kips): N/A (W ALL) Max Shear (kips): 40.78

Shear Strength Ratio: N/A (W ALL) Shear Strength Ratio: 1.266
Max Moment (kip-ft): N/A (W ALL) Max Moment (kip-ft): 470.45
Moment Strength Ratio: N/A (W ALL) Moment Strength Ratio: 1.251
Strength Check: N/A (WALL) Strength Check: NO GOOD
NOTES: 1. Loads and load effects correspond to LRFD load combinations.
2. Highlighted panel point loads may cause a local overstress, see notes in Ponding Instructions.

Fig. 5-5. Joist girder output for Example 2.

OPTIONAL INCREASED LOAD DATA Member I Ieff = I/1.15

4
Joist Load lb / f t 24K 7 247 215 in.
4
JG Panel Point Load 11.7 kips 36G8N11.7F 2406 2092 in.

Fig. 5-6. Optional increased load data (revised design) for Example 2.

36 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

Joist Girder Output

Top Joist Girder Bottom Joist Girder

Joist Joist React. Panel Point Load Joist React. Panel Point Load
Number kips kips kips kips
2 4.34 5.46 11.17
3 4.38 5.58 11.40
4 4.41 5.66 11.56
5 4.42 5.69 11.62
6 4.41 5.66 11.56
7 4.38 5.58 11.40
8 4.34 5.46 11.17

Max Shear (kips): N/A (W ALL) Max Shear (kips): 39.94

Shear Strength Ratio: N/A (W ALL) Shear Strength Ratio: 0.990
Max Moment (kip-ft): N/A (W ALL) Max Moment (kip-ft): 459.47
Moment Strength Ratio: N/A (W ALL) Moment Strength Ratio: 0.982
Strength Check: N/A (WALL) Strength Check: OKAY
NOTES: 1. Loads and load effects correspond to LRFD load combinations.

Fig. 5-7. Joist girder output (revised design) for Example 2.

General Input (Defined in Beams & WF Girders spreadsheet)

Design Methodology ASD
Minimum Design Loads ASCE 7-22
Primary Member Type WF Girder
Secondary Member Type Beam
Beam Span 40.00 ft
WF Girder Span 40.00 ft
Beam Size W12X19
WF Girder Size W24X55
Number of Beam Spaces 8
Dead Load Supported by Deck 10.00 psf
Dead Load Supported by Beams 18.00 psf
WF Girder Self Weight 40.00 lb/ft
Snow Load 24.10 psf

Fig. 5-8. General input for Example 3.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 37

 Ponding Specific Input
Water level relative to zero datum: -5.32 in.
Compute load on deformed roof: Y (Y or N)
Snow density, γs: 17.00 lb/ft3
Concurrent snow load factor, k s: 0.75 ---
Top of roof elevation:
Top Left 0.000 in.
Top Right 0.000 in.
Bottom Left -10.000 in.
Bottom Right -10.000 in.
Camber:
Beam 0.000 in.
override: in.
Top WF Girder 0.000 in.
override: in.
Bottom WF Girder 0.000 in.
override: in.
Bay is mirrored:
Left Y (Y or N)
Right Y (Y or N)
Top N (Y or N)
Bottom Y (Y or N)
Beam support is wall:
Top Y (Y or N)
Bottom N (Y or N)
Beam is rigid:
Beam 1 (Leftmost) N (Y or N)
Beam 9 (Rightmost) N (Y or N)
Effective moment of inertia:
Beam 104.0 in.4
override: in.4
WF Girder 1,080 in.4
override: in.4

Fig. 5-9. Ponding specific input for Example 3.

38 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

With the preliminary design complete, rain load and ponding will be evaluated next using the “Ponding Analysis” sheet. When
the “Beams & WF Girders” case is selected, general input will be automatically copied from the preliminary design as shown in
Figure 5-8.
The ponding specific input is defined using water level, snow density, roof elevations, and edge conditions as defined for Exam-
ple 2. It is assumed that the beams and girders are constructed without camber. A summary of the ponding specific input is shown
in Figure 5-9.
The analysis is performed by clicking the “Run Analysis” button. The beam output table (Figure 5-10) indicates that the beams
are sufficient for the ponding condition. The “WF Girder Output” table (Figure 5-11), on the contrary, indicates that the wide-
flange girder does not have sufficient strength for the ponding condition. Thus, the design needs to be revised.
The results of the ponding analysis show that the beam reactions are as large as 5.82 kips, resulting in a total point load on the
girder of 11.84 kips (Figure 5-11, note self-weight of the girder is included in the girder point load). Increasing the size of the
girder to accommodate this level of loading would be sufficient because a larger girder with greater strength will also have greater
stiffness (thus reducing the load amplification due to ponding). After iteration, setting the girder point load to 11.0 kips results in
a W24×68 girder, which is sufficient. Iterations are aided by the optional increased load data table on the “Beams & WF Gird-
ers” sheet.
The “WF Girder Output” table for the revised design (Figure 5-12) indicates that the strength of the W24×68 girder is sufficient.
The other tables also indicate a safe design, and thus, the ponding analysis is complete with W12×19 beams and W24×68 girders.

Beam Output
Beam Max Shear Equiv. Load Shear Max Moment Equiv. Load Moment Strength
Number kips lb/ft Strength Ratio kip-ft lb/ft Strength Ratio Check
1 4.78 239.1 0.083 44.40 222.0 0.720 OKAY
2 5.07 253.7 0.088 46.79 233.9 0.759 OKAY
3 5.46 273.2 0.095 50.05 250.2 0.812 OKAY
4 5.73 286.4 0.100 52.27 261.4 0.848 OKAY
5 5.82 291.1 0.102 53.06 265.3 0.861 OKAY
6 5.73 286.4 0.100 52.27 261.4 0.848 OKAY
7 5.46 273.2 0.095 50.05 250.2 0.812 OKAY
8 5.07 253.7 0.088 46.79 233.9 0.759 OKAY
9 4.78 239.1 0.083 44.40 222.0 0.720 OKAY

NOTES: 1. Loads and load effects correspond to ASD load combinations.

Fig. 5-10. Beam output for Example 3.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 39

WF Girder Output WF Girder size can be revised by inputting larger loads in the ''optional increased load
data'' section of the Beams & WF Girders spreadsheet
Top WF Girder Bottom WF Girder
Beam Beam React. Point Load Beam React. Point Load
Number kips kips kips kips
2 3.97 5.07 10.35
3 4.17 5.46 11.13
4 4.31 5.73 11.66
5 4.36 5.82 11.84
6 4.31 5.73 11.66
7 4.17 5.46 11.13
8 3.97 5.07 10.35

Max Shear (kips): N/A (WALL) Max Shear (kips): 39.06

Shear Strength Ratio: N/A (WALL) Shear Strength Ratio: 0.233
Max Moment (kip-ft): N/A (WALL) Max Moment (kip-ft): 456.33
Moment Strength Ratio: N/A (WALL) Moment Strength Ratio: 1.365
Strength Check: N/A (WALL) Strength Check: NO GOOD
NOTES: 1. Loads and load effects correspond to ASD load combinations.

Fig. 5-11. Wide-flange girder output for Example 3.

40 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

WF Girder Output
Top WF Girder Bottom WF Girder
Beam Beam React. Point Load Beam React. Point Load
Number kips kips kips kips
2 3.90 4.93 10.06
3 4.04 5.20 10.61
4 4.13 5.39 10.97
5 4.17 5.45 11.10
6 4.13 5.39 10.97
7 4.04 5.20 10.61
8 3.90 4.93 10.06

Max Shear (kips): N/A (WALL) Max Shear (kips): 37.19

Shear Strength Ratio: N/A (WALL) Shear Strength Ratio: 0.189
Max Moment (kip-ft): N/A (WALL) Max Moment (kip-ft): 431.95
Moment Strength Ratio: N/A (WALL) Moment Strength Ratio: 0.978
Strength Check: N/A (WALL) Strength Check: OKAY
NOTES: 1. Loads and load effects correspond to ASD load combinations.

Fig. 5-12. Joist girder output (revised design) for Example 3.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 41

5.4 EXAMPLE 4—BAY WITH LOW SLOPE AND SECONDARY MEMBERS
PERPENDICULAR TO A FREE-DRAINING EDGE

Given:
A warehouse structure is constructed with a monoslope roof as shown in Figure 5-13. To maximize interior volume while main-
taining building height restrictions, a variance was obtained for the slope to be x in./ft. The roof slopes down to a free-draining
edge. Loading on the roof consists of a uniform dead load of 20 lb/ft2 (which includes an allowance for the weight of the primary
and secondary members), a live load of 20 lb/ft2 (which is reducible), and a uniform snow load of 27.3 lb/ft2 for a ground snow
load of 39 lb/ft2 in accordance with the provisions of ASCE/SEI 7. The joists and joist girders are Fy = 50 ksi steel. Based on
preliminary design, not evaluating the effects of ponding, the joists are 30K11 and the joist girders are 52G8N12.3K. Note that
the lightest joist that does not require the use of X-bridging was selected to reduce total cost. Determine if this roof is adequate
for ponding.

Solution:
The undeflected surface of the roof has a positive slope to a free-draining edge; therefore, the static head is zero. However, water
can still accumulate if roof deflections cause a low point to form in the interior of the bay. ASCE/SEI 7, Section 8.3, identifies
conditions where an investigation to determine if low points form is required. The roof slope is x in./ft as given in the problem
statement, which is less than 4 in./ft; thus, further investigation is required.
A low point will form in the bay adjacent to the free-draining edge if the angle with respect to horizontal at the end of the joist in
the center of the bay is negative when load is applied. The joist end being evaluated is indicated as Point A in Figure 5-13. Roof

50'-0"
50'-0"

Point B
50'-0"

Point A

Fig. 5-13. Roof framing plan for Example 4.

42 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

slope, camber in the joist girder, and camber in the joist add to the angle, while deflection of the joist girder and deflection of the
joist subtract from the angle. Each source will be evaluated separately.
The roof has a slope of x in./ft. This corresponds to an angle of:
x in.
θ=
12 in.
= 0.0156 rad

Note that given the small angles in this example, use of trigonometric functions is unnecessary.
The joist is supported at one end by a joist girder, which will be constructed with camber, and at the other end the joist is sup-
ported by the wall, which does not have camber. According to the SJI Specifications, the approximate camber for a joist girder
with a span of 50 ft is 1 in. This corresponds to an angle at Point A of:
1 in.
θ=
50 ft (12 in./ft )
= 0.00167 rad

The joist girder will deflect causing the angle to decrease. Approximating the loading on the joist girder as uniform, the maxi-
mum deflection is:
5wL4
Δ max =
384EI

Loading should consider both dead and snow load. While not explicitly applicable to the current investigation, ASCE/SEI 7, Sec-
tion 8.2, requires ponding analyses be performed using unfactored loads. Therefore, a load combination of D + 0.7S (where 0.7
is the ASD load factor for snow) will be used for this analysis and the loading on the joist girder is:
w = ( D + 0.7S ) L p
= ⎡⎣( 20 lb/ft 2 ) + 0.7 ( 27.3 lb/ft 2 )⎤⎦ ( 50 ft )

⎛ 1 kip ⎞ ⎛ 1 ft ⎞
= (1,960 lb/ft )
⎝ 1,000 lb ⎠ ⎝ 12 in.⎠
= 0.163 kips/in.

The panel point loads on the joist girder are:

wLs
Pc =
N
( 0.163 kip/in.)( 50 ft )(12 in./ft )
=
8
= 12.2 kips

The gross moment of inertia of the joist girder can be computed using Equation 2-2. Note that in this equation L is the design
length, which is taken as the joist girder span − 3 ft as discussed in Section 2.8.
I jg = 0.027NPc Ld (2-2)
= 0.027 (8 ) (12.2 kips) ( 50 ft − 3 ft ) ( 52 in.)
= 6,810 in.4 

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 43

The effective moment of inertia of the joist girder is computed by dividing the gross moment of inertia by 1.15.
6,810 in.4
I=
1.15
= 5,920 in.4

The resulting maximum deflection of the joist girder is:

5wL4
Δ max =
384EI
4
5 ( 0.163 kips/in.) ⎡⎣( 50 ft ) (12 in./ft )⎤⎦
=
384 ( 29,000 ksi ) ( 5,920 in.4 )
= 1.60 in.

This corresponds to an angle at Point A of:

−1.60 in.
θ=
( 50 ft )(12 in./ft )
= − 0.00267 rad

The joist will also be constructed with camber. According to the SJI Specifications, the approximate camber for joists with a span
of 50 ft is 1 in. The shape of the camber is not specified; however, it can be approximated as a circular arc. The angle at the end
of a member with circular arc camber, span, L, and maximum camber at midspan, c, is approximately 4c/ L.
4c
θ=
L
4 (1 in.)
=
( 50 ft )(12 in./ft )
= 0.00667 rad

The end rotation of a simply supported beam with unform load is:
wL3
θ=
24EI

The load on the joist is computed in the same manner as for the joist girder.
⎛ 50 ft ⎞
w = ⎡⎣( 20 lb/ft 2 ) + 0.7 ( 27.3 lb/ft 2 )⎤⎦
⎝ 8 ⎠
⎛ 1 kip ⎞ ⎛ 1 ft ⎞
= ( 244 lb/ft )
⎝ 1,000 lb ⎠ ⎝ 12 in.⎠
= 0.0203 kips/in.

The gross moment of inertia of the joist can be computed using Equation 2-1. The load that causes L/ 360 deflection in the joist
can be determined from the SJI Load Tables as 190 lb/ft.
I j = 26.8 × 10 − 6 wL 360 L3 (2-1)
−6
= 26.8 × 10 (190 lb/ft ) ( 50 ft − 3 ft ) 3

= 624 in.4 

44 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

 Table 5-1. Summary Results for Angle with
Respect to Horizontal at Point A for Example 4
Source Angle, rad
Roof slope +0.0156
Primary member (joist girder) camber +0.00167
Primary member (joist girder) deflection −0.00267
Secondary member (joist) camber +0.00667
Secondary member (joist) deflection −0.0116
Total +0.00967

The effective moment of inertia of the joist is computed by dividing the gross moment of inertia by 1.15.
624 in.4
I=
1.15
= 543 in.4

The end rotation is:

wL3
θ=
24EI
− ( 0.0203 kips/in.) ⎡⎣( 50 ft ) (12 in./ft )⎤⎦
3

=
24 ( 29,000 ksi ) ( 543 in.4 )
= − 0.0116 rad

A summary of the calculations is presented in Table 5-1. Summing the angles from each of the five sources results in a positive
angle of 0.00967 rad. Because this angle is positive even under dead and snow loads, water will not accumulate on the bay adja-
cent to the free-draining edge.
The interior bays also need to be checked because their slope is also less than 4 in./ft. The process would be similar, computing
the angle with respect to the horizontal at Point B (Figure 5-13) of the joist at the center of the interior bay and ensuring that it
was positive, thus allowing water to flow out of the bay. However, by inspection of the results in Table 5-1, the angle at Point B
will be even greater than that at Point A because joists in the interior bays are supported by joist girders at both ends; the only
difference will be that the angles due to camber and deflection of the primary members will be zero.
Because it has been established that a positive slope toward the free edge will remain in the deformed configuration, rain loads
will be zero and ponding will not occur. However, even if the angle was negative, the bay may be strong enough to support the
resulting rain loads. The following example illustrates such an evaluation.

5.5 EXAMPLE 5—BAY WITH LOW SLOPE AND SECONDARY MEMBERS

PARALLEL TO A FREE-DRAINING EDGE

Given:
Determine if the roof from Example 4 is adequate for ponding if the joists run parallel to the free draining edge.

Solution:
Just as in the previous example, the undeflected surface of the roof has positive slope to a free-draining edge; therefore, the static
head is zero. However, water can still accumulate if roof deflections cause a low point to form in the interior of the bay. ASCE/
SEI 7, Section 8.3, identifies conditions where an investigation to determine if low points form is required. Because this bay is

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 45

adjacent to a free-draining edge, and the secondary members are parallel to the free-draining edge, the slope at which further
investigation is required is:
Ls S + π
β=
20
( 50 ft ) ( 6.25 ft ) + π
=
20
= 0.557 in./ft

The roof slope is x in./ft as given in the problem statement; thus, further investigation is required. Note that the roof slope is also
less than 4 in./ft, so the roof meets both conditions for further evaluation.
A low point will form in the bay adjacent to the free-draining edge if the elevation of Point B becomes lower than Point A, where
the points are identified in Figure 5-14. Roof slope, camber in the joist girder, and camber in the joist increase the elevation of
Point B with respect to Point A, while deflection of the joist girder and deflection of the joist decrease the elevation of Point B
with respect to Point A. Each source will be evaluated separately.
The roof has a slope of x in./ft, and the joists are spaced at 6 ft-3 in. This corresponds to a difference in elevation of:
z = ( x in./ft ) (6.25 ft)
= 1.17 in.

50'-0"
50'-0"
50'-0"

Point B
Point A

Fig. 5-14. Roof framing plan for Example 5.

46 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

The joist girder will be constructed with camber that further increases the elevation difference between points A and B. Accord-
ing to the SJI Specifications, the approximate camber for joist girders with a span of 50 ft is 1 in. Joists are constructed with
camber in the shape of a circular arc with radius, r = 3,600 ft = 43,200 in. Note that using this radius and the member length of
50 ft, the camber can be computed as 1.04 in. The approximate value of 1 in. will be used for these calculations because the dif-
ference is minor.
Based on the geometry of a circular arc, the camber at a distance x from a member end is:
2
⎛L ⎞
z = r2 − − x − (r − c )
⎝2 ⎠

Therefore, the camber at the location along the joist girder where the first joist frames in is:
2
⎡ ( 50 ft ) (12 in./ft ) ⎤
z = ( 43,200 in.) − ⎢ 2
− 75 in.⎥ − ( 43,200 in. − 1 in.)
⎢⎣ 2 ⎥⎦
= 0.414 in.

The joist girder will deflect, decreasing the elevation difference. Approximating the loading on the joist girder as uniform, the
deflection at the location along the joist girder where the first joist frames in is:
wx
z=− ( L3 − 2Lx 2 + x 3 )
24EI

=−
( 0.163 kips/in.)( 75 in.)
24 ( 29,000 ksi ) ( 5,920 in. )
{ ⎡⎣(50 ft )(12 in./ft )⎤⎦ − 2 ⎡⎣(50 ft )(12 in./ft )⎤⎦ (75 in.) + (75 in.) }
4
3 2 3

= − 0.622 in.

where the joist girder loading and the moment of inertia of the joist girder are the same as those calculated in Example 4.
The joist will also be constructed with camber. According to the SJI Specifications, the approximate camber for a joist with a span
of 50 ft is 1 in, which adds directly to the difference in elevation between Points A and B.
z = 1 in.

Finally, deflection of the joist will subtract from the difference in elevation between Points A and B. Using the loading and
moment of inertia calculated in Example 4, the deflection of the joist at Point B is:
5wL4
z=−
384EI
4
5 ( 0.0203 kips/in.) ⎡⎣( 50 ft ) (12 in./ft )⎤⎦
=−
384 ( 29,000 ksi ) ( 542 in.4 )
= − 2.18 in.

where the joist loading and the moment of inertia of the joist are the same as those calculated in Example 4.
A summary of the calculations is presented in Table 5-2. Summing the differences in elevation from each of the five sources
results in a negative difference in elevation of −0.218 in., indicating that Point B will be lower than Point A in the deformed con-
figuration. Thus, a positive slope to drainage does not exist, and water will accumulate on the bay adjacent to the free-draining
edge.
The finding that water will accumulate on the roof does not necessarily indicate that the design is insufficient. If it can be shown
that the roof has adequate strength and stiffness to preclude progressive deflection and resist the ponding loads, then the design
is adequate. The SJI Roof Bay Analysis Tool will be used to further investigate the design.
After entering the pertinent information on the “Joist & Joist Girders” sheet, the general input table on the “Ponding Analysis”
sheet will appear as shown in Figure 5-15.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 47

 Table 5-2. Summary Results for Difference in Elevation
Between Point A and Point B for Example 5
Source Difference in Elevation, in.
Roof slope +1.17
Primary member (joist girder) camber +0.414
Primary member (joist girder) deflection −0.622
Secondary member (joist) camber +1.00
Secondary member (joist) deflection −2.18
Total −0.218

The ponding specific input needs to be defined for this case. While the preceding analysis showed that water will accumulate in
the bay, water may still drain over the edge. Thus, the water level is defined as a nominal height of 0.5 in. above the height of the
free-draining edge. The snow density is computed based on the ground snow load. The snow density is computed from ASCE/
SEI 7, Equation 7.7-1, as:
γ s = 0.13pg + 14 ≤ 30 lb/ft 3
= 0.13 ( 39 lb/ft 2 ) + 14 ≤ 30 lb/ft 3
= 19.1 lb/ft 3 ≤ 30 lb/ft 3
= 19.1 lb/ft 3

The concurrent snow load factor may be defined as 0.75; however, it is conservatively defined in this example as 1.00 to be con-
sistent with the preceding calculations to determine if water would accumulate on this bay. The top of roof elevation is defined
based on the slope, and the “rightmost joist,” which is the wall for this example, is defined as rigid. Default computed values of
camber and effective moment of inertia are used. The ponding specific input is shown in Figure 5-16.
With all the input defined, the ponding analysis can be performed. The results of the analysis confirm that the bay does accumu-
late water; however, the impact on the strength of the bay is minor. As shown in the joist output table in Figure 5-17, only the
two joists nearest the free-draining edge support accumulated water (indicated by increased strength ratios); however, the joists
are strong enough to accommodate the increased loads. As shown in the joist girder output table in Figure 5-18, the joist gird-
ers are subjected to somewhat larger panel point loads at the two panel points nearest the free-draining edge. These larger panel
point loads do not result in required shear or flexural strengths that exceed the available strengths. However, the panel point load

General Input (Defined in Joists & Joist Girders spreadsheet)

Design Methodology ASD
Minimum Design Loads ASCE 7-22
Primary Member Type Joist Girder
Secondary Member Type Joist
Joist Span 50.00 ft
Joist Girder Span 50.00 ft
Joist Size 30K11
Joist Girder Size 52G8N12.3K
Number of Joist Spaces 8
Dead Load Supported by Deck 10.00 psf
Dead Load Supported by Joists 19.00 psf
Joist Girder Self Weight 50.00 lb/ft
Snow Load 27.30 psf

Fig. 5-15. General input for Example 5.

48 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

nearest the free-draining edge exceeds the specified panel point load. Panel point loads greater than the specified panel point load
can cause load overstress of web verticals; however, this overload is minor and does not require action. These results demonstrate
that the bay possesses adequate strength and stiffness to preclude progressive deflection and resist potential ponding loads; thus,
the design is compliant with the provisions in ASCE/SEI 7, Sections 7.11 and 8.3.

Ponding Specific Input

Water level relative to zero datum: 0.50 in.
Compute load on deformed roof: Y (Y or N)
Snow density, γs: 19.10 lb/ft3
Concurrent snow load factor, k s: 1.00 ---
Top of roof elevation:
Top Left 9.375 in.
Top Right 0.000 in.
Bottom Left 9.375 in.
Bottom Right 0.000 in.
Camber:
Joist 1.042 in.
override: in.
Top Joist Girder 1.042 in.
override: in.
Bottom Joist Girder 1.042 in.
override: in.
Bay is mirrored:
Left Y (Y or N)
Right N (Y or N)
Top Y (Y or N)
Bottom Y (Y or N)
Joist support is wall:
Top N (Y or N)
Bottom N (Y or N)
Joist is rigid:
Joist 1 (Leftmost) N (Y or N)
Joist 9 (Rightmost) Y (Y or N)
Effective moment of inertia:
Joist 541.8 in.4
override: in.4
Joist Girder 5,967 in.4
override: in.4

Fig. 5-16. Ponding specific input for Example 5.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 49

Joist Output
Joist Max Shear Equiv. Load Shear Max Moment Equiv. Load Moment Strength
Number kips lb/ft Strength Ratio kip-ft lb/ft Strength Ratio Check
1 5.95 238.2 0.715 74.43 238.2 0.715 OKAY
2 5.95 238.2 0.715 74.43 238.2 0.715 OKAY
3 5.95 238.2 0.715 74.43 238.2 0.715 OKAY
4 5.95 238.2 0.715 74.43 238.2 0.715 OKAY
5 5.95 238.2 0.715 74.43 238.2 0.715 OKAY
6 5.95 238.2 0.715 74.43 238.2 0.715 OKAY
7 5.96 238.5 0.716 74.58 238.6 0.717 OKAY
8 6.08 243.2 0.730 76.57 245.0 0.736 OKAY

NOTES: 1. Loads and load effects correspond to ASD load combinations.

2. Shear strength ratio computed assuming shear capacity equal to 12.5% of the end reaction for
shear reversals, see Note 15 on the Ponding Instructions spreadsheet.

Fig. 5-17. Joist output for Example 5.

50 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

Joist Girder Output
Top Joist Girder Bottom Joist Girder
Joist Joist React. Panel Point Load Joist React. Panel Point Load
Number kips kips kips kips
2 5.95 12.22 5.95 12.22
3 5.95 12.22 5.95 12.22
4 5.95 12.22 5.95 12.22
5 5.95 12.22 5.95 12.22
6 5.95 12.22 5.95 12.22
7 5.96 12.24 5.96 12.24
8 6.08 12.47 6.08 12.47

Max Shear (kips): 42.81 Max Shear (kips): 42.81

Shear Strength Ratio: 0.999 Shear Strength Ratio: 0.999
Max Moment (kip-ft): 611.94 Max Moment (kip-ft): 611.94
Moment Strength Ratio: 0.999 Moment Strength Ratio: 0.999
Strength Check: OKAY Strength Check: OKAY
NOTES: 1. Loads and load effects correspond to ASD load combinations.
2. Highlighted panel point loads may cause a local overstress, see notes in Ponding Instructions.

Fig. 5-18. Joist girder output for Example 5.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 51

52 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40
Chapter 6
Concluding Notes
Ponding is an important consideration in the design of roofs, Mitigation is achieved by providing sufficient strength and
one that poses unique challenges due to the nonlinear nature stiffness to support the ponding load. This strategy requires
of the loading. This Design Guide has presented relevant performing a ponding analysis. It is possible that a roof
code provisions for the design of ponding, a recommended designed for other loads will be adequate for ponding loads
design methodology, and several design examples. and that modifications are not necessary to accommodate
Key provisions related to rain loads and ponding include: the ponding loads. However, if the roof is found to be inad-
• Rain load is the weight of water that will accumulate on equate, the following modifications can be implemented:
the roof in a specific design condition and consists of • Stiffen the roof system to reduce the additional load
static head, hydraulic head, and ponding head. caused by ponding.
• The primary drainage system and some secondary drain- • Strengthen the roof system so that it can support the addi-
age systems are required to be blocked in the design tional load caused by ponding.
condition. • Adjust the shape of the roof surface to reduce the addi-
• The lowest point of discharge that is not assumed to be tional load caused by ponding (e.g., increasing roof slope,
blocked is referred to as the SDSL. specifying camber, or adding tapered insulation or sloping
• The water rises to a level equal to the hydraulic head fill).
above the inlet to the SDSL. • Increase drainage capacity (e.g., widen scuppers or add
• Roof deflections caused by dead load and rain load must drains near columns).
be considered when determining the ponding head. • A combination of the above.
• No specific method of analysis for ponding is prescribed. The recommended design methodology is implemented
• Ponding instability checks must be performed for snow in the SJI Roof Bay Analysis Tool. This easy-to-use tool
load also, but the method of evaluation is not prescribed. is applicable to rectangular bays with primary members
supporting equally spaced secondary members. It explic-
The recommended design methodology, a state-of-the-art itly accounts for roof slope, camber, and a variety of edge
procedure that directly assesses loads in the ponded condi- conditions.
tion based on the deformed shape of the roof, complies with Structural steel and open web steel joists are economi-
these requirements and permits either of the two general cal and efficient for roof systems, but their use requires that
strategies for addressing ponding in design: avoidance and design professionals give close attention to the potential for
mitigation. ponding. With ponding, aspects of design not typically asso-
Avoidance is achieved by providing sufficient slope to a ciated with strength—such as deflections, roof slopes, cam-
free-draining edge. This can be achieved using sloping roof ber, and the locations and sizes of roof drains—can strongly
members, tapered insulation, or sloping fill. Roofs with suf- influence the safety and reliability of the roof system. None-
ficient slope do not form low points where water can accu- theless, with proper attention and coordination, ponding can
mulate when subjected to dead and snow loads. be effectively addressed by the design team.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 53

54 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40
Symbols

A Tributary area for drain, ft2 Zx Plastic section modulus, in.3

Bp Amplification factor for ponding c Maximum camber at midspan, in.
C Flexibility coefficient for a member cd Coefficient of discharge for open-channel scuppers
Cp Flexibility coefficient for primary members d Depth of joist girder, in.
Cs Flexibility coefficient for secondary members dh Hydraulic head, in.
2
D Dead load, lb/ft dp Ponding head, in.
E Modulus of elasticity of steel, ksi ds Static head, in.
2
EI Flexural stiffness, kip-in. g Acceleration of gravity = 386 in./s2
Fy Steel yield stress, ksi i Rainfall intensity, in./hr
4
I Moment of inertia, in. ks Concurrent snow load factor
Id Moment of inertia of deck, in.4/ft pg Ground snow load, lb/ft2
Ij Approximate gross moment of inertia of joist, in.4 r Radius of camber, in.
Ijg Approximate gross moment of inertia of joist girder, w Uniform load, kips/in.
in.4
wL/360 Nominal live load that will produce an approximate
Ip Moment of inertia of primary member, in.4 deflection of 1/360 of the span, lb/ft
Is Moment of inertia of secondary member, in.4 wa Uniform load using ASD load combinations, lb/ft
L Member length, ft wi Initial uniform load (without ponding effect), lb/ft
L Design length of joist or joist girder, span − 3 ft, ft wu Uniform load using LRFD load combinations, lb/ft
Lp Primary member length, ft w(x) Total load on a beam at position x, kips/in.
Ls Secondary member length, ft x Position along the length of a beam, girder, joist, or
joist girder, in.
Ma Required flexural strength using ASD load combi-
nations, kip-in. y(x) Total deflection along beam span at position x,
kips/in.
Mn Nominal flexural strength, kip-in.
z Difference in elevation, in.
Mp Plastic moment strength, kip-in.
Δmax Maximum deflection, in.
Mu Required flexural strength using LRFD load combi-
nations, kip-in. Ω Safety factor
N Number of joist spaces β Roof slope of bay with secondary members parallel
to free draining edge, in./ft
Pc Specified panel point load, kips
γ Unit weight of water = 62.4 lb/ft3
Q Flow rate, gal/min
γs Density (unit weight) of snow, lb/ft3
R Rain load, lb/ft2
θ Roof slope at free draining edge, rad
S Snow load, lb/ft2
ϕ Resistance factor
S Secondary member spacing, ft
Ws Scupper width, in.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 55

56 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40
Glossary

Cricket. A built-up ridge structure on a roof designed to Secondary drainage system for structural loading (SDSL).
divert water away from areas that may collect water. Roof drainage system through which water is drained
Primary member. In a typical framing system, a structural from the roof when the drainage systems that are required
member that spans between columns or walls and sup- to be assumed to be blocked by ASCE/SEI 7, Section 8.2
ports secondary members—for example, girders or joist are blocked or not working.
girders. Secondary member. In a typical framing system, a structural
Scupper. An opening in the side of a building (typically member that spans between primary members or walls
through a parapet wall) for the purpose of draining water and supports decking—for example, beams and open web
off the roof. steel joists.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 57

58 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40
References

AISC (2016), Specification for Structural Steel Buildings, Denavit, M.D., Zhu, M., and Scott, M.H. (2022),
ANSI/AISC 360-16, American Institute of Steel “Verification and Validation of a Nonlinear Analysis
Construction, Chicago, Ill. Model for Ponding,” Proceedings of the Annual Stability
AISC (2022), Specification for Structural Steel Buildings, Conference, Structural Stability Research Council.
ANSI/AISC 360-22, American Institute of Steel Finnemore, E.J. and Franzini, J. (2001), Fluid Mechanics
Construction, Chicago, Ill. with Engineering Applications, McGraw-Hill Education,
AISC (2023), Steel Construction Manual, 16th Ed., Boston, Mass.
American Institute of Steel Construction, Chicago, Ill. Fisher, J.M. and Denavit, M.D. (2018), Structural Design
AISI (2016), North American Specification for the Design of of Steel Joist Roofs to Resist Ponding Loads, Technical
Cold-Formed Steel Structural Members, AISI S100-16w/ Digest 3, Steel Joist Institute, Florence, S.C.
S1-18, American Iron and Steel Institute, Washington, FM Global (2021), Property Loss Prevention Data Sheet
D.C. 1-54, Roof Loads and Drainage, FM Global, Johnston,
ASCE (2016), Minimum Design Loads and Associated R.I.
Criteria for Buildings and Other Structures, ASCE/SEI ICC (2021a), International Building Code, International
7-16, American Society of Civil Engineers, Reston, Va. Code Council, Inc., Washington, D.C.
ASCE (2022), Minimum Design Loads and Associated ICC (2021b), International Plumbing Code, International
Criteria for Buildings and Other Structures, ASCE/SEI Code Council, Inc., Washington, D.C.
7-22, American Society of Civil Engineers, Reston, Va. Kloiber, L.A. and Burmeister, S.B. (2020), Design
ASCE (2023), “ASCE Hazard Tool,” American Society of Considerations for Camber, Design Guide 36, AISC,
Civil Engineers, Reston, Va., https://ascehazardtool.org. Chicago, Ill.
ASTM (2019), Standard Specification for Carbon Structural Marino, F.J. (1966), “Ponding of Two-Way Roof Systems,”
Steel, ASTM A36/A36M-19, ASTM International, West Engineering Journal, AISC, Vol. 3, No. 3, pp. 93–100.
Conshohocken, Pa. NOAA (2023), “Precipitation Frequency Data Server
ASTM (2022), Standard Specification for Structural Steel (PFDS),” National Oceanic and Atmospheric Admin­
Shapes, ASTM A992/A992M-22, ASTM International, istration, Washtington, D.C., https://hdsc.nws.noaa.gov/
West Conshohocken, Pa. pfds.
Baber, T.T. and Rigsbee, E.D. (2010), “Noniterative Finite NRCC (2020), National Building Code of Canada 2020,
Element Analysis of Ponding,” Proceedings of the ASCE/ National Research Council Canada, Ottawa, Canada.
SEI Structures Congress, ASCE, pp. 1,150–1,159. O’Rourke, M. and Downey, C. (2001), “Rain-on-Snow
Carter, C.J. and Zuo, J. (1999), “Ponding Calculations in Surcharge for Roof Design,” Journal of Structural
LRFD and ASD,” Engineering Journal, AISC, Vol. 36, Engineering, ASCE, Vol. 127, No. 1, pp. 74–79.
No. 3, pp. 138–141. O’Rourke, M. and Lewis, A.R. (2020), Rain Loads: Guide
Colombi, P. (2006), “The Ponding Problem on Flat Steel to the Rain Load Provisions of ASCE 7-16, American
Roof Grids,” Journal of Constructional Steel Research, Society of Civil Engineers, Reston, Va.
Vol. 62, No. 7, pp. 647–655. SDI (2022), Standard for Steel Deck, ANSI/SDI SD-2022,
Denavit, M.D. (2019), “Approximate Ponding Analysis by Steel Deck Institute, Glenshaw, Pa.
Amplified First-Order Analysis,” Engineering Structures, Silver, E. (2010), “A Strength Design Approach to Ponding,”
Vol. 197, No. 7, pp. 109428. Engineering Journal, AISC, Vol. 47, No. 3, pp. 175–187.
Denavit, M.D. and Scott, M.H. (2021), “Strength and SJI (2003), 75-Year Steel Joist Manual: A Compilation of
Reliability of Structural Steel Roofs Subjected to Ponding Specifications and Load Tables Since 1928, Steel Joist
Loads,” Journal of Structural Engineering, ASCE, Institute, Myrtle Beach, S.C.
Vol. 147, No. 2, pp. 04020318.

AISC DESIGN GUIDE 40 / RAIN LOADS AND PONDING / 59

SJI (2020), Standard Specifications for K-Series, LH-Series,
and DLH-Series Open Web Steel Joists, and for Joist
Girders, ANSI/SJI 100-2020, Steel Joist Institute,
Florence, S.C.
SJI (2022), “Roof Bay Analysis Tool—With Ponding An­al­y­
sis,” version 5.1, Steel Joist Institute, Florence, S.C., https://
steeljoist.org/resources/roof-bay-analysis-tool-with-
ponding-analysis.
Wilen, J. (2012), “Still Water Runs Deep,” Professional
Roofing, July, pp. 34–38.

60 / RAIN LOADS AND PONDING / AISC DESIGN GUIDE 40

 Smarter. Stronger. Steel.
American Institute of Steel Construction
312.670.2400 | www.aisc.org

D840-24