For building owners, facilities and portfolio managers, and commercial roofing contractors

Drainage decides what the roof weighs. The edge decides whether it stays.

Low-slope commercial and industrial roofs · internal drains, scuppers, overflow provision, perimeter edge metal

Two systems that are cheap to specify properly, routinely trimmed in value engineering, and responsible out of all proportion for the losses that follow.

30-second answer

What actually decides whether a low-slope roof drains, and whether it stays attached?

Two line items decide most of the risk on a low-slope roof. Primary drainage fails in predictable, maintainable ways, and the overflow provision behind it is a structural safeguard, because standing water adds about 5.2 pounds per square foot for every inch of depth. At the perimeter and corners, uplift is highest, and the edge metal is supposed to be a tested assembly — if anyone kept the test data.

Stop — immediate hazard

Standing water over an occupied space, with sagging, cracking, or doors that have stopped closing

Water is heavy and a low-slope roof can hold a great deal of it. A visibly sagging deck, new cracking in a ceiling or a wall below the pond, deflection you can see from the floor, or doors and partitions that have gone out of square under the affected bay are signs of a structure being loaded, not a roofing defect to schedule.

The deflection makes it worse rather than better: as the deck sags, the pond deepens, and the deeper pond adds more load. That feedback is why this is an evacuate-and-call situation rather than a get-the-ladder situation.

Clear the floor area beneath the standing water, restrict access to the space below it, and get a structural engineer to the building. Do not send maintenance staff onto the roof to open the drain.

Learning paths and saved lessons

At a glance

The short versionSection link

Six numbers and definitions this page uses throughout. Every one of them is a planning figure or a model provision, not a determination for your building.

Weight of standing water: 5.2 lb per sq ft, per inch of depthArithmetic, not a citation: water at 62.4 lb per cubic foot, divided by twelve. It is the same 5.2 that appears as the coefficient in the rain-load equation.
Where the overflow inlet sits: Typically about 2 in. above the roof low pointThe design height published trade guidance describes as typical. The governing number is in your adopted code and in the structural design for this roof.
Ponding, as the trade defines it: Water still standing 48 hours after rainThe Asphalt Roofing Manufacturers Association's definition, published by a manufacturers' trade association. It is an industry convention rather than a code threshold.
What edge metal is tested to: ANSI/SPRI/FM 4435/ES-1Test Methods RE-1, RE-2 and RE-3. Gutters are outside the standard and are covered separately by ANSI/SPRI GT-1.
Model code section for edge systems: 2021 and 2024 IBC §1504.6 · 2015 and 2018 IBC §1504.5The number moved between the 2018 and 2021 editions. Ohio, on the 2021 IBC under a rule effective 1 March 2024, is on §1504.6 — a specification written to the older numbering would cite §1504.5.
Uplift by zone: Corner > perimeter > fieldCorners highest, then perimeter, then field. Zone widths and pressures are a wind-design output for a specific building; no table can supply them.

Tradeoffs

When to spend on this, and when the money belongs elsewhereSection link

The advice this page gives is: fix the drainage while the roof is open, and buy edge metal you can document. Here is where that advice is right, and where it is wrong.

Best when

A tear-off or full replacement is already scheduled. Sump correction, drain relocation, added overflow provision and a tested edge system are cheapest when the membrane is already off and the deck is already exposed.
The roof is parapet-bounded on all sides, so water that does not reach a drain has nowhere to go but up.
The building has occupied or high-value space directly under a bay that ponds, or houses tenants whose lease makes water intrusion your problem rather than theirs.
You are adding weight to the assembly anyway — a recover, thicker insulation, a solar array — and the rain load question is already being asked of an engineer.
The roof carries heavy debris load: mature trees upwind, an adjacent taller building shedding onto it, or a process exhaust that deposits solids.

Think twice if

The roof is nearing the end of its planning range anyway. Retrofit overflow work on a roof you will replace in three years is money spent twice — unless the load question is live, in which case it is a safety item and the economics stop being the point.
Adding overflow scuppers means cutting a structural parapet or a shear wall. That is a structural alteration and it is not a roofing decision.
Nothing on the roof actually ponds. Chronic dry roofs with generous slope and redundant drains are not where your capital does the most good, and this page is not an argument for spending on a roof that drains.
Your building sits in a jurisdiction whose adopted edition exempts your specific reroof scope from adding secondary drainage. That exemption may be real, and it may also be the wrong reason to skip it.
The edge is a low parapet with a properly anchored, documented system already in place and no history of wind damage. Replacing sound, tested edge metal to buy a bigger test number is not risk reduction.

What changes the answer

Whether an engineer has ever calculated the rain load for this roof with the primary drainage assumed blocked. If nobody has, the whole conversation is being had without its central number.
The adopted code edition in your jurisdiction, its local amendments, and whether the work is being permitted under the building code or the existing building code — the two have carried different reroofing exceptions.
Whether the deck is steel, concrete, or wood, and whether the framing was designed with ponding instability checked. Stiff decks and flexible decks behave very differently under the same pond.
How much of the drainage problem is insulation rather than structure. A drain that sits proud of the deck because tapered insulation was built up around it is a design fault that a new tapered package can fix.
The wind design for this building: exposure, height, geometry, risk category and enclosure. Those set the pressures at the edge, and they are the reason a detail that works on a two-storey warehouse is not automatically adequate on a twelve-storey one.
Whether anybody actually has the ES-1 test data for the edge profile that is on the roof. Not the manufacturer's brochure — the tested resistance for the tested profile.

How it works

Water leaves on a schedule, or it stays and gets heavySection link

Every height on a low-slope roof is measured from one datum, and every pound of water is measured from the same arithmetic. Start there and the rest of the page follows.

Section through a roof bay where the primary drain has been blocked by debris. The static head is the distance from the roof low point up to the overflow inlet; the hydraulic head is the water standing above that inlet, and it is what pushes water out through the overflow. Drawn to a vertical scale of one inch of water depth per twelve units, so the two-inch static head and the four-inch scupper opening are comparable by eye.Original diagram, Understanding Roofing. Callout text is repeated in full below the figure.

What the numbered callouts are

Debris on the strainer. The primary roof drain is blocked. Leaves, needles, wind-blown film, roofing debris left by the last trade on the roof, or hail and ice piled over the dome. The strainer is doing exactly what it is for — keeping solids out of the leader — and the cost of that is that it collects them.
The drain bowl, set in a sump. A correctly set drain sits below the surrounding roof surface, with the membrane clamped between the bowl and a clamping ring, and the bowl anchored through the deck. Where the bowl instead sits level with or proud of the finished insulation, water never reaches it until it is already several inches deep.
The roof low point. The datum. Every other height on the drawing is measured up from here.
The overflow inlet. Drawn two inches above the low point. The distance from the roof surface up to this inlet is the static head, written d_s. It is water that is on the roof before the secondary system does anything at all.
Water above the inlet. The hydraulic head, written d_h. This is the head that actually drives flow through the overflow, which is why an overflow that is set too high, or an opening that is too small, shows up as a deeper pond rather than as a leak.
The overflow scupper. An opening through the parapet, discharging over the side of the building where somebody can see it. That visibility is a design feature, not an accident: a secondary system that discharges quietly into the same storm line as the primary one tells nobody that the primary one has failed.
Deck deflection. The deck sags under the water, which deepens the pond, which adds load, which increases the sag. Trade guidance describes exactly this feedback: as water accumulates, deck deflections increase, and the increased deflection results in additional ponding water. The extra depth it produces is the ponding head, d_p.

Primary and secondary are different systems with different jobs

The primary system — internal drains, or through-wall scuppers at the roof surface — is the one that is supposed to work every time it rains. It is a plumbing system with a roofing interface, and it is maintained.

The secondary provision is not a second primary system. It is sized and set for a condition the primary system is assumed to be failing in. Trade guidance describes secondary systems as designed to remove excess water once it reaches a designed height — typically about two inches — before the roof structure is overloaded. It is also sized on a different basis, and the two authorities do not agree on which. Trade guidance records that US plumbing codes require both primary and secondary drainage to be designed for a hundred-year, sixty-minute rainfall rate, while the loading standard agrees on that rate for primary drainage and recommends a hundred-year, fifteen-minute rate for overflow — a shorter, more intense burst. Which one governs your roof is a question for the designer and the AHJ, not one this page can settle; the point is that the secondary system is deliberately sized against a harder event, because it is the last thing standing between a blocked drain and a structural problem.

The strainer is not the enemy

It is common to find strainer domes missing from commercial roofs, usually removed by someone who decided the drain would flow better without one. Trade guidance says the opposite twice over: strainers block debris from getting into the drain lines, and they can actually improve flow into the drain by breaking up the vortex that forms above it. A missing strainer moves the blockage from a place you can reach with a broom to a place that requires a plumber and a camera.

Debris is the failure mode, and it has been measured

The single best-documented statement of the problem opens a laboratory study of drain and scupper flow rates published through the building-enclosure consultants’ institute: drainage of rainwater has long been considered an essential attribute for the proper performance of any roof system, and long-term and excessive accumulation of water will contribute to the deterioration of most roofing systems and, in worst-case scenarios, has been responsible for excessive live loads that can lead to structural collapse. The study ran drains and scuppers at controlled flow rates with the inlets progressively obstructed and measured how deep the water stood. Only the abstract is publicly reachable, so this page does not reproduce the measured depths — but the direction of the finding is not in dispute, and it is the reason the secondary provision exists.

Failure modes

How primary drainage actually failsSection link

None of these are exotic. All of them are visible from the roof surface, and most of them are visible in a photograph.

The useful thing about primary drainage is that it fails in a small number of ways, and every one of them leaves evidence. The table below is this page’s own catalogue, built from the mechanisms described in the trade guidance cited at the foot of the page. The last column is the one that matters most for capital planning: some of these failures the secondary provision covers, and some of them it does not.

Nine ways a low-slope roof stops draining, and whether the overflow provision is any help. Original table, Understanding Roofing, built from mechanisms described in the trade guidance cited below.
Failure mode	Mechanism	What is visible	Does the overflow provision help?
Debris on the strainer	Leaves, needles, wind-blown film, gravel or trade debris collect on the dome. Flow area falls; water depth rises.	A dark tide-line ring around the drain in dry weather; standing water more than 48 hours after rain; vegetation at the low point.	Yes. This is the exact condition the secondary provision is designed for.
Strainer removed	Somebody removed the dome to make the drain flow faster. Debris now travels into the leader instead of stopping at the roof surface.	A bare clamping ring with no dome. Often several on the same roof, because whoever did it did it everywhere.	Yes, once the leader blocks — but the blockage has moved somewhere you cannot reach without a plumber.
Drain rim above the roof plane	Insulation and cover board are built up around a drain that was never reset, so the rim ends up level with or above the surrounding roof.	A permanent ring of standing water around a drain with a clean strainer. Wet roof, no debris.	Partly. The overflow still limits the maximum depth, but the roof now carries a permanent load the design did not assume.
Tapered insulation built up at the drain	A tapered package re-slopes the field without keeping a sump at the drain, or thickens the roof around it.	Water standing in a shallow saucer whose low point is not at the drain. Best caught on the tapered layout drawing, before it is built.	Partly, and the same caveat applies. Trade guidance warns directly that increases in insulation thickness can restrict drains and overflow systems.
Retrofit drain insert	A smaller drain body is dropped inside an existing leader to make a watertight connection to the new membrane, reducing flow area at the one point that is already the constraint.	Nothing, from the roof. It is a specification decision, visible on the submittal and in the drain schedule.	It raises the odds the overflow is needed. Ask what flow capacity the retrofit leaves and who calculated it.
Ice or hail over the inlet	Hail piles over the inlet, or the inlet and the scupper throat freeze. Primary and secondary can be blocked at the same time, in the season the roof is also carrying snow.	Obvious while it is happening and gone afterwards, which is why post-storm walks matter more than scheduled ones.	Not reliably. This is the case where both paths fail together, and it is the argument for clearing before a freeze rather than after.
Blocked or undersized leader	The roof drain is fine; the pipe inside the building is not. Undersized, partly blocked, or discharging into a storm system that is itself surcharged.	Nothing on the roof. Water backs up at a drain with a clear strainer — the same signature as a rim set too high.	Yes, and this is a case where the overflow discharging visibly outside the building is the only way anyone finds out.
Overflow provision blocked, too high, or absent	The secondary path is itself obstructed, set too high after a recover raised the field, or sized without coordination with the rain load the structure was designed for.	Measurable. The elevation of every overflow inlet relative to the current roof low point is a five-minute check after any recover.	No. This is the failure of the thing that was supposed to catch the other failures.
Ponding instability	A flexible deck sags under accumulated water; the sag deepens the pond; the deeper pond increases the sag.	A pond deeper in the middle of a bay than the slope explains; visible sag from the floor; new cracking in ceilings or partitions beneath.	Not on its own. This is a structural condition, and it is why the model code requires ponding instability to be checked.

Read this table one item at a time

Debris on the strainer

Mechanism: Leaves, needles, wind-blown film, gravel or trade debris collect on the dome. Flow area falls; water depth rises.
What is visible: A dark tide-line ring around the drain in dry weather; standing water more than 48 hours after rain; vegetation at the low point.
Does the overflow provision help?: Yes. This is the exact condition the secondary provision is designed for.

Strainer removed

Mechanism: Somebody removed the dome to make the drain flow faster. Debris now travels into the leader instead of stopping at the roof surface.
What is visible: A bare clamping ring with no dome. Often several on the same roof, because whoever did it did it everywhere.
Does the overflow provision help?: Yes, once the leader blocks — but the blockage has moved somewhere you cannot reach without a plumber.

Drain rim above the roof plane

Mechanism: Insulation and cover board are built up around a drain that was never reset, so the rim ends up level with or above the surrounding roof.
What is visible: A permanent ring of standing water around a drain with a clean strainer. Wet roof, no debris.
Does the overflow provision help?: Partly. The overflow still limits the maximum depth, but the roof now carries a permanent load the design did not assume.

Tapered insulation built up at the drain

Mechanism: A tapered package re-slopes the field without keeping a sump at the drain, or thickens the roof around it.
What is visible: Water standing in a shallow saucer whose low point is not at the drain. Best caught on the tapered layout drawing, before it is built.
Does the overflow provision help?: Partly, and the same caveat applies. Trade guidance warns directly that increases in insulation thickness can restrict drains and overflow systems.

Retrofit drain insert

Mechanism: A smaller drain body is dropped inside an existing leader to make a watertight connection to the new membrane, reducing flow area at the one point that is already the constraint.
What is visible: Nothing, from the roof. It is a specification decision, visible on the submittal and in the drain schedule.
Does the overflow provision help?: It raises the odds the overflow is needed. Ask what flow capacity the retrofit leaves and who calculated it.

Ice or hail over the inlet

Mechanism: Hail piles over the inlet, or the inlet and the scupper throat freeze. Primary and secondary can be blocked at the same time, in the season the roof is also carrying snow.
What is visible: Obvious while it is happening and gone afterwards, which is why post-storm walks matter more than scheduled ones.
Does the overflow provision help?: Not reliably. This is the case where both paths fail together, and it is the argument for clearing before a freeze rather than after.

Blocked or undersized leader

Mechanism: The roof drain is fine; the pipe inside the building is not. Undersized, partly blocked, or discharging into a storm system that is itself surcharged.
What is visible: Nothing on the roof. Water backs up at a drain with a clear strainer — the same signature as a rim set too high.
Does the overflow provision help?: Yes, and this is a case where the overflow discharging visibly outside the building is the only way anyone finds out.

Overflow provision blocked, too high, or absent

Mechanism: The secondary path is itself obstructed, set too high after a recover raised the field, or sized without coordination with the rain load the structure was designed for.
What is visible: Measurable. The elevation of every overflow inlet relative to the current roof low point is a five-minute check after any recover.
Does the overflow provision help?: No. This is the failure of the thing that was supposed to catch the other failures.

Ponding instability

Mechanism: A flexible deck sags under accumulated water; the sag deepens the pond; the deeper pond increases the sag.
What is visible: A pond deeper in the middle of a bay than the slope explains; visible sag from the floor; new cracking in ceilings or partitions beneath.
Does the overflow provision help?: Not on its own. This is a structural condition, and it is why the model code requires ponding instability to be checked.

Original table. Mechanisms and the maintenance points behind them are drawn from the IIBEC and ARMA guidance cited in the source list; the grouping, the visibility column and the overflow column are this page's own analysis and are not a determination about any roof.

The arithmetic

What the water on the roof weighsSection link

One number carries this whole argument, and it is not a code citation. It is the density of water.

Water weighs 62.4 pounds per cubic foot. A layer one inch deep over one square foot is one twelfth of a cubic foot, so it weighs 62.4 ÷ 12 = 5.2 pounds. That is the whole derivation, and it is why 5.2 turns up as the coefficient in the rain-load equation used for structural design.

The table below applies it. It is arithmetic, deliberately: nothing in it accounts for the shape of a real pond, the distribution of load across a bay, the deflection that deepens the pond, or the capacity of any particular structure. It exists to give a facilities manager a sense of scale before an engineer gives them a number.

Load added by standing water, by depth and by roof area. Arithmetic from water at 62.4 lb per cubic foot; not a structural determination. Original worked example, Understanding Roofing.
Depth of standing water	Added load	Over a 10,000 sq ft area	Over a 40,000 sq ft area
1 inch	5.2 lb per sq ft	52,000 lb · 26 tons	208,000 lb · 104 tons
2 inches	10.4 lb per sq ft	104,000 lb · 52 tons	416,000 lb · 208 tons
4 inches	20.8 lb per sq ft	208,000 lb · 104 tons	832,000 lb · 416 tons
6 inches	31.2 lb per sq ft	312,000 lb · 156 tons	1,248,000 lb · 624 tons
8 inches	41.6 lb per sq ft	416,000 lb · 208 tons	1,664,000 lb · 832 tons

Read this table one item at a time

1 inch

Added load: 5.2 lb per sq ft
Over a 10,000 sq ft area: 52,000 lb · 26 tons
Over a 40,000 sq ft area: 208,000 lb · 104 tons

2 inches

Added load: 10.4 lb per sq ft
Over a 10,000 sq ft area: 104,000 lb · 52 tons
Over a 40,000 sq ft area: 416,000 lb · 208 tons

4 inches

Added load: 20.8 lb per sq ft
Over a 10,000 sq ft area: 208,000 lb · 104 tons
Over a 40,000 sq ft area: 832,000 lb · 416 tons

6 inches

Added load: 31.2 lb per sq ft
Over a 10,000 sq ft area: 312,000 lb · 156 tons
Over a 40,000 sq ft area: 1,248,000 lb · 624 tons

8 inches

Added load: 41.6 lb per sq ft
Over a 10,000 sq ft area: 416,000 lb · 208 tons
Over a 40,000 sq ft area: 1,664,000 lb · 832 tons

Uniform depth over a uniform area, in US short tons. Real ponds are not uniform and real decks deflect, which concentrates load and deepens the pond. This table is a sense of scale, not a load case: rain load and ponding instability for your building are calculated by a licensed structural engineer under your adopted code and the referenced loading standard.

Set that against the two-inch static head that the overflow provision is normally designed around. Two inches over a 40,000 square foot roof is a little over 200 tons of water sitting on the building before the secondary system has moved a drop — and that is the designed-for condition, the one the structure is supposed to be able to carry. Everything above the overflow inlet is the head that drives it out. Everything the overflow cannot get rid of is accumulating on a structure whose designer never intended it.

This is also why raising the roof matters. If a recover adds two inches of insulation across the field and the overflow inlets stay where they are, the geometry the structural assumption rested on has changed, and nobody involved in the roofing scope necessarily noticed.

The perimeter

The edge is not the field, and it is not trimSection link

Wind does not load a roof evenly. It concentrates at the places where the geometry breaks, which is exactly where the least-engineered component usually sits.

Left: a section through a parapet edge, showing what a tested edge system consists of and the two directions it is loaded in at once — suction on the top surface and an outward pull on the face. Right: a conceptual plan showing why the same detail is not equally adequate everywhere on the roof.Original diagram, Understanding Roofing. Callout text is repeated in full below the figure.

What the numbered callouts are

The wood nailer, and its anchorage. The nailer is bolted or screwed down into the parapet, and everything above it transfers load through those fasteners. It is the first thing covered and the last thing anyone inspects. A nailer anchored into deteriorated masonry, or a nailer that was simply left in place from the previous roof, is a tested system resting on an untested foundation.
The coping and its continuous cleat. The formed metal that caps the parapet, hooked over a cleat on the outside face. This is the component ES-1’s RE-3 test loads: front-face and back-face pulls with simultaneous load on the top surface.
The membrane termination. The membrane turns up the inside face of the parapet as base flashing and is held mechanically at the top by a termination bar. Where a roof is mechanically attached or ballasted rather than fully adhered, the edge metal may be restraining unadhered membrane directly — which is the specific case ES-1’s RE-1 test exists for.
The peel path. Once the edge releases, the membrane behind it is free at the perimeter, air gets underneath, and the failure runs back into the field. The reason a small metal component gets this much attention is that its failure mode is not local.

On the plan: the middle of the roof is the field, the band around all four sides is the perimeter, and the four squares are the corners. Suction is lowest in the field, higher at the perimeter, and highest at the corners. Department of Energy building-science guidance gives the mechanism directly: during high wind events, vortices form along the edges of the roof creating areas of localised negative pressure above the roof. This page draws the zones with no dimensions and no pressures on purpose — both are outputs of a wind design for one specific building under your adopted code and the referenced loading standard, and publishing either would be publishing a design determination.

What ES-1 actually tests

Edge metal is one of the few roofing components the model building code sends to a named test standard by name. In the 2024 model text, metal edge systems on low-slope built-up, modified bitumen and single-ply roofs must be designed and installed for wind loads in accordance with the structural chapter and tested for resistance in accordance with Test Methods RE-1, RE-2 and RE-3 of ANSI/SPRI ES-1. The full designation of the standard is ANSI/SPRI/FM 4435/ES-1, Test Standard for Edge Systems Used With Low Slope Roofing Systems.

The three ES-1 test methods the model code references, plus the standard that covers gutters instead. Compiled by Understanding Roofing from NRCA and IIBEC technical guidance; see the source list.
Test method	What it loads	How the result is expressed	What it does not tell you
RE-1	A metal edge flashing's capacity to restrain unadhered roof membrane at the perimeter — the case that arises on ballasted and mechanically attached single-ply roofs without a peel stop near the edge.	Pounds force per linear foot.	Nothing about an adhered roof, and nothing about whether the membrane behind the edge is itself adequately attached.
RE-2	Resistance of fascias and gravel stops to horizontal load — the outward pull away from the building face.	Pounds force per square foot.	Nothing about uplift on a coping's top surface, which is a different test.
RE-3	Copings' resistance to separate front-face and back-face pulls, applied while the top surface is simultaneously loaded.	Pounds force per square foot.	Nothing about the anchorage of the nailer the coping is fixed to, which is outside the tested specimen.
ANSI/SPRI GT-1 — gutters	Gutters used to secure the perimeter edge of the membrane. The 2021 and 2024 model text puts these in their own subsection and sends them to Test Methods G-1 and G-2 of SPRI GT-1.	Per the GT-1 methods, which are outside ES-1 entirely.	ES-1 explicitly does not apply to gutters, so an ES-1 report is not evidence about one.

Read this table one item at a time

RE-1

What it loads: A metal edge flashing's capacity to restrain unadhered roof membrane at the perimeter — the case that arises on ballasted and mechanically attached single-ply roofs without a peel stop near the edge.
How the result is expressed: Pounds force per linear foot.
What it does not tell you: Nothing about an adhered roof, and nothing about whether the membrane behind the edge is itself adequately attached.

RE-2

What it loads: Resistance of fascias and gravel stops to horizontal load — the outward pull away from the building face.
How the result is expressed: Pounds force per square foot.
What it does not tell you: Nothing about uplift on a coping's top surface, which is a different test.

RE-3

What it loads: Copings' resistance to separate front-face and back-face pulls, applied while the top surface is simultaneously loaded.
How the result is expressed: Pounds force per square foot.
What it does not tell you: Nothing about the anchorage of the nailer the coping is fixed to, which is outside the tested specimen.

ANSI/SPRI GT-1 — gutters

What it loads: Gutters used to secure the perimeter edge of the membrane. The 2021 and 2024 model text puts these in their own subsection and sends them to Test Methods G-1 and G-2 of SPRI GT-1.
How the result is expressed: Per the GT-1 methods, which are outside ES-1 entirely.
What it does not tell you: ES-1 explicitly does not apply to gutters, so an ES-1 report is not evidence about one.

ES-1 does not address weathertightness of metal edge flashings; it is a wind-resistance standard. The standard itself prescribes a safety factor of 2.0, which trade guidance records as not incorporated into the code's reference to it — so a resistance figure has to be read together with the basis it was reported on. Edition matters: the 2011 edition was referenced by the 2018 IBC and the 2017 edition by the 2021 IBC.

Where this actually goes wrong on a project

The failure is rarely that somebody knowingly installed untested metal. It is that the chain of documentation breaks. A designer specifies “ES-1 compliant” without stating the design pressures. A bid comes back with a shop-fabricated profile at a better price. Nobody asks for the test report, because the phrase “compliant” appeared in the submittal. The metal goes on, the coping looks identical to the tested one from the ground, and the building has an edge system whose resistance nobody ever compared to anything.

Trade guidance is explicit about what compliance actually requires: a designer has to specify edge flashings with tested resistance loads equal to or greater than the applicable design wind loads, and the design wind loads have to be on the construction documents. That is two numbers and a comparison. A shop-fabricated shape is entirely legitimate — the national roofing contractors’ association runs ES-1 certification programmes for shop-fabricated shapes with two independent testing bodies — but the legitimacy comes from the report, not from the shape.

The other quiet gap is the gutter. Where a gutter carries the perimeter edge of the membrane, it is doing a structural job and ES-1 does not cover it. Ask which standard the gutter was tested to, and expect GT-1 or an honest “none”.

Owner checklist

What to go and find out about your own roofSection link

None of this requires anyone untrained to go on the roof. Most of it is answerable from documents, from a contractor already under contract, or from a photograph.

Does this roof have a secondary drainage provision at all? Look at the roof plan, or ask. Many older low-slope roofs have none, and nobody discovers it until a drain blocks.
Where are the overflow inlets, relative to the current roof surface? Not the original surface — the current one. If the roof has been recovered since the overflows were built, this is the single highest-value measurement on the list.
Where does the overflow discharge? If it discharges somewhere nobody looks, or into the same storm line as the primary system, it can be doing its job for years without anyone learning that the primary system has failed.
Do the drains have strainers, and were they there last year? A photograph of every drain, annually, is a cheap and surprisingly informative record.
Is there standing water 48 hours after rain, and where? Map it. The same three spots every time is a design fault; a new spot is a deflection or a blockage.
Has anyone ever calculated the rain load for this roof? With the primary drainage assumed blocked. If the answer is no, that is the finding.
Do you hold the ES-1 test report for the edge metal that is on the building? If the roof was installed under a specification, it should be in the close-out documents. If it is not, you now know something about the close-out documents too.
Is anything at the edge loose? Rattling coping, separated joints, missing cleat fasteners, oil-canning that has got worse. These are visible from the ground with binoculars and from a window on an adjacent building.
Who clears the drains, how often, and where is the record? If the answer involves a warranty, the record is part of your coverage.
What happens on this roof after a storm? A post-storm walk by a qualified contractor catches ice, hail debris and lifted edge metal in the window where each is still cheap.

Cost and lifecycle

The lifecycle argument, and why there is no price on this pageSection link

Drainage correction and edge securement are among the cheapest things on a low-slope roof to do at the right moment and among the most expensive to do at any other moment. That asymmetry, not a unit price, is the number that should drive the decision.

Units: None. This page publishes no dollar figure.
Scope included: What a drainage-and-edge scope has to itemise before two proposals for it are comparable — stated as scope, not as price.
Not included: Any dollar amount, unit rate, percentage of project cost, or regional multiplier for drains, scuppers, sumps, tapered packages or edge metal.
Geography: United States. Nothing on this page is specific to any market.
Data as of: 26 August 2026 — the date every source listed below was opened and read. No cost dataset was used, because none was found that separates drains, overflow provision, sumps, tapered correction or edge metal as line items at a national scale.
Confidence: None for dollars: this page publishes no figure, and any figure it published would be invented. High for sequencing, because the sequencing argument depends only on when the deck and the membrane are open, which is a schedule fact rather than a market one.
Method: How this figure is built

Why the money is a sequencing question

Almost everything on this page is buried. A sump is under the membrane. A drain bowl is clamped through the deck. Overflow provision through a parapet is a hole in a wall. Edge metal is under the coping and over the nailer. Each of these is a modest scope while a roof is off and a substantial one afterwards, because the afterwards version involves removing and reinstating work that was just paid for — and, where a warranty is in force, doing it in a way the warranty holder will accept.

That means the drainage decision is made when the recover-versus-tear-off decision is made, not after it. A recover that leaves the existing drain elevations alone and adds two inches of polyiso over the top has just made every drain on the roof two inches shallower relative to the field, unless the tapered package was designed to keep the sump. Trade guidance is explicit that simply re-sloping a roof with tapered insulation may not be adequate, and that increases in insulation thickness can restrict drains and overflow systems.

What a comparable scope has to itemise

Number, size, type and elevation of primary drains, and whether existing bowls are reused, reset, or replaced.
Sump construction at each drain: pre-manufactured sump pan, tapered sump, or field-built — and its dimensions.
Strainer type at every drain, and whether the existing ones are being reinstated or replaced.
Secondary provision: what exists today, what is proposed, and if nothing is proposed, the written reason.
The tapered insulation layout with crickets and saddles shown on a drawing, not described in a sentence, and the average R-value that layout produces.
Edge system: profile, gauge, metal, fabricator, and the ES-1 test data for the profile actually being installed.
Anchorage of the wood nailer, which is the part of the edge that is inside the parapet and invisible after the coping goes on.
Attachment density by zone, and the zone map that produced it.
A flood test or other verification at the drains, and who witnesses it.

Two proposals that agree on membrane and disagree on this list are not two prices for the same roof. Normalising scope before price is the whole exercise.

The asymmetry that actually matters

Edge metal is a linear-foot item on a roof priced by the square foot, which makes it a small fraction of a re-roof. Its failure mode is not small: once the metal releases, the membrane behind it is unrestrained at the edge and peels back into the field. The loss is not a length of coping. It is the roof, plus the water that follows it into the building, plus whatever the building was doing that week. That is the shape of the risk, and it is why documentation on a cheap item is worth arguing about.

A planning range is not a quote. It is a number to argue with a proposal about — the only price that binds anyone is the one in a signed scope of work for this building.

Considerations

What changes this on a real buildingSection link

The axes that genuinely change this on a real building. Each one ends where a qualified professional starts.

Code and jurisdiction

There is no nationwide building code for site-built construction. States and local governments adopt and amend model codes on their own schedules, which is why a section number is not a fact about your building until you know your edition.

A worked example: Ohio Administrative Code rule 4101:1-15-01, “Roof assemblies and rooftop structures”, effective 1 March 2024, reads in its entirety that Chapter 15 of the International Building Code, 2021 edition, is incorporated by reference and modified in Section 101.1.1 of the code. So an Ohio project in 2026 is working from the 2021 model text, where edge systems sit at Section 1504.6 and gutter securement at Section 1504.6.1 — the same numbers the 2024 model text uses, but not the numbers the 2015 and 2018 editions used, where the provision was Section 1504.5.

Roof drainage moved as well. The secondary-drainage requirement that sat in the 2015 text at Section 1503.4.1 is at Section 1502.2 in the 2021 and 2024 text. Same requirements, different numbers — and a specification written to an edition your jurisdiction has not adopted invites an argument at inspection. The lesson is not the numbers. It is that a section number without an edition and a jurisdiction attached to it is not a citation.

Jurisdiction: Ohio. Edition adopted: IBC 2021, Chapter 15. Rule effective 1 March 2024, five-year review due 1 March 2029; prior effective dates 2002, 2005, 2007, 2011 and 2017. Official source: codes.ohio.gov. This is the law in Ohio and nowhere else, it does not describe local amendments, and it is not a determination about any permit. Confirm your own adopted edition, its amendments, and the interpretation with your authority having jurisdiction.

Structural weight

Rain load is a structural load with its own code provisions. Model text requires each portion of a roof to sustain the load of rainwater that accumulates on it if the primary drainage system for that portion is blocked, plus the uniform load caused by water rising above the inlet of the secondary drainage system. Separately, roofs below a threshold slope must be checked for ponding instability — verification that the structure is stiff enough to preclude progressive deflection.

Both of those are engineering determinations. This page explains what the terms mean so you can tell whether they have been done. It does not do them, and neither does a contractor’s proposal.

Rain load, ponding instability, deck capacity and any change to a structural parapet are the work of a licensed structural engineer for this specific building. Nothing here is a structural determination, and the arithmetic on this page is arithmetic, not a design method.

Wind

Uplift is not uniform across a roof. Department of Energy building-science guidance puts the mechanism plainly: during high wind events, vortices form along the edges of the roof creating areas of localised negative pressure — suction — above the roof. That is why the same assembly is specified at one attachment density in the field, a higher one at the perimeter, and a higher one again at the corners.

It is also why an edge detail is a wind item rather than a trim item. The edge sits in the highest-pressure part of the roof and is loaded in two directions at once: suction on its top surface and an outward pull on its face.

Uplift performance belongs to a tested assembly on a specific building at a specific site. A membrane does not carry a wind rating by itself, and neither does a coping profile. Design pressures, zone widths and attachment densities are outputs of a wind design under the adopted code and standard for your building — not values this page can supply.

Maintenance

Drainage is the one part of a commercial roof whose performance degrades entirely through neglect rather than through age. A drain that was correct on the day it was installed is blocked four autumns later, and nothing about the roof has changed except what is lying on it.

A defensible maintenance programme names the cadence, names who does it, and produces a record: drains and strainers cleared, scupper openings clear, overflow inlets clear and at the right height, debris removed from the whole field rather than raked toward the drain, and a walk after every major storm and every trade visit that put people on the roof.

Debris load is site-specific. Trees, adjacent taller buildings, process exhaust and nearby construction all change the cadence, and no general interval substitutes for looking at what your roof actually collects.

Climate

Cold climates add a failure mode that the drawing on this page does not show: the inlet freezes. Ice over a strainer dome, or an ice plug in a scupper throat, blocks the primary and the secondary path at the same time, and it does it during the season when a roof is also carrying snow. Hail does something similar for a shorter time — hailstones pile over a drain inlet and act as debris until they melt.

Warm and wet climates load the other side of the problem: chronic ponding accelerates membrane deterioration, and dirt and debris in the ponded area damage the membrane surface as well as feeding the blockage.

Rainfall intensity is not stationary. Trade guidance presented in 2025 notes that extreme-precipitation days have risen sharply in several US regions since the mid-twentieth century while the rainfall-intensity maps in the plumbing code have remained unchanged across editions — so a drainage system sized to code is sized to a historical assumption, not to a guarantee.

Moisture and ventilation

Chronic ponding does not stay a surface problem. Water finds a seam or a fastener, saturates the insulation beneath it, and the saturated insulation is both heavier than the design assumed and substantially less insulating than the R-value on the submittal. Wet insulation is also why a small leak at a drain can produce a very large area of wet assembly before anyone sees a ceiling stain.

This is the point at which a drainage problem becomes an insulation and vapour-control problem and a moisture-survey problem, and the point at which core cuts stop being optional.

Access and site conditions

Everything described on this page happens on a commercial roof, which is a workplace. Under the federal general-industry standard, an employee on a walking-working surface with an unprotected side or edge four feet or more above a lower level must be protected from falling, and the standard sets distinct requirements for low-slope roof work within six feet of the edge, between six and fifteen feet, and beyond fifteen feet. Holes, including skylights, are separately covered.

The practical consequence for an owner is that drain clearing is occupational work with a fall-protection obligation attached, and the person who is closest to the drain — at the low point, in standing water, often near an edge or a scupper — is exactly the person the standard is written about.

Nothing on this page is an instruction to go on a roof. The fall-protection standard is addressed to employers; the fact that trained crews use it is a reason to keep untrained staff off the roof entirely rather than a procedure to copy.

Warranty and repair

Who warrants the edge, and what maintenance the coverage is conditioned onSection link

Warranty questions on drainage and edges are mostly questions about who owns which part of the assembly. Resolve them in writing before the work starts, because the answer after a failure is whichever answer the documents happen to support.

Which document covers what: A membrane manufacturer’s system warranty and a contractor’s workmanship warranty are different documents with different terms, different durations and different holders. A no-dollar-limit warranty and a material-only warranty are also different products in kind, not in degree: they differ on whether labour to repair is covered at all. Ask which document covers which component, and get the answer as a document rather than as an assurance.
The edge specifically: Perimeter edge metal is frequently supplied by a sheet-metal fabricator, installed by the roofing contractor, and specified by a designer, which is three parties before the membrane manufacturer is involved. Whether the edge is inside the membrane system warranty, inside the workmanship warranty, inside a fabricator’s product warranty, or in none of them is a question with a real answer for your project. Ask it before award.
Maintenance as a condition: Roofing warranties commonly carry maintenance and inspection conditions, and drainage is the obvious candidate for one. If clearing drains is your obligation under the document, then your maintenance record is part of your coverage, and an undocumented programme is a weak position in a claim regardless of whether the work was actually done.
The interfaces: Ask specifically how a leak at a drain, at a scupper throat, or at the perimeter termination will be characterised. These are the three locations where the roofing scope meets somebody else’s scope — plumbing, masonry, and the building structure — and they are where the sentence “that is not a roofing defect” is most often used.

Repairability

Drainage defects are among the more repairable things on a low-slope roof, and among the least often repaired. A blocked strainer is a maintenance visit. A sump that was built too shallow can usually be corrected locally with a tapered sump package during a re-roof. A drain bowl that sits at the wrong elevation is a bigger job, because correcting it means opening the deck.

Retrofit drain inserts — a smaller drain body dropped into an existing leader — are the tempting shortcut, and they are a deliberate reduction in the flow area of a drain that is already the constraint. Where one is proposed, ask what the resulting flow capacity is and who calculated it.

Edge metal repairs badly. A damaged coping run can be replaced, but matching a profile that was shop-fabricated years ago to a drawing nobody kept is a fabrication exercise, and the replacement will only be a tested system if the replacement profile is one that was tested. This is a reason to keep the submittal.

This page does not interpret any warranty and does not tell you what yours covers, whether it transfers, or whether it is enforceable. Those are questions about a specific document and about the law where the building stands. Warranty and contract language on this site is flagged for legal review; read the actual documents for your project, and take advice on them.

Ask before you sign

Questions to ask an installerSection link

Questions that separate a contractor who has thought about drainage and edges from one who has priced a membrane. Weak answers here are informative.

Show me the ES-1 test data for the edge profile you are actually installing, at the design pressure for this building.
A good answer produces a test report for a specific profile in a specific gauge and metal, and a design pressure from the wind design, and shows that the first number is at least the second. A weak answer offers a brochure, a manufacturer’s general claim of compliance, or the observation that the shop has been making this shape for twenty years. Note that the standard itself prescribes a safety factor of 2.0 that trade guidance says is not incorporated into the code reference — so ask which basis the number in front of you is on.
Where is the zone map for this roof, and what changes at the perimeter and the corners?
Attachment density, fastener pattern, and often the insulation board layout change by zone. A contractor who cannot show you the zones is not pricing them, which usually means the field density has been priced across the whole roof.
What is the existing secondary drainage provision on this roof, and what will it be when you finish?
Many older low-slope roofs have none. The honest answers are “there is none and here is what we propose”, “there is one and it is being retained at this elevation”, or “there is none, the code path we are permitting under does not require us to add one, and here is that in writing”. The answer you should not accept is a shrug.
Draw me the sump at a typical drain, with dimensions, and tell me who fabricates it.
A pre-manufactured tapered sump, a field-cut sump and no sump at all are three different roofs. The dimension that matters is how far below the surrounding finished surface the drain rim ends up.
How much thicker is the insulation package than what is there now, and what does that do to the drain and overflow elevations?
This is the question that catches a recover which quietly raises the whole roof relative to its drains. If the answer is that the tapered design keeps the sump, ask to see the tapered layout drawing.
How is the wood nailer anchored, and to what — and who verifies it before the coping goes on?
The nailer is the part of the edge assembly that becomes invisible first and matters most. Anchorage into a deteriorated masonry parapet or into a nailer that was already there is the failure nobody sees.
Are you flood-testing at the drains, and who is present when you do?
A flood test at a drain is a straightforward verification of the one detail that is both buried and structural. If it is not in the scope, ask what verification replaces it.
What are you doing about existing drain leaders inside the building?
A perfect roof drain above a partially blocked or undersized leader is still a blocked drain. This is plumbing scope, and it is routinely nobody’s.
If we find the deck is wet or deteriorated at the drains, what happens to the price and the schedule?
Drains are where chronic ponding has been sitting longest, so they are where deck damage is most likely. A unit price and an allowance agreed before the roof is open is a negotiation; the same conversation on the third day of a tear-off is not.
Who is responsible for clearing the drains during construction, and what happens if it rains on an open deck?
Tear-off debris is the most reliable drain blocker there is, and it arrives on a roof whose membrane is partly off. This is a written answer, not a verbal one.

Require these in writing

Drain schedule: quantity, size, type, manufacturer, and whether each is new, reset, or reused.
Sump detail at each drain type, dimensioned, with the finished rim elevation relative to the surrounding insulation.
Strainer type at each drain, and confirmation that strainers are installed and not omitted.
Secondary drainage: existing condition, proposed condition, inlet elevation, and opening size — or a written statement that none is being added and on what basis.
Tapered insulation layout drawing showing crickets, saddles and sumps, with the average R-value that layout achieves.
Wind design basis: the standard and edition used, risk category, exposure, and the resulting zone map with attachment densities by zone.
Edge system: fabricator, profile designation, metal and gauge, and the ES-1 test report for that profile with the resistance values stated.
Nailer anchorage detail: fastener type, spacing, substrate, and the inspection point before the coping is installed.
Gutter securement, where gutters carry the membrane edge, tested to ANSI/SPRI GT-1 rather than to ES-1.
Verification plan: flood test or equivalent at drains, seam probing, and who witnesses and signs each.
Debris and drain-protection plan during construction, including who clears drains daily and what happens to an open deck in rain.
Photographic close-out documentation of every drain, scupper and edge condition before the coping and cover go on.

What goes wrong

Misconceptions and failure modesSection link

Beliefs that survive on commercial roofs because nothing tests them until the day something does.

Common misconceptions

Common belief
The overflow scuppers are a code formality. They have never had water in them.
What is actually true
That is the point. A secondary provision that has never flowed is a secondary provision that has never been needed, not one that is surplus. It is sized and set for the day the primary system is blocked, and the whole design intent is that it is a nuisance — water pouring down the face of the building where somebody notices — rather than a collapse. Judging it by how often it runs is like judging a fire pump by how often the building has burned.
Common belief
A little standing water on a flat roof is normal.
What is actually true
Some water after rain is normal; water that is still there two days later is the trade’s definition of ponding. The Asphalt Roofing Manufacturers Association — a manufacturers’ trade association, so an industry convention rather than a code threshold — defines ponding water as water which remains on a roof 48 hours or longer. A roof that ponds is not a roof with a cosmetic issue; it is a roof carrying load it was probably not designed to carry indefinitely, with a membrane deteriorating faster underneath it.
Common belief
We took the strainer domes off so the drains would flow better.
What is actually true
Published trade guidance says strainers do two useful things: they block debris from reaching the drain lines, and they can improve flow into the drain by breaking up the vortex that forms above it. Removing them moves a blockage you could have cleared with a broom into a leader inside the building.
Common belief
We are re-covering, not replacing, so the drainage requirements do not apply.
What is actually true
Sometimes true, and it depends on which code you are permitting under and which edition your jurisdiction adopted. Trade guidance from 2017 documents a real mismatch in the 2015 model texts: the building code’s reroofing section carried an exception saying that recovering or replacing roof coverings shall not be required to meet the requirement for secondary drains or scuppers for roofs that provide for positive roof drainage, while the existing building code’s equivalent section carried no such exception.
The same guidance records the consultants’ institute’s stated opinion that the lack of such provisions in the 2015 building code will increase the probability of roof collapses. That is an opinion held by a qualified body, not a finding — but it is a reason to ask which code path your permit is on and to get the answer in writing rather than to assume the exemption.
Common belief
The edge metal is trim. It is a sheet-metal item.
What is actually true
The model building code disagrees. Metal edge systems on low-slope built-up, modified bitumen and single-ply roofs are required to be designed and installed for wind loads in accordance with the structural chapter and tested for resistance in accordance with Test Methods RE-1, RE-2 and RE-3 of ANSI/SPRI ES-1. It is a tested assembly with a code-referenced test standard behind it, and it sits in the highest-suction part of the roof.
Common belief
The gutter is part of the edge system, so it is covered by the same test.
What is actually true
It is not. ES-1 applies to perimeter edge-metal flashing systems such as embedded edge metal, metal fascia and metal copings, and it specifically does not apply to gutters. Where a gutter is used to secure the perimeter edge of the membrane, the 2021 and 2024 model text puts it in its own subsection and sends it to Test Methods G-1 and G-2 of SPRI GT-1. A gutter carrying a membrane edge with no test basis at all is a common and consequential gap.
Common belief
Our edge metal is compliant — the manufacturer says so.
What is actually true
Compliance is a comparison, not a property. Trade guidance is explicit that a designer has to specify edge flashings with tested resistance loads equal to or greater than the applicable design wind loads, which means somebody has to know both numbers for this building. A shop-fabricated shape can be perfectly legitimate — the national roofing contractors’ association runs ES-1 testing and certification programmes for shop-fabricated shapes with two independent testing bodies — but the legitimacy comes from the test report, not from the fabricator’s confidence.
Common belief
Wind uplift is the membrane manufacturer's problem.
What is actually true
Uplift performance belongs to a tested assembly on a specific building: deck, attachment, insulation, cover board, membrane and the edge together, at densities that differ between field, perimeter and corner. No membrane carries a wind rating by itself, and no coping profile does either. See the membrane comparison for the same point made about the sheet.

How it actually fails

Debris blockage at the primary drain: Leaves, needles, wind-blown film, gravel, and construction debris collect on the strainer dome, which is doing its job. Flow area falls, water backs up, and the depth on the roof rises until it reaches the secondary inlet — if there is one.What you can see: A dark tide-line ring on the membrane around the drain, visible in dry weather. Vegetation growing at the low point. Water still standing more than 48 hours after rain. From inside, nothing at all, right up until there is.
Drain rim above the finished roof plane: Insulation and cover board are built up around a drain that was never reset, or a tapered package raises the surrounding field without keeping a sump. The drain is then at the top of a shallow basin rather than at the bottom of one, and the roof holds an inch or two of water permanently before anything flows.What you can see: A permanent ring of standing water around a drain that has no debris on it at all. Clean strainer, wet roof.
Overflow provision defeated by its own roof: The overflow scupper was set correctly against the original roof surface, and then a recover added insulation over the whole field. The inlet is now closer to the roof surface than intended, or — more dangerously — the field has risen so the effective static head changed and the structural assumption behind it no longer holds. The mirror-image version is a scupper left too high, or partly blocked by a raised parapet detail.What you can see: Measure it. The elevation of every overflow inlet relative to the current roof low point is a five-minute check that almost nobody does after a recover.
Ponding instability — the pond that deepens itself: A flexible deck deflects under the weight of accumulated water, the deflection increases the depth, and the increased depth increases the deflection. Trade guidance describes exactly this: as water accumulates, deck deflections can increase, thereby resulting in additional ponding water. Where the structure lacks the stiffness to arrest the cycle it is a progressive condition, which is why the model code requires it to be checked.What you can see: A pond that is deeper in the middle of a bay than the slope explains. Visible sag from the floor below. Cracking in ceilings or partitions under the bay. Treat these as the safety callout at the top of this page describes, not as a maintenance ticket.
Edge release and progressive peel: Suction lifts the coping or fascia — the highest pressures on the roof are at the perimeter and corners, where vortices form along the edges — and once the metal releases, the membrane it was restraining is free at the edge. Air gets under the sheet and it peels back into the field. The failure that started as a length of lifted metal ends as a roof.What you can see: Before the event: loose, rattling or oil-canned coping; open or separated joints; missing fasteners on a cleat; a nailer you can move by hand. After the event: metal in the car park and a clean peel line running back from the edge.
Wet insulation under a chronic pond: Water sits, finds a seam or a fastener penetration, and saturates the boards beneath. The wet area is heavier than the design assumed and its thermal performance drops well below the specified R-value. It also spreads laterally far beyond the visible pond.What you can see: Soft or spongy areas underfoot reported by trades on the roof. Infrared or nuclear survey anomalies that do not match the pond outline. Core cuts are what settle it — see roof condition assessment.

Sources and further readingSection link

Understanding Roofing / Published August 26, 2026

Scope and limitations

It cannot tell you whether your roof drains.
That is answered by standing on it 48 hours after rain, by a survey of drain and overflow elevations against the current roof surface, and by a moisture survey — not by a page.
It does not perform, and must not be read as, a drainage design.
Drain quantity and sizing, scupper sizing, tapered layout, rain load and ponding instability are the work of a qualified design professional for this specific building.
It publishes no structural determination of any kind.
The 5.2 pounds per square foot per inch of depth is arithmetic from the density of water, and the worked example built on it is arithmetic.
Neither is a design method, neither accounts for a real pond's shape, and neither substitutes for an engineer.
It cannot tell you what your jurisdiction requires.
Every code provision here is model text, read either in the publisher's own public-access edition or in a reproduction, plus one worked example of an actual adoption in Ohio.
Your adopted edition, its local amendments, whether your work is permitted under the building code or the existing building code, and your authority having jurisdiction all govern.
It publishes no cost figure for drains, scuppers, sumps, tapered packages or edge metal, because no dataset was found that separates them.
The sequencing argument on this page is a scope argument, not a price.
It does not reproduce the measured flow-rate reductions from the laboratory study of debris on drains and scuppers.
Only the abstract of that paper was publicly reachable; the numeric results were not, and inventing them would be worse than omitting them.
It cites no FM Global Property Loss Prevention Data Sheet.
Data Sheet 1-49 on perimeter flashing is the obvious source for the loss history behind this page's edge argument, and every route to the current document was blocked at the time of writing.
Third-party mirrors of an unverifiable revision were not used.
It does not interpret any warranty, contract or insurance policy, and it does not tell you who is liable for anything.
Warranty and contract content on this site is flagged for legal review.

Ohio Administrative Code rule 4101:1-15-01 — Roof assemblies and rooftop structures
Ohio Legislative Service Commission, Ohio Administrative Code (adopted state rule) / Effective 1 March 2024
The worked example of what a code adoption actually is: “Chapter 15 of the International Building Code, 2021 edition, is incorporated by reference and modified in Section 101.1.1 of this code.” Promulgated under R.C. 119.03, statutory authority 3781.10(A)(1), effective 1 March 2024, five-year review due 1 March 2029, with prior effective dates of 2002, 2005, 2007, 2011 and 2017. It is the basis for this page's statement that an Ohio project works from the 2021 model text. The section numbers within that text are supported separately, by the NRCA and IIBEC guidance listed below and by the aggregator reproduction cited last.
This is the law in Ohio and nowhere else. It adopts the 2021 edition of the model code, not the 2024 edition quoted elsewhere on this page; the modifications referenced in Section 101.1.1 are not reproduced here; and local jurisdictions administer and may add requirements. It is cited to show a reader what an adoption record looks like and why an edition matters, not to state a requirement for any other building.
2024 International Building Code, Chapter 15: Roof Assemblies and Rooftop Structures
International Code Council — codes.iccsafe.org, the publisher's own public-access edition. MODEL CODE TEXT. / 2024 edition
The authoritative location of the model text this page describes: Section 1502.2 (secondary emergency-overflow roof drains or scuppers where the roof perimeter construction extends above the roof), Section 1504.6 (metal edge systems designed and installed for wind loads in accordance with Chapter 16 and tested to Test Methods RE-1, RE-2 and RE-3 of ANSI/SPRI ES-1) and Section 1504.6.1 (gutters securing the perimeter edge tested to Test Methods G-1 and G-2 of SPRI GT-1).
A model code published by a private standards organisation. It is not law anywhere until a jurisdiction enacts it, and jurisdictions amend. The chapter is delivered through a script-driven reader that could not be extracted at the time of writing, so the section text on this page was read in the reproductions and in the trade guidance listed below and checked against this edition's section numbering. Nothing here is a determination about any permit.
Testing of Shop-fabricated Edge Metal
National Roofing Contractors Association
That the 2024 IBC carries edge systems at Section 1504.6 and gutters at Section 1504.6.1 with the RE-1/RE-2/RE-3 and G-1/G-2 test references quoted on this page; the full designation “ANSI/SPRI/FM 4435/ES-1, Test Standard for Edge Systems Used With Low Slope Roofing Systems”; that “ES-1 applies to perimeter edge-metal flashing systems, such as embedded edge metal, metal fascia and metal copings; it specifically does not apply to gutters”; and that NRCA maintains ES-1 certification programmes for shop-fabricated edge metal with UL Solutions and with Intertek Testing Services NA Inc.
Trade guidance and a promotional page for a certification programme, not adopted law. It describes the model code rather than any jurisdiction's adopted edition, and NRCA's certification of a shape is not a determination that the shape is adequate for a given building.
A compendium of changes — the roofing-related provisions of IBC 2021
Mark S. Graham, Vice President of Technical Services, NRCA — in Professional Roofing / 1 April 2021
That in the 2021 edition of the model code the edge provision is at “Section 1504.6—Edge Systems for Low-slope Roofs”, where “the requirement for ANSI/SPRI ES-1 testing of metal edge systems has been expanded to include all built-up, polymer-modified bitumen and single-ply roof systems having slopes less than 2:12”, and that “Section 1504.6.1—Gutter Securement for Low-slope Roofs adds a requirement for gutters to be tested for wind resistance according to ANSI/SPRI GT-1's Test Methods G-1 and G-2”. It is the basis for this page's statement that the 2021 and 2024 editions carry the same numbers and that the 2015 and 2018 editions did not.
A trade-association account of what changed between two editions of a model code, not adopted law and not an adoption record for any jurisdiction. It describes IBC 2021 as published; local amendments are outside it. It is the source for the section numbering in the 2021 edition, not for any requirement applying to a specific building.
Understanding ES-1: Proper Specification and Implementation
Mark S. Graham, Vice President of Technical Services, NRCA — in IIBEC Interface (International Institute of Building Enclosure Consultants) / March 2020
What each ES-1 test method actually loads and how the result is expressed: RE-1 for a metal edge flashing's capacity to restrain unadhered membrane at the perimeter, stated in pounds force per linear foot; RE-2 for resistance to horizontal, outward loads on fascias and gravel stops, expressed in pounds force per square foot; RE-3 for copings' resistance to separate front-face and back-face pulls with simultaneous load on the top surface, expressed in pounds force per square foot. Also that ES-1 does not address weathertightness; that a designer must specify edge flashings with tested resistance loads equal to or greater than the applicable design wind loads; that ES-1 prescribes a safety factor of 2.0 which is not incorporated into the code reference; and that vertical and horizontal design loads come from different pressure zones.
Written against the 2018 IBC and ASCE 7-16 and against the 2011 edition of ES-1, with the 2017 edition noted as forthcoming in IBC 2021. Section numbers, standard editions and zone nomenclature have moved since; treat the substance as durable and every number as edition-dependent. It is trade guidance, not law, and not a design method for any building.
Secondary Drainage and Ponding Requirements in the IBC and IEBC
Wanda Edwards, PE, Senior Director of Technical Services, RCI — in IIBEC (International Institute of Building Enclosure Consultants) / 28 November 2017
The reroofing mismatch this page describes: that 2015 IBC Section 1511.1 Exception 2 provided that recovering or replacing roof coverings “shall not be required to meet the requirement for secondary (emergency overflow) drains or scuppers… for roofs that provide for positive roof drainage” while IEBC Section 706.1 carried no equivalent exception, with ICC's interpretation recorded as being that the exception does not apply under the IEBC. Also the secondary-drainage requirement at IBC Section 1503.4.1 of that edition; the rain-load requirement that each roof portion sustain the load of rainwater accumulating if the primary drainage for that portion is blocked plus the uniform load of water rising above the secondary inlet (Section 1611); the ponding-instability verification of adequate stiffness to preclude progressive deflection (Section 1611.2); and RCI's stated opinion that the lack of such provisions in the 2015 IBC will increase the probability of roof collapses.
Written about the 2015 model codes. Section numbers have moved since — secondary drainage was at Section 1503.4.1 then and is at Section 1502.2 in the 2021 and 2024 model text, and the reroofing section has been renumbered more than once. The collapse-probability statement is explicitly the institute's stated opinion, not a finding, and this page presents it as such. Not law anywhere.
Raising the Bar in Standards: The ASCE 7 Standard and Low-slope Roof Drainage
Stephen Patterson, RRC, PE, and Madan Mehta, PE, PhD — in IIBEC (International Institute of Building Enclosure Consultants) / 1 September 2020
The arithmetic anchor for this page: that in the rain-load equation R = 5.2(dₛ + dₕ), the constant 5.2 is “the weight of a 1-in.-thick layer of water in psf, obtained from the density of water being 62.4 pounds per cubic foot.” Also the definitions used throughout this page — that static head refers to the elevation of the inlet level of the drainage element above the roof surface and exists only with secondary drainage elements, and that hydraulic head is the height of water above the inlet level of a drainage element — and that US plumbing codes require both primary and secondary drainage for low-slope roofs to be designed for a 100-year, 60-minute rainfall rate while ASCE 7 agrees on that rate for primary drainage but recommends a 100-year, 15-minute rate for overflow drainage.
A standards commentary rather than a design method, written against the ASCE 7 editions current in 2020. It does not size drainage for any building, and the rain-load equation is quoted here only to explain where the 5.2 comes from — not as an instruction to use it.
Key Considerations for Roof Drainage Design
Clemente Zamarripa, PE, and Amrish Patel, PE — in IIBEC (International Institute of Building Enclosure Consultants) / 27 March 2026 (presented at the 2025 IIBEC International Convention)
That secondary systems are designed to remove excess water once it reaches a designed height, typically two inches, before the roof structure is overloaded; that poorly designed tapered insulation is a source of inadequate drainage; the worked failure this page alludes to, in which a scupper size designed by the architect, a drain flow analysis by the plumbing engineer, and a rain load used by the truss engineer of record were not coordinated with one another; and the observation that the rainfall-intensity maps in the plumbing code have remained unchanged from the 1995 edition through the 2024 edition while extreme-precipitation days have risen substantially in the Northeast and Midwest since 1958.
Convention-paper guidance for design professionals, not a code and not a design for any building. Its precipitation-trend figures are drawn from national climate assessments and describe regional averages rather than any site.
Roof Drainage
Stephen Patterson — in IIBEC (International Institute of Building Enclosure Consultants) / Proceedings of the RCI 22nd International Convention
That strainers are important because they block debris from getting into the drain lines and can actually improve flow into drains by breaking up the vortex of water flowing into the drain; that overflow drains and scuppers are located above the low point of the roof — approximately two inches — to help prevent them becoming blocked; and that simply re-sloping a roof with tapered insulation may not be adequate because increases in insulation thickness can restrict drains and overflow systems.
Convention-proceedings guidance, not a code or a design. The two-inch figure it gives for overflow placement is described as typical practice, not as a requirement for any jurisdiction.
The Effects of Debris on the Flow Rates of Roof Drains and Scuppers
Jim D. Koontz, RRC, PE — in IIBEC Interface (International Institute of Building Enclosure Consultants) / December 2013; originally presented at the RCI 25th International Convention, 2010
The framing statement quoted on this page: “Drainage of rainwater has long been considered an essential attribute for the proper performance of any roof system. Long-term and excessive accumulation of water will contribute to the deterioration of most roofing systems and, in worst-case scenarios, has been responsible for excessive live loads that can lead to structural collapse.” Also that the study was laboratory work measuring water accumulation depth at controlled flow rates with drain and scupper inlets progressively obstructed.
Only the publicly reachable abstract of this paper could be read. The measured flow rates and water depths from the laboratory work are not reproduced on this page because they could not be verified in the source document, and the full text sits behind the institute's technical library.
Ponding Water Basics: Proper Drainage Design and Low-Slope Roofs
Asphalt Roofing Manufacturers Association / 29 January 2019
The definition used on this page — “Ponding water is defined as the water which remains on a roof 48 hours or longer” — and the deflection feedback loop: “Ponding water can substantially increase the load on roof decks. As water accumulates, deck deflections can increase, thereby resulting in additional ponding water.” Also that dirt, debris and other contaminants can affect and damage the membrane surface.
Published by a manufacturers' trade association, and written around asphaltic systems rather than around every low-slope assembly. Its 48-hour definition is an industry convention, not a code threshold, and its statements about what model codes require are the association's characterisation rather than a citation to an adopted edition.
29 CFR 1910.28 — Duty to have fall protection and falling object protection
U.S. Occupational Safety and Health Administration
That employees on a walking-working surface with an unprotected side or edge four feet or more above a lower level must be protected from falling by a guardrail system, safety net system, or personal fall protection system; that holes, including skylights, require covers, guardrails or fall protection; and that the standard sets distinct requirements for low-slope roof work less than six feet from the edge, between six and fifteen feet, and fifteen feet or more.
A general-industry occupational-safety standard addressed to employers, not guidance for building owners about roof condition. The fact that trained workers use fall protection is a reason to keep untrained staff off the roof entirely rather than a procedure for anyone to copy, and construction work on a roof is governed by a separate standard.
Low-Slope (“Flat”) Roofs
U.S. Department of Energy, Building America Solution Center (Pacific Northwest National Laboratory)
The wind mechanism this page relies on at the edge: “During high wind events, vortices form along the edges of the roof creating areas of localized negative pressure ('suction') above the roof”, and that in high-wind zones fully adhered membranes are recommended.
Written primarily for residential and light-commercial low-slope construction as best-practice guidance, not as adopted law and not as a wind-design method for a specific building. It does not address drains, scuppers or overflow provision at all.
Code text reproduced for cross-checking: 2024 IBC Chapter 15, Sections 1502.2, 1504.6 and 1504.6.1
UpCodes — a commercial code aggregator. The page consulted reproduces the Mississippi Building Code 2024. / 2024 edition as reproduced
Reading the 2024 model section text verbatim where the publisher's own reader could not be extracted, and cross-checking the numbering used on this page: Section 1502.2 on secondary emergency-overflow roof drains or scuppers where the roof perimeter construction extends above the roof in such a manner that water will be entrapped if the primary drains allow buildup for any reason; Section 1504.6 on metal edge systems and the RE-1/RE-2/RE-3 references; and Section 1504.6.1 on gutters securing the perimeter edge and the SPRI GT-1 references.
A commercial code aggregator, not an official jurisdiction source and not an adoption record for Mississippi or anywhere else. The characterisation that the Mississippi Building Code 2024 adopts the IBC 2024 without amendments is the aggregator's own, and was not verified against Mississippi's official adoption record. Used here only for reading the section text and confirming the numbering, never as the sole support for a code claim.
Code text reproduced for cross-checking: Ohio Building Code 2024 (IBC 2021) Chapter 15, Sections 1502.2, 1504.6 and 1504.6.1
UpCodes — a commercial code aggregator. The page consulted reproduces the Ohio Building Code 2024. / Ohio Building Code 2024, based on IBC 2021, as reproduced
Cross-checking the section numbering in the 2021-edition text Ohio has adopted, against the NRCA guidance listed above: Section 1502.2 on secondary emergency-overflow roof drains or scuppers, Section 1504.6 on metal edge systems with the RE-1/RE-2/RE-3 references, and Section 1504.6.1 on gutter securement with the G-1/G-2 references — the same numbers as the 2024 edition, and the correction of the numbering this page carried in its first draft.
The same commercial code aggregator, and the same limits: not an official jurisdiction source, not an adoption record, and never the sole support for a code claim. It is cited here because the Ohio-adopted 2021 text is the edition the worked example on this page turns on, and the section numbering in it had to be read rather than assumed. The Ohio amendments referenced in Section 101.1.1 are not reproduced here.

Suggest a correction