What a roof moisture survey finds, and what the cores confirm

Is the roof wet, and how bad?

Two questions, and they need different answers.

A roof moisture survey answers the first one. A camera that reads heat, flown at night after a sunny day, finds where the insulation is holding water. It marks the spots. It says how big they are.

It doesn’t answer the second. How bad is it? Is the deck rotting, is the leak still open, does the roof come off or get patched? That’s a different job. It takes cores cut through the membrane, and it takes judgment about what the cores show.

A lot of vendor pages describe the camera and skip the cores. This piece doesn’t. It’s about what the camera finds and what conditions it needs. What confirms the finding, and what the finding can’t tell you. And about the number everybody quotes for replacing a roof, which nobody can source.

If you came for one thing:

• What the camera sees
• When it works
• What confirms it
• What it can’t tell you
• The number everyone quotes
• What you receive

What the camera sees

Wet insulation holds heat. Dry insulation doesn’t.

Over a sunny day, the whole roof soaks up heat. After sunset, the dry parts give it back fast and cool off. The wet parts hold on to it. For a few hours, a wet patch is warmer than the dry roof around it. That’s what the camera sees.

The camera doesn’t see water. It reads the surface temperature of the membrane. The difference between a wet patch and the dry roof beside it is the whole signal. And a temperature on its own doesn’t tell you how much water is there — only that something is different.

Four things decide how strong that difference is.

How much sun the roof got. More sun, more stored heat, more contrast. Summer works better than winter. A clear day works better than an overcast one.

How fast the roof cooled. A sharp drop after sunset pulls the dry roof down quickly and leaves the wet patches standing out.

The wind. Wind cools everything evenly and flattens the difference.

The difference between inside and outside. Heat from the building below rises through the roof. Wet insulation lets it through faster than dry. In cold weather, when the sun is weak, that’s the signal that does the work.

Dew or frost on the surface dulls all of it.

One reversal worth knowing. Over a cold room — a freezer, a refrigerated warehouse — the whole thing flips. Cold comes up through the wet insulation faster than through the dry. The wet patch reads cooler than the roof around it. Same physics, opposite sign. Anyone who tells you wet always reads warm hasn’t flown over a cold-storage roof.

We don’t read warm or cold. We read a difference, and we write down the conditions that made it.

When it works

We fly at night. The window opens about an hour after sunset and runs until sunrise. Our procedure follows ASTM C1153-23, and that’s the window it describes — broad, not narrow. You’ll see “one to three hours after sunset” quoted. That’s tighter than the practice, and nobody wrote it.

Before the night, the day matters. The roof needs sun. A warm, clear day loads the heat the survey reads. After a warm, cloudy day we wait longer to start. And before we leave, we go back and re-shoot the first area we did. If the contrast held, the night was good enough.

No real rain in the previous 24 hours. Rain on the surface reads as a wet roof, and rain in the insulation is what we’re trying to find. A roof that’s just been rained on isn’t ready.

Wind under about 15 miles an hour at the time of the survey. Above that, the wind cools everything the same and the signal goes flat.

Every one of those gets written down. The weather on the night. The weather for the 24 hours before. The exact time we flew. That’s not paperwork. It’s what lets next year’s survey be compared to this one, and it’s what makes the finding hold up.

If the night doesn’t meet the conditions, we don’t fly. We reschedule. A survey flown on a bad night produces a report that looks fine and isn’t.

What roof it works on, and where it doesn’t

The method is built for the common commercial roof. Rigid insulation boards on the deck, a membrane on top of the boards. Perlite, wood fiber, glass fiber, polyiso, polystyrene, cellular glass, foam sprayed in place. If your roof is one of those, the survey applies.

Some roofs defeat it, and we say so before the flight.

Anything under pavers, stone ballast, or gravel. The cover buries the signal. Heat from wet insulation has to come up through rock before the camera sees it, and most of it doesn’t.

A protected membrane roof. Extruded polystyrene sitting on top of the membrane, under ballast. The insulation is above the waterproofing, the ballast is above that, and the camera sees the ballast. Very hard to read.

Lightweight concrete and gypsum decks. They hold water from the day they were poured. A survey reads them as wet whether they’re leaking or not.

A dry layer over a wet one. Insulation stacked in layers, with the wet board under a dry one. The dry layer masks it.

None of that means the roof can’t be inspected. It means infrared isn’t the right first tool. A survey we know will read poorly is one we’d rather not sell you.

What the patterns mean

Wet insulation shows up in three shapes, and the shape tells you something about the roof.

Solid rectangles. Boards wetting one at a time. Water gets into one board, soaks it through, and the whole board reads warm with sharp edges. Perlite, wood fiber, glass fiber, and some foam boards do this.

Rectangle outlines. A warm border with a cooler middle. Slow-wetting boards like cellular glass and some foams take on water at the edges first. The outline lights up before the center does.

Here’s the trap. Boards that don’t sit tight against each other make the same outline, dry. The gap between boards is a small thermal path, and it draws the same picture as a wet edge. An outline is a candidate, not a finding.

Irregular blobs. No edges, no pattern. Poured or sprayed insulation — lightweight concrete, gypsum, foam in place — wets in whatever shape the water takes. So does water sitting on top of the insulation or pooled underneath it, whatever the insulation is.

The shape gets recorded with every anomaly. It’s the first thing a core checks. A core that comes up dry inside a rectangle outline tells you exactly which trap you fell into.

What else makes a warm patch

Water is one reason a patch of roof reads different from the roof around it. The practice we follow treats the survey as pattern recognition. Telling wet insulation apart from everything else that makes a warm patch. Here’s the everything else, in plain words.

The roof isn’t the same everywhere. Thicker insulation here, thinner there. A different board in a repaired section. A membrane that’s thicker in one area, or patched with a different material. Deeper gravel on one side.

The building isn’t the same underneath. A warm office under one section and a cold warehouse under the next. Equipment putting heat up through the deck. Ductwork. A wall under the roof that changes how it cools.

Things stick up through it. Fasteners, flashings, curbs, drains, vents. Each one is a small thermal path.

The surface isn’t uniform. A shiny patch next to a dull one reads different under the same conditions. A coating. An overlay. Debris. Water sitting on the surface from a drain that didn’t clear.

Something nearby is radiating. A parapet wall, a rooftop unit, a neighboring building. Heat from the side, not from below.

Every one of those can look like wet insulation from above.

That’s why we look at the underside of the roof when we can get to it. And it’s why the daylight visual pass matters. Half that list shows up plainly in a daytime photo — the patch, the drain, the unit, the pond. You rule those out before anyone cuts a hole.

What confirms it

The camera finds candidates. Cores confirm them. That’s not our preference. It’s the practice.

A core is a cut through the membrane and the insulation, down to the deck. It comes out as a plug. From it you learn what the roof is actually built of. How many layers, what kind of insulation, what’s under it. And you learn how much water is in the insulation. A camera can’t tell you either of those. A core tells you both.

The practice sets a minimum, and it’s more coring than most people expect.

At least one core in every roof section, just to establish what that section is made of. A section is any area separated by a wall or an expansion joint. One core in dry insulation per section, so there’s a dry reference to compare against. At least one in dry insulation for every different roof construction. And one core in every kind of anomaly the survey found, for every different construction. Three anomaly shapes across two roof constructions means six cores in the anomalies alone.

A moisture probe — a pin pushed through the membrane — can stand in for some of those readings. But only after it’s been checked against a core on that roof. Probe first and you don’t know what the number means.

A capacitance meter or a nuclear gauge, the kind you roll across the surface, adds information. It doesn’t verify. Under the practice, non-contact instruments complement a core. They never replace one.

Every hole gets patched, in a way that doesn’t hurt the waterproofing. A verification that leaves the roof worse than it found it isn’t a verification.

Your roofing contractor cuts and patches the cores, at the spots we mark. Or we coordinate a contractor before the visit. Either way: we image, we mark the roof, and we document the core locations and what came out of them. We don’t cut roofs. Cutting a roof is roofing work, and it’s done by the people who’ll warranty the patch.

What it can’t tell you

Plainly, so it’s settled before you hire anyone.

Where the water got in. A wet patch is where water ended up. Water travels under a membrane. The entry point can be a long way from the patch — a seam, a flashing, a penetration, a puncture. Finding it is a different inspection.

Whether the leak is still open. Wet insulation stays wet long after the entry point is fixed. A warm patch can be an old leak sealed years ago or an active one. The camera can’t tell which.

How long it’s been wet. No clock in the image.

Whether the deck is compromised. Water sitting on a steel deck rusts it. Water on a wood deck rots it. The camera sees the insulation above the deck. A core sees the deck. A structural engineer says whether it’s still holding the roof up.

How long the roof has left. Wet area is one input. Not the answer.

Whether to replace it. That’s the next section, and it’s the one people most want the camera to answer.

What the survey does give you is a map. Where the insulation is holding water, how big each area is, what shape it takes. And a marked roof your roofer can core without guessing. Everything after that is somebody else’s expertise, aimed at a spot instead of a whole roof.

The number everyone quotes

Ask three roofers when a roof needs replacing instead of patching, and you’ll hear a percentage. Twenty percent wet. Twenty-five. Thirty. Over that, tear it off.

We went looking for where the number comes from. It isn’t in the standard. It isn’t in anything published by ASTM, IIBEC, NRCA, or SPRI. We couldn’t find it in a peer-reviewed study. The earliest place we could find it stated is a roofing contractor’s web page from 2021. It says tear-off can usually be avoided under 25 percent. The same company, on a later page, calls 25 percent a threshold a lot of contractors use.

That’s what it is. A rule of thumb. Maybe a useful one. Not a standard.

There is a 20 percent in the standard, and it may be where the folklore started. It’s a camera spec. The imaging system should be able to detect insulation that has lost a fifth of its insulating ability. That’s how sensitive the equipment has to be. It says nothing about when to replace a roof.

There is no percentage in the standard.

What actually decides repair, recover, or replace is a list, and none of it is a percentage. What the deck looks like when the cores come out. What condition the membrane is in across the whole roof, not just over the wet parts. How many roof layers are already stacked up there, because that’s weight and it’s tear-off cost. Whether the roof drains. What the warranty says about cutting into it. And the arithmetic. Cutting out and replacing the wet sections, against taking the whole thing off.

A survey feeds that decision. It tells you how much is wet and where, so the arithmetic has real numbers in it. It doesn’t make the decision, and nobody should tell you it does.

Where the standard fits

Our procedure follows ASTM C1153-23, the recognized practice for locating wet insulation with infrared imaging. It’s the 2023 edition, and it’s current.

The practice covers walkover surveys and aerial ones. Ours are flown. Two things come with that. An aerial survey gets a daylight visual pass within two days, by air photographs or by walking the roof. That puts the surface conditions that affect the reading on record. And the report states how high above the roof the camera was. That sets how small a wet area it can pick out.

We don’t reproduce the standard here, and we don’t tell you it requires anything of you. It’s the practice we follow. This is a statement of alignment, not a certification.

There’s no such thing as a C1153 certification. The standard describes a procedure. It doesn’t certify people or companies.

What you receive

The report is built around what the practice asks for, stated as what we deliver.

The conditions. Wind, outside temperature, and cloud cover on the night. Cloud and precipitation for the 24 hours before. The date of the last real rain. The date and time of the survey. What the roof surface looked like.

The equipment and the flight. The imaging system used, and how high it was above the roof.

The map. A scaled drawing of the roof, or an air photo of it. Every wet area marked by size and location. Every core location.

The cores. What each one showed. What the roof is built of at that spot, and how much water was in the insulation.

The thermograms. Representative images from the survey, so your reviewer sees what we saw.

Two things about that list. The map and the cores are the deliverable. The images support them, not the other way around. And the report names what was found and where. It doesn’t say what caused it or what to do about it. That’s the next conversation, with your roofer and, if the deck’s in question, an engineer.

If you want to talk about a roof

A short call covers three things. How the roof is built: sections, layers, what’s under the membrane. That decides whether the survey applies and how many cores it needs. Who’s cutting the cores. And what decision you’re trying to make, so the report is built for it.

We’re owner-operated. The person you talk to is the person who flies the roof and signs the report.

If your roof is ballasted, or the deck is lightweight concrete, we’ll tell you the survey will read poorly. Before you pay for it.