A roof can be perfectly dry on top and soaked underneath. That is the problem an infrared moisture scan exists to solve.
Water that gets through a membrane does not stay where it entered. It spreads sideways through the insulation, and from above there is nothing to see. The roof looks the same. Meanwhile the insulation has stopped insulating, the deck is getting wet, and any new roof laid over it is being laid over a problem.
How it works
Infrared does not see water. It sees heat.
Through a sunny day the roof absorbs heat. After sunset it releases it, and dry material releases it quickly while wet material holds on to it. For a few hours after dark the wet areas of a roof are measurably warmer than the dry ones, and a thermal camera reads that difference as a pattern.
Which is why the conditions are not a scheduling detail, they are the method. The scan needs a day with real sun, a dry evening, and not too much wind. Run it on the wrong night and it does not fail visibly; it simply shows nothing, and nothing looks exactly like a dry roof.
Why cores are not optional
A warm patch is a suspicion, not a finding.
Plenty of things make part of a roof warmer than its neighbors: a duct underneath, a vent, a patch of a different material, a pile of debris, a section that was repaired with something else. Every one of those shows up as an anomaly.
A core cut through a representative anomaly settles it. It shows whether the material is actually wet, how wet, and what the assembly is made of while you are in there.
So when you are handed a scan report by anyone, including us, the question to ask is how many cores were taken and where. A report of thermal images with no cores is a set of suspicions presented as conclusions.
What you get
A map of the roof with the wet areas located and, crucially, measured. Not just where, but how much as a proportion of the roof.
That figure is the one the decision turns on. A small wet share can be cut out and replaced as part of a recover. As the share grows, removing the wet parts starts to cost what removing everything would cost, and a tear off becomes the honest answer. Where exactly that line falls is a judgment, but it cannot be made at all without the number.
The report also has to be biddable. Areas located and quantified, so a contractor can price replacing the wet sections without commissioning a second survey.
When it is worth it
Three situations, and they are all decisions rather than curiosity.
Before a reroof. A recover over wet insulation seals water into the assembly, keeps it working on the deck, and will usually void the new warranty. Knowing the wet share before anyone bids is what makes the bids comparable.
Before buying a building. A low slope roof is a capital item and its condition is mostly invisible. This is the survey that tells you whether you are buying a roof with years in it or a project.
When leaks appear in several unrelated places. That pattern usually means the problem is spread rather than local, and the map is how you find out.
A simple residential tile or shingle roof rarely needs one.
When we would say no
Some roofs do not read. Ballast and aggregate obscure the signal. Certain insulation types and assemblies do not produce a usable difference. Standing water on the surface masks what is underneath.
We will say so before booking rather than afterward. On those roofs, impedance or nuclear methods, or simply a planned grid of core cuts, may give a better answer for similar money.
Next: what a full roof inspection covers, or flat roof inspection for the roofs this is most used on.
Common questions
What does an infrared roof scan actually detect?
It detects heat, not water. Wet material holds more heat than dry material and releases it more slowly, so after sunset the wet areas of a roof stay warmer than the dry ones for a while. The camera reads that difference and the pattern shows where the moisture is.
- The camera sees surface temperature, nothing else
- Wet insulation cools more slowly than dry insulation
- The useful window is after sunset while the roof is cooling
- A warm patch is an anomaly, not yet a confirmed fact
- Cores are what turn an anomaly into evidence
This matters because plenty of things other than moisture make a patch of roof warmer: a vent below, an air conditioning duct, a patch of different material, debris. A scan report that presents thermal images as findings without a single core cut is showing you suspicions and calling them conclusions. Ask how many cores were taken, always. A scan with no cores is cheaper for a reason, and the reason is that it commits to nothing.
Why does it have to be done at night?
Because during the day the sun heats everything equally and swamps the difference the camera is looking for. The roof has to absorb heat through the day and then cool, and wet areas give that heat back more slowly, which only becomes visible once the sun is off it.
- A sunny day loads the roof with heat
- After sunset the dry areas cool faster
- The window is a few hours, not all night
- Rain or heavy cloud during the day ruins it
- Wind over the surface also flattens the difference
That is why a scan is scheduled around the weather rather than around the calendar, and why we would rather move a date than run a survey in conditions that produce an unreadable result. A scan run on the wrong evening does not fail obviously; it simply shows nothing, and that is easy to mistake for a dry roof. We would rather tell you the evening is wrong than hand you a clean looking map that means nothing.
When is a scan worth paying for?
Before a reroof decision, before buying a building with a low slope roof, and when a roof leaks in several places without an obvious cause. In all three the question is how much of the insulation is wet, and nothing visible on the surface answers it.
- Before deciding between a recover and a tear off
- Before buying a building with a flat or low slope roof
- When leaks appear in several unrelated places
- After a long slow leak, to find how far the water travelled
- Not usually needed on a simple residential tile or shingle roof
The decision it informs is worth far more than the scan. A recover over wet insulation seals the water in, keeps it working on the deck and voids the new warranty, so the cost of getting this wrong is a roof that fails early. Knowing the wet share before anyone bids is what makes the bids comparable. The scan costs a fraction of the decision it informs, which is what makes it worth doing properly.
What are the limits of infrared?
It reads the surface, so it struggles where the construction hides the signal. Ballasted roofs, some tile assemblies and roofs with certain insulation types do not produce a clear reading, and a scan on one of those is not worth the fee.
- Ballast and aggregate obscure the thermal signal
- Some assemblies simply do not read usefully
- Standing water on the surface masks what is below
- It finds moisture, not the hole that let it in
- Other survey methods suit some roofs better
We will tell you before booking whether your roof is a sensible candidate, and sometimes the answer is no. Where infrared will not read, nuclear or impedance methods or a straightforward grid of core cuts may give a better answer for the same money, and recommending one of those instead is not a loss of a sale. Saying so costs us the booking and saves you a fee for a survey that could not have worked.
Does the scan tell me where the leak is?
Not directly. It shows where water has ended up, which is often some distance from where it got in, because water travels between layers before it settles. The map narrows the search considerably but the entry point is still found by inspection.
- Wet areas show where water collected, not where it entered
- Water travels along the deck and between layers
- The map usually points toward the entry point
- Details above a wet area are the first place to look
- A visual inspection still does the final work
Used together the two are far stronger than either alone. The scan says this quarter of the roof is wet; the inspection then concentrates on the details above that quarter rather than treating the whole roof as equally likely. On a large roof that is the difference between finding a leak and guessing at it. Neither tool replaces the other, and anyone selling the scan as a leak finder is overselling it.