Infrared thermography can cover a 100,000 sq ft roof in an evening and identify probable moisture zones across the full area — a capability that core sampling alone cannot match for large buildings. We use IR to guide the core sampling program, not to replace it.
Infrared scanning for commercial roofing works on a specific physical principle: wet insulation retains heat longer than dry insulation after the sun goes down. A thermal camera scanning the roof surface in the hours after sunset sees the moisture zones as warmer anomalies against the cooling dry field. A trained thermographer can map these warm anomalies across the full roof area in a single evening, producing a candidate moisture map that shows where the building's insulation may be saturated.
That is a powerful tool. It is also a tool with specific operational limits. Dallas conditions create both good IR seasons and bad ones. Summer nights in June through September cool slowly — the roof surface holds heat from the 160°F daytime peak long after sunset, which reduces the temperature differential that makes wet-zone anomalies detectable. The best IR windows in Dallas are October through April, when daytime solar loading is sufficient to charge the insulation and evening temperatures drop fast enough to create a clear differential between wet and dry zones. A hot August night with slow cool-down produces ambiguous thermograms that are not a sound basis for a capital decision.
We use it to produce a probability map that tells us where to concentrate core sampling — which lets us produce a higher-confidence moisture assessment at lower total cost than a dense grid-pattern core survey would require.
Conditions that produce clear, usable thermograms: At least 4 hours of direct solar loading on the roof surface during the day, a temperature differential of at least 15-20°F between the roof surface and the ambient air at scan time, no wind above 15 mph (wind equalizes surface temperatures and suppresses the anomaly contrast), no precipitation in the prior 48 hours (wet membrane surface from rain creates false-positive anomalies across the whole field), and cloud cover that is not blocking the solar loading during the day.
Conditions that produce ambiguous or useless thermograms: Overcast days with insufficient solar loading, scan timing that is too early in the evening before sufficient differential has developed, scan timing that is too late when the roof has cooled uniformly, reflective membrane surfaces (white TPO) that absorb less solar energy and therefore charge less relative to the insulation below, and buildings with unusual thermal mass in the deck assembly that changes the cool-down curve.
Dallas-specific note: Many Dallas commercial buildings have white or light-grey TPO or EPDM membranes installed for the energy code compliance benefit of high solar reflectance. These reflective surfaces absorb less solar energy and produce a weaker IR signal — the temperature differential between wet and dry zones is smaller and the thermogram is harder to read. On highly reflective membranes, core sampling at strategic density often produces better data than IR at lower cost.
The IR scan produces a thermogram with warm anomaly zones marked as probable moisture locations. We then pull cores at each significant anomaly zone to confirm the finding — dry or damp or wet. We also pull cores at several presumed-dry zones identified in the thermogram to establish the baseline and confirm the IR read is accurate on this building.
This approach — IR to map, cores to confirm — lets us cover a large roof (100,000+ sq ft) with higher confidence than either tool alone. IR alone produces probable locations without confirmation. Cores alone at a grid density sufficient to find all significant wet zones on a 100,000 sq ft building would require 30-40 core pulls. IR plus targeted cores typically produces equivalent confidence at 15-20 core pulls, which is both faster and less disruptive to the membrane.
The combined deliverable is a moisture distribution map with each zone coded as confirmed-wet (core-verified), probable-wet (IR anomaly, not yet core-confirmed), or dry (IR and core both clear). This three-level classification tells the owner where the confidence is high versus where it is inferred.
