How Moisture Moves Through a Building

Why this matters

Two houses hand you an identical dark stain on the underside of the roof sheathing. In one, the right repair is a tube of sealant and a length of duct, done in an afternoon. In the other, that same repair changes nothing, because the water is arriving by a completely different route. The stain does not tell you which house you are in. The timing and the shape of it do, and they do it reliably, because each transport route runs on its own driving force and leaves its own signature. Reading the route before you quote the repair is what separates a fix from a return visit.

A sibling article covers the control strategies and vapor retarder classes. This one is about identifying which route is actually running.

Before you go into the attic or open a wet assembly

Weight first: a saturated insulation batt sitting on a wet ceiling is heavy, and opening that ceiling from below drops the whole load on you. Probe from the side, support the area, and stand out from under it.

In the attic, step only on framing or on walk boards you placed. Fibrous insulation is an inhalation hazard, so wear a fitted particulate respirator rather than a nuisance mask. If you find loose granular fill, stop and back out; treat it as presumed asbestos-containing until it is tested, under 29 CFR 1910.1001 for general industry work or 29 CFR 1926.1101 for construction work, and bring in abatement rather than disturbing it to get a better look.

Attics reach temperatures that put a person down fast. Work them early, carry water, tell someone you are up there, and come out at the first sign of headache, cramping, or a stumble.

Any splice, junction box, or older wiring in a wet attic is an electrical hazard, so de-energize the circuits serving that space at the panel, lock and tag, and prove the conductors dead with a meter checked on a known live source before and after, per 29 CFR 1910.333(b)(2) and the live-dead-live sequence at NFPA 70E-2021, 120.5.

Four routes, four driving forces

Two of these move liquid water and two move vapor, and the split matters because a vapor retarder does nothing about a hole and an air barrier does nothing about capillary rise. Matching the repair to the route is the whole game.

One property of air transport is worth saying separately because it breaks people's intuition: air travels, so the water it carries can be deposited a long way from where it entered. A leak at a wall's bottom plate can wet sheathing near the roof line, because the air went up a stud bay first. Diffusion has no such freedom. It deposits where the temperature profile puts the condensing plane.

The rate gap, and why it sets your investigation order

The four routes do not move comparable amounts of water. Bulk water moves the most by a wide margin, which is why drainage details outrank everything. Between the two vapor routes the gap is also enormous, and it is the one people get wrong.

The comparison commonly cited in building-science literature puts roughly a third of a quart of water moving by diffusion through a 32-square-foot sheet of gypsum over a heating season, against roughly 30 quarts through a 1-square-inch hole under a 10-pascal pressure difference over that same season. Those figures carry their conditions with them, and the pressure difference in particular is what makes the second number large; at a lower driving pressure the hole moves less. But the shape of the result holds across the assumptions: air leakage moves moisture on the order of two magnitudes faster than diffusion through an intact plane.

So the investigation order follows the rates, not the ease of testing. Rule out bulk water, then air transport, then capillary, then diffusion. A tech who starts with a vapor retarder question is starting at the smallest term.

The gate

Does the wetting track a rain event, a pressure event, or the season?

Rain event points to bulk. A pressure event, meaning fan operation, stack effect on cold days, or a door that changes the pressure regime, points to air transport. The season, meaning a slow appearance over weeks with no correlation to any switch, points to diffusion or to a source that runs continuously. Capillary declares itself by continuing in dry weather with a stable upper edge.

Two houses below, same stain, and the gate sends them to different repairs.

House one: the stain that tracked a switch

Roof sheathing discolored in a rough circle, perhaps two feet across, centered above the main bathroom, with a second smaller patch at the attic hatch. The rest of the deck is clean. Complaint appeared in the third week of a cold stretch.

The bath fan discharges into the attic through a short length of flex that ends in mid-air, which is common enough that it is worth checking first on any attic moisture call. A smoke source at the hatch on a cold morning showed air moving up out of the house.

Put a number on it. The fan is rated 80 cubic feet per minute and the household runs it about an hour a day across roughly a 120-day heating season. Indoor air at 70 F and 40 percent relative humidity carries about 0.0062 pounds of water per pound of dry air.

• Mass flow: 80 cubic feet per minute x 0.075 pounds per cubic foot = 6 pounds per minute, so 360 pounds per hour.
• Water carried: 360 x 0.0062 = about 2.2 pounds per hour.
• Over the season: 2.2 x 120 = about 270 pounds, which is on the order of 32 gallons of water delivered into the attic.

And that is a floor, not an estimate of the worst case, because the air leaving a bathroom right after a shower is far wetter than 40 percent relative humidity. Against the diffusion comparison above, this single defect is moving orders of magnitude more water than the entire ceiling plane's diffusion.

Repair: duct the fan to a proper exterior termination in insulated duct, seal the attic hatch, and air seal the top plates and penetrations in that ceiling area. The vapor retarder question never comes up, because diffusion was never the term that mattered here.

House two: the same stain, uniform, and a different answer

Discoloration across the entire north slope of the sheathing, even from ridge to eave, with no concentration at the hatch, the plumbing stack, or the can lights. It appeared gradually over two winters. The ceiling plane had already been air sealed and blower-door tested by a previous contractor, and a smoke check at the accessible penetrations showed nothing moving.

Two findings explained it. The house runs around 55 percent relative humidity through the winter, which is a high indoor dew point, and the source is an open dirt-floor crawl space feeding the whole house through the floor. And the eave vents were buried under insulation when the attic was topped up, so the attic had lost most of its ability to carry moisture away.

Route: a continuous moisture source raising the whole house's dew point, arriving at the deck by diffusion and by whatever residual leakage remains, with no ventilation path to remove it.

Repair, in the order the rates dictate: seal the crawl space floor with a proper vapor barrier and address the source, then restore the eave ventilation with baffles, and only then consider the interior vapor retarder. Reversing that order, which is what "put a vapor barrier on the ceiling" amounts to, addresses the smallest term and leaves the source running.

The distinguishing evidence between the two houses was not the stain, and it was not a moisture reading. It was uniform against concentrated, and gradual against tracking a switch.

When the pattern says diffusion and the damage says otherwise

This is the check that keeps house two from becoming a wrong answer.

Diffusion through an intact plane is slow, and it rarely produces structural decay on its own in a normally ventilated assembly. So if the pattern reads uniform but the damage is severe, delaminated sheathing, soft framing, or standing water, do not accept diffusion as the explanation. Go back and hunt for the air path you missed, because a missed leak is far more likely than a diffusion failure that outran the assembly's drying.

The specific places a leak hides from a casual check: a chase around a chimney or a plumbing stack, the wall between a house and an attached garage, a dropped soffit over cabinets, an interior partition top plate that was never sealed, and a duct chase that connects two floors. A blower door with the attic hatch open and a smoke source at each of those, on a day with a real indoor to outdoor temperature difference, finds what a walk-through misses.

What to record so the next visit can tell them apart

The gate needs history, and history is only available if somebody wrote it down. Three fields on the first visit make the second visit short.

Date and weather at the time of observation, specifically the outdoor temperature and whether there had been rain in the previous 48 hours. Without this the rain-event branch of the gate is unusable.

A photograph with something in it for scale, showing the extent and the edges of the staining. Extent is the evidence, and the edges are what separate uniform from concentrated. A close-up of the stain alone is worthless for this purpose.

The house's indoor conditions as a dew point, not as a relative humidity, plus the same for any attic, crawl space, or basement. Dew point is the figure that stays comparable when the next tech measures it at a different temperature.

Then verify the diagnosis rather than assuming the repair worked: return during the same season, under similar outdoor conditions, and take a moisture content reading in the same spot on the sheathing. A number that has fallen confirms the route you named. A number that has not means you fixed a real defect that was not the one delivering the water.

References

• ASHRAE Handbook, Fundamentals, for vapor transport, air leakage, and psychrometric relationships
• Building-science literature comparing seasonal moisture transport by air leakage against diffusion through an intact plane, stated with its assumed pressure difference and season
• 29 CFR 1910.1001 (general industry) and 29 CFR 1926.1101 (construction), OSHA asbestos standards, for suspect loose fill and insulation materials
• 29 CFR 1910.333(b)(2) and NFPA 70E-2021, 120.5, for de-energizing and proving dead before work in a wet attic
• See related: Moisture Control in Residential Construction; How to Work Out Where Condensation Will Form; What Dew Point Predicts