Roof Leaks Only During Wind-Driven Rain Decision Tree

Why this matters

A roof that stays dry through a steady all-day soaker but drips during a 30 mph squall is the single hardest leak to pin down, because the failure only opens up when rain arrives at an angle the roof was never designed to shed. Vertical rain runs down the slope and off the eave the way the shingles intend; wind-driven rain travels horizontally, gets under courses, packs into vent louvers, and pressurizes the wall cavity behind siding. If you only test with a garden hose pointed down the slope, you will leave clean and the customer will call back after the next gusty front. Diagnosing this class correctly means reproducing wind, not just water, and that changes both your test method and where you look.

Symptom presentation

The customer reports leaks that correlate with storms carrying wind, not total rainfall: a brief but windy thunderstorm wets a ceiling that a calm tropical downpour leaves dry. Stains are often near walls, at gable ends, around windows, or at the rake edge rather than mid-field. The leak may quit the instant the wind shifts even if rain continues. In coastal and high-wind regions the pattern tracks prevailing storm direction: leaks show on the south or west face of the house and never on the lee side.

Quick checks

Before you set a ladder, get the wind story. Ask which direction the storm came from, whether the leak stops when the wind dies, and whether it ever leaks in calm rain. Walk the interior and mark every stain, then go to the attic and trace the wet path uphill with a moisture meter; wind-driven entry almost always tracks well above the visible ceiling spot. Note any wall-to-roof intersections, sidewall step flashing, and penetrations on the windward face. Check whether the leak coincides with a specific exposure, that single fact narrows the search to one or two roof planes.

Isolation tree

Start at the windward elevation and work the suspects in order of likelihood.

  • Leak shows near a wall-roof junction. Suspect counterflashing or step flashing that is reverse-lapped or short. Wind packs water up behind a step flashing leg that is fine under gravity flow. GO TO step-flashing test.
  • Leak shows at the rake or gable. Suspect missing or undersized rake-edge metal and shingles not sealed at the rake. Wind lifts the rake course and drives water under. Inspect for unsealed tabs and gaps behind drip edge.
  • Leak shows around a vent, pipe boot, or louvered gable vent. Suspect that the louver or boot admits horizontal rain it sheds vertical rain. Gable louvers without an interior baffle are a classic windward source. GO TO penetration test.
  • Leak shows at the ridge during wind from one bearing only. Suspect ridge-vent external baffle missing or end plugs absent. Covered separately, but confirm here.
  • Leak shows mid-field with no junction nearby. Suspect lifted or unsealed shingle tabs from prior wind, or fasteners driven high (overdriven nails above the sealant line). Lift adjacent tabs by hand and check seal-strip adhesion.

Step-flashing test: direct a spray bottle or low-volume hose at the wall ABOVE the roof line, simulating rain that hits the siding and runs down behind the flashing. If you reproduce the leak this way but not by wetting the roof field, the wall water-management is the path, not the roof plane.

Penetration test: have a helper run a leaf blower or you wait for a windy day, then wet the suspect louver or boot from the side. Reproduction confirms horizontal entry.

Confirming diagnosis

Reproduce the leak under simulated wind. The most reliable field method is two people: one inside watching the wet point with a flashlight and a moisture meter, one outside running water in a defined sequence from low to high and ALWAYS spraying the windward face from the side and from above the roof-wall junction, never just down-slope. Isolate one suspect per test cycle and give each five to ten minutes; wind-driven paths are slow to show because water has to back up before it overcomes the lap. A thermal camera helps spot the cool wet track inside before liquid water appears. If a calm down-slope spray never leaks but a side-on spray at the wall does, you have confirmed wind-driven entry rather than a field defect, and you are done isolating.

Remediation

Fix the actual entry path, not the stain. Reverse-lapped or short step flashing must be removed and reinstalled with each piece woven shingle-over-flashing-over-shingle and counterflashing let into a reglet or surface-mounted and sealed at masonry. Per FEMA and IBHS high-wind guidance, hand-seal lifted tabs and seal the rake course; use a starter strip with factory adhesive at eaves and rakes. Replace gable louvers admitting rain with baffled or storm-rated vents, or add interior baffles. Re-seat or replace pipe boots with a storm collar. Where shingles were nailed high, the only durable fix is to replace those courses; you cannot re-seal a tab whose sealant strip never contacted the course below because the fastener line is wrong. In hurricane zones verify the field nailing pattern meets the local high-velocity wind requirement (six nails per shingle in many HVHZ jurisdictions).

Do not certify a wind-driven leak repaired after a calm-water hose test alone. A down-slope hose will not reproduce horizontal entry, so a "passing" calm test gives false confidence and guarantees a callback after the next windy storm. Reproduce wind direction with a side-on spray before you call it fixed.

References

  • IRC R905.2 and R905.2.6 (asphalt shingle application, flashing, and high-wind fastening)
  • IBHS FORTIFIED Roof technical requirements (sealed roof deck and edge details for wind-driven rain)
  • FEMA P-499 Technical Fact Sheet 7.1 and 7.2 (roof-to-wall and flashing details for wind resistance)
  • ASTM D3161 / D7158 (shingle wind-resistance test classifications)