Atrium roofs leak for a long list of reasons, and a surprising share of them trace back to water that was never meant to be there in the first place. Not water coming through the glass. Water sitting on top of it because the route off the roof stopped working.
Drainage is the part of a glazed roof nobody thinks about until it’s the only thing they can think about.
A glazed roof drains on two levels
The obvious level is the surface. Rain lands on the glass, runs down the pitch, reaches a gutter at the eaves or a channel at the perimeter, and leaves through an outlet.
The second level is inside the framework, and it’s the one that gets forgotten. Glazing bars in a well-designed system carry small drainage channels and weep holes, because the design assumes a certain amount of water will get past the outer seal. It isn’t a failure when that happens. The system is built to collect that water and take it back out to the surface, below the glass but above the inner line of defence.
That second level only works if the channels are clear and the weeps are open. Block them and water that the design expected to drain away has nowhere to go. It backs up, finds the inner seal, and arrives in the building as a leak that looks like a glass failure and isn’t.
Where the design goes wrong at the outset
Some atria were never going to drain well. The common causes are visible on the drawings if anyone looks.
Too little pitch is the first. A shallow fall needs everything downstream of it to be perfect, and nothing stays perfect for twenty years. Flat sections hold water, and standing water on a glazed roof works on the seals continuously rather than occasionally.
Too few outlets is the second, usually paired with outlets that are too small for the catchment they serve. An atrium roof can collect a lot of water in a short, heavy burst, and that’s exactly the condition an undersized outlet fails in. Rainfall in the UK has got lumpier, and roofs designed decades ago were sized for the pattern of their time.
Then there’s the absence of a second route. Good design assumes the primary outlet will block at some point and gives the water an overflow, somewhere it can escape without entering the building. Plenty of atria have no overflow at all, so the first blockage becomes an internal leak rather than an obvious stain down an external wall.
Junctions are the other weak point. Where a glazed roof meets a parapet, an abutment or an older roof covering, two systems have to be joined and the water has to be handled across that joint. Those details are where the original design effort runs out most often. The Centre for Window and Cladding Technology publishes guidance covering how glazed envelopes should handle water, and it’s a reasonable reference when you’re trying to work out whether what you own was sensible.
Where it goes wrong afterwards
Most drainage failures aren’t design failures. They’re maintenance failures on a design that was fine.
Leaves and silt collect in gutters, and in a city centre that silt is mostly road dust and it sets hard. Moss grows in channels. Bird nesting material blocks an outlet completely in a single season. Pigeons choose the sheltered corner of a glazed roof for exactly the same reason water collects there.
Weep holes get sealed up deliberately, which is the one that hurts. A contractor chasing a leak runs a bead of silicone along a bar, closes the weeps, and converts a working drainage path into a reservoir. The leak stops for a while, because the water is now accumulating inside the framework instead of arriving in the lobby. Then it finds another exit, somewhere worse, and the trail back to the cause is now buried under remedial sealant.
Settlement and thermal movement do slower damage. Falls flatten, joints open, and a roof that drained properly on day one carries a puddle in the same place every time it rains.
What to actually check, and how often
Twice a year is a sensible baseline, scheduled after leaf fall and before the winter. The checks themselves are not complicated.
Are the gutters and channels clear to the full depth rather than skimmed on top. Are the outlets flowing, tested with water rather than looked at. Are the weep holes open, which means finding them on the drawings or the system details first. Is there standing water anywhere after rain, and is it in the same place as last time. Is there an overflow, and would it work.
The useful habit is to do this at the same visit as the glass clean, since someone is already up there with close access and the planning of that access sits under the Health and Safety Executive’s work at height guidance either way. Treating maintenance and safety as one routine rather than two separate exercises is the same logic applied to plant, and it works for the same reason.
When a blockage has already caused ingress, the trace comes before the repair, and firms doing this properly, skyaccessglazing.co.uk among them, will separate the two on the quotation.
If your atrium has had three leak repairs in five years and nobody has ever cleared the gutters to the base, the drainage is the first thing to look at rather than the last.
