What a Roof Covering Assumes About the Layers Beneath It

The tile and slate record gives the test directly. A lapped covering shows repeating open joints that rely on overlap and gravity, and a seamed metal roof shows continuous trays whose joints are folded shut. That single observation, the record says, explains why one type carries a genuine secondary drainage plane beneath it and the other does not need one in the same sense.

It is worth making that observation before any other question is asked, because almost everything else follows from it. The way each family sheds water decides what has to exist underneath, what that layer has to be able to do, and which trade has to have been thinking about it while the roof was being built.

A covering of overlapping units sheds most of the water and expects the rest to be handled. The record describes the underlay as a drainage plane rather than a waterproofing layer, whose purpose is to catch what the covering lets through, including wind-driven rain, snow entry, condensate and debris, and carry it out at the eaves. The difference between those two descriptions shows up wherever the underlay is trimmed short, punctured, or lapped against the direction of flow.

That plane only works if water reaching it can keep moving. Where the underlay is draped between rafters, the sag below the batten line is the channel water runs in. Where it is laid over a continuous deck or sarking there is no drape, the batten bears along its whole length and each crossing becomes a small dam, so a counter-batten is what restores a drainage path and an air space above the sheet. The published junction record calls that crossing the point at which the path is either created or absent, concealed for the roof's whole life.

The underlay only works if it stays continuous to the outside. Where it is cut back behind a gutter, stopped short at a valley, or lapped against the flow, water that the covering handed over correctly is delivered into the structure instead of out of it, and the record notes that the symptom looks like a covering problem and is not one.

The eaves junction record describes the same requirement from the other side: water leaving the roof has to be delivered into the gutter, far enough out that it clears the fascia and with the underlay behind it discharging into the gutter rather than behind it. When that goes wrong, the damage appears on the fascia, on the wall or at the base of the wall rather than at the roof.

A standing seam roof asks for different things and asks for them firmly. The record states that everything the metal sits on is chosen with the covering's behaviour in mind: the deck has to be continuous, even and dry enough to receive clips and to avoid telegraphing its irregularities into a reflective surface, and the separation layer between metal and deck manages contact, abrasion and incidental water. A depression in the deck becomes a visible dish, a proud fastener becomes a wear point as the tray slides, and a deck closed in wet condition releases moisture upward into the interface where it has no easy route out.

The other requirement is inboard rather than outboard. Because the covering is thin, conductive and impermeable, it runs cold on clear nights, so the record treats the continuity of the air and vapour control layers below the deck as mattering more here than under a covering that can breathe through its joints. That is a requirement on the build-up, arriving from a decision about the visible surface.

A composite panel roof carries weathering, insulating and lining duties in one component, and the record observes that there is remarkably little to discuss about the field as a result. Design attention moves outward to the longitudinal joint, the end joint, the fixings and every point where a panel abuts something that is not a panel. Within the panel the liner face closes the air path without any site-formed lap, which is why the joints between panels carry that duty entirely.

A built-up twin-skin roof answers the same brief in the opposite way. There the liner sheet is the air and vapour control element as well as the visible ceiling, so its sealed side laps, end laps and closed perimeter are what the assembly depends on. The record notes that a liner which looks faultless from below and is open at its side laps has failed at the duty that matters, and the symptom appears later, inside the build-up.

The covering entries state that the build-up beneath is somebody else's subject, and they are right to: the same tiles can sit over a ventilated cold roof, over insulation held between and under rafters, or over a warm deck, and nothing about the units settles that. The build-up entries state the converse, that the covering is chosen separately. Read singly, each is a complete and accurate account of one thing.

Read together, they describe a dependency that neither is permitted to assert. A covering choice is also a decision about whether a secondary drainage plane is needed, whether a counter-batten path has to exist, whether a continuous deck is required, and where the air and vapour control lines have to be. Asking that question at the point the covering is chosen is cheaper than discovering it when the battens are being ordered.

The useful questions are about the layers rather than the units. The tile and slate record asks what secondary drainage path is assumed at underlay level and where exactly it discharges, whether a counter-batten void is part of the build-up and, if it is not, what takes over its function, and whether the build-up beneath relies on air paths the covering arrangement has to keep open. The standing seam record asks whether the chosen deck can receive clips across the whole roof including at the perimeter zones, and which vapour and air control layers sit below the deck.

Both records also ask about metals, and the question is the same in each case: which metals will sit in contact at fixings, flashings and rainwater goods, and which metals will the run-off from this roof pass over before it leaves the building. Compatibility along the whole drainage path is described as part of the design rather than a site decision.