The short answer: clear the outlets before the leaves fall, get the laps and upstands looked at while the weather still allows a repair, note where water stands before it freezes, and watch the ceilings once the heating goes on. Autumn work is cheap because it is planned; the same faults in January are not.
What should be done before the leaves fall?
Clear every route water uses to leave the roof, and do it before the trees make the job urgent. Outlets, gratings, gutters and the falls that feed them are what turn a wet week into a non-event, and they are the one part of a roof that stops working entirely rather than gradually.
Timing is the whole argument. A blocked outlet found in October is ten minutes with a gloved hand; the same blockage found in January has spent three months holding water against laps and upstands, and whatever it has softened is now a repair in weather that does not want one. On a portfolio this is exactly why the visits in a planned maintenance programme are dated rather than requested.
What is worth checking while the weather still allows a repair?
Everything that would need fixing on a dry, mild day. The seams between sheets, the upstands where the roof turns up against a wall, the collars around pipes and vents, the edges and the kerbs of anything glazed: these are the details that let go first, and every one of them is easier to put right in September than in December.
The point is not that autumn reveals faults winter would hide. It is that autumn still offers the weather to act on what is found. A lifted lap discovered now is a small planned repair; the same lap discovered when it is already letting water through is a holding measure, a wait, and then the repair anyway.
Why note where water stands before it freezes?
Because a puddle you have recorded in November tells you something a puddle you notice in March does not. Standing water is common and not always a fault, but where it sits, how long it takes to clear and whether it is spreading are the facts that decide whether it matters, and they are worth capturing while you can still see them.
Freezing changes the arithmetic. Water that has worked into a lap or a crack expands when it freezes and opens what it is sitting in, so a detail that has been tolerating standing water all summer can fail over a single cold snap. What ponding does and does not mean is set out in what a puddle on a roof actually means; the autumn job is simply to photograph it, dated, so the spring comparison is real rather than remembered.
What changes indoors once the heating goes on?
Two things start at once, and they look identical on a ceiling. Rainfall arrives more often, and every home starts producing warm moist air that has to go somewhere. That is why the first marks of the season are so often blamed on the wrong cause.
Watch for staining that appears with rain and dries back afterwards, which points outside, and for marks that arrive on cold still nights around the edges of a ceiling, which do not. Telling them apart before anybody is called to the roof saves a wasted visit and is covered in condensation or a leak. Either way, note the date and the weather each time something changes: by February that record is worth more than any description.
What genuinely cannot be done in cold or wet conditions?
Anything that needs a dry surface, and anything that has to cure. Bitumen membranes will not bond to a wet roof, so rain stops the work rather than slowing it. Torch-applied systems otherwise cope with cold well, with adjusted technique and material kept warm, which is why winter reroofs are ordinary rather than exceptional.
Cold-applied liquid systems are the ones the calendar really governs. They cure chemically and need temperature and humidity held steady until they have hardened, through the night as well, so the forecast decides the date rather than the diary. The exact conditions are listed on our liquid waterproofing page. The practical consequence is worth knowing in October rather than being told in December: if a roof needs liquid work, the useful weather is running out.
What is a fix-now problem, and what waits for spring?
The test is whether water is already getting in, or getting into the build-up. Anything actively leaking, any open lap, any failed upstand and any blocked outlet is a now problem regardless of the month, because every week of winter makes it larger and the deck underneath is what pays.
What can wait is work that is about condition rather than watertightness: a covering that is tired but intact, surface wear with no split in it, a detail that is ageing but still closed. Those are better done in settled weather, priced properly and planned, and a condition survey in autumn is how the two lists get separated before anybody has to decide in a hurry.
The honest summary is that winter does not break many roofs on its own. It finds the ones that were already going, and it charges more to fix them because the weather is against everybody. Anything that moves work from January to September is worth doing, and most of it is a small repair rather than a project.