Winter working on site: the compliance jobs that change when the clocks go back

Most construction phase plans are written in reasonable weather, for a programme that assumes reasonable weather, by someone picturing the site in daylight. Then the clocks go back, the site has four usable daylight hours in the middle of the day, the ground stops draining, and a document that was accurate in September is quietly describing a site that no longer exists.

None of the duties change. CDM is the same in December as in June, and so is the Work at Height hierarchy and everything else. What changes is the set of conditions those duties are being met in, and the honest position is that a plan which does not acknowledge winter is a plan that is now partly fiction.

The construction phase plan is a live document, and this is when it proves it

The construction phase plan is supposed to be reviewed and updated as the work and the risks change — it is not a document produced for the file at the start of a job. Seasonal change is one of the clearest cases for a review, and one of the most commonly skipped, because nothing has obviously happened. No new trade has arrived, no new activity has started. The site has simply become a different place to work.

A sensible winter review is short and specific. What activities now happen in artificial light that did not before. Which access routes become unsafe when frozen or waterlogged. What the trigger is for stopping work on weather grounds and who makes that call. Where people warm up and dry out, because they will do it somewhere whether it is planned or not. What the construction phase plan needs to contain covers the structure; the seasonal addition is largely about access, lighting, weather limits and welfare, and it should be dated so that it is visible as a review rather than looking like the original was always that thorough.

Lighting changes the risk profile of work that was previously routine

Task lighting on a site is generally thought about for the task. What gets missed is everything between the tasks: the route from the gate to the welfare unit, the edge of an excavation that is lit from one side, the stack of materials in a yard that is now a trip hazard in shadow.

Temporary lighting also creates a new category of problem. Glare and hard shadow can be worse than uniform low light, because they hide the thing immediately outside the lit cone. A floodlight pointed across a scaffold or an excavation can leave the far edge in deep shadow while convincing everyone the area is lit. Walking the site after dark, at the time people actually work, is the only reliable way to find this, and it is a different exercise from a daytime inspection of the same ground.

That walk is worth recording as an inspection in its own right rather than being folded into the normal cycle. What a site inspection record needs to capture applies equally; a record that says which areas were walked, at what time, and in what conditions is considerably more useful six months later than one that says "site inspected".

Permits written for summer conditions stop fitting

Permit systems tend to carry assumptions that are invisible until the weather breaks them. A hot works permit assumes a fire watch that can actually see; in poor visibility that assumption needs revisiting. A confined space entry assumes a rescue plan that works, and cold, wet conditions affect both the casualty's deterioration rate and the rescuers' dexterity. Excavation permits assume ground conditions that freeze–thaw cycles change week to week.

The practical move is not to rewrite the permit system. It is to add weather and light conditions as explicit fields on the permit itself, so the person authorising it has to consider them and the record shows what the conditions were. How permit to work systems are structured covers the mechanics. The small change of making conditions a recorded input rather than an unstated assumption does most of the work.

Weather holds need a named decision maker and a written trigger

Almost every site has an understanding that work stops in high winds. Far fewer have a written trigger, a named person who makes the call, and a record of the decision. The absence matters in two directions: without it, the call gets made late because nobody wants to own it, and afterwards there is nothing to show the decision was made on a basis rather than on mood.

Wind limits are the obvious case, particularly for lifting and for work at height, and the thresholds are generally set by the equipment manufacturer and the lift plan rather than by a site rule of thumb — those documents are the authority. But a trigger is also worth defining for ground conditions, for ice on access routes and scaffolds, and for visibility. The decision record can be two lines in the site diary: conditions, decision, who made it. That is enough.

The records get thinner exactly when they should get thicker

This is the pattern worth watching for. In winter, everything that happens outdoors becomes less pleasant, including writing things down. Inspections get shorter. Toolbox talks move indoors, get briefer and lose the walk-round element that made them useful. Briefings happen in the dark at the gate with people who want to get moving.

The result is that the season with the highest risk generates the thinnest evidence. If you look back at a site's records across a year, the winter months are frequently the sparsest, which is the opposite of what the conditions justify. What makes a toolbox talk record adequate rather than decorative is worth re-reading in October, because winter talks are exactly the ones that decay into a signature sheet.

Short, specific and seasonal beats long and generic here. A five-minute talk about the particular icy ramp on this site is worth more than a twenty-minute generic winter safety briefing, and it produces a record that means something because it names the hazard it was about.

Welfare is a duty, not a nicety

Welfare provision carries requirements year round, and cold and wet weather makes the drying and warming elements materially more important rather than optional. A site where people cannot get dry has a productivity problem, a morale problem and a health problem, and the last of those is the one that is a legal matter rather than a management preference. Facilities that were adequate in summer may not be adequate for a wet shift in February, and that is worth checking rather than assuming — the current HSE guidance on welfare requirements is the place to check it against.

Subcontractors arriving mid-winter need a different induction

A trade arriving in November has never seen the site in good conditions and has no baseline for what is normal here. They do not know which route floods, which gate ices over, or which area was fine in September and is not now. An induction written for a dry site in daylight will not tell them.

Adding a short seasonal section to the induction — current access routes, current lighting arrangements, the weather hold trigger and who calls it, where to dry out — takes very little time and addresses the group most likely to be caught out. It is also easy evidence of a competent induction, which matters when the question is asked later.