The heating season is getting shorter, and by how much depends where you live

The useful question about a warming climate, for somebody deciding what to put in their house, is not how much the summers have changed. It is how much less heating they are going to buy.

That has a proper answer, because heating demand has a standard unit, and this site holds 77 years of daily temperatures for 25 British towns. So here is the arithmetic, computed from the series rather than quoted from anybody.

What a degree day is

A heating degree day counts how far below a base temperature a day sat. The UK convention is a base of 15.5 degrees: a day averaging 5 degrees contributes 10.5 degree days, a day averaging 16 contributes nothing.

Add them up across a heating year, July to June, and you get one number for how much heating that year demanded. It is roughly proportional to the fuel a building burns, which is why energy managers have used it for decades to compare one winter with another without being fooled by the weather.

The base of 15.5 rather than 18 reflects an old and still reasonable assumption: a house at 19 degrees inside gets the last few from people, cooking, lighting and appliances. The absolute figure matters less than consistency, and everything below is on the same base.

The headline

Across the 25 towns, averaged:

The 2020s row covers six complete heating years, which is not a decade, so treat it as a direction rather than a settled figure.

Two things are worth drawing out. The first is that the 1960s, not the 1950s, were the coldest decade in this series in almost every town, so anybody comparing today with “the fifties” is using a slightly milder baseline than they think.

The second is that every one of the 25 towns fell. Not most of them, not on average: all of them. That is unusual in a dataset this size and it is what makes the number worth acting on.

By town, and the pattern in it

The gradient is clean: the further south and the further from the Atlantic, the larger the proportional fall. The south coast has shed about a sixth of its heating demand. Belfast has shed less than a tenth.

That is the signature of a maritime climate. Northern Ireland and western Scotland sit in air that has crossed a great deal of ocean, and the ocean warms slowly and buffers everything downwind of it. Brighton and Cambridge are much more exposed to continental air, which has warmed faster.

It is worth separating the proportional change from the absolute number, because they point different ways. Dundee has changed least in percentage terms among the mainland towns and still needs 1,054 more degree days a year than Plymouth, which is 65 percent more heating. A milder Dundee is a long way from a mild one.

What this changes, and what it does not

It changes running cost, not sizing

This is the distinction that matters most and the one most easily got wrong.

A heat loss calculation sizes a system against the coldest conditions it has to handle: a design external temperature of around -2 or -3 degrees in much of England, lower in the Highlands. That design condition has moved far less than the seasonal total, because it is set by cold snaps rather than by averages, and a 14 percent milder season can still contain the coldest week in a decade.

So nothing here is an argument for a smaller boiler or a smaller air-to-water heat pump. What it is an argument for is lower expected annual consumption at the same capacity: fewer hours of running, not less power while running.

If you want that in pounds for your own house, the running cost calculator takes your own rate and your own hours rather than assuming either.

It shifts the arithmetic on heating investments slightly

Anything whose payback depends on annual heating consumption gets marginally worse as the seasons mild: insulation, glazing, a heat pump against a gas boiler. Less demand means less saving per year on the same capital cost.

The size of that effect is a sixth of the heating bill in the south over sixty years, which is real but is not the thing that decides any of these investments. Fuel prices move faster in a single year than the climate has moved in a decade, so this belongs in the background of the decision rather than in the foreground.

It does not mean a milder winter is a cheaper one

Degree days measure weather, not bills. A mild winter at high unit rates costs more than a cold one at low rates, which is roughly what happened between 2020 and 2023. See what the October 2026 price cap changes for the other half of that equation, and when to put the heating on for the decision this data sits underneath.

The two directions are the same fabric conversation

A building that loses less heat in January gains less in July. Everything that reduces heating degree day consumption, insulation, airtightness, better glazing, also reduces summer overheating, which is why this site treats them as one subject rather than two. What it costs to make an old house cope with heat is the summer half of the same spend.

How the numbers were produced

Daily mean temperatures for each town, 1950 to 2026, from this site’s stored series. Degree days computed per day against a base of 15.5 degrees, summed over heating years running 1 July to 30 June and labelled by the starting year. A heating year counts only when it holds at least 360 daily readings, and a decade is reported only where at least seven complete heating years survive that test.

The source is ERA5-Land reanalysis, pinned to one model across the whole series so no decade is being compared against a different grid. It is a modelled reconstruction averaged over cells of roughly 11 kilometres, not a thermometer in each town centre, which smooths extremes. For a decadal seasonal total that smoothing is close to immaterial; for a single record day it would not be.

Per town figures for the towns themselves, including the warmest days in each series and the per decade tables, are on the town pages.

Weather data from Open-Meteo, licensed CC BY 4.0 and processed here.