Solar irradiance in the UK: the real numbers

Irradiance, irradiation, and why the distinction matters

Irradiance is power arriving now, in watts per square metre. Irradiation is energy arriving over a period, in kilowatt hours per square metre.

Almost everyone, including us in casual use, says “irradiance” for both. The distinction earns its keep in two places:

• Peak irradiance decides whether an array ever reaches its rating, and whether an inverter clips. On the brightest June day in London, a 35° south roof peaks at about 602 W/m² on our modelling — 60% of the 1,000 W/m² a panel is rated at.
• Annual irradiation decides how much electricity you get, which is the question everyone is actually asking.

The resource, location by location

Annual global irradiation on a horizontal surface. Horizontal because it is orientation-neutral: it describes the place, not somebody’s roof.

Source: our own modelling from PVGIS 5.3, European Commission Joint Research Centre. Year-to-year variation is one standard deviation of annual output as reported by PVGIS. Full dataset: data/irradiance-detail.yaml.

The whole country sits inside a 1.5× band. That is the first surprise: the latitude difference between Brighton and Shetland is nearly ten degrees, and the annual resource differs by a third.

Tilt a roof into it and the numbers rise

A flat surface is not what anyone installs on. Pitch the same square metre to 35° and point it south and it intercepts substantially more, because it faces the sun more squarely for more of the day and through more of the year.

The gain is consistently 16–19%, everywhere. Which is a useful hint about something we cover in full on roof orientation and pitch: the ratios between orientations barely change across the UK, even though the totals do.

The shape of the year is the real story

Here is what the annual figure hides.

Source: our own modelling from PVGIS 5.3, European Commission Joint Research Centre, at 20% system losses. Full dataset: data/irradiance-detail.yaml.

Read the last two columns first. A 4 kWp array in Lerwick produces about 0.9 kWh on an average December day. The same array in Brighton produces about 4.6 kWh. Both are in the same country.

And the seasonal concentration climbs steeply with latitude. Brighton takes 68% of its annual output in the six months from April to September; Lerwick takes 78%, and only 8% of its year arrives across November, December, January and February combined.

That has practical consequences:

• A battery does not fix winter. It shifts electricity by hours, not months. In a Lerwick December there is very little to shift.
• Sizing to your annual consumption oversizes for summer. An array matched to your yearly demand will overshoot heavily from May to August and undershoot from November to February. That is the surplus-export problem our rooftop calculator warns about.
• Solar and heating are not substitutes. A heat pump’s demand peaks in the months solar output troughs.

Is the UK too far north for this?

The most common objection, and the numbers do not support it. At an identical 35° south pitch, with identical settings:

Source: our own modelling from PVGIS 5.3, European Commission Joint Research Centre. Holding the pitch at 35° understates the southern locations, each of which would do better at its own optimum tilt. Datasets: data/irradiance.yaml and data/irradiance-detail.yaml.

Brighton out-yields Munich. Berlin sits between our figures for Manchester and Norwich. Germany built one of the world’s largest solar fleets on a resource comparable to southern England’s.

Madrid is genuinely much better — around 40% above Brighton — so the answer to “is Spain sunnier” is plainly yes. The answer to “is the UK viable” is that the country with the most instructive deployment record has roughly our sunshine.

How much does a year vary?

Less than the folklore suggests. PVGIS publishes a standard deviation of annual output from year-to-year weather variation, and across our 25 locations it runs from 2.4% to 4.1% — Plymouth the steadiest, Aberystwyth the most variable. The figure for every location is in the table above.

So a dull year costs a few per cent, not a third. Two consequences:

• An annual average is a defensible planning figure. It is not a forecast of any particular year, but the error bar is small.
• A single year’s output tells you very little about whether a system is underperforming. A 15% shortfall against a modelled figure is not weather. It is shading, soiling, a fault, or a modelling assumption that did not hold.

There is no clean latitude pattern in the variability, which is worth saying because it would be an easy story to tell. Plymouth and Kirkwall are both among the steadiest and they are 700 miles apart; Aberystwyth and Inverness are both among the most variable. Whatever drives it is more local than latitude, and we are not going to invent a mechanism for it.

What irradiance data cannot tell you

Your roof. These are modelled figures for reference locations, on unshaded surfaces. Shading is site-specific and can cost far more than latitude does.

Your microclimate. A valley, a hill, a coastal fog belt or an urban heat island all move the answer, and a 25-point national grid cannot see them.

Soiling and snow. Dirt, lichen, bird mess and snow cover all reduce output, and none is in the irradiance figure.

What the electricity is worth. A kilowatt hour generated in June and exported is worth a small fraction of one generated in December and used. Irradiance data is silent on that, and it is the question that decides the economics.