Why solar produces so little in winter

Opening the app in December is, for many owners, the first moment they suspect something is wrong. Next to summer, the numbers look like a fault.

They aren't. The winter drop is the most predictable thing about solar — and also the period from which the least can be diagnosed. Both are worth explaining, because together they determine what to look at in winter and what to ignore.

How big the drop normally is

At our latitudes, December output tends to be roughly a fifth to a tenth of July's. The exact ratio depends on latitude, tilt and orientation — the further south and the steeper the roof, the gentler the drop.

Over a full year, December and January together account for only a few per cent of annual output. A system that makes over a thousand kilowatt-hours in a summer month may make a few dozen in December and still be perfectly healthy.

Why it happens

Four things act at once:

• The sun is low. Its rays travel a longer path through the atmosphere, so less energy reaches the panel at any given moment than in summer.
• The day is short. In December the sun is above the horizon roughly half as long as in June — and the useful part, when it's high enough, is shorter still.
• Roof tilt is optimised for the year, not for winter. A tilt around 30–35° yields the most annually. Winter sun would prefer something steeper, but that would cost output in summer, when there's far more energy to be had.
• It's overcast more often. European winter cloud is more persistent than summer cloud.

One detail that runs against intuition: cold is good for panels. A photovoltaic cell loses efficiency as it heats up, so a clear frosty February day produces more per hour than a muggy July one. The winter drop isn't caused by cold — it's caused by a lack of light.

Snow

An unbroken layer of snow cuts output to practically nothing. It usually doesn't last long — panels are smooth, tilted, and even the small output underneath warms them enough for the layer to slide off by itself.

Don't clear it off. The risk of falling from a roof is real, so is scratching the glass, and the kilowatt-hours gained aren't worth it — we're talking about single-digit kWh that winter will hand back within days anyway.

Why winter supports almost no conclusions

This is the heart of it. The model of expected output (PVGIS, in our case) is far less certain in winter than in the rest of the year, and from two directions at once:

1. Reality varies more. Where a summer month differs year to year by something like 15–25 %, a winter month can differ by 40 %. One long inversion episode makes a bigger difference in December than a whole summer does.
2. The model tends to overestimate winter. With a low sun, the effects of the horizon, ground reflection and diffuse radiation are harder to compute, so models tend to overshoot winter months.

Put those together and you get a conclusion that sounds inconvenient but is honest: December and January cannot decide whether a system is healthy.

Which is why we treat them differently. Outside winter we flag a suspicious month below 70 % of expectation; in December and January only below 50 %. That isn't leniency for the sake of a quiet life — it follows from how large the winter noise is. A stricter threshold would generate false alarms on healthy systems.

What winter does reveal

Winter isn't diagnostically blind. Three things show up reliably:

An outage. The difference between "little" and "nothing" is clear even in December. A month below 2 % of expectation is an outage regardless of season — and in winter it's the easiest to miss, because a low number surprises nobody.

Shade that's invisible in summer. A low sun stretches shadows far enough that an obstacle the June sun clears starts shading for hours in December. When output falls only in part of the year and only in part of the day, shade is the first suspect — our shading check needs nothing but an address.

A fault the inverter reports itself. A status code means the same in January as in July; see inverter status codes.

Winter and money

The winter drop hurts financially more than the lost kilowatt-hours suggest — winter is exactly when you consume most. The ratio flips: in summer most output goes to the grid at a low export price, while in winter you use most of the little you make at home, displacing expensive imported electricity.

A kilowatt-hour produced in December is therefore worth more to your wallet than a July one. The mechanics are covered in self-consumption.

What to actually do

1. Don't judge a system by winter. Wait for spring, or look at a full year.
2. Distinguish "little" from "nothing". Nothing is always a signal, even in January.
3. Leave the snow alone.
4. If winter worried you, check the whole year, not December — the quick check compares annual or monthly output against a model for your location and says whether the gap sits inside normal variation.

And the sentence that matters most: if December looks low to you but the data doesn't point at a problem, we'll say so. Winter alone never brings down a verdict here.