Shade on solar panels: how to tell it's costing you output

Shade is the most underrated cause of low output. It has three properties that make it the perfect culprit: it costs more than you'd expect, it arrives slowly, and it isn't visible from the window at midday.

Why it costs more than the area suggests

Panels in a string are wired in series. The current flowing through the string is set by its weakest element — so a partly shaded panel doesn't reduce output by its own share, it drags the whole string down.

Modern panels soften this with bypass diodes: usually three per panel, each able to route around a shaded third. That limits the loss but doesn't remove it. The practical consequence stands: a shaded tenth of the area does not mean a tenth of the output lost. It's usually considerably more, and the exact figure can't be estimated from a photo.

Two kinds of shade, two different giveaways

Far shading (the horizon). A hill, a treeline on the skyline, continuous built-up area. It affects the whole array and mainly the start and end of the day. It's stable — it doesn't change from year to year.

Near shading. A chimney, an aerial, a tree, a dormer, a neighbour's extension. It affects part of the array locally, and it moves — across the roof through the day, and lengthening through the year. This is the diagnostically interesting kind, because it changes and because something can be done about it.

The signature that gives it away

What separates shade from most other causes is that it's regular:

• It repeats at the same time of day. Output runs normally, then drops at a particular hour even under a clear sky.
• It changes through the year. In June, with the sun high, everything may be fine; from September, as the sun drops, the obstacle starts shading earlier and for longer. The winter dip is therefore deeper on shaded systems than it should be.
• It's a smooth curve, not a step. A fault tends to be a step ("normal until yesterday, half of it since"). Shade draws a notch — a gradual dip and recovery.

If output fell in a step and permanently, shade is unlikely. That case is covered in your inverter isn't producing.

How to verify it without climbing on the roof

A horizon profile from a model. For any point on Earth it's possible to compute how high the terrain stands around it and how much sunlight that costs. That's exactly what our shading check does — it needs only an address and returns how much the horizon takes. It has one important limit: the model knows the terrain, not your tree or your chimney. It finds far shading, not near shading.

Comparing panel strings against each other. This is the strongest available evidence of near shading — and it only works on connected inverters that report each string's power separately. If two strings persistently fail to deliver comparably under a clear midday sun, something is in the way of one of them.

Our rule for it carries four safeguards against false alarms:

• it compares the ratio of the weaker to the stronger string, not absolute values (we don't know how many panels are on each),
• it reports below a ratio of 0.75; below 0.40 it's more likely a disconnected or heavily shaded string,
• only readings where the stronger string delivers at least 300 W count — at dusk or under cloud, ratios mean nothing,
• the effect must appear in at least 60 % of readings, and there must be at least five, so that a cloud drifting over half the roof doesn't count.

And even then it's only a warning with medium confidence, never a serious problem. The reason is honest: string imbalance is explained just as well by shade as by strings legitimately having different panel counts or different orientations (typically east–west). Inverter data can't tell those apart, so claiming a fault would be claiming more than the data supports.

Your own observation. The cheapest method: photograph the roof on the same day in the morning, at midday and in the afternoon. Then again six months later — the shift between seasons is precisely what reveals what matters.

What can be done

Cheapest first:

1. Trim the greenery. Most near shading is trees, and trees grow. Regular trimming is the cheapest measure in all of solar.
2. Rewire the strings. If shaded panels are scattered across strings, grouping them into one string confines the damage to it. That's work for an installer and may not be feasible.
3. Optimisers or microinverters on the shaded panels. These decouple the problem panel from the fate of the string. They address the root cause, but cost money and installation — worth it only where the shade is permanent and can't be removed. Do the economics from the actual loss, not from a generic promise.
4. Move or remove the obstacle. An aerial or a dish can be moved; a chimney cannot.

What shade is not: dirt. Cleaning and shading are different problems, and clean glass under a tree produces just as little as dirty glass.

When to suspect shade first

• The system never produced what it should have — from day one. A fault that developed later would show up later.
• The gap is deeper in autumn and winter than in summer.
• The array is on a roof with a chimney or dormer, or among taller buildings.
• There are mature trees around the house that were smaller at installation.

Where Slunora helps

The shading check computes a horizon profile for your address and says how much the terrain takes — without connecting to anything and without your production data.

The quick check compares actual output against expectation and shows whether the gap is larger than weather and orientation explain.

And if you've connected an inverter, we compare strings against each other continuously — after a few sunny days that becomes a usable finding. If the data shows no imbalance, we say that too.