Shading: how much output you really lose
We modelled fifteen shading scenarios rather than repeating rules of thumb. The headline result is not the annual loss — it is that shading is overwhelmingly a winter problem. A building due south costs 2% of summer output and 75% of winter output.
Where the obstruction is matters far more than how big it is.
A wide building due south at 30° above the horizon costs about 20% of annual output. The same obstruction due north costs nothing at all — we modelled it at 60° and the loss was zero.
And the annual figure hides the important part. That southern building removes 1.5% of summer output and 75% of winter output. Shading is not a uniform tax on generation; it is a winter phenomenon, and winter is when the electricity is worth most.
Why direction dominates
The sun in the UK is always in the southern half of the sky. It rises north of east and sets north of west in high summer, but it spends the middle of every day — and the whole of every winter day — to the south.
So an obstruction’s cost is set by how much of that southern arc it covers. This is the same geometry that makes a north-facing roof lose more than half its output, and it produces a result that surprises people: a tall building due north of your house is irrelevant to your solar panels.
| Obstruction lies | Output | Loss | Relative cost |
|---|---|---|---|
| Due north | 919 | none | |
| Due east | 889 | 3.3% | |
| Due south | 814 | 11.4% | |
| Due west | 894 | 2.7% |
Source: our own modelling from
PVGIS 5.3,
European Commission Joint Research Centre. This run also confirms that PVGIS's
horizon sectors are ordered clockwise from due north, which we checked rather
than assumed. Dataset: data/shading.yaml.
The scenarios
Every figure is our own modelling. Shade factor is MCS’s SF — one minus the loss — so an SF of 0.80 means you keep 80% of the unshaded output.
| What is blocking the sky | SF | Annual loss | Range across the UK | Summer | Winter |
|---|---|---|---|---|---|
| Low urban skyline all round, 10° | 0.98 | 2.2% | 1.8–3.1% | 0.1% | 10.3% |
| Dense urban or valley, 20° all round | 0.87 | 13.5% | 12.1–13.5% | 1.9% | 57.8% |
| Steep valley or tall buildings all round, 30° | 0.75 | 25.2% | 24.0–25.2% | 8.4% | 75.5% |
| Building or hedge due south, 20° high, wide | 0.89 | 11.5% | 10.2–11.5% | 0.1% | 57.8% |
| Building due south, 30° high, wide | 0.80 | 19.8% | 18.6–19.8% | 1.5% | 75.5% |
| Tall building due south, 45° high, wide | 0.67 | 32.8% | 31.4–32.8% | 14.8% | 75.6% |
| Single tall tree due south, 45°, narrow | 0.92 | 7.9% | 7.4–8.0% | 1.2% | 28.1% |
| Single very tall tree due south, 60°, narrow | 0.89 | 11.4% | 11.4–12.0% | 5.8% | 30.7% |
| Tall obstruction to the south-east, 45° | 0.83 | 16.9% | 16.2–17.1% | 12.4% | 26.5% |
| Tall obstruction to the south-west, 45° | 0.84 | 16.4% | 15.9–17.1% | 9.3% | 33.7% |
| Tall obstruction due east, 45° | 0.93 | 7.0% | 6.7–7.4% | 8.8% | 0.9% |
| Tall obstruction due west, 45° | 0.93 | 7.3% | 6.9–8.7% | 9.5% | 1.1% |
| Terraced row: tall obstructions east and west, 45° | 0.86 | 14.3% | 13.7–15.4% | 18.3% | 2.1% |
| Tall obstruction due north, 60° | 1.00 | 0.0% | 0.0–0.0% | 0.0% | 0.0% |
Source: our own modelling from
PVGIS 5.3,
European Commission Joint Research Centre, using the shade factor method MCS
MIS 3002 sanctions: generation with the obstruction divided by generation without
it. Unshaded baseline for London is
919 kWh/kWp. Summer is April to September;
winter is November to February. Dataset: data/shading.yaml.
The winter finding
Read the last two columns of that table together and a pattern appears that the annual figure conceals entirely.
| Scenario | Annual loss | Summer loss | Winter loss |
|---|---|---|---|
| Low urban skyline, 10° all round | 2.2% | 0.1% | 10.3% |
| Dense urban, 20° all round | 13.5% | 1.9% | 57.8% |
| Building due south, 30° | 19.8% | 1.5% | 75.5% |
| Tall building due south, 45° | 32.8% | 14.8% | 75.6% |
| Obstruction due east, 45° | 7.0% | 8.8% | 0.9% |
| Obstruction due west, 45° | 7.3% | 9.5% | 1.1% |
| Terraced row, east and west, 45° | 14.3% | 18.3% | 2.1% |
Southern obstructions are a winter problem. In winter the sun never climbs high and never leaves the southern sector, so anything blocking the south blocks everything. A 30° building takes three quarters of November-to-February output and essentially none of the summer.
Eastern and western obstructions are a summer problem, and the mirror image. In winter the sun does not get round far enough east or west to be blocked at all, so a 45° obstruction due east costs under 1% of winter output — while costing 8.8% of summer.
This matters more than the annual percentages, for two reasons set out elsewhere on this site. Winter output is already scarce: on our figures a London array takes only 18% of its year between November and February, and a Shetland array 8%. And winter units are the valuable ones — most likely to be used rather than exported, when a used unit is worth six times an exported one.
Width matters as much as height
Two scenarios at the same height and the same direction, differing only in how wide they are:
| Due south at 45° above the horizon | Annual loss |
|---|---|
| A wide building, covering 130° of the horizon | 32.8% |
| A single narrow tree, covering 30° | 7.9% |
Four times the loss for the same height. The sun crosses a narrow obstruction in minutes; a wide one blocks it for hours. So “there is a tall tree to the south” and “there is a terrace of houses to the south” are not the same problem, and a survey that records only obstruction heights is recording half the information.
A single tree also raises a possibility a building does not: pruning. A tree that can be reduced is often the cheapest available improvement to a solar array, which is why our orientation page puts removing shading above every geometry change.
The ratios hold across the country
Each scenario was modelled at Brighton, London, Manchester and Glasgow. The spread between locations is in the table above, and it is narrow: a median of 1.2 percentage points, 1.8 at the widest.
That means these are usable as national figures. Brighton generates 35% more than Glasgow per kWp unshaded; the proportional cost of a given obstruction barely moves. Same conclusion as the orientation matrix, for the same underlying reason: geometry is geometry everywhere.
What MCS requires of an installer
If you are buying from an MCS installer, shading is not optional paperwork. MIS 3002 requires it in the written performance estimate:
- The estimate must state a shade factor (SF), and annual output is calculated as kWp × Kk × SF. Where the horizon is clear, SF is 1.00 and the assessment may be omitted — but it has to be stated.
- Where SF is less than 1 and the MCS method is used, the estimate must carry a specific note: the factor “was calculated using the MCS shading methodology and we believe that this will yield results within 10% of the actual energy estimate stated for most systems.”
- Where another method is used, the note instead confirms the system “will deliver at least 90% of the energy (in kWh) as set out in this performance estimate.”
- Where SF is below 1.00 and the MCS method is used, you must also be given the sunpath diagram used to calculate it.
Two things worth taking from that. The standard’s own accuracy claim is ±10%, which is a useful calibration against any figure presented to you as precise. And you are entitled to the sunpath diagram — if a quote states a shade factor without one, ask.
Optimisers, microinverters and what they actually solve
There are two distinct shading problems and they need different answers.
Far shading — the horizon, the terrace opposite, the hill. This is what we have modelled. Sunlight does not arrive. No electronics recover it, and no product claim should suggest otherwise.
Near shading — a chimney, a vent, an aerial, a dormer, throwing a hard shadow across part of the array while the rest is in full sun. Here the loss can be much worse than the shaded area implies, because panels wired in series are limited by their weakest member. Power optimiser devices and Microinverter units address exactly this, by letting each panel operate independently.
So the honest summary is that module-level electronics can be well worth it on a roof with hard near shading and buy you nothing on a roof whose problem is a blocked southern horizon. Which of those you have is the question to answer before paying for either.
Practical steps
Stand on the roof, or as close as you can get, and look south. Everything on this page follows from what covers the southern arc between about 10° and 45° above the horizon.
Note width as well as height. A compass bearing for each edge of an obstruction is more useful than its height alone.
Think about January, not June. A shadow that clears by 10am in June may cover the array all day in December, because the winter sun never gets above it.
Ask for the sunpath diagram and the shade factor, in writing, before you buy. You are entitled to both.
Price the pruning before pricing the optimisers. If the obstruction is vegetation you control, that is the intervention with the best return, and it is the one nobody selling equipment will suggest.
Sources
Sorry to hear that. What was the problem?