Shade, trees and chimneys: what actually costs you output

The threshold, and why the convergence matters

Shading is usually discussed qualitatively: some is bad, more is worse, get a site survey. There is a number, and two independent bodies reached it separately.

A state building code defines Total Solar Resource Fraction, a measure of how much of the available solar resource a surface actually receives once obstructions are accounted for. It then uses 80% as a cut-off: below that, a roof is excused from the solar-ready construction requirements that would otherwise apply. The regulator's position is that at that point the roof is not worth requiring readiness on.

A national laboratory's methodology for estimating national rooftop solar potential uses the same figure for the opposite purpose. It excludes from its estimates any roof area that cannot "produce 80% of the energy produced by an unshaded system of the same orientation". Area below that is not counted as potential at all.

Neither cites the other, and they are doing different jobs. One is deciding what to require of new construction; the other is deciding what to count in a resource estimate. That they land on the same figure is the strongest available signal that 80% is where usefulness genuinely falls away rather than an arbitrary line.

What that gives you is a way to ask a specific question. Not "is my roof shaded" but "what is the solar resource fraction on the planes you propose to use, and how much of the array falls below 80%". A site survey produces that number, and if nobody will give it to you, nobody has measured it.

The other exclusions, which rule out more roof than shade does

The same laboratory methodology applies several other filters before shade is considered, and they are worth knowing because together they exclude more roof than shading typically does.

North-facing planes are excluded outright, specifically those with an azimuth between 292.5 and 67.5 degrees. Not derated: excluded. In the northern hemisphere a north-facing plane receives too little direct sun to be worth counting.

Anything steeper than 60 degrees is excluded. A very steep plane sees the sun too obliquely through the productive part of the day.

Planes smaller than 10 square metres are excluded, on a contiguous basis. That is roughly 108 square feet, which is around five modern modules. Below that, the fixed costs of getting an array onto a plane outweigh what it can carry.

And on flat roofs a module-to-roof ratio of 0.70 is applied, accounting for roughly 2.5% self-shading between rows. Panels on a flat roof are tilted and therefore shade each other, so a flat roof cannot be filled the way a pitched one can.

One more filter is worth naming because it is the one homeowners apply themselves and get wrong. People tend to rule out a plane because it faces east or west, on the belief that only south counts. The methodology does not: it excludes only the northerly arc, which leaves east and west firmly inside what is worth counting. An east or west plane produces less than a south one and it produces it at different hours, which on a time-of-use tariff can be worth more rather than less. Ruling it out by instinct is the most common way a homeowner shrinks their own system.

The practical use of this list is as a sanity check on a proposal. If a design puts modules on a north-facing plane, on a very steep plane, or on a small isolated section, that is not necessarily wrong, but it is outside what a national laboratory counts as usable and it deserves a specific explanation.

What we are not going to tell you

There is a claim made constantly in this area that we went looking for and could not verify, and it is worth being explicit because it is the claim most likely to be made to you.

How much a module-level architecture recovers from partial shading. Microinverters and power optimizers are sold on the basis that a shaded module no longer drags down its whole string, and the mechanism is real. What we could not find is a primary source quantifying the recovery.

We looked. It is precisely the figure a manufacturer would benefit from publishing and precisely the figure we could not locate in laboratory work.

So we state the mechanism and decline the number. If your roof has shading that varies across it, per-panel management has a real argument behind it. If someone tells you it recovers a specific percentage, ask where that figure comes from, because we could not find its source.

And we give no rule of thumb about a single shaded panel. The old advice was that one shaded module drags an entire string down to its level, which was a reasonable description of an older architecture and is not a description of a modern one with per-panel electronics. How much of that original penalty remains, and under what shading patterns, is exactly the unpublished quantity above. Anyone still repeating the original rule as though nothing has changed is describing equipment nobody is selling you.

We also give no figure for what shade costs in general, because the question is malformed. Shade at nine in the morning in winter costs almost nothing; shade across midday in summer costs a great deal. The quantity that matters is the resource fraction on the specific planes, which is what a survey measures and what the 80% threshold speaks to.

The four obstructions, and how differently they behave

Shading is treated as one problem and it is at least four, distinguished by whether the obstruction moves, grows, or can be removed.

Fixed obstructions on your own roof: a chimney, a vent stack, a dormer, a satellite dish. These cast a shadow that sweeps predictably and they are permanent, but they are also small and close, which means their shadow is sharp and moves quickly. A survey captures them exactly, and the design works around them by leaving that area empty.

Fixed obstructions off your property: a neighbour's roofline, a tall building, a hill. Permanent, not yours to change, and the reason a site can simply be unsuitable. These are the ones worth establishing before you get attached to the idea.

Removing a tree is the one response that is not purely technical. It is irreversible, it can be governed by a local tree ordinance or a preservation order, and the tree may not be yours to cut. Nothing in the codes and methodologies we read grants a solar owner any right to remove or trim vegetation on another property, and nothing obliges a neighbour to keep a view clear. Establish who owns the tree and what the local ordinance says before the array design assumes the shadow will disappear.

Trees, which are the difficult case, because they are the only obstruction that grows. A survey measures the shadow that exists on the day it is done. The system is expected to run for decades, and nothing in a standard analysis models the tree at maturity. A young tree south of the array is a future problem that today's measurement will not show.

And seasonal or transient shading: a deciduous tree that is bare in winter and dense in summer, or snow lying on a lower row. These vary through the year in ways an annual figure averages away, and the direction matters, because summer shading costs far more than winter shading does.

The useful question about any obstruction is which of these four it is, because the answer determines whether the response is design around it, accept it, remove it, or plan for it to get worse.