Reflectance assumptions in a daylight model: an audit

Reflectance is the input a reviewer should check first, because the average daylight factor formula in BRE 209 divides by one minus the square of the room's average surface reflectance. A model run with an average internal reflectance of 0.5 and the same model run at 0.4 produce materially different answers from identical geometry. Nothing in the drawings changes. The compliance conclusion can.

So the question for anyone reviewing a daylight report is not whether the software is right. It is whether the numbers fed into it match the building that will be built.

Why reflectance drives the result

In the average daylight factor method the denominator contains the term one minus R squared, where R is the area weighted average reflectance of the room's internal surfaces. As R falls, the denominator rises and the reported daylight factor falls. The relationship is not linear, so a small reduction in assumed reflectance at the upper end of the range has a larger effect than the same reduction lower down.

Climate based methods behave differently but are not immune. Where illuminance is computed by ray tracing, interreflected light contributes a substantial share of the result in a deep plan room or a room facing an obstruction, and that share is governed entirely by assumed surface reflectances.

Two consequences follow. In a shallow, well glazed room facing open sky, reflectance assumptions matter relatively little because most of the light arrives directly. In a deep room, a room behind a balcony, or a lower floor facing a close obstruction, reflectance assumptions can be the difference between pass and fail. Those are exactly the rooms that end up in dispute, which is covered from the design side in assessing daylight for a proposed building's own rooms.

Values commonly assumed for surfaces

The widely used set of internal surface reflectances is a ceiling at around 0.7, walls at around 0.5 and a floor at around 0.2. Applied to a typical room these give an area weighted average in the region of 0.5, which is the figure BRE guidance treats as representative of a domestic interior.

Those values assume light finishes. They do not describe:

• exposed concrete soffits, which sit well below a painted plasterboard ceiling
• dark feature walls, timber panelling or exposed brickwork
• dark timber or dark tiled flooring, which can fall below 0.1
• fitted kitchens with dark units occupying a large share of the wall area
• rooms where a large wardrobe or bookcase covers a wall

Externally, ground and obstruction reflectances are usually assumed at around 0.2, representing typical paving and brickwork. A model that assumes a higher figure for a facing elevation is claiming bounced light from a surface that may be dark brick, glass at an unhelpful angle, or in deep shadow for most of the day.

Where a model simplifies the urban surroundings into uniform reflecting surfaces, the assumed reflectance of those surfaces does a large part of the work. Check what was assumed and whether it matches the context photographs.

Glazing transmittance and frame factors

The second family of inputs is the glazing, and here the gap between assumption and specification is usually wider.

The transmittance point deserves emphasis. Specifying solar control glazing to satisfy overheating criteria reduces visible light transmittance, and a daylight model run on a clear double glazed assumption will then overstate the result. The two assessments have to be run on the same glass, which is one of the more common coordination failures between the overheating and daylight workstreams.

Where an optimistic assumption shows up

A reviewer can usually spot a generous input set without rerunning anything.

Look for a reported result that clears the criterion by a narrow margin across many rooms. Genuine building performance varies; a cluster of results sitting just above a threshold suggests inputs were adjusted until they did.

Look for an inputs table that gives a single average reflectance figure without naming the surfaces, or that gives no figure at all. Look for a glazing transmittance quoted to two decimal places with no product reference. Look for the absence of any maintenance or dirt correction. Look for rooms with balconies where the balcony soffit is absent from the model, which is visible by comparing the rendered geometry with the elevations.

And look at the room naming. A room described as a study or a utility space, excluded from assessment, that is drawn at bedroom size with a bedroom window, is a scoping decision rather than a modelling one, and the scope question is dealt with separately in the choice of which rooms to assess.

Running a sensitivity test

The honest answer to input uncertainty is a sensitivity run, and it costs very little once the model is built.

1. Run the base case with the assumed values, documented.
2. Run a pessimistic case: internal reflectances reduced by roughly a fifth, external reflectance reduced, glazing transmittance taken from the specified product rather than a generic value, and a dirt correction applied.
3. Report both sets of results in the same table, by room.
4. Identify the rooms whose conclusion changes between the two cases, and say what the design response is.
5. Where the pessimistic case fails, either change the design or commit to a finishes specification by condition.

The last point is the practical resolution. If a result depends on light internal finishes, the finishes should be secured, and a condition requiring internal surface reflectances no lower than stated values is enforceable in a way that a sentence in a report is not. Reviewers should ask for it and applicants should offer it.

Documenting the inputs

A daylight report that can be checked has an inputs table near the front. It should name, for each case:

The software and version, the calculation method and the standard applied with its edition. The sky model used, or the weather file for a climate based assessment. Internal surface reflectances for ceiling, walls and floor, with the resulting area weighted average. External ground and obstruction reflectances. Glazing visible light transmittance with the product or specification reference. The frame factor or a statement that frames are modelled geometrically. Any maintenance or dirt correction. The working plane height and the grid spacing and offset from walls. The source and date of the geometry, including the survey of neighbouring buildings.

Also record what was left out. A model that omits trees, future consented schemes on adjoining land, or internal partitions not yet designed should say so. Omissions stated are defensible; omissions discovered are not.

What a reviewer checks against the drawings

Inputs are only half the audit. The other half is whether the model is the building.

Compare the glazed area per room in the model against the window schedule, not against the elevations, because elevations show structural openings and schedules show glass. Compare room depths and ceiling heights against the sections. Check balcony depths, reveal depths and any brise soleil. Check that the floor to be assessed is the floor that is built, since a mezzanine or a raised ground floor changes the sky angle. Check the neighbouring context against a topographical survey rather than mapping data, particularly parapet heights.

Then look at the conclusion and ask whether it would survive the pessimistic case. Where it would not, the report is a statement of intent rather than an assessment, and a reviewer signing it off is taking on the gap. Anyone being asked to review and sign a draft assessment should set that expectation before starting, which is part of what the professional directory review route is for.

A short closing observation. Reflectance assumptions are not dishonest in themselves. They are necessary, because nobody knows the paint colour at planning stage. What makes them a problem is leaving them unstated, and the fix takes one table and a second model run. For the wider effect those values have on reported figures, see reflectance assumptions and how they change daylight results, and for the method itself the daylight and sunlight assessment page sets out what the report covers.

Enter your postcode on PlanScope to see which reports a scheme with tight daylight margins is likely to need.

PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval. Daylight modelling is a specialist discipline and councils usually expect a suitably qualified author, so professional review is strongly advised.