Setting up a daylight model that survives scrutiny

A daylight model is only as good as the assumptions written down beside it, and most of the daylight model assumptions a reviewer ends up querying are the ones nobody recorded. Reflectances taken from software defaults, neighbouring room depths guessed from the street, a baseline that quietly omits a consented scheme next door. None of those is fatal on its own. Together they make a result impossible to rely on, and a reviewer who cannot reproduce the numbers has to treat the assessment as untested.

Survey data the model has to be built from

Start with what the geometry is measured from, because everything downstream inherits that error. A topographic survey tied to Ordnance Survey datum, a measured survey of the existing building if it is being retained, and an elevational survey of every neighbouring facade carrying a window under assessment.

The awkward part is the inside of other people's homes. You cannot measure room depth, internal layout or room use behind a neighbour's window, so some of the model is inference. The honest way to handle it is to record the source of every assumed room, in a table, window by window:

• Measured externally, with window head, cill and reveal dimensions and the date of the survey.
• Room depth and use taken from drawings on the council's planning register, with the application reference.
• Room depth and use assumed, with the basis stated, for example a standard depth for the building typology.

That table is the single most useful thing in a daylight report. It tells a reviewer which results are measured and which are inferred, and it lets an objector challenge an assumption rather than the whole assessment. The alternative, a blanket sentence saying rooms were assumed to be of typical dimensions, invites the response that the evidence carries less weight, which is roughly what happens when objecting on daylight grounds is done well.

Window references must match the drawings. Numbering that changes between the plan, the table and the results is the most common reason a reviewer asks for a resubmission rather than reading the conclusion.

Assumed reflectances and where they come from

Reflectance values belong in the method section, not in the software defaults. Internal daylight results move materially when they change, so state the figure used for every surface and name its source.

Values widely used in UK practice sit around 0.2 for ground surfaces and external obstructions, 0.5 for internal walls, 0.2 for floors and 0.7 for ceilings. Those are starting points from the daylight literature rather than fixed requirements, and the correct value for a dark brick courtyard is not 0.2 because the software offered it. Where a scheme leans on reflected light from its own facades to get a borderline room over the line, the reflectance of those facades becomes a design commitment that should be conditioned, not an assumption.

Vertical sky component is unaffected by any of this. It is measured at the centre of the window on the outside face of the wall against a uniform overcast sky, so reflectance and internal layout do not change it. That is exactly why VSC remains the workhorse metric for effects on existing neighbours while internal daylight metrics argue about surfaces.

Reveals, frames, glazing and the sky model

Four input sets decide internal results and all four get fudged.

Clear double glazing in common use transmits a little under 0.7 of incident light, and triple glazing less again, so a scheme that switches to triple glazing for Part L reasons after the daylight report was written has invalidated it. If the specification is not fixed at application stage, model the worst reasonable case and say so.

The sky model matters as much. A uniform overcast sky underpins VSC and daylight factor work. Climate-based assessment against illuminance targets needs a weather file, and the file should be the nearest representative one to the site rather than whichever city the software opens with. Name the file in the report.

The existing baseline and consented schemes

The baseline is the condition at the date of the application, which usually means the existing buildings standing on and around the site. Two mistakes recur.

The first is modelling a cleared site as the baseline where the existing building is to be demolished. That flatters the proposal by comparing it against open sky. If a notional or policy-compliant baseline is used for a secondary comparison, label it and keep it separate from the primary results.

The second is omitting consented but unbuilt schemes. A reviewer checking the council's register will find them, and a cumulative scenario that only appears after an objection looks like concealment even when it was an oversight. Run the cumulative case, report it alongside the proposed case, and note the application references of the schemes included. Where the loss is temporary rather than permanent, that distinction has its own treatment, which is the subject of temporary loss of daylight during construction.

Sunlight follows the same discipline with a different period. Annual probable sunlight hours are assessed for windows within 90 degrees of due south, with the winter period running from 21 September to 21 March. Modelling the summer solstice instead of the equinox and the winter period is a straightforward error that discredits everything else in the document.

Sensitivity testing the assumptions

A single set of numbers from a single set of assumptions is a claim. And a range is evidence.

Pick the three or four inputs that are actually uncertain, usually neighbouring room depth, the reflectance of the proposed facades, glazing transmittance and the extent of any obstruction you could not survey. Rerun the affected windows with each varied to a reasonable bound and report how far the result moves. Where a window passes on central assumptions and fails on pessimistic ones, say so in the conclusion rather than burying it in an appendix.

This is also how to answer the policy question honestly. The Framework asks authorities to apply daylight and sunlight guidance flexibly where a scheme would still give acceptable living conditions, and the current edition uses policy codes rather than paragraph numbers, so check the code on gov.uk before citing it. Flexibility is easier to argue for a window that is marginal across a tested range than for one presented as a precise pass. The way an authority actually words that balance comes from its own policy, which is why local plan amenity policies and how they use BRE 209 is worth checking before the modelling starts.

Documenting the model so a reviewer can rebuild it

A reviewer should be able to reconstruct the model from the report without the model file. That means recording, in one place: the survey source and date, the software and version, the sky model or weather file, every reflectance and glazing value, the window and room assumption table, the massing model revision and date, the datum and coordinate system, and the list of cumulative schemes included.

Add the drawings the massing was built from, by revision number. Schemes change, and a daylight result that corresponds to an earlier massing is worse than no result at all because it reads as current.

Where a Daylight and Sunlight Assessment sits alongside sun-on-ground work for gardens or amenity space, keep the model shared and the reporting separate, since an Overshadowing Study uses the same geometry but different test dates and criteria.

Errors that get a result rejected

The ones that come up again and again, in rough order of how often they appear:

1. Window references that do not match the drawings, so results cannot be located.
2. Assessing windows to non-habitable rooms and reporting them alongside habitable ones without distinction.
3. A missing or wrong datum, so levels are out by a floor height.
4. Balconies, brises-soleil and recessed terraces left out of the model.
5. Measuring from the cladding face rather than the glass line.
6. Treating BRE numerical guidelines as absolute standards, or presenting ratios without the absolute values behind them.
7. No cumulative scenario where consented schemes exist.
8. A conclusion that aggregates pass rates across a whole block without identifying which rooms fail and by how much.

The last one is the real tell. Reporting that most windows comply tells a reviewer nothing about the two flats that will be dark, and those two are what the officer's report will be written about. Single aspect units are usually where the failures cluster, and there are design responses, as set out in daylight to single aspect flats and how to fix them.

A daylight assessment drafted with PlanScope sets out the method, the assumptions table and the policy framing, and a qualified professional then runs and signs the modelling. For specialist technical reports, qualified professional input may be required. PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval.