Surface water flood maps against a local model

National surface water mapping is a screening tool, and a local hydraulic model is evidence. Treating the first as the second is how a flood risk assessment ends up arguing that a site floods when it does not, or that it does not when the drainage records say otherwise. The question of whether to rely on the surface water flood map or a local model has a straightforward answer: use the national mapping to establish whether there is an issue, and a local model where the layout depends on the answer.

What the national mapping is built from

The Risk of Flooding from Surface Water dataset is a national scale product, produced by applying rainfall directly to a digital terrain model and routing the resulting flow across the surface in two dimensions. It reports three bands of annual exceedance probability, commonly described as high, medium and low, corresponding to roughly a 1 in 30, a 1 in 100 and a 1 in 1000 chance in any year.

Two features of the method explain most of its behaviour. Rainfall is applied to the ground surface rather than routed through a drainage network, and drainage removal enters the calculation as a simplified allowance rather than as the actual sewers, gullies and culverts under the street. So the mapping shows where water would collect and flow if the drainage system contributed only a uniform amount of capacity.

The underlying terrain model comes from airborne survey data at a national grid resolution, which is excellent for a catchment and coarse for a plot. The national assessment has also been re-run and updated, with revised outputs and additional information such as depth and velocity in more recent versions, so the vintage of the dataset you are citing matters.

It is the right tool for its purpose. It tells a case officer, a lead local flood authority and an applicant that a site sits on a flow path or in a depression, and that the matter needs looking at.

Where its resolution misleads

The failures are predictable, and a reviewer will recognise them.

Small features disappear. Kerbs, walls, fences, raised thresholds, speed tables and narrow ditches all influence where water goes, and most of them sit below the resolution of a national terrain model. A flow route shown crossing a garden may in reality be stopped by a boundary wall built in 1935.

Buildings are treated inconsistently across dataset versions, so a flow path may run straight through a terrace in one version and around it in another. Culverts and piped watercourses do not appear as conveyance, which produces apparent ponding upstream of a culvert that actually works.

Ground level changes since the survey are invisible. A site raised or regraded in the intervening years, a demolished building, a new highway scheme or a filled quarry will not appear.

And the absence of mapped risk is weaker evidence than its presence. A site can flood from a blocked gully or a surcharging sewer without appearing on the map at all, which is why recorded incident history matters as much as modelled output.

What a local model adds

A local hydraulic model does the things the national product cannot.

It couples the drainage network with overland flow, usually by linking the sewerage undertaker's network model to a two dimensional surface model, so surcharge and exceedance are calculated rather than assumed. It uses local topographic survey, including levels taken on the site itself, so kerbs, thresholds and boundary features are in the geometry. Culverts, screens and structures appear with their real dimensions, and they can be tested blocked. And it can be calibrated against recorded flooding, which is the step that converts a model into evidence.

The cost is not trivial, and the programme runs to weeks rather than days once survey and network data have been obtained. That is why the decision to commission one should be taken on the basis of what the layout depends on rather than as a default.

Where a basement or other below ground accommodation is proposed, the case for a local model is at its strongest, because the consequence of an error is severe and the objection is predictable. The way that argument is usually framed is set out in basement development and the flood risk objection.

Lead local flood authority studies and their status

Between the national dataset and a bespoke model sits a body of work that applicants routinely overlook.

Lead local flood authorities commission surface water management plans and integrated urban drainage studies, often jointly with the sewerage undertaker, and these define Critical Drainage Areas and identify known flooding mechanisms street by street. Authorities also hold flooding incident records under their duty to investigate, and highway authorities hold records of carriageway flooding and gully performance.

The strategic flood risk assessment prepared by the local planning authority draws these together and is the evidence base against which applications are judged. It also defines things national mapping does not show at all, including the functional floodplain.

Read those documents before writing an assessment. Where a local study identifies a mechanism affecting your site, an assessment that relies only on national mapping and reaches a different conclusion will be contradicted by the authority's own evidence, which is the worst position to be in.

Reconciling conflicting sources in one assessment

Conflict between sources is normal, and the assessment should resolve it explicitly rather than quietly preferring the convenient one.

The structure that works is to present each source, state what it shows, and then explain the reconciliation in a short reasoned paragraph. Where a local model shows less risk than the national mapping, say why: better topography, the sewer network represented, calibration against recorded events, a boundary feature surveyed on site. Where the national mapping shows less risk than the records, say that too, and treat the records as governing.

Three habits make this credible. Show the national mapping extract even where you are departing from it, because omitting it looks evasive. Quote incident records with dates and sources rather than summarising them. And where sources cannot be reconciled, take the more conservative position for the design and explain the uncertainty, rather than presenting a confidence the evidence does not support.

That approach also serves the policy tests. Where the exception test is engaged, the safety limb depends on the residual risk position, and the way that evidence is marshalled is described in exception test evidence and the two limbs in practice.

When commissioning a model is proportionate

Proportionality is the word the guidance uses, and it means matching the effort to the decision the evidence has to support.

A model is usually proportionate where a mapped flow route crosses the developable area and the layout has to respond to it, where below ground accommodation is proposed, where the scheme must demonstrate no increase in flood risk elsewhere and the mechanism is complex, where the site sits within a Critical Drainage Area with an identified mechanism, or where an objection has already been raised on modelled grounds and has to be answered in kind.

It is usually not proportionate on a single plot, an extension, or a site where the mapped risk sits comfortably outside the development footprint. There the answer is a topographic survey, a levels based assessment, threshold levels set with freeboard, and a clear exceedance route. The expectations for smaller schemes are set out in drainage evidence for a single new dwelling.

A middle option is often overlooked. A topographic survey plus a simple levels and volumes assessment, using the national mapping to identify the mechanism and local survey to test whether it reaches the site, answers a great many queries at a fraction of the cost of a coupled model. Where it does not resolve the question, it at least scopes what the model needs to do.

Citing the source so it can be checked

Vague citation is what turns a defensible assessment into a contested one. For every dataset used, state the dataset name, the publisher, the version or vintage, the date you obtained it, and the scenario or band being reported. Include the extract as a figure with a north point, a scale bar and the site boundary drawn on it.

For a local model, state who built it, when, what software, what terrain data, what network data, what events were run, whether it was calibrated and against what. For incident records, name the holder and the reference. For a strategic flood risk assessment or drainage study, give the document title, author and date, and quote the relevant passage rather than paraphrasing it.

Then keep the Flood Risk Assessment and the Drainage Strategy consistent with each other, because a contradiction between the two is read as an error in both. The flood risk assessment establishes the risk and the safety case; the drainage strategy manages the rainfall the development itself generates, and each should reference the other's conclusions rather than restating them differently.

PlanScope drafts both documents structured to UK planning policy, with the dataset references, survey levels and model outputs flagged as inputs to be supplied rather than assumed. Reports should be reviewed before submission, and PlanScope does not guarantee planning approval. For specialist technical reports, qualified professional input may be required.

This is general guidance, not legal advice. Check the current national datasets and your lead local flood authority's own evidence base before relying on it.