Factual errors that recur in drafted flood sections

The recurring AI flood risk section errors are not writing problems. They are input problems, and they repeat because the flood section of a planning document pulls data from four separate sources that all look interchangeable and are not: the Environment Agency Flood Map for Planning, the long term flood risk service, the Planning Practice Guidance vulnerability table, and the published climate change allowances. A drafted section that reads fluently can still state the wrong flood zone, apply the wrong allowance band and confuse the two flood tests, all in the same page.

Each of these is checkable in a few minutes if you know what to check. Work through them in order before anything goes to a council.

Flood zone taken from the wrong source

The most common error is a flood zone pulled from the wrong map. There are two public services and they answer different questions.

The Flood Map for Planning shows Flood Zones 2 and 3 for rivers and the sea, ignoring the presence of flood defences. That is the map planning policy uses, and it is the one a flood zone statement must cite. The long term flood risk service shows risk from rivers, the sea and surface water as a likelihood category, taking defences into account. It is useful for context and it does not give you a planning flood zone.

A draft that says the site is at very low risk, or quotes a percentage chance from the long term service, has used the wrong source. So has one that says a site is in Flood Zone 1 on the strength of a postcode rather than a drawn site boundary, which matters on any site that straddles a zone boundary.

Check three things: that the zone is stated as Flood Zone 1, 2, 3a or 3b with the Flood Map for Planning named as the source, that the date of the map check is given, and that the site boundary rather than a postcode centroid was used. Where part of the site is in a higher zone, the section should say which part and what is proposed there. The report itself is a Flood Risk Assessment, and the zone statement is its foundation.

Vulnerability classification misapplied

Flood risk vulnerability classification is a table in the Planning Practice Guidance, and it decides whether development is permitted in a given zone and whether the exception test is needed.

The classifications run from water-compatible development through less vulnerable, more vulnerable, to highly vulnerable, and the mistakes are predictable. Dwellings are more vulnerable, and drafts sometimes describe housing as less vulnerable because the scheme is small. Basement dwellings are highly vulnerable, which is a different row of the table with much tighter consequences. Caravan and camping sites are treated differently from permanent residential. And a mixed use scheme takes the classification of each use, so a scheme with flats over shops has to address the more vulnerable classification for the residential element even where the ground floor is fine.

The error that causes the most trouble is quietly classifying a scheme one row lower to avoid the exception test. An officer checking the table will spot it, and the credibility cost extends to the rest of the document.

State the classification, name the guidance table it comes from, and apply it to each use in the scheme separately.

Climate change allowance band wrong

Climate change allowances are published by the Environment Agency and are selected by river basin district, by epoch and by the allowance category appropriate to the development's lifetime and vulnerability. A drafted section that gives a single national percentage has not engaged with any of that.

Three errors recur. Using an out of date set of allowances, which matters because the allowances have been revised and are expressed differently for peak river flow and for peak rainfall intensity. Selecting the wrong epoch, usually by applying a 2050s allowance to a residential development with a lifetime running past 2100. And applying a river flow allowance to a surface water calculation or the reverse.

The check is mechanical. Name the river basin district or management catchment, state the development lifetime and the epoch that follows from it, state the allowance category, give the percentage used, and cite the published allowances with the date they were accessed. Any drafted section that cannot produce those five items has not done the work. Urban creep is a separate uplift applied to area rather than rainfall, and conflating the two is its own error.

Finished floor levels without survey data

Finished floor levels are where a drafted section invents numbers if it is allowed to. The level has to be set relative to a modelled flood level, usually the 1 in 100 year level with the climate change allowance applied, plus a freeboard allowance agreed with the authority, and expressed in metres above Ordnance Datum.

That requires three inputs the document cannot generate: a modelled flood level for the site, a topographical survey tied to Ordnance Datum, and the authority's freeboard expectation. Where any of them is missing, the correct output is a flagged gap rather than a plausible figure, which is exactly why PlanScope drafts insert a "[TO BE CONFIRMED]" marker instead of filling the space.

Leaving the marker visible in the working draft and resolving it before submission is the discipline. Deleting the marker and rounding to a tidy number is how an unsupported level reaches a council.

The same applies to flood depths, velocities and safe access routes. If no model output exists, the section says so and the report recommends what is needed.

Sequential and exception tests confused

These two tests do different jobs and drafted sections regularly merge them.

The sequential test asks whether there are reasonably available sites at a lower risk of flooding that could accommodate the development. It is about site selection, it is applied by the local planning authority, and it needs a defined search area, a search methodology and a list of sites considered and discounted with reasons. A Sequential Test that simply asserts no alternative sites exist is not a sequential test.

The exception test applies where the sequential test cannot be satisfied and the vulnerability classification and zone combination requires it. It has two parts, both of which must be passed: that the development provides wider sustainability benefits to the community that outweigh the flood risk, and that it will be safe for its lifetime without increasing flood risk elsewhere.

Three errors to look for. A document that runs the exception test without having applied the sequential test first. A document that treats the exception test as a single judgement rather than two limbs. And a document that applies either test to a scheme where policy does not require it, which wastes everybody's time and signals that the classification work was not done.

Surface water and fluvial risk merged

Flood zones describe risk from rivers and the sea. Surface water flooding is a separate source with a separate map, a separate consultee and a separate design response.

A site in Flood Zone 1 can have a serious surface water problem, and a drafted section that concludes the site is at low risk on the strength of the flood zone alone misses it entirely. Sites in a critical drainage area, sites at the bottom of a catchment and sites with a historic ponding record all need the surface water position addressed separately.

The sources to address are fluvial and tidal, surface water and overland flow, groundwater, sewer flooding, and artificial sources such as reservoirs and canals. A competent section covers each briefly and says what evidence it relied on, with the drainage response set out in the Drainage Strategy rather than buried in the flood section.

Illustrative example. A draft describes a site as Flood Zone 1 and at very low risk, concluding that no further assessment is required. The site sits in a critical drainage area identified by the lead local flood authority, with a recorded surface water flow path along its eastern boundary. The planning application is validated and then held pending a drainage strategy, losing six weeks.

A checking order for the section

Work through it in this sequence, which follows the dependencies rather than the document order.

1. Confirm the flood zone from the Flood Map for Planning using the drawn site boundary, and record the date of the check.
2. Classify each use in the scheme against the vulnerability table and state which policy consequence follows.
3. Decide whether the sequential test is required, and if so whether it has been carried out by the authority or needs to be supplied.
4. Decide whether the exception test is engaged, and address both limbs separately if it is.
5. Select the climate change allowance by river basin district, epoch and category, and cite the published source.
6. Check every level, depth and velocity back to a model output or a survey, and flag what is missing.
7. Address surface water, groundwater, sewer and artificial sources separately from the flood zone conclusion.

Seven checks, and most of them take minutes. The reason they are worth doing in this order is that an error at step one invalidates everything after it, and an error at step five is invisible unless you know which band should have been used. Where the draft structure is sound and the evidence is thin, the pattern described in what drafts get right about structure and wrong about evidence is exactly what you are looking at, and version discipline matters when sections are regenerated, as set out in version control when report sections are regenerated.

See PlanScope pricing for how the drafting and review tiers work if you want the checking time built in rather than added on.

This is general guidance, not legal advice. Flood risk assessment for a planning application should be reviewed by a suitably qualified professional, and climate change allowances and flood maps are revised, so check the current published sources. PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval.