Establishing site levels from survey and elevation data

Get the site levels survey planning drawings right and half the awkward questions disappear. Ridge heights are measured from existing ground, not from your finished floor. Finished floor levels in a flood zone are set against a modelled water level to ordnance datum. Drainage falls, retaining walls and the relationship to a neighbour's garden all follow from the same set of numbers.

Published elevation data will get you to a feasibility answer in an afternoon. It will not get you to a submission on any site where a level change matters.

Why levels drive so many planning judgements

Permitted development limits are written against ground level. The General Permitted Development Order measures the height of an extension from the highest part of the surface of the ground next to the building, which on a sloping plot can differ from the height a builder measures by half a metre or more.

Amenity assessments work the same way. A 45-degree test applied from a neighbour's window, an overshadowing study, a judgement about overlooking from a raised terrace: each depends on where the ground actually sits, not on where the drawing assumes it sits.

And in flood risk work, levels are the whole argument. A finished floor level is set with freeboard above a modelled flood level, and both are expressed to ordnance datum. Get the datum wrong and the assessment is meaningless.

Published elevation data and its resolution

The Environment Agency's national LIDAR programme is the workhorse. Its composite datasets cover most of England and are published as open data through the Defra environmental data service, at resolutions down to a metre and finer in some areas.

Two products matter, and confusing them is the commonest error in a feasibility study. The digital terrain model represents bare earth with vegetation and buildings removed. The digital surface model represents the top of whatever is there, including tree canopy, roofs and parked cars. For ground levels you want the terrain model.

Accuracy is good for a free dataset but it is not survey grade. Vertical accuracy is typically within a small number of centimetres under open conditions, and worse under dense canopy, on steep slopes and in urban canyons where the algorithm has had to interpolate. Treat it as a reliable picture of the shape of the ground and an approximate statement of its height.

The national terrain products derived from mapping contours are coarser again, useful for catchment-scale work and of little use on a single plot.

When a topographical survey is unavoidable

Commission a topographical survey where any of the following applies.

1. The ground falls by more than about half a metre across the footprint, so permitted development limits or ridge height calculations turn on the measurement.
2. The scheme is in Flood Zone 2 or 3, or on a site of one hectare or more, because finished floor levels and threshold levels have to be surveyed rather than estimated.
3. Drainage is being designed, since invert levels, falls and the connection point cannot be interpolated from open data.
4. There is a retaining structure, a basement or a change in level at a boundary.
5. The relationship with a neighbouring window, garden or party wall is contested or likely to be.

A survey to the usual specification gives spot heights on a grid, levels along boundaries and at the base of walls, tree positions with stem diameters, manhole covers with invert levels, and services where they are visible. Ask for it in CAD as well as PDF, and ask for it tied to ordnance datum rather than to an arbitrary site datum.

Ordnance datum and local benchmarks

Heights in Great Britain are referenced to Ordnance Datum Newlyn, expressed as metres above ordnance datum. Historic benchmarks cut into walls and bridges are still visible everywhere, but the network is no longer maintained, and values from old benchmarks should be treated as indicative rather than authoritative.

Modern surveys establish levels using satellite positioning tied to the national correction network, then set a temporary benchmark on site: a nail in a manhole cover, a mark on a boundary wall, something that will still be there when the contractor arrives.

State the datum on every drawing. A survey on an arbitrary datum where the site datum was set at 100.000 is perfectly usable, but if the flood risk assessment then quotes a modelled level in metres above ordnance datum and nobody converts, the two documents describe different worlds. That mismatch reaches decision notices more often than it should.

Existing and proposed levels on the drawings

What a case officer needs to see is straightforward, and most validation queries about levels come from its absence.

On the block plan: existing spot heights across the site and at the boundaries, and proposed finished floor levels for each building.

On the sections: the existing ground line drawn through the proposal, the proposed ground line, and the difference between them shaded or annotated. If you are cutting into a slope, show the cut.

On the elevations: existing ground level as a dashed line behind the proposed elevation, with ridge and eaves heights dimensioned from it. Where the ground falls along an elevation, dimension the height at both ends.

Retaining structures, steps, ramps and level access to the entrance should all appear, because an officer reviewing accessibility will look for them.

Levels in flood risk and drainage work

Finished floor levels in a flood risk assessment are set above the modelled design flood level with freeboard, and the figure is agreed with the Environment Agency or the lead local flood authority rather than chosen. Check the current published position, because climate change allowances are revised.

Threshold levels matter as much as floor levels: a raised floor with a doorway at the old level does not work. Safe access and egress is assessed on ground levels along the route out of the site, which is why a topographical survey usually needs to extend beyond the red line along the access.

For drainage, the levels fix what is possible. A soakaway needs its base above the seasonally high groundwater level with a margin, a gravity connection needs fall to the point of discharge, and an attenuation tank needs cover depth. The drainage strategy has to reconcile all three, and it cannot do that on interpolated data. Where the site sits in a mapped flood zone, the flood risk assessment sets the levels the drainage then has to work within.

Common errors that reach the decision notice

The surface model used as a terrain model, so ground levels under trees or against a building are reported a metre or two too high.

Ridge height measured from the finished platform rather than from the highest part of the adjacent existing ground, which turns a compliant extension into an unlawful one.

Levels assumed across a boundary. Your neighbour's garden may be 600 millimetres higher than yours, and an assumption in either direction affects the overlooking and overshadowing arguments.

Two datums in one application, usually because the topographical survey and the flood modelling were done by different people at different times.

Contours scaled off a small-scale plan and presented as spot heights, which gives false precision to numbers that were never that accurate.

An illustrative example: an architect working from open terrain data designed a single-storey rear extension to a house on a plot falling 900 millimetres from front to back. The data suggested the rear ground level was 200 millimetres higher than it turned out to be, and the eaves height as built would have exceeded the permitted development limit measured from the true adjacent ground. A half-day topographical survey before the design was fixed would have caught it. This example is illustrative rather than a real case.

Run the desk check first, by all means. Checking the wider constraints at the same time is efficient, and using MAGIC to check designations around a site covers that sweep, while checking whether a site is on a brownfield land register picks up allocations. Then get the survey before the drawings harden.

This is general guidance, not legal advice. Check the position with your local planning authority or a qualified professional before relying on it.

PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval.