Geophysical survey methods and which suits your site

Choosing an archaeological geophysical survey method is a question about your geology and your likely targets, not about which technique is best. Magnetometry covers ground quickly and finds cut features such as ditches, pits and kilns, so it is the default for open farmland on chalk, limestone or gravel. Ground penetrating radar finds masonry, floors and voids and works where magnetometry is defeated. Earth resistance sits between the two and is sensitive to the season. Get the method wrong and you pay for a survey that produces a clean, empty plot of nothing while the archaeology is still there.

The county archaeologist or the local authority's heritage adviser will usually want the method agreed in advance. Go to them with a proposal, not a quotation.

What each technique detects

Magnetometry measures small variations in the earth's magnetic field caused by buried material. Ditches and pits filled with topsoil read as positive anomalies because the fill is more magnetically susceptible than the surrounding subsoil. Hearths, kilns and areas of burning read strongly because firing alters the magnetic properties of clay. Stone walls, by contrast, are often invisible, because limestone and chalk have little magnetic contrast with the soil around them.

Ground penetrating radar sends a pulse into the ground and records reflections from interfaces. It picks up walls, floor surfaces, rubble spreads, culverts and voids, and it produces results in depth slices rather than a single plan, which is why it can distinguish a Roman floor from a modern service trench above it.

Earth resistance passes a current through the ground between probes and measures how readily it flows. Stone walls and rubble resist, so they read high. Ditches holding moisture conduct, so they read low. The technique therefore sees both structures and cut features, but only when there is a moisture contrast to see them with.

Ground conditions that defeat magnetometry

Magnetometry is quick and low cost per hectare, which makes it tempting everywhere. Several site conditions make it close to useless:

• Igneous and some metamorphic geology, where the background magnetic variation swamps any archaeological signal.
• Made ground, demolition rubble and former industrial land, where modern debris produces dense clutter.
• Proximity to ferrous infrastructure. Steel fences, buried pipelines, overhead lines, reinforced concrete and vehicle hardstanding all generate anomalies that mask a wide strip either side.
• Highly magnetic soils over certain sandstones and ironstones.
• Recent ploughing with ferrous debris, or a field that has been used for storage.

A desk assessment of the geology from the national geological mapping, together with a walkover noting fences, poles and surfaces, answers most of this before anyone quotes. Where a field is bounded by steel post-and-wire on three sides, expect a usable survey area noticeably smaller than the red line.

Where radar earns its cost

Radar is slower and more expensive per hectare, and it is the right answer in four situations.

Urban and previously developed sites, where ferrous clutter rules out magnetometry but hard surfaces give radar a clean coupling surface. Sites where the target is masonry rather than cut features, such as a suspected church, villa or industrial structure. Surveys inside standing buildings or beneath floors, where no other technique is practical. And sites where depth information matters, because a layout at 0.4 metres and one at 1.2 metres tell different stories about truncation.

Radar is defeated by clay. High soil conductivity attenuates the signal, so a heavy clay site may return useful data only in the top few hundred millimetres. Waterlogged ground is worse. If the geological mapping shows clay at the surface, ask the contractor for a trial transect before committing to a full survey.

Resistivity and its seasonal limits

Earth resistance depends on a moisture contrast between the feature and its surroundings, and that contrast changes through the year. Late spring and summer generally give the clearest results for both walls and ditches. Waterlogged winter ground flattens the contrast because everything conducts. A severe summer drought can flatten it the other way.

That seasonality is worth programming for. A resistance survey booked for January on a clay site is likely to be repeated.

Resistance is also slower than magnetometry, because probes have to be inserted at every reading. On a large site it is often used as a targeted follow-up over anomalies identified by magnetometry rather than as a primary survey.

Resolution and what it buys

Resolution is set by the sample interval along each traverse and the spacing between traverses. A standard magnetometer survey reads at close intervals along lines spaced a metre apart, which resolves ditches and larger pits reliably but can miss small features such as individual postholes or cremation burials.

High resolution means closer traverse spacing, more data and more cost. The decision should follow the target. If the desk assessment suggests a prehistoric settlement with possible postholes, standard resolution may understate the site. If the question is whether a known enclosure extends into the field, standard resolution answers it.

State the sample interval and traverse spacing in the written scheme of investigation, because a report that omits them cannot be assessed properly by a reviewer.

Reporting and interpretation standards

Two documents govern quality. Historic England's guidance on geophysical survey in archaeological field evaluation sets out method selection, data processing and presentation. The Chartered Institute for Archaeologists publishes a standard and guidance for archaeological geophysical survey that covers project design, personnel competence and reporting.

A report that meets them will include the survey method and instrument, the sample interval and traverse spacing, the processing steps applied, greyscale and trace plots at a stated scale, an interpretation plan distinguishing archaeological, geological and modern anomalies, and a statement of confidence. It will also be deposited with the Historic Environment Record.

The interpretation plan is the part to read carefully. Anomalies labelled as possible or uncertain are doing real work, and a developer scoping a programme needs to know which of them will drive trenching.

Agreeing the method with the adviser

Bring the proposal to the heritage adviser with the desk evidence attached, because the method follows the evidence. The Archaeological Desk-Based Assessment should already have set out the known record, the geology and the historic land use, and the historic mapping work described in the note on using historic maps in an assessment often identifies the ferrous clutter that rules a technique out.

Expect the written scheme of investigation to name the technique, the area, the resolution and the reporting standard, and to say what happens next. Geophysics does not prove absence. A negative survey on clay proves very little, and advisers know it, which is why trial trenching usually follows regardless on sites of real potential. The difference between an evaluation and a watching brief is covered in the comparison of a watching brief and an evaluation.

For programme purposes, treat survey, reporting and adviser response as separate items. The field element may take days. The report and the response to it take longer, and on a major application the results feed into the Heritage Statement that accompanies the submission.

Enter your postcode on PlanScope to see which heritage and archaeology reports a site is likely to need before you commission a survey.

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. For specialist technical reports, qualified professional input may be required.