Before anyone digs, someone has to know where to dig. Some sites are obvious, such as standing ruins or mounds. Many are found by chance, when a road crew or a farmer turns up bones or pottery. But most archaeological knowledge of whole landscapes, rather than single sites, comes from survey: the systematic search for traces of the past across the land, often without lifting a spade.
Walking the fields
The simplest method is pedestrian survey, sometimes called fieldwalking. A team spreads out across a plowed field or open ground, walking parallel lines called transects, set at fixed intervals from one another, and records or collects every piece of pottery, flint or brick they see. Plowing brings buried material to the surface each year, so the scatter of finds on top can reveal where settlements once stood below. The pottery types in each cluster suggest when it was occupied.
No team can walk everything. Survey therefore relies on sampling. Probability sampling, choosing areas at random or by a fixed plan, lets a team generalize from a fraction of the region to the whole, much as a pollster generalizes from a sample of voters. A stratified sample might cover river valleys, hillsides and plateaus in proportion to their size, so that no landscape type is ignored simply because it is hard to reach.
Looking from above
Aerial photography changed survey in the twentieth century. After the First World War, the British archaeologist O. G. S. Crawford showed how photographs taken from aircraft could reveal ancient field systems and enclosures invisible on the ground. The key is the crop mark. Crops grow taller and greener over buried ditches, which hold more moisture, and stunted over buried walls, where the soil is thin. From the air, these differences draw the plans of vanished buildings. Droughts make the effect stronger: in the dry summer of 2018, flyers in Britain and Ireland spotted hundreds of previously unknown sites in parched grass.
Laser scanning from aircraft, known as LiDAR, has transformed survey in forests. An aircraft fires hundreds of thousands of laser pulses per second at the ground and times their return. Most bounce off leaves, but some pulses slip through gaps in the canopy, and the returns from the ground are used to build a bare-earth map. Low walls, terraces, roads and reservoirs that would take years to find on foot appear in days. One early success came at Caracol in Belize in 2009, where a few days of flights mapped more than a quarter century of ground survey had managed.
Seeing underground
Geophysical survey reads the ground's physical properties without digging. Each technique responds to different buried features:
Magnetometry detects small changes in the earth's magnetism caused by burning, kilns, hearths and ditches filled with topsoil. It is fast and can cover several hectares a day, which makes it the workhorse of large projects.
Earth resistance survey passes a weak current through the ground: damp ditches carry it easily, while stone walls resist it. Mapping resistance shows buried masonry as bright lines against darker soil.
Ground-penetrating radar sends radar pulses into the ground and times their echoes, which lets it estimate how deep a buried feature lies. Processed into horizontal "time slices," it can show a building's plan at several depths.
None of these methods identifies an object directly. They produce anomalies, patches that differ from their surroundings. A magnetic anomaly might be a kiln, a bonfire from last century or a buried tractor part. That is why every anomaly needs ground-truthing, testing by excavation or coring to find out what it really is. Remote sensing tells archaeologists where to look; it does not tell them what they will find.
Worked example: the buried stones that were not stones
Durrington Walls, a vast circular earthwork about 3 kilometers from Stonehenge, has been surveyed more intensively than almost any site in Britain. In 2015 the Stonehenge Hidden Landscapes Project announced that ground-penetrating radar had detected a row of large buried objects, perhaps as many as 90, beneath the bank of the henge. Newspapers reported a buried "superhenge" of standing stones.
The next year a team excavated two of the anomalies to test the idea. They found no stones. Instead they found huge pits that had once held timber posts, which had been pulled out and the holes packed with chalk rubble before the bank was built over them. The radar had been right that something lay there, but wrong about what it was. Only digging could tell timber from stone.
In 2020 a different team reported a ring of about twenty enormous pits, each roughly 10 meters wide and 5 meters deep, forming a circle about 2 kilometers across around Durrington Walls. They had been spotted with remote sensing and tested with cores that reached the bottom of several. Some scholars suggested they might be natural hollows, and the debate prompted further work. The pattern is typical of good survey: remote methods find the candidates, cores and trenches test them, and claims grow firmer only as different kinds of evidence agree.
