Ground-Penetrating Radar in Archaeology: Seeing Below the Surface
Ground-penetrating radar (GPR) is one of the most powerful tools in the non-invasive archaeology toolkit. By sending radar pulses into the ground and recording the reflections from buried interfaces, it can detect walls, floors, voids, pits, graves, and other archaeological features without breaking the surface. The technology was developed for engineering applications — detecting pipes, cables, and structural voids — and was adapted for archaeological use from the 1980s onwards. The combination of GPR with advanced data processing and three-dimensional visualisation software has transformed its utility for archaeological survey, particularly for the mapping of complex buried landscapes at high resolution.
How GPR Works
A GPR system consists of a transmitting antenna, a receiving antenna, and a control unit. The transmitting antenna sends short pulses of electromagnetic energy (typically in the frequency range of 50 MHz to 2.5 GHz) into the ground. When the pulses encounter a boundary between materials of different dielectric properties — such as the interface between soil and a stone wall, a floor surface, or the edge of a filled pit — part of the energy is reflected back to the receiving antenna. The time taken for the pulse to travel to the reflector and back (the two-way travel time) indicates the depth of the reflector, given the known velocity of electromagnetic waves in the soil type being surveyed.
A single GPR pass produces a two-dimensional profile (a radargram) showing reflections at different depths. To produce a three-dimensional map of a site, the antenna is moved in parallel transects across the survey area, and the resulting profiles are combined and processed to generate horizontal depth slices — essentially maps of reflective features at successive depths below the surface. Modern GPR carts, pulled by hand or towed by a small vehicle, can survey a hectare in a day; multi-channel systems can survey faster.
What GPR Can and Cannot Detect
GPR is most effective for detecting features that create clear dielectric contrasts: stone walls (higher dielectric constant than surrounding soil), fired clay floors (denser than pit fills), and voids in masonry or below-surface cavities. It is particularly well suited to sites with sandy or loam soils with relatively low clay content, as clay absorbs radar energy and limits penetration depth. In heavy clay, GPR may penetrate only 0.3–0.5 metres; in dry sand, it can reach 5–10 metres.
GPR does not see organic material well in most conditions — waterlogged wood, peat, and plant remains produce limited reflections unless they create a strong contrast with the surrounding matrix. It also does not directly identify the nature of what it detects; a strong reflection could be a wall, a buried concrete service, a tree root mass, or a geological feature. Interpretation of GPR data requires experience, familiarity with the local soil conditions, and, ideally, calibration against known excavated contexts.
Stonehenge: The Hidden Landscapes Project
The Stonehenge Hidden Landscapes Project, led by Vince Gaffney and a team from the University of Bradford and Vienna between 2010 and 2014, applied GPR (alongside magnetometry, earth resistance, and electromagnetic survey) to a 12 square kilometre area around Stonehenge. The results were transformative. The survey identified at least 17 previously unknown monuments, including a line of up to 90 buried stone or timber posts on the south side of Durrington Walls, a large oval monument with a surrounding ditch, and a range of burial mounds and pits not visible on the surface. The data collectively revealed Stonehenge not as an isolated monument but as the focal point of a densely monument-filled ceremonial landscape.
The Durrington posts, when their extent was appreciated, were interpreted as a palisade or processional avenue associated with the Durrington Walls henge — the settlement associated with the monument-builders. The discovery rewrote understanding of the scale and organisation of the monument complex.
Pompeii
GPR survey at Pompeii has produced detailed sub-surface maps of areas not yet excavated — approximately a third of the site remains sealed under volcanic tephra. The survey data shows building outlines, street surfaces, garden layouts, and even the positions of trees from root voids preserved in the ash. The data has guided recent targeted excavations in the Regio V area, where the dig team has used GPR results to open precisely targeted windows into the unexcavated areas rather than stripping large surfaces blindly. The finds — a fast-food bar (thermopolium) with intact painted frescoes and food residues still in the serving containers — received international media attention in 2020.
The Nazca Lines, Peru
GPR survey along the Nazca lines and figures in southern Peru has investigated the sub-surface structure of the line construction and identified buried features along the geoglyph networks. The survey confirmed that the lines were made by removing surface cobbles and stacking them in low walls along the edges, with the pale subsurface exposed as the line surface. It also identified pits and post-holes along some lines, consistent with the interpretation that some geoglyphs marked processional routes to water sources or ritual gathering points.
Current Developments
Developments in antenna design and data processing are increasing the speed, depth, and resolution of GPR survey. Drone-mounted GPR systems are now being tested for survey of sites where vehicle access is restricted. Machine learning algorithms are being applied to GPR radargrams to automate the identification of archaeological features, reducing the expert processing time that has historically been the main bottleneck in large-scale surveys. The combination of GPR with other geophysical methods — particularly magnetometry, which is faster for large-area survey — allows each method to compensate for the other's limitations and produces more complete sub-surface images than either alone.