Brief overview

The presence of buried archaeological features can often be detected by magnetometer survey. Small local anomalies in the earth's magnetic field result from the contrasting levels of magnetic susceptibility which exist between infilled 'cut' features/structures and the local substrate or bedrock. This effect is principally due to the varying iron content in the soil and rock forming minerals. Certain features, notably fired clay structures such as kilns, hearths and furnaces acquire their own magnetic identity and display thermoremanent magnetism which is induced by the geomagnetic field in the relatively iron rich clays on cooling from high firing temperatures.

The majority of archaeological magnetometer surveys are undertaken with fluxgate gradiometers which measure the magnetic gradient between two sensors arranged on a vertical separation, normally 0.5 or 1m apart, with the lower sensor being carried approx 25cm above the ground. Readings can be automatically logged at walking pace, allowing large amounts of data to be recorded and stored quite rapidly. Our surveys are usually undertaken within grids at 1m or 0.5m traverse separation logging at 4 readings per metre, totalling between 40,000 and 80,000 readings per hectare. Computer processing of data in the field allows the survey to be monitored and modified as required.

Magnetometers are also used in a broader prospecting role to initially locate areas of 'magnetic activity' and also to help determine the extent of anomalies associated with previously identified features. The presence of more subtle anomalies however can only revealed by detailed coverage. In prospecting sites which extend over many hectares or where only selective detailed magnetometer survey may be required target areas can be often be chosen by reference to a topsoil magnetic susceptibility map combined with the results of other archaeological assessments.

Examples

The following images show the results of gradiometer surveys. The unit of measurement of the vertical magnetic gradient is the nanotesla (nT) = 10^-9 tesla (T) . Most have been carried out with the instrument set to a sensitivity of 0.1 nT. Given a reasonable response, resolution of features of less than 1m in width can be expected. The 1m x 0.25m configuration is the one we most widely use for area evaluations. Under optimal conditions, and using configurations of two or more gradiometers, coverage of several hectares per day can be achieved. A traverse separation of 0.5m may be deployed to locate smaller features or to provide additional clarity especially where survey is confined within narrow corridors (e.g.. proposed pipelines). The obvious merits of close traverse survey (0.5m or less) need to be carefully weighed against consideration of time and cost. For the purposes of assessment knowing the location, geometry and extent of buried archaeological features usually provides sufficient information and more extensive coverage may be considered more important than high resolution.

The plots in this section have been prepared as inverted grey shades, with positive values represented as the dark and negative as the light element. This representation is preferred by many archaeologists as it gives the impression of infilled intrusive features not dissimilar to patterns observed during excavation. The gradiometer data can be imaged in various ways including grey scale, stacked trace, dot density, contour, 'wire frame' and 'false relief' plots; each presentation allowing the data to be visualized with emphasis on a particular aspect. For location purposes a plot which displays an accurate plan of the buried features is important. A further level of information which give a representation of the dynamic range of the recorded signal is also required, grey scale images, as shown here, offer a compromise in displaying both plan and relatively subtle variations in anomaly strength on the same plot. Stacked ( X-Y) trace plots are used to both display the raw data and aid interpretation; other types of presentation are prepared where useful, with colour plots where appropriate.

Below: A Neolithic and later hill top enclosure, Raddon Hill, Devon, UK .Survey area: 390 x 180 m. Grid 30m. Geology: Permian breccia.

The example above shows a detailed gradiometer survey of a 5 hectare hilltop. The survey objective was to provide a context for a series of prehistoric features recorded by excavation during the construction of an access road to a reservoir site (the new road runs through the centre of the area from left to right). Gradiometer survey proved the features to form part of a complex series of enclosure ditches and associated pits. Selective excavation demonstrated that the site originated as a Neolithic causewayed enclosure.

Below: A Roman military installation at Ide, overlooking Exeter, Devon, UK. 

Survey area: 150 x 150 m. Geology: Devonian sandstone.

.

The location and plan of former stone structures can frequently be retrieved by magnetometer survey (two examples are shown above). In many cases the stonework will have been thoroughly robbed for subsequent use leaving only a rubble- or earth-filled trench where once the wall footings stood. If this trench contains material of sufficient magnetic contrast to produce a local magnetic anomaly the original ground plan of the building will be apparent on a detailed gradiometer plot. Normally the anomaly generated by the 'robbing trench' will be positive where topsoil or other material of relatively higher magnetic susceptibility infills the trench. Occasionally infilling (either deliberate, or from alluvial silts) with material of relatively low magnetic susceptibility will cause the trench to produce a negative magnetic anomaly. In the second example above, although the walls of the buildings are no longer extant their removal has not continued down to the base of the footings, and sufficient stone (in this case limestone) of relatively low magnetic susceptibility survives to generate a pattern of local negative anomalies. The thermoremanent properties of fired clay also make brick walls and buildings a suitable target for magnetic survey. Structural elements built of of fired clay products are also responsive, notably the hypocausts of Roman buildings, together with fired clay flooring tiles and concentrations of fallen clay roof tiles. Specific internal features such as hearths can often be located, and where fire may have destroyed timber buildings which had clay plastered hurdlework walls or panels (wattle and daub construction) a good local magnetic response may be anticipated. It is also possible for certain stone to acquire thermoremanence on burning. In many areas much early building utilized soft sedimentary rocks, which tend to be of relatively low magnetic susceptibility. However, in igneous zones more magnetic rock may be used, some stone, notably basalt, is particularly thermoremanent.

Below. Enclosures, trackways, pits, quarry pits and structural features. Left: Romano-British village complex near Dover, Kent, UK. Geology: chalk. Right: multi-period site, Lincolnshire, UK. Geology: gravels.

Composite views of extensive archaeological features can often be retrieved cost effectively by magnetometer survey. Both of the sites above were partially known from aerial photography and fieldwalking. In these examples the results of one day's survey (left) and two days' survey (right) have defined enclosures, trackways, pits and suggested the location of several structures.

Below: Iron Age hut circles and ditch underlying a well preserved medieval ridge and furrow landscape, subsequently sealed by wind blown sand, Harlyn Bay, Cornwall, UK. Geology: wind-blown sand over middle-upper Devonian mudstones. This plot forms part of a seven kilometer long pipeline corridor surveyed at a 0.5 m traverse interval which identified, and allowed for the protection of, numerous prehistoric sites.

Medieval documentary sources record a major inundation of windblown sand at this coastal site. There are no surviving surface indications today of the buried ridge and furrow cultivation (represented graphically on the gradiometer plot by the broad light and dark banding) which excavation has shown to be in an excellent state of preservation. The presence of prehistoric roundhouses was also confirmed. Situated close to a group of nationally important Iron Age burials, which were excavated in the early years of the twentieth century, this survey has provided the first evidence for contemporary prehistoric settlement.

Below: A gradiometer plot showing an enigmatic riverside complex of Romano-British date overlain by medieval ridge and furrow, Trent Valley, Derbyshire, UK. Geology: River gravels with local alluviation. The plot shows silted ditch circuits, together with pits and numerous other intrusive elements. A former natural channel edge shows clearly (projected along the dotted line below).

In some situations natural formations such as palaeochannels of river systems can be seen on magnetometer plots. The example above clearly shows the edge of a former channel (dotted line) against which an enclosure, demonstrated by excavation to be of Romano-British date, has been appended. Biological activity can also alter the magnetic properties of the deposits within man-made features: the fills of certain fenland drains, for example, can become highly visible to magnetometer survey due to processes associated with organic decomposition.