Geophysical Measurements

One of the key elements of research was making an archaeological-geophysical survey which was intended to locate the surviving archaeological relics and specify the conditions of their deposition.

Measurements in the field were made of an area of c. 3.5 ha, in the central part of the plateau. Individual areas of investigation were delimited by former field boundaries which are marked by low walls of stone (most probably collected from the surface of the fields) in most cases, overgrown by natural hedges. In the eastern and central part of the investigated area there were surviving traces of archaeological trenches with unearthed and conserved remains of town buildings from different periods. The ground surface was uneven, with numerous cut features and spoil heaps, especially, in the vicinity of the old trenches. There was also need to take into account that some of the spoil heaps had been levelled and scattered over a sizable area and are not visible at present on the surface of the terrain. This topography of the ground surface may have affected the results of measurements of the intensity of magnetic field, producing anomalies in distribution of the values measured, which were particularly in evidence when gradient measurements were taken. Another difficulty was posed by the field boundary walls. We cannot rule out that in a part of the investigated area may survive relics of land drainage systems (historical and modern; e.g, we observed field drainage wells and terracotta pipes visible on the surface in the area south of the study area). During the surveying period parts of the investigated area were under tall vegetation which also affected the data obtained because it changed the zero settings of lower magnetometer sensors.

The survey was made using two different instruments – a Geometrics G-858 caesium magnetometer with two sensors, to measure the vertical and horizontal pseudo-gradient of the magnetic field, and a Bartington Grad 601-2  single axis dual sensor fluxgate gradiometer. Both devices were operated in similar conditions, using a uniform grid with samples taken moving from west to east, sensors positioned at a height of 0.3 m. The line spacing was either 0.5 or 1 m (depending on the size of the sought for archaeological features), and the sample interval 0.25 m. In processing the obtained data and during visualisation a 25 x 25 cm raster was interpolated.

Exact location of the archaeological remains responsible for magnetic sampling anomaly (even as deep as 2-4 m below the modern ground level) was specified by measuring the vertical gradient (in case of the fluxgate device) and pseudo-gradient (caesium magnetometer) by setting the sensors in a horizontal and vertical position. We were aware that the penetration range of the fluxgate device is limited in general to a depth which corresponds to the distance between the fluxgate sensors (in this particular study, 1.2 m). The prospecting depth for the caesium magnetometer was much greater but it would supply also data on geological changes of the substrate, which in many cases were similar to anomalies caused by buried archaeological features. Tests were made also using an additional magnetometer as a base station to make corrections, ie, record variation in the strength of the magnetic field not dependent on natural causes – long-term (24 h) and short-term, sudden interferences. This helped us distinguish with high precision only the anomalies caused by architectural remains underground. At the same time the described operations made it necessary to integrate in the survey different types of measuring devices and substantially reduced the efficiency of prospecting increasing at the same time the weight of the portable system. For this reason we decided not to make this type of survey of the entire investigated site, taking into consideration that the brief sampling time (0.1 s for the caesium magnetometer, 0.2 s for the fluxgate gradiometer) practically makes us independent from natural changes of the magnetic field. Especially the caesium magnetometer which measures the values of all vectors of total magnetic field was sufficiently effective, although in the rough terrain we were unable to use this heavy equipment and GPS. The Geometrics G-858 with two sensors was used to measure pseudo-gradient; sensors were configured horizontally, 0.5 m apart, and vertically - 0.3 and 1.3 m above the ground. In both settings the sampling rate at normal walking speed made possible observation at points 10 cm apart. The line spacing was 0.5 m. This made it possible for us to cover, in days’ work, an area of c. 1.5 - 2 ha and repeat the surveys of the same area with a different sensor configuration and using a different sampling grid. We were able to work out the best possible method of running the investigation to guarantee recording of the highest possible number of anomalies caused by buried archaeological remains.

In most cases the presence of the features caused magnetic anomalies recorded at both configurations of sensors – in horizontal and vertical gradient. Consequently, for most areas we selected an option in which we recorded the horizontal pseudo-gradient, as quicker and easier to perform. At the same time we continued to record the variation of the vertical gradient, especially where fluxgate gradiometer readout was inconclusive.

During survey made with the Bartington gradiometer we recorded the value of the vertical component of the vector of total strength of the magnetic field in the interval of  – 115 to + 115 nT/m. All the obtained values, marked on the map in white, clustered in the northern area of the investigated site. The cause of this are dissimilar magnetic properties of the ground surface (higher value of magnetic susceptibility of the subsurface layer) and also from the presence of buried archaeological relics directly underneath the surface. In the southern  part of the investigated area the dynamics of change are much smaller and anomalies similar in character as in the north appear only as distinct closed areas within metres E35-50: N30-70, E220-240; N10-30 of the grid. Additionally visible on the map are two linear anomalies: a wider strip running diagonally S-E, in metres E 180-240; N120-130, and a narrower strip aligned N-S, in metres E180-200; N 0-100. The first zone of anomaly corresponds to the remains of a street which had been uncovered by excavation. In the area of the second anomaly there were no traces are visible on the ground surface to explain its origin.

A different map would be obtained if we limited analysis to an interval of -3 + 3 nT/m discarding interference with the highest dynamics, caused by differences in magnetic susceptibility mentioned earlier noted in the northern part of the study area, but mostly, also by the presence of modern metal objects which were responsible for characteristic and easily interpreted dipolar anomaly.

The view obtained in this case is typical for a multi-layer site containing architectural relics. Architectural remains causing magnetic anomaly may be expected to lie at a different depth and vary in their degree of preservation; moreover, next to fragments of building foundations we may hope to find entire buildings, mostly filled with rubble. Linear anomalies were also present, similar to those detected in the area of street exposed during earlier excavation. Their position at right angles to these relics suggest that the anomalies may be caused by the remains of a street grid. Introduction of colour to the conducted analysis makes it possible to obtain information about the different depth of deposition of the remains causing the anomaly in the distribution of values of the magnetic field. In case of most linear anomalies the vertical component vector was in the range of +2.5 -3 nT/m and was marked in brown. Similar values were recorded in case of anomalies detected in the northern part of the investigated site, where they were also marked in brown. However, it needs stressing that the dynamics of change in that area is much greater and is apparent in particular on a 3D model of distribution of the recorded values. Other linear anomalies occur in a number of much smaller areas, forming more or less geometric shapes – rectangles and squares, with clearly preserved right angles. These may be expected to be relics of buildings. It seems that the lower values of the vertical component of the vector of total strength of the magnetic field recorded in the area of this particular anomaly (up to + 0.5 up to + 2 nT/m) are the result of a greater depth of deposition of the relics responsible for causing the abovementioned anomaly. When interpreting the origins of anomaly we also must take into account the already noted fact that the condition of the remains which causes them may differ. Next to solid foundations of stone, with overlying walls preserved to different height, there may be isolated stones or even only impressions of foundations, with an earth fill. To account for this, on an interpretation map of sources of magnetic anomaly we used lines of different thickness to describe the level of preservation of the relics causing the anomalies. Not everywhere it was possible to distinguish linear structures as sources of anomaly. It was so in case where these occur in a surrounding of layers of rubble, at times, of substantial thickness. In such case, on the interpretation map we marked irregular spots with locally higher recorded values of the strength of the magnetic field, marked brown, yellow and green.

Let us note that the information given above refers to layers found at the maximum depth of penetration, which in case of the fluxgate gradiometer was 1.2-1.5 m below the modern ground level. These values are calculated theoretically and are dependent from many factors, mainly from the magnetic susceptibility of the ground and the buried remains. Moreover, according to the information from the producer and some of the analysis results, in case of exceptionally strong anomalies it is possible to locate features found deeper than the theoretical boundary of penetration.

A more complex distribution of values of strength of the magnetic field was obtained from the survey made with the caesium magnetometer. In the first place, because of its greater depth range. Most samples were taken with sensors positioned at a height of c. 30 cm above the ground level, parallel to each other, 1 m apart. In this way we obtained samples useful for simulating the value of the pseudo-gradient of the horizontal component of the vector of total strength of the magnetic field.

The obtained range of values was smaller:  -25 to + 25 nT/m. Similarly as with data from the fluxgate gradiometer survey we distinguished two zones of dissimilar magnetic susceptibility – a northern and a southern. In the northern zone, the dynamics of change is greater and it seems that this area could have suffered greatly from fire and shows anomaly typical for the heavy burned features. Due to heavy burning most of the anomalies, recorded in particular in the grid squares E-150-210; N270-360, have a dipolar character even though they retain their linear character. We are convinced that their source is the presence of architectural remains. This is evidenced both by the linear character of changes in the distribution of magnetic data and presence of right angles and parallel lines which form closed geometric structures. A closer analysis of the obtained data made for an interval of – 3 to + 3 nT/m, using the same colour key as for Bartington gradiometer samples, not only made it  possible to hypothetically distinguish structures – sources of recorded anomaly, but also furnished information on the character of the terrain. Practically over the entire area there are apparent anomalies characteristic for areas with a heavy deposit of rubble. Most anomalies are observable in exactly the same locations, irrespective of surveying device. The map of results obtained with the caesium magnetometer is on the whole complementary to data from the fluxgate gradiometer ; useful for identifying the course of the main roads and locating the remains of individual buildings. This procedure gives us grounds for planning potential archaeological excavation, and its value   shall be demonstrated using as an example the remains which cause anomaly within grid metres E100-130; N295-320.

In this particular area we detected a series of parallel linear anomalies aligned SE-NW  which from the west and the east connected with two other areas of anomaly at right angles to them. The anomaly expands on the western side where it forms zones typical for areas filled with building detritus. The described magnetic anomaly most probably is caused by remains of foundations of a large rectangular building which is in alignment with the city’s street plan. Its conjectured dimensions are 15 x 20 m. The building had an extension on its NE side and apparently was partitioned into a number of spaces the foundations of the inner walls the cause of the weaker anomalies observable within this rectangular structure. The western part of the building is probably filled with rubble, as suggested by the expansion of a zone of higher values of the strength of the magnetic field and less distinct boundaries of anomaly (this is because the contrast between the surviving remains and their surroundings is smaller).

Another striking anomaly which lends itself to a tentative reconstruction of the layout of the buried features responsible for magnetic anomaly, is near the eastern boundary of the surveyed area – in grid squares E 330-370; N270-300. Here multiple-layered remains caused anomaly were detected during sampling made with both devices. This suggests that in this area the upper level of the archaeological deposit lies directly below the ground level although no traces on the surface indicate the presence of archaeological remains. Anomalies visible on the map and the 3D model of distribution of the measured values of the gradient of the vertical component of the vector of total strength of the magnetic field measured with the Bartington gradiometer A very similar pattern of the magnetic field was obtained for the same location with the caesium magnetometer. Among the narrow lines of anomaly recorded here, caused by the presence of architectural remains, stands out a circular structure c. 18 m in diameter, closed off from the south by a linear anomaly (street boundary ?), with a ca. 8 x 10 m rectangular structure abutting from the north-east. The described pattern of anomalies is quite similar to the remains exposed during excavation east of the surveyed area. We can expect that also in the area of this particular anomaly are buried relics of public buildings, but it is also possible that this linear anomaly designates the boundary of a large robber trench (?).

Similar interpretations can be made for most structures detected in the area covered by our project. Provisional results from the magnetic survey presented here will be analysed as an element of a system of data about the site based on databases in association with maps, aerial photographs, numerical spatial models, and other data from spatial analysis obtained thanks to a broad spectrum of non-destructive methods used at the introductory stage of fieldwork.