Report on geophysical surveys, October 1998 and May 1999

Introduction

An initial earth resistance survey was conducted in October 1998 in advance of excavation by the University College London Field Archaeology Unit the following summer. This survey covered an area of approximately 0.45ha, both within the scheduled monument (SAM No. WS 73) and into the agricultural land immediately to the west, where a previous earth resistance survey (Davis and Hasler pers comm) had located significant anomalies. In addition to the earth resistance survey a single resistivity pseudo-section was recorded over the planned area for the excavation to provide an indication of the possible depth of surviving wall footings.

During a return visit to the site in May 1999, field testing and calibration of a Ground Penetrating Radar (GPR) system and high sensitivity caesium magnetometer were conducted. Whilst on a site a request was made by the custodian to investigate the covered geometric mosaic that had suffered recent buckling due to an unknown cause, though possibly ground subsidence. This mosaic represents an original Roman floor covering that has never been lifted or re-laid. It was hoped that the GPR survey would confirm the presence or absence of a hypocaust beneath the floor to better inform the ongoing conservation of the mosaic.

The site (SU 98 14) lies on soils of the Bignor association (Soil Survey of England and Wales 1983) developed over Upper Greensand and Gault (Institute of Geological Sciences 1957) and at the time of survey, grid square 6 was under grass. The Scintrex Smartmag was used to survey all of grid square 6. The Pulse Ekko radar was used to survey the southern half of grid square 6 and the geometric mosaic housed in one of the museum buildings.

Method

A total of five 30 x 30m squares were surveyed during the initial earth resistance survey (Figure 1; grid squares 1-5) and the resistivity pseudo-section was conducted within the centre of grid square 4. Both the GPR and Scintrex Smartmag field tests were conducted over a single grid square (Figure 1; grid square 6) previously tested through excavation and known to contain archaeological remains. Figure 1 also indicates the approximate location of the earth resistance survey conducted by Davis and Hasler and the location of the geometric mosaic. A more detailed location plan for the latter survey is provided in Figure 6.

Earth Resistance survey

Earth resistance survey was conducted over grid squares 1-5 with a Geoscan RM15 resistance meter, MPX15 multiplexer and PA5 mobile probe array. The twin-electrode configuration was used with mobile probe spacings of 0.5m and 1.0m at a sample interval of 0.5m x 1.0m and 1.0m x 1.0m respectively. The data is presented in Plan A as greytone images of both the raw and enhanced data following processing with a directional shadow filter to better define linear anomalies. In addition, data from two mobile probe spacings has been combined to enhance both near surface and deeper origin residual anomalies (Plan A, 3 and 6).

The resistivity pseudo-section was collected with a Campus GeoPulse multiplexed resistivity meter from a linear array of 25 electrodes separated by 0.5m. A dipole-dipole array was used to collect readings from all non-reciprocal combinations of the 25 electrodes. The data was subsequently reduced to a subsurface resistivity model using the RES2DINV software package. Elevation measurements were recorded with a Nikon DTM-1 Total Station along the line of the pseudo-section to provide a final topographic correction to the data presented in Figure 7.

Ground Penetrating Radar

A Pulse Ekko PE1000 console was used to collect data for a CMP test (for velocity calculations) from a trial traverse with antennas of centre frequency of 900MHz and 450Mhz. From this data the velocity of subsurface radiowaves at the site was determined to be ~0.081m/nS. Subsequent processing of the GPR data involved the removal of any low frequency transient response (dewow) and the application of a spreading and exponential compensation (SEC) gain function to enhance late arrivals. The data is presented both as a series of individual vertical profiles and as instantaneous amplitude time-slices at selected time windows below the initial time-zero reflection from the ground surface.

Grid Square 6

The southern half of this grid was surveyed with a 450MHz antenna along 30m parallel traverses spaced 0.5m apart. Representative amplitude profiles and time-slices are presented in Figures 2 and 3. Estimations of the velocity of the radar wavefronts in the ground (necessary for calibration of the recorded two-way travel times into absolute depths) are always difficult. In this instance, given the complex media involved, the depth axes of figures 2 and 4 (calculated from an average subsurface velocity of 0.081m/nS) should be regarded with caution as a best estimate only.

Geometric Mosaic

This was surveyed with a 900MHz antenna along parallel traverses separated by 0.1m from within the protective fencing surrounding the mosaic.

The data is presented as a series of representative profiles (Figure 4) and also as horizontal "time-slices" (Figure 5) showing the intensity of radar reflections at increasing depths from the surface.

Caesium Magnetometer.

Grid Square 6

Parallel traverses were made in a west-east direction at 1.0m intervals, with readings taken every 0.1m along each traverse. The results are presented in Plan B as both a greyscale and a traceplot of the data. Both plots have had high magnitude anomalies (caused by near surface iron objects) removed through the application of a 2m radius thresholding median filter (Scollar et al 1990).

Results

Earth Resistance

The most notable anomalies are the three linear high resistance responses found in the agricultural field to the west of the scheduled monument (grid squares 1-3). These anomalies are apparently due to near surface features and would normally, from their high resistance response, be interpreted as possible wall footings (Plan A6). However, previous earth resistance survey conducted within the scheduled monument identified similar high resistance anomalies, that on excavation were revealed to be ditches with a gravely fill that had apparently lost moisture more rapidly than the surrounding soil (Clark 1993, Rudling pers comm).

Within the scheduled monument (grid squares 4 and 5) the survey has located a number of significant responses including an area of high resistance in the north-east corner of grid square 4 and two high resistance linear anomalies; the first running east-west across the centre of grid square 4 and the second running north-south at the very foot of grid square 5. Additional more intense high resistance anomalies are also found to the east of grid square 5. Here, modern paving stones marking the original building foundations have no doubt influenced the geophysical response. It is, however, reasonable to suggest that all of the latter high resistance anomalies may well be related to Roman wall foundations.

The linear high resistance anomaly crossing grid square 4 was investigated further with a resistivity pseudo-section (Figure 7). However, there is little evidence for this anomaly at the appropriate location along the pseudo-section array (~7.0m) beyond a slight increase in the depth of the surface high resistance layers. A more definitive response is recorded to the southern end of the pseudo-section (10.0-12.0m) where two wall-type high resistance anomalies are found extending to an apparent depth of ~0.3m and 0.65m respectively.

Ground Penetrating Radar

Grid Square 6

Few anomalies are evident within this data other than a number of near-surface reflections (possibly due to small stones) and a more continuous reflection at approximately 15nS that may well be related to the interface with the underlying natural greensand. Examination of the time-slices shows areas of disturbance in the very near-surface that correlate with the approximate location of trenches 85W and 85Z. However, it is not clear whether this disturbance is due to surviving Roman masonry or the trenches themselves. In addition, the data from later reflection times contains a considerable degree of noise that has led to strong east-west anomalies concurrent with the sample transects. These are most evident in the amplitude time slices >20nS (Figure 3).

Geometric Mosaic

Air wave reflections from the surrounding fencing are evident in the data (e.g. Figure 4, Line 5) as shallow diagonal lines crossing the profile. However, analysis of the time-slices produced from these profiles (Figure 5) shows an interesting anomaly (A) apparently representing a very near surface void which is also evident as an abrupt anomaly through the profiles of Lines 1 and 3 (approximately 1-2m on the horizontal scale). It is possible that this anomaly represents reflections from an in situ hypocaust system at the eastern end of the mosaic. Unfortunately, data in the vicinity of the damaged area is too limited to suggest the presence or absence of a similar feature at the western end of the mosaic.

Caesium Magnetometer.

Grid Square 6

There are no distinct archaeological anomalies from this data set. However, there is an area of disturbance in the south eastern part of the grid square that may relate to the previous excavations. Other anomalies are attributable to modern interference such as wire fencing.

Conclusion

The initial earth resistance survey conducted in October 1998 has successfully identified a number of anomalies both within the scheduled area and in the agricultural field immediately to the west. Whilst the high resistance response of the anomalies within the agricultural field would be expected from buried wall footings, their considerable extent (>30m) and results of previous earth resistance surveys suggests these may well arise from ditch-type features. Furthermore, there is no evidence for an enclosing wall to the villa complex within the area surveyed. Within the scheduled area a number of high resistance responses were recorded due to either buried Roman wall footings or the influence of more recent paving stones laid on the ground surface. Results from the resistivity pseudo-section suggest that the maximum extent of these wall foundations is approximately 0.65m below the current ground surface.

The second visit to the site in May 1999 provided an opportunity to field test newly acquired GPR equipment and a high sensitivity caesium magnetometer. Both trial surveys failed to detect notable archaeological anomalies, but did locate areas of disturbance associated with previous excavations of the site. In addition, the GPR survey of the geometric mosaic revealed a potential void that may possibly indicate the presence of a hypocaust to the east of this feature.

Bibliography

British Geological Survey, 1957, Chichester, England and Wales, Sheet 317, Drift Geology, 1:50,000.

Scollar, I. Tabbagh, A. Hesse, A. and Herzog, I. (eds.), 1990, Archaeological Prospecting and Remote Sensing. Cambridge.

Soil Survey of England and Wales, 1983, Soils of England and Wales, Sheet 6, South East England.

List of enclosed figures

Figure 1 Location plan of survey grid squares over excavation plan (1:1250).

Figure 2 Radar profiles from the survey of grid square 6.

Figure 3 Amplitude GPR time slices from grid square 6.

Figure 4 Radar profiles from the survey the geometric mosaic.

Figure 5 Amplitude GPR time slices from the geometric mosaic.

Figure 6 Sketch plan of the survey area for the geometric mosaic.

Figure 7 Subsurface resistance model calculated from the resistivity pseudo-section.

Plan A Greyscales of the earth resistance data from the October 1998 survey (1:1000).

Plan B Linear greyscale and traceplot of the caesium magnetometer data (1:300).