Report on geophysical survey, March 1999

Introduction

A geophysical survey of approximately 2ha was conducted over the location of a Roman building at Church Farm, Ashington, West Sussex (TQ 12 15). Buried masonry was originally discovered on this site in 1947 when limited excavation took place following surface finds of Roman building material, pottery and vitrified flints (examples of which can still be seen today). The site was scheduled (SAM No. WS410), but no further action was taken to define the extent of the structural remains. A geophysical survey was requested, through the Monument Protection Programme, in order to clarify the nature of the site and its extent.

The aim of the survey was therefore to investigate the geophysical response of the buried structural materials and examine the extent of the site in relation to the current SAM boundary.

The site lies on the crest of a stream valley slope that rises to the north. Typical stagnogleyic soils of the Wickham 1 association (Soil Association of England and Wales 1983) have developed over Hythe Beds and bands of Atherfield and Weald Clay (Institute of Geological Sciences 1984). At the time of the survey the eastern field and scheduled area was uncultivated (set-aside), with the remainder of land under grass. The ground to the north and west of the scheduled area was planted with a rape crop preventing the extension of the survey in these directions.

Method

Magnetometer survey

An initial magnetometer survey was carried out over the scheduled area and available land directly adjacent to it. The aim was to rapidly detect the focus of archaeological activity at the site, which, from the excavation findings, was presumed to include ceramic building material and other fired artefacts. The survey was conducted over all the numbered grid squares in Figure 1 using the standard method outlined in note 2 of Annex 1. The results are presented at a scale of 1:1000 in Plan A. A.1 shows a stacked trace plot of the data, and A.2 a linear greyscale. The measured values have been corrected to zero the median of each instrument traverse to remove heading errors. The high magnitude anomalies caused by near surface iron objects have been removed using a 2m radius thresholding median filter (Scollar et al 1990) to improve the visual appearance of the plots. Grid squares 9-12 were further processed using a fourier domain band-reject filter to remove periodic anomalies in the data produced by the operator's stride length.

Magnetic susceptibility

Topsoil magnetic susceptibility (MS) measurements were also made. Although ploughing may have affected the preservation of the building foundations it should also have brought magnetically enhanced material into the topsoil (Cole et al 1993; 144). Magnetic susceptibility readings were taken on a 10m grid in numbered squares 5-6, 9-14 and 17-22 (Figure 1) using a Bartington MS2-D search loop and MS2 susceptibility meter. At each sample interval the sensor was zeroed in the air before at least four readings were taken to eliminate spurious signals derived from ferrous objects or poor sensor-ground contact. An average of the readings was noted and later used to create a plot for comparison with the magnetic and resistivity data (see Plan C). The MS results are presented as a linear greyscale plot superimposed over the base OS map at a scale of 1:2500 in Figure 2.

Resistivity survey.

Resistivity survey (see Annex 1, note 1) was applied over the areas of most intense activity identified by the magnetometer survey in order to further define the remains of the Roman buildings. The survey was conducted over grid squares 6, 11-13 and 19-21 (Figure 1) using a twin probe configuration with a Geoscan RM15 resistivity meter, an MPX 15 multiplexer and an adjustable PA5 electrode frame. Data was collected using both 0.5m and 1.0m mobile probe spacings over each 30m grid square. Grid square 20 was initially surveyed using traverses spaced at 0.5m, but time pressure resulted in the remaining six grid squares being surveyed using a 1.0m traverse interval. The results are presented at a scale of 1:1000 in Plan B. Both traceplots and linear greyscales of the data from each mobile probe spacing are presented in plots B.1-4. B.5 shows the greyscale of data processed to enhance near-surface features (Clark 1990, 156).

A graphical summary of all significant anomalies discussed in the following text is provided in Plan C. Numerals in square brackets (e.g. [1]) refer to anomalies marked on this plan.

Results

Modern interference

Due to the absence of ferrous field boundaries, little modern interference is evident beyond the areas of magnetic disturbance [1] caused by discarded corrugated iron and [2], possibly a drain to the adjacent stream. Broad anomalies traversing the site on a north-east to south-west alignment are visible in the magnetic data at [3] and on the resistivity data at [4]. With no obvious archaeological explanation these may instead reflect the presence of deeper geological variation (eg between bands of clay and greensand).

Several anomalies can be related to the differing agricultural use of the fields. The magnetometer data shows the edge of the set-aside scheduled area and crop at [5] and between the cropped and grassed area at [6]. The resistivity data shows the greater resistance [7] of the soil planted with a young crop.

Significant anomalies

Magnetometry

Three areas of positive magnetic readings [8], [9] and [10] could possibly represent the position of buildings that contained fired clay objects, such as roof or floor tiles or a hypocaust structure. Between the northern edges of [8] and [10] is a possible thermoremanent anomaly [11] perhaps a furnace for heating the villa or kiln. Adjacent to this is one of several pit-type anomalies [12], perhaps related to rubbish disposal in this area (where numerous oyster shells have been found). A small number of other isolated anomalies [13-16] may also be pits.

Further positive magnetic anomalies are the alignments at [17], [18] and [19] in the western field and [20] in the eastern one, representing ditches. [20] coincides with an area of very slightly increased magnetic noise [21] in which it may be possible to discern three broad sides of a rectilinear configuration. The latter is on a similar alignment to, but on a larger scale than, the more intense responses of [8] - [10]. These patterns [21] may represent a second complex of Roman buildings but a lack of scattered fired clay responses (and a low MS - see below) suggests that this was an area of less intense activity. The north-western edge of [21] appears to be bounded by negative magnetic linear anomalies [22] and [23]. To the north of these are two areas of positive magnetic anomalies [24], perhaps containing pits.

Various areas of negative magnetism [25] - [27] may not be of archaeological significance.

Magnetic susceptibility

The majority of the area surveyed with the field loop produced very low magnetic susceptibility values (8-10 x10-5 SI), with little variation from background levels. However, there is greater enhancement (20 x10-5 SI) over the central area [28], which includes anomalies [8] and [9]. This reinforces the interpretation of the latter as occupation features. No such effect was observed at [10], however, and the overall response in the eastern field was uniformly low.

Resistivity

The high susceptibility anomaly [28] also coincides with an area of high resistance values [29]. Within the latter are slight indications of patterning [40] identifiable from the greyscales in Plan B, very tentatively suggestive of structural remains. Further areas of high resistance are visible at [30] and [31], which roughly correspond to the positive magnetic anomalies [9] and [10]. The high resistance near-surface anomaly [32] also relates to the magnetic response [15] although the adjacent parallel anomalies of [33] and [34] are only observed in the resistance data. The final anomalies that appear to coincide with the magnetic results are [35] and [36] (cf [20] and [21]). These, along with [37] appear to be more deeply buried than many of the high resistance anomalies as they are not evident on Plan B.5.

Two sets of parallel low resistance linear responses [38] and [39] appear to cut across the high resistance anomalies [29] and [30] and are presumably more modern in date. The near-surface anomalies at [39] are most likely to be the tracks or ruts created by a tractor turning in this field.

Conclusions

Despite the known presence of structural remains and archaeological deposits, the geophysical response over this site has been disappointing.

Magnetometer survey has identified a number of areas of anomalous activity focusing on the known location of building remains. Surface finds in the western field in this area included a quantity of red tile and pottery, as well as several oyster shells indicating a high status Roman settlement such as a villa. The results from resistivity and magnetic susceptibility also demonstrate activity in this area (grid squares 12 and 20), although no actual structural remains can be outlined. This may indicate the effects of plough damage or, in the case of resistivity, the lack of significant electrical contrast with the natural soil.

Further ill-defined areas of reduced magnetic activity could be present to the north-east of the main area and could indicate the presence of further remains. However, such an interpretation must be very tentative as the topsoil MS here is low and the magnetic response is very weak.

Several linear anomalies have been detected and could be components of field systems associated with the Roman occupation on the site.

The focus of the archaeological activity seems therefore to be at the very eastern edge of the scheduled area, extending partly into the neighbouring unscheduled field. There are few significant responses in the protected western end of the site, and perhaps some reconsideration of the scheduled area here would be appropriate. The geophysical survey results alone, however, do not provide a sufficiently definitive delimitation of the site and some intrusive evaluation may yet be necessary to do so.

Acknowledgments

The author wishes to express the thanks of English Heritage to Mr R. Duke of Church Farm for allowing access to the site.

References

British Geological Survey, 1984, Brighton and Worthing, England and Wales Sheet 318/333, Solid and Drift Geology, 1:50,000.

Clark, A. 1996, Seeing Beneath the Soil; Prospecting Methods in Archaeology. Batsford.

Cole, M. A. Linford, N. T. Payne, A. W. Linford, P. K. 1995, Soil Magnetic Susceptibility Measurements and Their Application to Archaeological Site Investigation. Science and Site: Evaluation and Conservation, The Archaeological Sciences Conference 1993. pp 144-162.

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, 1:25000.

List of enclosed figures and plans:

Figure 1 Location plan of survey grid squares over base OS map (1:2500).

Figure 2 Greyscale plot of raw magnetic susceptibility data superimposed upon OS base map (1:2500).

Plan A Traceplot and greyscale of raw magnetometer data (1:1000).

Plan B Traceplots and greyscales of raw resistivity data for both 0.5m and 1.0m mobile probe spacings and a greyscale of data enhanced to accentuate near surface features.

Plan C Summary of significant geophysical anomalies.