Equipment and Theory

Ground Penetrating Radar

The data for this investigation was acquired using a GSSI SIR3000 GPR data collection system running RADAN control software with a 400MHz centre frequency ground coupled antenna housed in a survey cart.  Antennas of high frequency provide high resolution data but only penetrate to shallow depths, whilst low frequency antennas provide deeper penetration with decreased resolution. The depth of penetration achievable with an antenna of a particular frequency is also dependant on the local subsurface conditions.  The 400MHz antenna was used after considering the expected subsurface conditions and the type and depth of the targets to be resolved.  The effective depth achieved using the selected antenna was approximately 1.6m.

 

Data Processing

The collected GPR profile data was processed and analysed with ReflexW developed by Sandmeier Software.  Data processing included: static correction to set the surface reflection interface to zero time, band pass filtering to improve the signal to noise ratio, and 2D background removal filtering to eliminate temporally consistent flat noise bands.

 

The 2D processed profiles or radargrams were viewed and analysed sequentially to determine the location and depth of structures and material changes in the ground.  Any subsurface changes and targets of interest were plotted onto scaled drawings of the survey site.

 

 

Magnetic Gradiometry

Magnetometer Gradiometry is a passive geophysical technique based on the detection of contrasts in the magnetic properties of different materials. It is among the most effective geophysical techniques used for archaeological investigations since many targets such as fired bricks and ferrous objects have distinctive magnetic properties resulting in a detectable magnetic response.

 

The magnetic response from naturally occurring rocks is due to the presence of the most common magnetic mineral, magnetite or its related minerals. All rocks contain some magnetite, from very small fractions of a per cent to several percent. Of archaeological significance, magnetic responses are observed in fired bricks which obtain a magnetic field during the cooling process. However the strongest magnetic responses are observed from man-made iron objects which exhibit a high degree of ferromagnetism.

 

Magnetic anomalies are composed of a positive and negative part known as a magnetic dipole (shown in figure 5). Its shape is related to the structural geometry of the object as well as its direction of magnetization, and the regional magnetic field. The characteristics of an anomaly also depend on the distance between the object making the anomaly and the measuring sensor, with increased distance resulting in broader, lower amplitude anomalies.

Magnetic equipotential lines created by a magnetized body

The data for this investigation was collected using a Geometrics G-858 Magnetic Gradiometer. The system is composed of two caesium vapour sensors orientated vertically (see figure 7) which measure small changes in the local magnetic field. The bottom sensor being closer to the subsurface is typically more affected by near surface anomalies, whilst the top sensor gives the broader magnetic response of the subsurface. The magnetic vertical gradient can be calculated from the two sensors giving a more precise representation of the magnetic response in the near surface, and also negating diurnal effects.

 

Positioning of the magnetic data was acquired with a Trimble Ag114 DGPS System and data logger. Data from this unit was streamed into the G-858 providing real time differential GPS positioning of the collected magnetic data. Magnetic base station data was collected using a Geometrics G-856 Proton Magnetometer. This was used to account for diurnal variations in the earth’s magnetic field.

 

The magnetic data was processed and analysed with MagMap 2000 developed by Geometrics Inc., and Surfer developed by Golden Software Inc.

 

The position data streamed from the DGPS to the magnetometer was firstly plotted and checked. Since there was a constant positional offset between the DGPS beacon and the magnetometer sensors, an offset algorithm was applied to correct the GPS readings to the actual reading positions.

 

A range de-spike filter was applied to the magnetic data from the top and bottom sensors. Values exceeding normal variations experienced in near surface investigations were removed. Using the de-spiked data from the top and bottom sensors the vertical gradient at each measurement point was calculated. The vertical gradient results enable the true shape of the magnetic anomalies created by shallow features to be determined. Furthermore vertical gradient measurements suppress local geological responses enhancing near surface anomalies, as well as removing diurnal changes in the total magnetic field.

 

The measurements from the magnetic base station were not used during data processing due to the difficulty in locating a magnetic quiet zone during data collection. This was not an issue since data collection extended over a relatively short time period and hence no significant time-dependent variations would have occurred.

The top sensor, bottom sensor and vertical gradient data sets were interpolated using a kriging algorithm. Kriging is a geo-statistical gridding method that produces visually appealing maps from irregularly spaced data sets. After gridding the magnetic data is transformed to an X, Y, Z data set on a constantly spaced grid, X and Y being the GPS position (Easting and Northing position using UTM WGS-84 coordinates) and Z being the magnetic reading.

 

The gridded data for the three magnetic data sets were plotted as a raster image map, which represents the Z values as a continuous colour spectrum. The colour spectrum was centred on the regional magnetic field intensity and the range was set according to surrounding variations in magnetic field intensity.