Multi‐offset ground penetrating radar methods to image buried foundations of a medieval town wall, Great Yarmouth, UK

Abstract

AbstractIn conventional common‐offset (CO) usage, ground‐penetrating radar (GPR) methods may be problematic where a target is hosted within a structurally complex subsurface, or where coherent noise energy obscures the response to that target. This often occurs when imaging archaeology in urban environments, where reflected and diffracted air‐wave energy can be pervasive in the presence of cars, walls and other nearby features. To overcome such issues we show a multi‐offset (MO) GPR approach to urban surveying. Multi‐offset velocity analysis is used as a filter by which ground‐propagating signal energy is separated from airborne arrivals, with the latter suppressed on application of normal moveout corrections and stacking. Pre‐stack migration is applied and yields considerable improvement to target resolution compared with post‐stack routines. The MO‐derived model of GPR velocity can also show the subsurface distribution of archaeological layering. Methods are demonstrated for a 25‐fold MO GPR profile acquired, using unshielded antennas of 200 MHz centre‐frequency, over foundations of the medieval town wall of Great Yarmouth, UK. The buried foundations represent a vertical discontinuity in the ground, which truncates both natural geological and archaeologically prospective layers. In addition to this complex geometry, which itself hinders GPR imaging, coherent air‐wave noise is problematic because there are numerous above‐surface features (e.g. parked cars, above‐surface walls) in the immediate vicinity of the profile. The improvement offered by MO techniques is benchmarked against a conventional CO routine and compared with co‐located borehole records for a comprehensive subsurface interpretation. Target foundations appear ca. 1.5 m beneath the ground surface, and GPR data support artefactual evidence of the construction of a sixteenth century rampart on their interior side. Despite the increased survey effort, we recommend MO methods at any location where CO data are dominated with coherent noise energy, and where complex subsurface geometries degrade the output from standard migration algorithms. Copyright © 2010 John Wiley & Sons, Ltd.