Instability analysis of the Villa Arianna site at Castellammare di Stabia (Naples)

• Published in: Near surface geophysics (Online) 2012-02, Vol.10 (1), p.89-100
• Main Author: Orlando, Luciana
• Format: Article
• Language: English
• ISSN: 1569-4445; 1873-0604
• DOI: 10.3997/1873-0604.2011049

ABSTRACT The detection of fractures using ground‐penetrating radar is usually obtained by analysing the attributes of reflected and scattered energy. In this paper, I test advanced processing of the GPR direct ground wave in order to successfully detect and map near‐vertical fractures that generally affect structures located in unstable areas. I show that the advanced processing applied to the GPR time window of a direct ground wave (embedding the early reflection arrivals) overcomes the problem that often occurs when detecting near‐vertical distensive fractures on the basis of reflected and scattered energy. For the validation of the processing, I use a synthetic model simulating an oblique fracture in a homogeneous and stratified host material. The theoretical radargrams show that an oblique vertical fracture induces variations in the ground direct wave, in terms of both two‐way traveltime and shape. Taking into account such evidences, I relate the fracture to the attributes of the delay time of the ground direct wave and the spectral analysis of a time window embedding the direct waves. I test the advanced processing on real data acquired in the Villa Arianna archaeological complex (Naples) by analysing three kinds of floors composed of heterogeneous material, concrete and ancient mosaic. The last two floor types are affected by extensive cracks, which are visible on the floor. The analysis of attributes of actual data indicates that both attributes are effective for the detection of fractures in an homogeneous near‐surface layer and that the spectral analysis is more effective than the delay time of the ground direct wave for an heterogeneous layer. The selected GPR attributes are quickly and easily applicable in the detection and monitoring of near‐vertical distensive fractures because they can be calculated using GPR data acquired with a system deployed in reflection and continuous mode.