Premium
High resolution GPR imaging and joint characterization in limestone
Author(s) -
Pipan M.,
Forte E.,
Guangyou F.,
Finetti I.
Publication year - 2003
Publication title -
near surface geophysics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.639
H-Index - 39
eISSN - 1873-0604
pISSN - 1569-4445
DOI - 10.3997/1873-0604.2002006
Subject(s) - geology , ground penetrating radar , bed , azimuth , bedding , offset (computer science) , porosity , mineralogy , radar , wavelet , remote sensing , anisotropy , geotechnical engineering , geometry , optics , telecommunications , physics , mathematics , artificial intelligence , computer science , horticulture , biology , programming language
We focus on the application of Ground Penetrating Radar (GPR) to evaluate limestone characteristics of interest in environmental and engineering studies, and in particular: a) to image joints, bedding planes and cavities; b) to improve accuracy and resolution of the method; c) to evaluate joint/bedding planes characteristics which affect the radar response with particular reference to thickness, sedimentary infilling, water/clay content and spatial frequency. The work is based on experiments carried out close to road cuts and cavities, where the exposed rock face allows calibration and validation of results. The test‐sites are located in NE‐Italy and are part of the Peri‐Adriatic carbonate Platform. The rocks in the area of study date back to the Paleocene and are mainly peritidal regressive limestone sequences consisting of more than 90% carbonates. Joints and bedding planes surfaces, as results from direct inspection at the rock face, are sufficiently smooth to neglect roughness effects. We demonstrate that imaging can be improved by non‐conventional data acquisition paradigms, and in particular by multi‐offset (linear multi‐fold, LMF) methods. Multi‐offset/multi‐azimuth (azimuthal multi‐fold, AMF) techniques were exploited to select optimum grid orientation and offset range for LMF application. Multi‐fold velocity analysis on CMP gathers shows rather constant wave propagation velocities. They slightly vary between different test‐sites (between 10 cm/ns and 12 cm/ns) primarily due to macro/microscopic characteristics of limestone, such as spatial frequency of joints and porosity respectively, which affect water content in the rock mass. Constant velocities allowed application of post‐stack time migration algorithms. Such algorithms attained imaging accuracy below 3% in the reconstruction of the detectable discontinuities at most test sites. A Kirchhoff algorithm proved to be the optimum solution as it effectively handled the steep dips (up to 70°) that characterize most of the examined sets of joints. Enhanced data quality further results from the application of original processing techniques, such as, in particular, Hough Transform based coherent noise/background removal and Wavelet Transform based instantaneous parameters computation and analysis. Maximum penetration depth at the examined test sites ranges between 15 m for 250 MHz (central frequency) bow‐tie shielded antennas and 23 m for 50 MHz unshielded resistively loaded linear dipoles. As for resolution, antennas in the range of 200‐250 MHz apparently provide effective discrimination of joints/bedding planes spaced not less than 40 cm, and seem therefore the adequate choice at the examined test sites. Low frequency antennas, in the range of 50 to 100 MHz (central frequency), provide a maximum 55% increment in penetration depth at the expenses of a substantially diminished resolution (around 20% of that provided by 200‐250 MHz antennas). The comparison of modelling results to multi‐fold data allows discrimination of radar response from joints filled by air and clayey deposits. Such results were validated by geological evidence at the exposed rock face.