ANALYSIS OF SEISMIC LANDSTREAMER DATA FOR EMBANKMENT INTEGRITY ASSESSMENT

Floods, seepages, and animal activities are known to be the main causes of internal damage and erosion of embankments, all these factors lead to partial or complete collapse. Embankment integrity losses may progress without any superficial or easy detectable evidence, in consequence visual surveillance or localized techniques are not effective. Geophysical techniques for characterization of embankments are constantly improving. In literature, there is evidence of numerous advances on the application of electrical and electromagnetic surveys. Refraction and surface wave methods are less considered because they can be more time consuming. They usually work best to assess vertical profiles of a limited section of the subsurface, without considering lateral variations of waves travelling through the media, which may lead to erroneous characterization. Nevertheless, in the last two decades, several studies have been published referring to the application of other surface waves methods to assess lateral characterization by detection and location of inhomogeneities in the subsoil. In this work, four surface wave processing techniques for embankments integrity assessment applied to landstreamer data are tested: energy, energy decay exponent, attenuation coefficient and autospectrum computation. To achieve this objective, classical seismic refraction analyses and the proposed surface wave methodologies were applied and compared. Some limitations related to landstreamer data acquisition were considered during the processing steps: the energy of the source was not constant for all the common shot gathers (CSGs) and noise was expected to be high. To overcome these limitations, the traces of each CSG were normalized by the maximum amplitude of the trace closest to the source, the stacked results were normalized for the data coverage in order to further balance each shot contribution and the impact of noise in the quality of results was evaluated. Data interpretation was performed using an empty pipeline within the embankment as a detection target, disregarding other geological information out of the scope of this work. Results showed that detection of the pipeline in the velocity model obtained from refraction tomography was precariously accurate and influenced by another anomaly located above the target. On the contrary, the proposed surface-wave methodology worked effectively, detecting the pipeline within the embankment. Energy, attenuation coefficient and autospectrum stacked plots depicted high energy concentrations, centered on the pipe, likely caused by the propagation from a high to a low-velocity material. Energy decay exponent, instead, appeared to be more sensitive to additional material contrasts within the embankment such as the loose materials, the concrete structure surrounding the pipe, and the empty pipeline. Trace normalization improved the interpretation and noise deriving from the acquisition setup had probably a significant impact on the negative offset data quality. Horizontal gradient computation allowed to estimate the location of the pipeline between 28 and 35 m along the reference investigated line. The combination of these four techniques applied to landstreamer data represents an efficient and rapid tool for monitoring embankments. The effectiveness of these tools was verified for the first time in this work, and it is presented as a novel and fast approach for the assessment of embankment integrity.