Rockfall localization from seismic polarization considering multiple triaxial geophones and frequency bands

• Published in: Journal of mountain science Vol. 17; no. 7; pp. 1541 - 1552
• Main Authors: Feng, Liang, Pazzi, Veronica, Intrieri, Emanuele, Gracchi, Teresa, Gigli, Giovanni, Tucci, Grazia
• Format: Journal Article
• Language: English
• Published: Heidelberg Science Press 01-07-2020; Springer Nature B.V
• Subjects: Attenuation; Azimuth; Earth and Environmental Science; Earth Sciences; Ecology; Environment; Field cameras; Frequencies; Geography; Localization; Mathematical analysis; P-waves; Polarization; Rockfall; Seismic attenuation; Seismic waves; Seismometers; Vibration effects; Vibrations; Early warning; Seismic polarization; Localization; Seismic monitoring; Rockfall
• ISSN: 1672-6316; 1993-0321; 1008-2786
• DOI: 10.1007/s11629-020-6132-1

Boulder/rock mass movements generate ground vibrations that can be recorded by geophone networks. Generally, there are two methods applied to rockfall trajectory reconstruction or rockfall seismic localization. One method uses seismic wave arrival times and is achieved by minimizing the differences in signal arrival times between multiple stations by grid map searching. The other method uses seismic polarization and is achieved by calculating event-source back azimuths from the seismic polarizations of rockfall signals. In this study, we proposed the use of an overdetermined matrix for joint localization based on the polarization method. The overdetermined matrix considers the contributions of all geophones in the network, and at each geophone is assigned a different weight according to the recorded signal qualities and the reliability of the calibrated back azimuths. This method shows a great advantage relative to the case in which only two sensors are employed. Besides, we suggested three marker parameters for proper frequency band selection in back azimuth calculations: energy, rectilinearity, and a special permanent frequency band (SPF). We found that the back azimuths calculated with energy and an SPF are generally close to the real back azimuths measured in the field, while the SPF is limited by seismic attenuation due to a long-distance propagation. The localization results of rockfalls were validated by using field camera videos and in situ calibrations. Three typical cases and 43 artificially released rockfalls are presented in this paper. The proposed method provides an interesting way to locate rockfall events and track rockfall trajectories and avoids the difficulties of obtaining accurate arrival times, as required by the arrival times method.