[S-SS07] Environmental Seismology: from deep earth to surface process

[SSS07-14] Seismic Detection of Rock Exfoliation events at Arabia Mountain, Georgia

★Invited Papers

*Xu Si¹、Zhigang Peng¹、Aislin Reynolds¹、Yunyi Qian²、Phuc Mach¹、Chang Ding¹、Karl Lang¹、Martha Cary Eppes³、Sidao Ni⁴ (1.School of Earth and Atmospheric Sciences, Georgia Institute of Technology、2.School of Resources and Safety Engineering, Chongqing University、3.University of North Carolina at Charlotte、4.Innovation Academy for Precision Measurement Science and Technology, Chinese Academy of Sciences)

キーワード：Rock Exfoliation, Seismic Detection, Deep Learning

Rock domes around the world are known for their exfoliation sheets, which represent peeling of their outermost layers. Over the past two years, several rock exfoliation events have been recorded in a former Lithonia Gneiss quarry near Arabia Mountain, Georgia, including a major event occurred around noon time on July 17^th 2023. Field measurements and high-resolution DEM differencing revealed that exfoliation fractures associated with this event extended across an area of approximately 250 m², with vertical displacements of up to 30 cm locally, forming compressive tent-like structures along the fracture perimeter. To monitor additional exfoliation events, a range of sensors were deployed since May 2024 and remain in operation until now. These include time-lapse cameras, air and rock surface/subsurface temperature sensors, and 26 Smartsolo geophones to record ongoing fracture events. Geophones were both buried in shallow soils surrounding the exfoliation fracture as well as anchored directly to the rock face. Additionally, larger events were also recorded by a nearby broadband seismic station (N4.Y52A) and its auxiliary pressure sensor. We applied deep learning methods to burst-like events from the continuous waveform recordings of the nearby geophones and the broadband seismic station. Detected events were subsequently classified using clustering methods into categories such as rock cracking and exfoliation events, mine blasts, and human activities such as traffic noises. Based on confirmed exfoliation events, we identified additional exfoliation events and compared their occurrence with stress and temperature measurements. Our methods offer new insights into the mechanisms of progressive, low-stress rock fracturing, enhancing our understanding of the physical processes driving time-dependent cracking and exfoliation in rock domes.