11:00 AM - 1:00 PM
[SSS07-P10] Evaluation of deformation characteristics around a rough fault through stick-slip sequences
Keywords:stick-slip, laboratory rock friction experiments, faults, precursory phenomenon
One of the physical models of phenomena precursory to earthquakes is the dilatancy-diffusion model, which assumes that dilatancy of rocks around the focal zone. For understanding the precursory phenomenon, it is important to investigate characteristics of rock deformation near faults through stick-slip, which is expected to be a mechanism of repeated seismic activity. In this study, laboratory rock friction experiments were conducted to clarify the deformation characteristics of near faults through stick-slip. In general, specimens with smooth fault surfaces are used in laboratory rock friction experiments. However, a surface topography of a fault is expected to critically affect the rock deformation near the fault. Hence, rock specimens with rough frictional surfaces like natural faults were used for the friction experiment in this study. By using such specimens, it is also expected to evaluate an effect of smoothing processes of fault surface topographies due to stick-slip sequences.
We used cylindrical Aji granite with a diameter of 20 mm and a length of 50 mm as experimental specimens. The specimen was deformed using a triaxial rock deformation apparatus in Toho University under a confining pressure of 75 MPa and an axial displacement rate of 0.03 mm/min, and a rough fault was created between two preformed saw-cut notches (30° angle to the axis of the specimen). Then, the confining pressure was increased to 100 MPa, and the axial displacement was resumed to generate stick-slip (“the stick-slip process”). The experimental conditions were determined with reference to Goebel et al. (2012, Journal of Geophysical Research). Four strain gages (1 mm in length) were attached to the specimen at 6 mm intervals along the axial line across the expected fault plane to measure circumferential strain.
Characteristic stick-slip behavior was observed in two experiments (ME01, ME02): in ME01, six slips with a stress drop of 131-153 MPa (L-slip) and three slips with a stress drop of 12-31 MPa (S-slip) were observed, and in ME02, seven L-slips (a stress drop of 50-109 MPa) and seven S-slips (a stress drop of 1 to 31 MPa) were observed. In both experiments, S-slips were concentrated in the initial stage of the stick-slip process. There was a clear gap between the amounts of stress decrease for L-slips and S-slips. Also, stress decrease and strain change at slips had a positive correlation.
The circumferential strain tended to increase relatively linearly at the stick-slip process (strain is positive for tension). At some slips, increase rates of strain and differential stress were decreased after a certain point just before slips in some cases. At the onset of a slip, the strain was increased suddenly and temporarily, and then transitioned to different value. There were three patterns of transitions: a decrease on both sides of the fault (pattern a), an increase on both sides of the fault (pattern b), and an increase on one side and a decrease on the other side (pattern c). For L-slip, 11 out of 13 times were pattern a, while for S-slip, 5 out of 10 times were pattern b and 4 out of 10 times were pattern a. These results support a model that S-slips represent the local slip to equalize the unevenness of the fault surfaces, while L-slips represents whole plane slip.
From the results, we proposed a model for the deformation characteristics near faults through stick-slip sequences as follows. In the initial stage of inter-slip period, a rock is extended circumferentially due to elastic deformation. Before slip, local disengagements are initiated on the fault plane, which is expected from the observed decreases in the increase rate of stress and strain. Then, slip starts at a certain point on the fault plane, leading to local slip or whole plane slip. Strain increases after slips observed in some cases are possibly because of expansions due to crack creations (or extension) and riding up uneven fault surface topographies in association with slip.
We used cylindrical Aji granite with a diameter of 20 mm and a length of 50 mm as experimental specimens. The specimen was deformed using a triaxial rock deformation apparatus in Toho University under a confining pressure of 75 MPa and an axial displacement rate of 0.03 mm/min, and a rough fault was created between two preformed saw-cut notches (30° angle to the axis of the specimen). Then, the confining pressure was increased to 100 MPa, and the axial displacement was resumed to generate stick-slip (“the stick-slip process”). The experimental conditions were determined with reference to Goebel et al. (2012, Journal of Geophysical Research). Four strain gages (1 mm in length) were attached to the specimen at 6 mm intervals along the axial line across the expected fault plane to measure circumferential strain.
Characteristic stick-slip behavior was observed in two experiments (ME01, ME02): in ME01, six slips with a stress drop of 131-153 MPa (L-slip) and three slips with a stress drop of 12-31 MPa (S-slip) were observed, and in ME02, seven L-slips (a stress drop of 50-109 MPa) and seven S-slips (a stress drop of 1 to 31 MPa) were observed. In both experiments, S-slips were concentrated in the initial stage of the stick-slip process. There was a clear gap between the amounts of stress decrease for L-slips and S-slips. Also, stress decrease and strain change at slips had a positive correlation.
The circumferential strain tended to increase relatively linearly at the stick-slip process (strain is positive for tension). At some slips, increase rates of strain and differential stress were decreased after a certain point just before slips in some cases. At the onset of a slip, the strain was increased suddenly and temporarily, and then transitioned to different value. There were three patterns of transitions: a decrease on both sides of the fault (pattern a), an increase on both sides of the fault (pattern b), and an increase on one side and a decrease on the other side (pattern c). For L-slip, 11 out of 13 times were pattern a, while for S-slip, 5 out of 10 times were pattern b and 4 out of 10 times were pattern a. These results support a model that S-slips represent the local slip to equalize the unevenness of the fault surfaces, while L-slips represents whole plane slip.
From the results, we proposed a model for the deformation characteristics near faults through stick-slip sequences as follows. In the initial stage of inter-slip period, a rock is extended circumferentially due to elastic deformation. Before slip, local disengagements are initiated on the fault plane, which is expected from the observed decreases in the increase rate of stress and strain. Then, slip starts at a certain point on the fault plane, leading to local slip or whole plane slip. Strain increases after slips observed in some cases are possibly because of expansions due to crack creations (or extension) and riding up uneven fault surface topographies in association with slip.