Episodic tremor and slip

GPS measurements from Victoria, British Columbia show periodic reversals in crustal deformation in the North American region of the Cascadia Subduction Zone.

In the Cascadia subduction zone, the Juan de Fuca Plate, a relic of the ancient Farallon Plate, is actively subducting eastward underneath the North American Plate. The boundary between the Juan de Fuca and North American plates is generally "locked" due to interplate friction. A GPS marker on the surface of the North American plate above the locked region will trend eastward as it is dragged by the subduction process. Geodetic measurements show periodic reversals in the motion (i.e., westward movement) of the overthrusting North American Plate.[4] During these reversals, the GPS marker will be displaced to the west over a period of days to weeks. Because these events occur over a much longer duration than earthquakes, they are termed "slow slip events".

Slow slip events have been observed to occur in the Cascadia, Japan, and Mexico subduction zones.[5] Unique characteristics of slow slip events include periodicity on timescales of months to years, focus near or down-dip of the locked zone, and along-strike propagation of 5 to 15 km/d.[5] In contrast, a typical earthquake rupture velocity is 70 to 90% of the S-wave velocity, or approximately 3.5 km/s.

Because slow slip events occur in subduction zones, their relationship to megathrust earthquakes is of economic, human, and scientific importance. The seismic hazard posed by ETS events is dependent on their focus. If the slow slip event extends into the seismogenic zone, accumulated stress would be released, decreasing the risk of a catastrophic earthquake.[8][9] However, if the slow slip event occurs down-dip of the seismogenic zone, it may "load" the region with stress.[8][10] The probability of a great earthquake (moment magnitude scale ${\displaystyle M_{w}\geq 8.0}$) occurring has been suggested to be 30 times greater during an ETS event than otherwise,[11] but more recent observations have shown this theory to be simplistic.[12] One factor is that tremor occurs in many segments at different times along the plate boundary; another factor is that rarely have tremor and large earthquakes been observed to correlate in timing .[13]

Slow slip events are frequently linked to non-volcanic seismological "rumbling", or tremor. Tremor is distinguished from earthquakes in several key respects: frequency, duration, and origin. Seismic waves generated by earthquakes are high-frequency and short-lived. These characteristics allow seismologists to determine the hypocentre of an earthquake using first-arrival methods. In contrast, tremor signals are weak and extended in duration.[14] Furthermore, while earthquakes are caused by the rupture of faults, tremor is generally attributed to underground movement of fluids (magmatic or hydrothermal).[15] As well as in subduction zones, tremor has been detected in transform faults such as the San Andreas.[16]

In both the Cascadia and Nankai subduction zones, slow slip events are directly associated with tremor.[4][17] In the Cascadia subduction zone, slip events and seismological tremor signals are spatially and temporally coincident,[18] but this relationship does not extend to the Mexican subduction zone.[19] Furthermore, this association is not an intrinsic characteristic of slow slip events. In the Hikurangi Subduction Zone, New Zealand, episodic slip events are associated with distinct, reverse-faulted microearthquakes.[6]

Two types of tremor have been identified: one associated with geodetic deformation (as described above), and one associated with 5 to 10 second bursts excited by distant earthquakes. The second type of tremor has been detected worldwide; for example, it has been triggered in the San Andreas Fault by the 2002 Denali earthquake and in Taiwan by the 2001 Kunlun earthquake.[20][21]