The lithospheric plates (also known as tectonic plates) that make up the Earth’s crust are hundreds to thousands of kilometers across and tens of  kilometers thick. One can imagine that  something as large as two such plates sliding past one another, let alone one  being pushed ("subducted") under  another, could create a terrific din.  Indeed, such a "noise" is what we typically feel and record as earthquakes, at least when this motion occurs quickly. What has become one of the new and  interesting areas of research is that  sometimes these adjustments happen in ways that do not generate earthquake  signals, but still can be measured on human time scales. One such event type, called slow slip events or silent earthquakes, is recorded on regional  global positioning system (GPS) networks rather than seismic networks  (Dragert et al., 2002). In the case of the Cascadia subduction zone in the Pacific Northwest (where the Juan de Fuca plate is being subducted under the  North American plate) these slow slip events seem to have occurred at regular intervals of 14 ± 2 months for the last 7 years (Miller et al., 2002).

Even more interesting to me and newer still are unusual tremor-like seismic signals that are occurring at depth along subduction zones. These non-volcanic, deep (30 km), low-frequency tremor signals were first identified by Dr. Obara in Southwest Japan along the Nankai subduction zone, presumably near the Moho discontinuity (Science, 2002). What is unique about these signals is that until their discovery in Japan, tremors had only been identified in volcanic areas. Now they have been identified in at least two non-volcanic settings at unusual depths and are likely associated with the processes of subduction.

Beginning in February of 2003, we started observing strange, long-lasting, weak signals on seismic stations of the Pacific Northwest Seismograph Network (PNSN) in the north Puget Sound area (Fig. 1). The tremor continued in bursts for roughly two weeks, and occasional shorter-lived events have happened in April and May. These tremor-like signals have a frequency content around 1-6 Hz and are visible without filtering just above the noise on some seismic stations, but they can be seen on more stations if the signals are bandpass filtered (Fig 2). Scientists at the Pacific Geoscience Center (PGC) in British Columbia report observing similar signals on their stations and, in fact, have observed these sorts of tremors in previous years, but they were dismissed as local noise until the Japanese discovery. Thus far, the events appear to be the same type of tremor as observed in Japan (Science, 2002).

These signals have very different  characteristics than typical tectonic earthquakes in that they do not have distinct P- and S-wave phases that make them difficult to identify and locate with traditional seismological methods. They most resemble volcanic tremor: a low frequency seismic (1-5 Hz) signal that has an emergent onset, lasts from minutes to hours, and is typically associated with the movements of fluids (aqueous or magmatic) through volcanic systems.

The occurrence of tremors begs that we ask many questions. What is the source of the signals we are seeing? How does the source propagate with time? If it is related to geologic processes, do those occur at the slab interface  or above it in the mantle wedge? Are they related to the movement of aqueous fluids at that depth? Could fluids lubricate the fault surface allowing for slip at lower rates of stress? Do they reflect a change due to the hydration of the mantle wedge? Do they reflect the propagation of a slow slip front? What is their association with slow slip events? Are they a series of small earthquakes over a large surface? What is the relationship, if any, between tremors and regional earthquakes? Can they be triggered or do they trigger tectonic events?

Unfortunately in much of geophysics and geology, we aren’t able to directly observe the physical processes being studied. Instead we are left to look at the evidence left behind (mountain ranges, fault and fracture zones, volcanic deposits, etc.), the signals generated by processes we can measure (e.g., earthquakes), measurable properties using imaging techniques (e.g., gravity, electrical resistivity, magnetic surveys), or the results from laboratory experiments. There is also an enormous range of time scales over which different processes occur, often leaving us with just a few examples to study and compare.

What we do know is that we are seeing emergent seismic signals that rise out of the noise and move across the network of stations with speeds similar to shear waves. We are seeing most of the silent slip events and in particular the largest ones with a similar source location. In fact, the scientists at the PGC see the events almost continually for periods of minutes to days, associate them both spatially and temporally with slow slip events as detected by the geodetic network, and term them Episodic Tremor and Slip (Science, 2003). However, the tremors in Japan are thus far anti-correlated to slow slip events. The tremors are occurring at depths that place them at or below the base of the crust, but above the subducting slab in a region called the mantle wedge. They occur above the part of the slab that is in transition between a locked zone and a region of stable sliding, as defined by earthquakes. So far, however, they are not occurring along the full extent of the subduction zone but from the southern end of the Puget Sound up into Canada, underneath Vancouver Island.

Our next step will be to use three arrays of seismometers as antennas so that we can use frequency-wavenumber methods to analyze the properties of the wavefield, to obtain more accurate locations, and to track the signals and locations as they evolve with time. Combining this with what we know from geochemistry, petrology and geophysical data about the source region should help us better understand the cause of the tremor signals. Not all the questions I have raised can necessarily be answered; I imagine I will come up with more as I continue my research as I certainly have much work left to do for my dissertation.