The pattern of surface strain accumulation in subduction zones is sensitive to the pattern of locking on the underlying plate interface during the interseismic phase, and the pattern of strain release during earthquakes. Both are of interest in the study of subduction seismogenic zones, which produce Earth’s largest earthquakes and most of its tsunamis. Both can be precisely measured with GPS. The pattern of strain accumulation and release may be sensitive to a variety of parameters, the determination of which is critical to understanding the seismic process in subduction zones, including fault friction, fluid flow and thermal state.

One of the remarkable discoveries of the past decade is the existence of periodic slow-slip events at subduction zones. In 2001, transient slow slip was recognized to occur in the northern Cascadia Subduction Zone (Dragert et al., 2001). The long data record and the fortuitous location of the continuous GPS station in Victoria, British Columbia, and many UNAVCO-built and supported GPS stations on the U.S. side of the bor¬der, revealed that slip events in this region repeat with surprising regularity, having a recurrence interval of 14 to 15 months (Miller et al., 2002). Intrigued by Obara’s (2002) finding of deep, non-volcanic, seismic tremors in the Nankai Subduction Zone, Rogers and Dragert (2003) established that the repeated Cascadia slip events were accompanied by prolonged episodes of similar seis¬mic tremor. This discovery of the correlation of tremor and slip (Figure 16) led to the empirical definition of “Episodic Tremor and Slip” (ETS): Repeated, transient ground motions at a plate margin, roughly opposite to longer-term deformation, accompanied by distinct, low-frequency, emergent seismic signals. Since then, other studies have reported slow slip events at several other subduction zones (e.g., New Zealand, Beavan et al, 1999; Central America, Lowry et al, 2001). ETS therefore appears to be a fundamental stress release process at subduction zones, with implications for the dynamics and earthquake potential of subduction zones. Without precise GPS observations, these processes would have gone unnoticed for much longer. Although the source mechanism of ETS remains to be fully understood, these events are likely indicative of transient creep below the locked portion of the subduction interface. Because of the implications of ETS for stress transfer and possible earthquake triggering, it is critically import to understand how this process affects the strain and stress budget at subduction zones. This connection will undoubtedly be the subject of intense research in the years to come.

In the past five years an increasing number of continuous GPS sites have been recording data at high sampling rates (>1 Hz) and several studies have produced significant advances in strategies for processing and using the resulting data to better understand coseismic and early postseismic processes . In 2003, Larson et al. showed that 1 Hz GPS records could complement accelerometer data in studies of the earthquake rupture process by providing direct measurement of ground motion during the passage of seismic waves without saturating, even in the near-field. Thirteen continuous GPS stations recording data at 1 Hz were installed near Parkfield, California, as part of the Southern California Integrated GPS Network (Langbein and Bock, 2004). These sites recorded data during the 2003 San Simeon and 2004 Parkfield earthquakes. These observations have formed the basis for several studies. Ji et al. (2004) showed that the GPS data provided important constraints in a joint inversion with strong motion and teleseismic data for the slip history of the San Simeon event. Choi et al. (2004) used the San Simeon data to demonstrate that a technique called “modified sidereal filtering” can significantly improve the precision of high rate position time series by mitigating the effects of multipath. High-rate GPS data recorded during and after the Parkfield earthquake enabled the separation of coseismic displacement from the immediate postseismic signal, demonstrating that the coseismic moment estimated from GPS agreed well with estimates from seismic data and that a strong postseismic deformation signal began minutes after the event (Langbein et al., 2006). Wang et al. (2007) combined accelerograph and 1 Hz GPS data recorded at collocated or neighboring sites during the San Simeon event to produce broadband time series, which provide valuable information for fault rupture studies and engineering applications.

Of all the stages in the earthquake cycle (coseismic, postseismic, and interseismic), advances in GPS capabilities during the past several years have had the most profound impact on the measurement and understanding of postseismic processes. We now understand that any earthquake may now signal the beginning of a new lithosphere-scale rock mechanics experiment, where geodetic data is acquired and used to infer the mechanical properties of faults and the rheology of the lower crust and lithospheric mantle. Assessing these properties is one of the current challenges in continental dynamics, because they control the temporal and spatial distribution of surface strain at all scales.