GPS technology has advanced to the point that obtaining sub-centimeter point positioning accuracies within regional networks is reasonably routine (see Yunck, this issue), establishing GPS as an important tool in earthquake geodesy and hazard monitoring. Recent papers have reviewed the principals of GPS and its application to crustal deformation research (e.g., Dixon, 1991; Hager et al., 1991; Larson and Agnew, 1991). Operation of continuously recording GPS receivers at fixed sites is becoming increasingly accurate and automated as well (e.g., Shimada and Bock, 1992; Blewitt et al., 1993; Bock et al., 1993). Techniques for rapid point positioning with GPS have also matured considerably, pushing the higher-frequency limits of current receiver and antenna technology (e.g., Genrich and Bock, 1992; Mader, 1992). Progress has also been made in methods of applying geodetic data towards detecting anomalies in secular rates of deformation (Langbein et al., 1993).
Application of geodetic technology towards hazard evaluation has advanced on several fronts. Interseismic deformation data have provided some crucial information in determining the long-term seismogenic potential of the Cascadia subduction zone, for example (e.g., Savage et al., 1991; Snay et al., 1991; and Dragert et al., 1994). Also, the geodetically determined rate of motion across the Ventura Basin formed the basis for revising the hazard estimate for the Oak Ridge fault, an eastern extension of which ruptured in the 1994 Northridge earthquake (Donnellan et al., 1994). That study indicated that earthquakes up to Mw=6.4 were more likely than less frequent Mw>7 events in the Ventura Basin because the deeper portions of the fault system slip aseismically according to the GPS results. Interestingly, Wesnousky (1986), had inferred events of Mw=6.9 for the Oak Ridge fault system on the basis of geological data in his synthesis of regional earthquake hazards.
Geodetic data are now being used in constructing synthetic models of long-term seismic hazard (e.g., Ward, 1994) and a variant of these methods is employed in the SCEC Phase II Report (in press). Geodetic data are useful in these circumstances for defining the expected rate of slip on a system of faults that exist within a volume of the earth's crust. This is important in shear zones and thrust sheet systems, where definition of slip rates by paleoseismic methods is particularly challenging (see Sieh, this issue). In shear zones, the geodetically determined rate may be higher than the summed slip rates across the many individual faults within the zone, implying a higher cumulative seismic hazard. This is evidently the case in the Eastern Mojave Shear Zone, where the rate of strain accumulation has recently been re-estimated, and on this basis Sauber et al. (1994) re-evaluated the long-term earthquake hazard of the shear zone. Furthermore, in regions where thrust sheet systems include faults that do not reach the surface (called ``blind'' thrust faults), geodetic data provide constraints on seismic hazard. In particular, this is true for the Los Angeles metropolitan region (e.g., SCEC Phase 2 Report), but also for less heavily populated areas such as the San Joaquin Valley of California (e.g., Stein and Ekstrom, 1992). Furthermore, models of fault slip in earthquakes computed from seismological and geodetic data apply to estimating regional stress changes associated with an earthquake. Several groups used this approach recently to study the 1992 Landers earthquake sequence, in particular its affect on stress along the San Andreas fault system (Harris and Simpson, 1993; Jaume and Sykes, 1993; and Stein et al., 1993). These stress change calculations became part of the SCEC's `Phase I' report on southern California earthquake hazards (SCEC Report, 1992). Stein et al. (1994) investigated possible stress triggering of the Northridge event by previous earthquakes.
Also, monitoring of ongoing deformation in active volcanic regions has proven important in understanding the temporal evolution of non-erupting or effusively erupting magma movement (e.g., Dvorak and Berrino, 1991; Holdahl and Dzurisin, 1991; Meertens and Smith, 1991; Delaney et al., 1993; Dvorak et al., 1994; and Langbein et al., 1993). Continuing deformation studies of major active volcanic complexes, such as Yellowstone and Long Valley calderas, augment seismicity monitoring and help to quantify activity levels in these potentially very hazardous areas. It is encouraging that recordings of the 1989 submarine eruption near Ito, Japan and the 1991 eruption of Hekla volcano in Iceland have provided intriguing and compelling geodetic evidence that anomalous deformation occurs prior to some volcanic eruptions. In the case of the eruption near Ito, anomalous deformation began to occur several days prior to commencement of the eruption (Shimada et al., 1990). Precursory deformation associated with seismicity preceded the eruption. Continuous GPS, strain and tilt measurements document the precursory deformation, and Okada and Yamamoto (1991) interpreted the sequence thoroughly. For the 1991 Hekla eruption, Sigmundsson et al. (1992) measured the displacement field associated with the eruption with GPS, and placed some constraints on the eruptive process. Continuous strain data from five instruments within 50 km of the eruption provided more details of the eruption's source (Linde et al., 1993). These strain data showed that it took 30 minutes for magma to rise 4 km to the ground surface, also indicating that in some cases, early warning of imminent eruption is possible.