As in past years, the most comprehensive series of crustal deformation measurements were made in California, particularly southern California. Given that deformation across the San Andreas Fault had already been well established by trilateration data [ Lisowski et al., 1991], most of the GPS measurements reported between 1991-1994 were concentrated to the west of the San Andreas fault. Many of these areas were inaccessible to trilateration surveys, and thus there was little geodetic information to constrain present day slip rates on faults. The Bird and Rosenstock [1984] and Weldon and Humphreys [1986] models discussed in the previous section made predictions for southern California which could be tested by GPS. These models disagreed in many regions, in particular, the offshore, the Mojave region, and the western Transverse Ranges
A GPS geodetic network was established along the coast of southern California 1986 by a group of universities interested in these questions (see Figure 2). They made annual observations at these sites for approximately five years. This network included the first GPS measurements on the offshore islands, including Santa Catalina, Santa Cruz, San Clemente, and San Nicolas Islands. In addition to these repeated GPS measurements, this group endeavored to recover historic triangulation and trilateration marks. In some cases, the triangulation record extended to the late 19th century. The results of this cooperative effort were published in a series of papers, separated by subject area: the Santa Maria Fold and Thrust Belt [ Feigl et al., 1990], the Ventura Basin [ Donnellan et al., 1993], the southern Coast Ranges [ Shen and Jackson, 1993], the Santa Barbara Channel [ Larson and Webb, 1992; Larsen et al., 1993], and the Gulf of Santa Catalina [ Larson, 1993]. A summary paper which included both GPS and VLBI measurements in the region was published by Feigl et al. [1993]. The VLBI observations included both fixed sites, e.g. Mojave and Owens Valley, which are situated along the Eastern California Shear Zone [ Savage et al., 1990], and several mobile VLBI sites.
The velocity determinations by Feigl et al. [1993] are shown in Figure 2. The velocities are with respect to the Pacific plate. Note the velocity of Vandenberg (VNDN) and the accompanying hypothesis that Vandenberg is very nearly on the Pacific plate. The island sites (TWIN, BLUF, CENT) also show no significant deformation with respect to the Pacific plate. In addition to the strong shear signal seen at sites east of the San Andreas Fault, compression throughout the western Transverse Ranges and the Los Angeles basin is also observed. Geodetic measurements across these active faults is particularly important, as the faults are not observable at the surface. Donnellan et al. [1993] found significant north-south compression across the Ventura Basin, 5-8 mm/yr, and also inferred block like rotations. Although the patterns of deformation across the Ventura Basin agree with geologic observations, the geodetic measurements of convergence rate are lower in magnitude by about one third [ Yeats, 1983].
Compression of a comparable magnitude is also seen across the Santa Barbara Channel, as
measured by both trilateration and GPS [ Larsen et al., 1993; Larson and Webb, 1992], consistent
with both seismicity and geologic indicators [ Lee et al., 1979, Yerkes et al., 1980]. The Los
Angeles Basin is also an area of active compression, 5
1 mm/yr. There also appear
to be active shear structures
in the Gulf of Santa Catalina (between BLUF and NIGU in Figure 2), although the uncertainties are large
[ Larson, 1993]. One accomplishment of these GPS observations in southern California has been
to quantify the style and variety of deformation throughout the region, where there had been little
or no geodetic information. Further GPS measurements will allow finer resolution of rotations and
slip rates.
While Feigl et al. [1993] and others summarize the results of annual geodetic observations, these types of measurements have in many cases been supplanted in southern California by continuous monitoring networks. Beginning in 1990 with three GPS receivers in Pasadena, Pinyon Flat, and La Jolla, the number of continuous operating receivers in southern California has increased to 20 at the present time [Y. Bock, written communication]. To date, there have been few published reports from these networks, with the exception of co-seismic estimates of deformation. Results from one of the southern California continuously operating GPS sites can be seen in Hudnut [this volume].
In addition to these studies which concentrated on the GPS technique, there has also been significant progress in combining different kinds of geodetic data and/or using older data. Johnson et al. [1994] have concentrated their efforts on the 20 year record of trilateration data from southern California. Instead of displaying crustal deformation via relative velocity, as was shown in Figure 2, they calculate strain rates across subnets. This approach allows them to identify the spatial patterns of strain throughout southern California. They have concluded that the San Jacinto fault, rather than the San Andreas fault, accommodates the bulk of shear deformation in the southernmost part of California. This contradicts the geologic record, which finds higher slip along the San Andreas (see summary in Lisowski et al. [1991] for slip rates). Saucier and Humphreys [1993] took a different approach, inverting a database of geologic slip rates and geodetic observations. From these, they were able to confirm many of the general features which have been observed in the GPS surveys described by Feigl et al. [1993]. For example, Saucier and Humphreys model significant deformation across the Santa Barbara Channel, and also across the Gulf of Santa Catalina. This result agrees with independent geodetic observations [ Larsen et al., 1993; Larson, 1993].