FIGURE 11.7 Height change data across San Andreas Fault at San Juan Bautista, 1975–1983. Bench mark 7328 is arbitrarily held fixed.

shaped array at Anza (Figure 11.12) detects minor vertical movements across a narrow, well-defined rift zone of the San Jacinto Fault (Figure 11.13); a closed quadrilateral across the San Andreas Fault at Wallace Creek (Figure 11.14) provides us two measurements of vertical offset across the fault as well as of the tilt of the fault blocks; lines in Death Valley and Long Valley are across youthful, normal faults; and the quadrilateral at the Pinyon Flat Geophysical Observatory (Figure 11.15) provides geodetic control for arrays of long-base fluid tiltmeters and other strain-measurement devices (Sylvester and Jackson, 1982; Sylvester, 1984).

We choose our surveying sites where geomorphic evidence shows that significant vertical movements have occurred in the recent geologic past, and we rely on other investigators to provide medium-range precursory information that identifies faults that may be in the preparation stages for a major earthquake. Then we establish arrays in the target area and resurvey them as frequently as possible and practical.

Like other investigators (e.g., Sharp and Lienkaemper, 1982), our procedure is to repeat precise leveling surveys of arrays of permanent bench marks established across active and potentially active faults. Comparison of surveys reveals height changes that then may be related spatially to surface faults and temporally to occurrences of earthquakes.

All our leveling arrays are relatively short—line lengths range from 200 to 2600 m and contain as many as 70 bench marks. Geometry of arrays is generally dictated by the terrain and property access and includes L-, Z-, W-, and closed, quadrilateral-shaped arrays (Sylvester, 1982). Some of the straight-line segments across faults are also aligned with a theodolite to document horizontal movement. Many of the fault-crossing arrays are quadrilateral-shaped to determine the tilt of each fault block independently by analyzing L-shaped subsets of bench marks (Figure 11.14).

Following Tryggvason (1968) and our accumulated

FIGURE 11.8 Site map of W-shaped leveling array across subsidence fault in Fremont Valley, California.