Spatial Variation of Ground Motion Determined from Accelerograms Recorded on a Highway Bridge by John C. Wilson* and Paul C. Jennings

A set of time-synchronized strong-motion accelerograms, obtained on the San Juan Bautista 156/101 Separation Bridge in California during the 6 August 1979 Coyote Lake earthquake (M, = 5.9), are used to study the spatial variation of ground motion at the bridge site, including traveling wave effects and the influence of multiple-support excitation. Analysis of the ground motion recorded at the base of two of the bridge supports (32.6 m apart) revealed the presence of a differential support excitation having a period of ~ 3 sec, much longer than any structural periods of the bridge. This signal also appeared as a noticeable long-period component in the superstructure displacements. Analysis of the vertical and radial components of the 3-sec ground motion indicated that ground displacements were retrograde for the duration of strong shaking, with several cycles exhibiting elliptical particle motions. These findings suggest that longperiod differential support motions were induced by phase delays in a Rayleigh wave traveling across the bridge site. Further support to this premise is given by the location of the bridge site near a maxima of the Rayleigh wave radiation pattern for the Coyote Lake earthquake (based on published focal mechanism data). Traveling wave effects were also detected for compressional body waves by a correlation analysis which indicated a time delay of ~ 7 msec between Pwave arrivals at two of the bridge supports. INTRODUCTION In dynamic analysis for earthquake engineering, it is common to assume that the entire base of a structure is uniformly subjected to the same ground motion. That is, the amplitude and phase characteristics of the ground motion are identical at all points where the structure is attached to the ground. This assumes that the ground motion is a result of spatially uniform, vertically propagating shear waves, or, that the wavelengths of the ground motions are long with respect to the dimensions of the structure. The assumption of uniform ground motion has usually been adopted for several reasons: (1) the scarcity of observed data on the spatial variations of seismic ground motions over distances comparable to the base dimensions of most structures; (2) the expediency of solving equations of motion for uniform base excitation (as opposed to equations for differential excitations); and (3) physical reasoning that indicates such an assumption will be very nearly satisfied for many common engineering structures. However, for structures of large spatial extent, such as bridges and pipelines, variation in ground motion over the length of support of the structure may be great enough to make the assumption of uniform ground motion inappropriate. In this case, the different ground motions occurring at each support create what is often referred to as the problem of "multiple-support excitation." The measurement and analysis of differential seismic ground motions occurring over distances of several tens to several hundreds of meters is of significant interest to both seismologists and earthquake engineers, and in particular, to bridge engi* Present address: Department of Civil Engineering and Engineering Mechanics, McMaster University, Hamilton, Ontario, Canada L8S 4L7. 1515 1516 JOHN C. WILSON AND PAUL C. JENNINGS neers. One of the earliest studies of the effects of traveling seismic waves on bridge structures was conducted by Bogdanoff et al. (1965) who examined the case of a seismic motion propagating along the length of a bridge foundation. The bridge responses were found to be noticeably different from those due to a uniform, rigid base excitation. Werner et al. (1977) and Werner and Lee (1980), investigating the effects of traveling seismic waves on the response of a single-span bridge, report that both the type of seismic wave as well as the angle of approach may substantially influence a bridge's dynamic response. Abdel-Ghaffar (1977) has also studied the problem and reports similar results. More recently, Smith et al. (1982) have examined some of the seismological aspects of the spatial variations of seismic ground motions using closely spaced array data for the 1979 Imperial Valley earthquake. Although many modern buildings are instrumented with strong-motion accelerographs, and many excellent records have been obtained from these installations, it was not until the mid-1970's that a program of strong-motion instrumentation of bridges and other transportation structures was initiated in California. The first set of accelerograms was obtained from this program when the San Juan Bautista 156/ 101 Separation Bridge was shaken by the 6 August 1979 Coyote Lake, California, earthquake. In this paper, records of the ground motion at the San Juan Bautista bridge site (provided by the California Division of Mines and Geology) are used to examine the nature of the seismic excitation to which the bridge was subjected during the 1979 Coyote Lake earthquake. It must be mentioned that, with the limited amount of data available, simplifying assumptions and approximate analyses were needed in order to assess the contribution of differential support motions. In an attempt to offset some of these limitations, several different lines of evidence are used in support of the various observations and conclusions. THE SAN JUAN BAUTISTA BRIDGE The purpose of this and the following section is to provide a general description of the San Juan Bautista 156/101 Separation Bridge and a discussion of the strongmotion instrumentation system deployed on the bridge. The availability of strong ground motion records at two separate stations at the bridge site provides the basis for subsequent analyses in this paper. The San Juan Bautista 156/101 Separation Bridge is located approximately 3 km northwest of the town of San Juan Bautista in San Benito County, California (see Figure 1). This two-lane, six-span bridge, constructed in 1959 and owned by the California Department of Transportation (Caltrans), carries a moderate amount of automobile and truck traffic on California State Highway 156 over U.S. Highway 101, and is typical of the late 1950's, early 1960's style of highway bridge design in the United States. Views of the bridge and typical dimensions are shown in Figure 2. Since this paper is concerned with an examination of the characteristics of the ground motions at the bridge site, only those structural features which directly pertain to the foundation substructure will be discussed here. Foundation support for the bridge consists of spread footings bearing directly on horizontal beds of Pliocene alluvial deposits estimated to be approximately 15 m in thickness, which in turn overlie granitic basement rock (Porter et al., 1983). Soil tests at the bridge site prior to construction gave Standard Penetration Test values of N of approximately 50. Values of N this high are indicative of a very dense soil (Scott, 1981). SPATIAL VARIATION OF GROUND MOTION 1517 The left abutment, denoted as A1 on Figure 2, was constructed on a naturally occurring rise of the ground surface while the right abutment (A7 on Figure 2) was constructed on fill material. The abutments and bents are skewed at 34.8 ° with respect to the bridge deck. For later discussions, a global X-Y-Z coordinate system is defined such that the X axis points in the longitudinal direction (parallel to the