Effect of Spatial Variation of Ground Motion on Highway Structures

The objectives of this study were to estimate the effect of spatial variability of ground motion on the seismic response of highway bridges and to provide guidelines for the seismic analysis and design of such structures to account for this effect. To accomplish these objectives, a methodology was developed to generate spatially varying ground motion time histories at the different supports of a bridge. Spatial variability of seismic ground motion can be mainly attributed to the following three mechanisms: 1) difference in arrival times of seismic waves at different locations, commonly known as the "wave passage effect"; 2) loss of coherence of seismic waves due to multiple reflections and refractions as they propagate through the highly inhomogeneous soil medium, referred to as the "incoherence effect"; and 3) change in the amplitude and frequency content of seismic ground motion due to different local soil conditions, known as the "local soil effect". The "wave passage effect" is characterized by the apparent velocity of wave propagation, the "incoherence effect" is characterized by the so-called coherence function, while the "local soil effect" is characterized by defining different response (or power) spectra at different locations corresponding to their local soil conditions. The methodology that was developed in this study to generate spatially varying seismic ground motion time histories at different locations reflects all three of these effects. A set of representative highway bridges was then selected to be analyzed using identical support ground motions and differential (asynchronous) support ground motions, in order to estimate the effect of spatial variability of seismic ground motion. Extensive preliminary as well as rigorous sensitivity analyses were performed using a large variety of different spatially varying ground motions. A Monte Carlo simulation approach was followed throughout. Based mainly on the general conclusions of the analyses performed in this study, and taking into account the observations and conclusions of other researchers' work, the following guidelines are proposed for the analysis and design of highway bridges to account for the spatial variation of seismic ground motion: (I) For bridges that are less than approximately 1,000 to 1,500 ft in total length, and have all their supports on the same local soil conditions, it is recommended that they be analyzed and designed using currently available seismic design practices; and (II) For bridges that are more than approximately 1,000 to 1,500 ft in total length, or bridges of any length that have supports on different local soil conditions, it is recommended (a) to perform time history dynamic analyses for design purposes, involving response spectrum compatible asynchronous support ground motion time histories reflecting the different local soil conditions, wave propagation and loss of coherence effects and (b) to consider a minimum of 20 time history analyses of the bridge to get a reliable estimate of the peak response that will be used for design purposes.