GROUND-MOTION VARIABILITY AT THE HIGHWAY-14 AND I-5 INTERCHANGE IN THE NORTHERN SAN-FERNANDO VALLEY

The purpose of this article is to analyze possible causes for failure of the structures at the Highway 14 and I-5 interchange in the norther n San Fernando Valley following the magnitude M 6.7 San Fernando earth quake on 9 February 1971 and the M 6.7 Northridge earthquake on 17 Jan uary 1994. This interchange consisted of a series of structures up to 600-m long; during the 1971 and 1994 earthquakes, several of these str uctures separated along expansion joints and collapsed. To analyze the possible causes of failure, we recorded aftershocks at and near the i nterchange for 2 weeks following the 1994 earthquake. The aftershock r ecordings reveal a substantial relative spectral amplification (in the 1- to 10-Hz range) of a factor of 3 to 6 for sites located at the int erchange (which is situated on sandstone) relative to a nearby rock (g neiss) outcrop site. Sites within the interchange area have a variatio n in amplification of a factor of 2 to 4 in the 1- to 10-Hz range. Als o, substantial secondary arrivals were generated at two locations in t he interchange area. We attribute variations in ground motion at the i nterchange to be due to variations in the composition of the sandstone and to topographic effects. We computed the linear-response strong gr ound motion for three sites at the interchange from the 1971 and 1994 main events using aftershock recordings of the Northridge earthquake f rom three sites (the south and north abutments of the southbound conne ctor of Hwy 14 to I-5 and a central location to the interchange) as em pirical Green's functions. The same source events are used for all thr ee sites for a particular synthesis so that differences between synthe sized strong ground motion are due to differences in location, geology , and topography of the three sites. Source rupture models are obtaine d from independent studies. We estimate (1) peak acceleration to have been near 1 g for both earthquakes, due to the topographic and geologi c conditions and the geometry of the fault rupture; (2) differential d isplacement across the north and south extremes of the area from the N orthridge earthquake to have been about 70 cm longitudinal and 40 cm t ransverse to the Hwy 14 to I-5 connector; (3) differential displacemen t across the north and south extremes of the area from the San Fernand o earthquake to have been about 25 cm longitudinal and 20 cm transvers e; and (4) relative motion to be fairly incoherent for frequencies abo ve 0.5 Hz. These observations are attributed to local geologic and top ographic conditions and to wave passage effects and could have contrib uted to collapse of structures during these earthquakes.