Duration characteristics of horizontal components of strong-motion earthquake records

abstract The duration characteristics of horizontal components of strong ground motion records obtained in the western U.S. were examined for engineering applications. A total of 84 accelerograms were used in the study, corresponding to a range of earthquake magnitudes, M , between 4.7 and 7.6, to distances to the source between 0.1 and 130 km, and to local subsurface conditions ranging from rock to soft clay. The definition of significant duration, D (sec), used was the time needed to build up between 5 and 95 per cent of the total Arias intensity of the record I A = π 2 g ∫ 0 t f a 2 ( t ) d t ( t f = total length of record ) . For rock sites, a consistent correlation was obtained between D and M : log D = 0.432 M 1.83. Values of D at soil sites show a much larger scatter, with the duration on rock being a lower bound. Detailed analyses were conducted on the accelerograms and their associated Husid plots, I A (t)/I A (t f ) versus t . It was found that a main parameter controlling the duration of strong ground motion was the total duration of rupture at the fault, d . Rock records are dominated by a strong part, having a duration, Δ, which is essentially a straight line in the Husid plot, and which is Δ ≲ D . This strong part coincides approximately with the arrivals of S waves which followed a direct path between source and station (direct S waves). The values of both Δ and D are mainly controlled by the duration of rupture at the fault, d , for M ≲ 7. Many soil records have, in addition to the strong part, another part which is moderately strong, has longer periods and appears after the direct S wave arrives. This additional part is responsible for the increased values of D observed for soil, and also for the curved shape of the Husid plot observed in many soil records. This part also contributes significantly to the observed richer content of long periods in soil records, reported by several authors. Some evidence suggests this part of the record may be mostly associated with the amplification by the soil of indirect, multipath body-wave arrivals, and with surface wave effects.