Near-Source Vertical and Horizontal Strong Ground Motion from the 20 April 2013 MW 6.8 Lushan Earthquake in China

The 20 April 2013 Lushan Mw 6.8 earthquake was well recorded by the NSMONS of China. Over 92 groups of three-component accelerograms are recorded by strong-motion stations in Sichuan, Yunnan, Gansu, and Shanxi provinces during the mainshock. In this paper, we investigated the spatial distribution and attenuation characteristics of PGAs, PGVs, and spectral accelerations of both horizontal and vertical strong motion from the Lushan earthquake. The observations are compared with NGA prediction models to investigate attenuation characteristics of near-source ground motion on different site classes, and the observed response spectra are compared with those from the Wenchuan earthquake and the design spectra of China. Spatial distribution of near-source strong ground motions shows a clear dependence on both azimuth and distance, mostly affected by the faulting mechanism and source effects. Large horizontal PGAs and PGVs are located within the isoseismic line of intensity VI, especially in a narrow belt of the rupture fault along the north-northeast (N-NE) direction. PGAs and PGVs attenuate rapidly with fault distance. Although only 70 km from the epicenter, the recorded ground motion is relatively weak at 51CDZ station on the footwall. The horizontal PGA and PGV are only 52.3 and 3:0 cm/s, respectively, indicating relatively weak ground motion in this area, with a good coincidence with the damage spatial variation pattern. Shortperiod ground shaking in the near-source region is strong. The largest spectral acceleration at period T = 0:2 s can reach 1699:5 cm/s2, whereas spectral acceleration values decrease sharply with the increasing of period. At periods T = 1:0 and 3.0 s, the largest spectral accelerations are only 155.7 and 25:4 cm/s2. The recorded PGVs are generally lower than the NGA predictions, whereas PGAs on sites B, C, and D generally agree well with the predicted results. Observed spectral accelerations at 0.2 s on site classes B and C are reasonably well described by the median for all azimuths. With the increasing of period, the discrepancy between NGA predictions and observations becomes larger. At 3.0 s, the observed values are less than 1/2 of the median predictions. Long-period ground shaking with T > 1:0 s are relatively weak, which could explain the fact that earthquake damage are mainly on the low-story buildings as shown by the previous study (Xu et al., 2013). Our results suggested that although with a similar focal mechanism as the Wenchuan and Chi-Chi earthquake, the vertical ground motion is relatively weak for this event. This is probably due to the fact that energy dissipation of this event focused in the depth of 12-25 km and the rupture did not break through the ground surface (Wang et al., 2013; Zhang et al., 2013). Both class C site and class B site show remarkable amplifications at short period about 0.2-0.5 s, whereas class C site generally have much wider amplification response period range compared with site class B for both distance range RJB < 30 km and 30 km < RJB < 90 km. After comparing observed response spectra with the Wenchuan earthquake records at the same station, we found that the observed response spectra had much narrower response range with less energy release at longer period. A comparison of the observed acceleration response spectra with the recently proposed design spectra in China for both horizontaland vertical components shows that the near-fault motions generally have higher response values at short period range 0.1-0.2 s whereas much lower response at period over 0.4 s.