EJSE Special Issue: Earthquake Engineering in the low and moderate seismic regions of Southeast Asia and Australia (2008) Seismic Hazard of Hong Kong J.W. Pappin 1, R.C.H. Koo 1 & M.W. Free 1, H.H. Tsang 2 Ove Arup and Partners Hong Kong Ltd, Hong Kong 1 Department of Civil Engineering, University of Hong Kong, Hong Kong, Email: [email protected] 2 ABSTRACT: This paper reviews the probabilistic seismic hazard assessment studies undertaken in recent years to estimate the potential seismic ground motion levels on bedrock in Hong Kong. A detailed catalogue of historical and recent seismicity within the South China region has been compiled. A suite of published em- pirical and stochastic attenuation relationships have been used with alternative source models and source pa- rameters in a logic tree hazard analysis. Uniform hazard bedrock ground-motion spectra having various prob- abilities of being exceeded in 50 years have been calculated. The results have been de-aggregated to investigate what earthquake magnitude and distance combinations have contributed most to the hazard levels for the different probabilities and structural periods. The obtained uniform hazard spectra have been compared to the study using an alternative assessment approach developed by the University of Hong Kong. Recent re- corded earthquake ground motions in Hong Kong are also presented with the uniform hazard spectra. Keywords: Seismic hazard, earthquake, ground motion, attenuation, response spectra, Hong Kong fault zones are interpreted to lie along the Southeast China coast. The Changle-Nanao Fault Zone, runs 1 INTRODUCTION along the coast and is offshore northeast of Hong Kong. The Lianhuashan Fault Zone, runs inland, pa- The Hong Kong Special Administrative Region is rallel to the coast from Shanghai to Hong Kong. located in an area of low to moderate seismicity. The Most of the tectonic deformation that is evident in current codes of practice for building design do not the rocks of Hong Kong today was caused by events require any seismic considerations. This paper re- that occurred during the Late Jurassic to Late Creta- views the assessments of the potential seismic ceous periods (160 to 90Ma) during what is referred ground motion levels on rock in Hong Kong. The to as the Yanshanian Orogeny (Sewell et al. 2000). evaluation of the potential site response effects in Sewell et al. (2000) state that there is no direct evi- Hong Kong are described in a companion paper in dence of fault displacements in either the offshore or Pappin et al. (2008). onshore Quaternary age (less than 2Ma) superficial deposits. They state that no fault displacements have been identified in the immediate vicinity of Hong 2 GEOLOGY AND TECTONICS OF THE HONG Kong from many hundreds of kilometres of offshore KONG REGION seismic lines of Quaternary age offshore alluvium and marine sequences. In recent years, thermo- 2.1 Regional Geology and Tectonic Setting luminescence (TL) dating has been used for the dat- Hong Kong is situated in Southeast China near the ing of fault gouge in southeast China, Ding & Lai south-eastern margin of the Eurasian Continental (1997), with possible peaks in fault activity identi- Plate in a stable continental intraplate region about fied at 270,000, 190,000 and 100,000 years before 700 km from the nearest plate boundary, which un- present. Lee et al. (1998) state that, based upon TL derlies Taiwan and trends south to the Philippines dating, the last major fault activity in Hong Kong and northeast to Japan [Sewell et al. (2000) and Fyfe can be interpreted to have occurred between 80,000 et al. (2000)]. The regional tectonic setting is shown to 100,000 years ago. in Figure 1. Two major, regional northeast trending 42 EJSE Special Issue: Earthquake Engineering in the low and moderate seismic regions of Southeast Asia and Australia (2008) will be exceeded if an earthquake of a certain magnitude at a certain distance occurs and mul- tiplying this likelihood by the annual frequency of such an event occurring in any of the source zones. By summing the annual frequencies of the ground motion level being exceeded from all specified earthquake distances and magnitudes the overall frequency is established. The use of this seismic hazard assessment meth- odology to estimate the potential seismic ground motion levels on bedrock in Hong Kong has been previously published in Free et al. (2004). Figure 1. Tectonic setting of Hong Kong An alternative methodology, namely the direct amplitude-based approach, has been developed in the University of Hong Kong (HKU) to assess seis- 2.2 Geology of Hong Kong mic hazard. This approach has been shown to be consistent with the results using conventional ap- The geology of Hong Kong is described by Sewell et proach in Tsang and Chandler (2006). Also, “site- al. (2000) and Fyfe et al. (2000). More than three- specific” and “event-specific” ground motion model quarters of the land area of Hong Kong is underlain can be adopted. by igneous rocks predominantly volcanic tuffs and Epistemic uncertainty arising from differences in granites of Late Jurassic to Early Cretaceous age expert opinion on a range of modelling assumptions (140 to 120Ma). Older, Late Paleozoic (420 to has been addressed through the use of a logic tree 240Ma) sedimentary rocks and younger Late Meso- [Kulkarni et al. (1984), Coppersmith and Youngs zoic to Tertiary (140 to 2Ma) sedimentary rocks un- (1986), National Research Council (1988)]. Aleatory derlie the majority of the remaining land area. Su- uncertainty, arising from natural physical variability, perficial deposits comprising Quaternary (less than has been addressed by allowing for the normal varia- 2Ma) alluvium and other unconsolidated deposits tion, represented by its standard deviation “sigma”, are also present throughout the territory. Large areas of the ground motion attenuation relationships in the of reclamation have been formed around the coastal hazard computation. areas of the territory. 3.2 Seismological Data for the Hong Kong Region 3 SEISMIC HAZARD ASSESSMENT Historical earthquake data for the Southeast China 3.1 Seismic Hazard Assessment Methodology region (Guangdong Province of China) has been ob- The conventional probabilistic seismic hazard as- tained from a range of sources. The Directory of sessment (PSHA) methodology, e.g. Cornell (1968), Earthquakes in China (BC 1831 to AD 1969) as McGuire (1993), has been applied using Oasys listed in Gu et al. (1983) and the Guangdong Seis- SISMIC, the in-house PSHA program of Arup. The mological Bureau (1991) database provide the most PSHA methodology used the following steps: extensive catalogues of historical earthquake data for the region. These have been supplemented by the • Potential seismic sources have been defined on data from the Geotechnical Control Office (1991). the basis of regional geology and seismicity, For historical earthquakes, where there is no instru- • Seismicity parameters defining the rate of earth- mental records, the event magnitude has been de- quake activity have been derived for each of the termined by Gu et al. (1983) using an empirical for- potential seismic sources, mula, M = 0.58 I +1.5, where I is the intensity at • Ground motion attenuation relationships, con- O O the epicentre of the event. The historical dataset sidered to be appropriate for the region, have comprises 199 events between 1067 and 1970. For a been defined, and number of the larger historical events, the magni- • The frequency of specified ground motion levels tudes defined by Johnston (1996a, 1996b) have been being exceeded has been derived by first deter- used. mining the likelihood that the ground motion 43 EJSE Special Issue: Earthquake Engineering in the low and moderate seismic regions of Southeast Asia and Australia (2008) A set of 3-component, long period seismographs, 3.3 Seismic Source Zones and Parameters was established at the Hong Kong Observatory in Southeast China is located within a stable continen- 1921. The first felt earthquake recorded by these in- th tal intraplate region and the association of earth- struments occurred on the 10 January 1924. During quakes. A number of previous studies have defined World War II, the original seismographs were lost seismic source zone models for Southeast China and no local observations were made between 1941 (Pun & Ambraseys 1992, Wong et al. 1998 and Lee and 1950. Recording resumed in 1951. The GSB has et al. 1998) and these models have been included in maintained a database of earthquakes within the this study using the logic tree described below. In Guangdong Province with magnitude, M L ≥ 2.0 addition, the source zone model shown in Figure 2 since 1970. A subset of this data, for the region has been developed and is incorporated into the log- within 500 km of Hong Kong, was obtained for this ic tree with higher weighting. The model extends out study. to a distance of 500km from Hong Kong. A more For the PSHA, all events are required to be statis- distant seismic source zone was also included for the tically independent and therefore foreshocks and af- region of Taiwan (not shown on Figure 2). The tershocks have been removed from the catalogue us- seismic activity parameters for the model shown in ing the methodology of Gardener and Knopoff Figure 2 are summarised in Table 1. (1974). Man-induced events have also been identi- The earthquake catalogue for the Hong Kong re- fied and removed from the catalogue. A reservoir gion includes three earthquakes with magnitudes induced earthquake swarm commenced in 1962 dur- greater than M = 7. Earthquakes with magnitudes M ing the filling of the Xinfengjiang Reservoir. The = 7.5 occurred in 1604 and 1605 and a magnitude M mainshock of the swarm had a local magnitude M L = = 7.4 event occurred in 1918.