Affordable Access

Access to the full text

Selection of ground motion attenuation model for Peninsular Malaysia due to far-field Sumatra earthquakes

Authors
  • Van, Tze Che1
  • Lau, Tze Liang1
  • Mok, Chai Fung1
  • 1 Universiti Sains Malaysia, School of Civil Engineering, Penang, Malaysia , Penang (Malaysia)
Type
Published Article
Journal
Natural Hazards
Publisher
Springer-Verlag
Publication Date
Oct 19, 2015
Volume
80
Issue
3
Pages
1865–1889
Identifiers
DOI: 10.1007/s11069-015-2036-8
Source
Springer Nature
Keywords
License
Yellow

Abstract

Establishment of quality seismic hazard assessment is governed by capability of an attenuation model to produce closest predictions to the actual ground motion. Therefore, a robust ground motion attenuation model must be selected to feed a logic tree. The scarcity of historical strong ground motion data hinders the development of attenuation model for the low-seismicity area such as Peninsular Malaysia. This paper aims to determine an appropriate ground motion attenuation model for Peninsular Malaysia out of 28 pre-selected ground motion attenuation models. Evaluation of pre-selected models respective to actual weak ground motion records in Peninsular Malaysia resulted from Sumatra earthquakes was conducted. A total of 327 seismic records from 44 distant subduction and strike-slip Sumatra earthquakes with moment magnitude ranging from 5.2 to 9.1 spanning in a distance range of 284–1292 km were obtained from 19 seismic stations operated by Malaysian Meteorological Department. The multi-channel analysis of surface waves was conducted on all seismic stations to characterise the sites. Based on graphs plotting and calculation of quantification measure, the best fitting model for distant subduction earthquakes is Nabilah and Balendra (2012) with RMSE value as low as 0.182 and 0.107 for interface and intraslab events, respectively. Si and Midorikawa (2000) and Somerville et al. (2009) models give closest prediction to distant strike-slip earthquakes.

Report this publication

Statistics

Seen <100 times