A comparison of two regional seismic damage estimation methodologies Tuna Onur, Carlos E. Ventura, and W.D. Liam Finn Abstract: This paper presents a comparison of the two main regional damage estimation methodologies currently in use, namely the modified Mercalli intensity (MMI) based approach and the spectral parameter based approach. In the first methodology, expected damage is related to ground shaking intensity in terms of MMI through damage probability matrices. In the second methodology, the ground motion intensity is described in terms of spectral acceleration (SA), and building response in terms of spectral displacement (SD). Both methodologies were applied to buildings in Van- couver of three different construction types: single-family wood-frame houses, low-rise unreinforced masonry buildings, and high-rise concrete frame structures with concrete shear walls. The two methodologies predict damage that lies in the same general damage categories of light and moderate, which are defined by fairly broad ranges in mean damage factors. The specific mean damage factors predicted by the two methods for a given location are significantly different, however. The significant differences in mean damage factors imply significant differences in damage costs and hence in seismic risk. Key words: earthquake, damage, seismic risk, vulnerability, modified Mercalli intensity (MMI), spectral response, dis- placement, acceleration. Résumé : Cet article compare les deux méthodes principales d’estimation régionale des dommages communément utili- sées : l’approche basée sur l’échelle d’intensité de Mercalli modifiée (MM) et l’approche basée sur le paramètre de ré- ponse spectrale. Dans la première, les dommages prévus sont reliés à l’intensité du tremblement du sol en termes de MM par des matrices de probabilité des dommages. Dans la seconde, l’intensité du mouvement du sol est décrite en termes d’accélération spectrale et les effets sur les bâtiments sont décrits en termes de déplacement spectral. Les deux méthodes ont été appliquées à trois genres de bâtiments de Vancouver : des maisons unifamiliales à ossature de bois, des bâtiments bas en maçonnerie non armée et des structures à charpente en béton de grande hauteur munies de murs de refend en béton. Les deux méthodes permettent de prédire les dommages dans les mêmes catégories générales de dommages légers et modérés qui sont définies par des plages relativement larges de facteurs de dommages moyens. Ce- pendant, les facteurs spécifiques de dommages moyens prédits par les deux méthodes pour un endroit donné montrent des différences significatives. Ces différences dans les facteurs de dommages moyens impliquent des différences impor- tantes dans les coûts des dommages et ainsi dans le risque sismique. Mots clés : séisme, dommage, risque sismique, vulnérabilité, intensité de Mercalli modifiée (MM), réponse spectrale, déplacement, accélération. [Traduit par la Rédaction] Onur et al. 1409 Introduction Estimating damage to structures is essential for regional seismic risk assessments, but it is a challenging task because of the large uncertainties involved. The challenge is even more pronounced in areas where records of earthquake dam- age from past earthquakes are scarce, such as southwestern British Columbia (BC). In the absence of past damage data with which to compare or calibrate the estimates, it is of in- terest to compare estimates from two methodologies widely used in practice to assess the stability of damage estimation. The conventional modified Mercalli intensity (MMI) based damage matrices have been widely used in regional earthquake damage estimations in urban areas since they were developed 20 years ago (ATC 1985). Although damage matrices are still commonly used, an alternative methodol- ogy has been applied within the framework of a major loss estimation project in the USA (FEMA–NIBS 1997). This project was contracted by the US National Institute of Build- ing Sciences (NIBS) to Risk Management Solutions Inc. (RMS) and a consortium of 30 earthquake engineering ex- perts. The methodology was calibrated by available damage data from past earthquakes, including Northridge and Loma Prieta (Whitman et al. 1997), and was implemented in the loss estimation software HAZUS (FEMA–NIBS 1997). The level of shaking is expressed in significantly different ways in the two methodologies. HAZUS abandons MMI as Can. J. Civ. Eng. 33: 1401–1409 (2006) doi:10.1139/L06-084 © 2006 NRC Canada 1401 Received 22 December 2004. Revision accepted 5 May 2006. Published on the NRC Research Press Web site at http://cjce.nrc.ca on 26 January 2007. T. Onur, 1 C.E. Ventura, 2 and W.D.L. Finn. Department of Civil Engineering, The University of British Columbia, 6250 Applied Science Lane, Vancouver, BC V6T 1Z4, Canada. Written discussion of this article is welcomed and will be received by the Editor until 31 March 2007 1 Present address: Risk Management Solutions, Inc., 7015 Gateway Boulevard, Newark, CA 94560, USA. 2 Corresponding author (e-mail: ventura@civil.ubc.ca).