Seismic vulnerability assessment of New Zealand unreinforced masonry churches

Churches are an important part of the New Zealand historical and architectural heritage, and the extensive damage occurred to stone and clay-brick unreinforced masonry portfolio after the 2010-2011 Canterbury earthquakes emphasises the necessity to better understand this structural type. An effort was undertaken to identify the national stock of unreinforced masonry churches and to interpret the damage observed in the area affected by the earthquakes: of 309 religious buildings recognized and surveyed nationwide, a sample of 80 churches belonging to the Canterbury region is studied and their performance analysed statistically. Structural behaviour is described in terms of mechanisms affecting the so-called macro-elements, and discrete local damage levels are correlated firstly with macroseismic intensity through Damage Probability Matrices, computed for the whole building and for each mechanism. The results show that the severity of shaking alone is not capable to fully explain the damage, strongly influenced by structural details that can worsen the seismic performance or improve it through earthquake-resistant elements. Simple-linear regressions, correlating the mean damage of each mechanism with the macroseismic intensity, but neglecting the difference in the vulnerability of different churches subjected to the same level of shaking, are then improved through use of multiple-linear regressions accounting for vulnerability modifiers. Several statistical procedures are considered in order to select the best regression equation and to assess which parameters have closer relationships with damage. Results show good consistency between observed and expected damage, and the proposed regression models can be used as predictive tools to help determine appropriate seismic retrofit measure to be taken. The conclusions drawn for the Canterbury region are then extended to the whole national stock and a quantitative seismic risk assessment for existing unreinforced masonry churches in New Zealand is presented, using different intensity measures to model the seismic hazard. Seismic risk is first computed mechanism by mechanism, highlighting how some mechanisms are more frequent than others, and that very large damage levels are expected for some New Zealand regions. Whereupon, an alternative synthetic damage index purely based on observed data is proposed to summarise damage related to several mechanisms and it is used to validate the choice of the best index for describing the global damage of a church when dealing with a territorial assessment. Territorial scale assessment of the seismic vulnerability of churches can assist emergency management efforts and facilitate the identification of priorities for more in-depth analysis of individual buildings. Finally, a preliminary attempt for dynamically characterize the response of unreinforced masonry church is conducted.