These Guidelines provide state-of-the-art guidance on the construction of vulnerability relationships from post-earthquake survey data. The Guidelines build on and extend procedures for empirical fragility and vulnerability curve construction found in the literature, and present a flexible framework for the construction of these relationships that allows for a number of curve-fitting methods and ground motion intensity measure types (IMTs) to be adopted. The philosophy behind the design of the framework is that the characteristics of the data should determine the most appropriate statistical model and intensity measure type used to represent them. Hence, several combinations of these must be trialled in the determination of an optimum fragility or vulnerability curve, where the optimum curve is defined by the statistical model that provides the best fit to the data as determined by a number of goodness-of-fit tests. The Guidelines are essentially a roadmap for the process, providing recommendations and help in deciding which statistical model to attempt, and promote trialling of models and IMTs.
The Guidelines are targeted at analysts with Master’s level training in a numerate subject that includes some level of statistics. The Authors recognise that the statistical analysis understanding of analysts varies. To accommodate for these differences, two levels of statistical approaches for constructing empirical fragility functions that include procedures of increasing complexity, are proposed.
All stages of the fragility and vulnerability curve construction are reviewed and presented in the Guidelines with practical advice given for the preparation of empirical data for use in the construction of these curves, for the identification of sources of uncertainty in the data and in the chosen intensity measures, and where possible, for uncertainty quantification and modelling.
To facilitate adoption of the Guidelines, the code and commands required for the implementation of the described statistical models are provided for the open source software R (2008). Appendices B to G also provide example applications of the guidelines, where each step of the guideline is illustrated for empirical datasets deriving from the 1980 Irpinia, Italy, earthquake, the 1978 Thessaloniki, Greece, Earthquake, the 1989 Newcastle and 2010 Kalgoorlie, Australia, earthquakes and for two earthquakes that affected the town of Christchurch New Zealand in 2010 and 2011. The fragility and vulnerability curves developed from these applications are all presented using a reporting template (presented in Appendix A) designed to facilitate the evaluation and inclusion of empirical fragility curves derived using these Guidelines into the Global Earthquake Model (GEM).