Selection of local case studies with fully probabilistic hazard and risk assessments

Abstract

For the Global Risk Assessment of the UNISDR s Global Assessment Report on Disaster Risk Reduction 2015 (GAR15), a coarse grain fully probabilistic risk assessment has been conducted at country level, obtaining in this version the full loss exceedance curve (LEC) from where relevant risk metrics such as the average annual loss (AAL) and probable maximum loss (PML), among others, can be derived. What is also important to highlight, is that following the exact same arithmetic, physical risk can be also assessed at higher resolution levels such as sub-national level using a proxy exposure database and even at city level using detailed building by building information. The selection of the resolution level lies, besides the available information, on if the question that is to be answered requires a detailed approach (e.g. for insurance, cost-benefit of retrofitting and land-use planning).Since physical risk is being calculated as the convolution of the hazard and vulnerability the selection of a higher resolution level may have influence in the way the exposure and in some cases the hazard are modelled. That may imply an abrupt change in scope, approach and resources associated to the risk assessment, especially in countries and regions where there is little information.By increasing the resolution level, it is also possible to consider other hazards such as floods and landslides that have a more relative punctual affection whilst also allowing considering more details that are relevant such as the site effects (due to the dynamic soil response) in the seismic hazard and risk analysis.This background paper presents a selection of fully probabilistic hazard and risk assessments conducted at local (city) level where the same methodology and model used for the GAR13 and GAR15 have been followed, that is, hazard is represented through a set of stochastic scenarios and physical vulnerability is quantified by means of probabilistic vulnerability functions. Even more, all the cases have been calculated using the CAPRA Platform (Cardona et al. 2010, 2012) the same used for the Global Risk Model of UNISDR s GAR from 2013. The information presented in each case is a summary of the complete studies and/or reports. Many of them have also been associated with peer-reviewed papers and proceedings of international conferences and all the relevant references are given in case the reader wants to know more details about any of them.For all the cases, physical risk, in terms of the same metrics obtained for the 216 countries, is assessed, but in some cases, in addition, those physical risk results have been used as the input data for holistic risk assessments to include risk and resilience drivers. These local cases have been calculated in the framework of both, consultancy activities and academic research. For each case a brief description of the area under analysis and the question that is attempting to answer is presented. Each of these studies constitutes a base for future updates and recalculation since many of the aspects considered in this kind of modelling, especially those related to the exposure and the socioeconomic conditions are dynamic and should be evaluated on a regular basis. Therefore, all these results may be different if evaluated later, for example in 5 or 10 years.