TY - JOUR A1 - Gomez-Zapata, Juan Camilo A1 - Pittore, Massimiliano A1 - Cotton, Fabrice A1 - Lilienkamp, Henning A1 - Shinde, Simantini A1 - Aguirre, Paula A1 - Santa Maria, Hernan T1 - Epistemic uncertainty of probabilistic building exposure compositions in scenario-based earthquake loss models JF - Bulletin of Earthquake Engineering N2 - In seismic risk assessment, the sources of uncertainty associated with building exposure modelling have not received as much attention as other components related to hazard and vulnerability. Conventional practices such as assuming absolute portfolio compositions (i.e., proportions per building class) from expert-based assumptions over aggregated data crudely disregard the contribution of uncertainty of the exposure upon earthquake loss models. In this work, we introduce the concept that the degree of knowledge of a building stock can be described within a Bayesian probabilistic approach that integrates both expert-based prior distributions and data collection on individual buildings. We investigate the impact of the epistemic uncertainty in the portfolio composition on scenario-based earthquake loss models through an exposure-oriented logic tree arrangement based on synthetic building portfolios. For illustrative purposes, we consider the residential building stock of Valparaiso (Chile) subjected to seismic ground-shaking from one subduction earthquake. We have found that building class reconnaissance, either from prior assumptions by desktop studies with aggregated data (top-down approach), or from building-by-building data collection (bottom-up approach), plays a fundamental role in the statistical modelling of exposure. To model the vulnerability of such a heterogeneous building stock, we require that their associated set of structural fragility functions handle multiple spectral periods. Thereby, we also discuss the relevance and specific uncertainty upon generating either uncorrelated or spatially cross-correlated ground motion fields within this framework. We successively show how various epistemic uncertainties embedded within these probabilistic exposure models are differently propagated throughout the computed direct financial losses. This work calls for further efforts to redesign desktop exposure studies, while also highlighting the importance of exposure data collection with standardized and iterative approaches. KW - Epistemic uncertainty KW - Sensitivity analysis KW - Scheme KW - Faceted taxonomy KW - Probabilistic exposure modelling KW - Earthquake scenario KW - Data collection KW - Earthquake loss modelling KW - Spatially cross-correlated ground motion KW - fields Y1 - 2022 U6 - https://doi.org/10.1007/s10518-021-01312-9 SN - 1570-761X SN - 1573-1456 N1 - Update notice Correction to: Epistemic uncertainty of probabilistic building exposure compositions in scenario-based earthquake loss models (Bulletin of Earthquake Engineering, (2022), 20, 5, (2401-2438), https://doi.org/10.1007/s10518-021-01312-9) Bulletin of Earthquake Engineering, Volume 20, Issue 5, Pages 2439, March 2022, https://doi.org/10.1007/s10518-022-01340-z VL - 20 IS - 5 SP - 2401 EP - 2438 PB - Springer CY - Dordrecht ER - TY - JOUR A1 - Hermkes, Marcel A1 - Kühn, Nicolas M. A1 - Riggelsen, Carsten T1 - Simultaneous quantification of epistemic and aleatory uncertainty in GMPEs using Gaussian process regression JF - Bulletin of earthquake engineering : official publication of the European Association for Earthquake Engineering N2 - This paper presents a Bayesian non-parametric method based on Gaussian Process (GP) regression to derive ground-motion models for peak-ground parameters and response spectral ordinates. Due to its non-parametric nature there is no need to specify any fixed functional form as in parametric regression models. A GP defines a distribution over functions, which implicitly expresses the uncertainty over the underlying data generating process. An advantage of GP regression is that it is possible to capture the whole uncertainty involved in ground-motion modeling, both in terms of aleatory variability as well as epistemic uncertainty associated with the underlying functional form and data coverage. The distribution over functions is updated in a Bayesian way by computing the posterior distribution of the GP after observing ground-motion data, which in turn can be used to make predictions. The proposed GP regression models is evaluated on a subset of the RESORCE data base for the SIGMA project. The experiments show that GP models have a better generalization error than a simple parametric regression model. A visual assessment of different scenarios demonstrates that the inferred GP models are physically plausible. KW - Gaussian Process regression KW - Epistemic uncertainty KW - Aleatory variability KW - Empirical ground-motion models KW - Bayesian non-parametrics KW - GMPE KW - Generalization error Y1 - 2014 U6 - https://doi.org/10.1007/s10518-013-9507-7 SN - 1570-761X SN - 1573-1456 VL - 12 IS - 1 SP - 449 EP - 466 PB - Springer CY - Dordrecht ER - TY - JOUR A1 - Douglas, John A1 - Akkar, Sinan A1 - Ameri, Gabriele A1 - Bard, Pierre-Yves A1 - Bindi, Dino A1 - Bommer, Julian J. A1 - Bora, Sanjay Singh A1 - Cotton, Fabrice Pierre A1 - Derras, Boumediene A1 - Hermkes, Marcel A1 - Kuehn, Nicolas Martin A1 - Luzi, Lucia A1 - Massa, Marco A1 - Pacor, Francesca A1 - Riggelsen, Carsten A1 - Sandikkaya, M. Abdullah A1 - Scherbaum, Frank A1 - Stafford, Peter J. A1 - Traversa, Paola T1 - Comparisons among the five ground-motion models developed using RESORCE for the prediction of response spectral accelerations due to earthquakes in Europe and the Middle East JF - Bulletin of earthquake engineering : official publication of the European Association for Earthquake Engineering N2 - This article presents comparisons among the five ground-motion models described in other articles within this special issue, in terms of data selection criteria, characteristics of the models and predicted peak ground and response spectral accelerations. Comparisons are also made with predictions from the Next Generation Attenuation (NGA) models to which the models presented here have similarities (e.g. a common master database has been used) but also differences (e.g. some models in this issue are nonparametric). As a result of the differing data selection criteria and derivation techniques the predicted median ground motions show considerable differences (up to a factor of two for certain scenarios), particularly for magnitudes and distances close to or beyond the range of the available observations. The predicted influence of style-of-faulting shows much variation among models whereas site amplification factors are more similar, with peak amplification at around 1s. These differences are greater than those among predictions from the NGA models. The models for aleatory variability (sigma), however, are similar and suggest that ground-motion variability from this region is slightly higher than that predicted by the NGA models, based primarily on data from California and Taiwan. KW - Strong-motion data KW - Ground-motion models KW - Ground-motion prediction equations KW - Style of faulting KW - Site amplification KW - Aleatory variability KW - Epistemic uncertainty KW - Europe KW - Middle East Y1 - 2014 U6 - https://doi.org/10.1007/s10518-013-9522-8 SN - 1570-761X SN - 1573-1456 VL - 12 IS - 1 SP - 341 EP - 358 PB - Springer CY - Dordrecht ER -