Quantification of model uncertainty in debris flow vulnerability assessment

Autor: Unni Eidsvig, Bjørn Vidar Vangelsten, Maria Papathoma-Köhle, Thomas Glade, J. Du
Rok vydání: 2014
Předmět:
Zdroj: Engineering Geology. 181:15-26
ISSN: 0013-7952
DOI: 10.1016/j.enggeo.2014.08.006
Popis: An appropriate treatment of uncertainties constitutes an important part of the risk assessment process. According to ISO 31000:2009 , risk is the “effect of uncertainty on objectives”. In order to quantify uncertainties in a risk estimate, the ambiguity in hazard, vulnerability and elements at risk need to be quantified. This paper presents a methodology to quantify uncertainty in vulnerability assessment for buildings affected by debris flow, focusing on model uncertainty. Two models for vulnerability assessment expressing the vulnerability of buildings to debris flow and the corresponding model uncertainty are proposed: The first model describes the degree of loss as a function of intensity representing uncertainty by uncertainty bands. Uncertainty bands give a good overview of the uncertainty in the model and are easy to understand by stakeholders. The second model describes the probability of exceeding different damage states as a function of intensity. This model is less intuitive, but more suitable for use in a probabilistic vulnerability assessment, as it provides a probability distribution of damage for all values of intensity. The models were developed using empirical data on debris height and degree of loss of buildings following a debris flow event in the Martell valley in South Tyrol, in August 1987. The proposed models were developed for typical South Tyrolean buildings. Although the results of the study are not transferable to other areas in the world, the model development procedure may be applied for other types of architecture and building types in different locations. The paper demonstrates the value of the proposed models as essential tools for both simple and advanced treatment of uncertainties associated with vulnerability and potential losses.
Databáze: OpenAIRE
Externí odkaz: