Bertelli, Silvia; (2025) Soil Liquefaction in Urban Areas: an Investigation into Risk Assessment and Modelling for the Insurance Sector. Doctoral thesis (Eng.D), UCL (University College London).

Text Thesis_edits_2025.pdf - Accepted Version Access restricted to UCL open access staff until 1 May 2026. Download (21MB)

Abstract

Liquefaction risk assessment modelling is gaining increased attention in the insurance industry. The consequences of liquefaction have been made evident during the past decade, with devastating events like the Mw6.3 2011 Christchurch Earthquake (New Zealand) costing more than US$15 billion and insurance claims of US$12 billion. The insurance industry was caught unprepared by such events, with catastrophe models underestimating the extent and severity of liquefaction manifestations. Despite the relevance of the loss experienced, the liquefaction risk assessment still represents an underestimated step in the catastrophe modelling process, with significant uncertainties persisting. In particular, the potential loss and risk estimates of spatially distributed infrastructure networks and portfolios of buildings have posed significant challenges to hazard and risk analysts. Indeed, to date, the industry lacks a coherent and straightforward approach, and existing methods do not adequately capture the true extent of liquefaction losses within the insurance and risk modelling sectors. The thesis explores and reviews geotechnical engineering practices existing in the literature from the perspective of the insurance industry needs. State-of-the-art procedures to evaluate the triggering and consequences of liquefaction may not be directly applicable or may need additional considerations when applied to the insurance sector. Interviews with catastrophe model developers, risk analysts, and model vendors in the industry have informed the implementation of a multi-criteria decision-making analysis. This analysis considers leading criteria such as "model performance", "data availability", "data quality", "user skills", "software licenses", and "time", thereby facilitating the business decision on which model to select. Different liquefaction risk assessment procedures – encompassing geospatial approaches, simplified procedures, and numerical and deep learning models - are critically examined, adopting the 2011 Christchurch earthquake as the initial case study for evaluating these models. The proposed analyses also attempt to explain discrepancies between the observed manifestations, simplified methods, and advanced constitutive numerical models when analysing soils with high fines content, like Christchurch's silty sands. The main contribution of this thesis is a decision-making framework for the liquefaction risk assessment methodologies that meet the insurance industry's needs. The derived framework is validated using the 20 May 2012 Emilia Earthquake in Italy as a second case study, a typical instance of an earthquake with liquefaction manifestation where data has been collected and recorded. The proposed framework is designed to adapt to the diverse range and quality of data encountered in catastrophe modelling. Lastly, it is emphasized that this framework offers valuable insights for loss and risk modellers, geotechnical engineers, and researchers, enabling them to more effectively understand, interpret, and model liquefaction, leading to the development of more accurate hazard maps and risk assessments.

Download activity - last month

Export as XMLCSVJSON

Download activity - last 12 months

Export as JSONCSVXML

Downloads by country - last 12 months

Export as JSONCSVXML

Archive Staff Only

View Item