Analysis of lava flow impacts for use in risk assessments

• Main Author: Meredith, Elinor Siwan
• Other Authors: Susanna Jenkins
• Format: Thesis-Doctor of Philosophy
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Science::Geology::Volcanoes and earthquakes
• Online Access: https://hdl.handle.net/10356/169501
• Institution: Nanyang Technological University

Growing populations in volcanically active regions place communities at risk from volcanic hazards. Some volcanic hazards leave little deposit but can kill everyone in their path, while others move slowly enough to avoid causing fatalities but leave thick solidified deposits that can render land unusable for decades. This thesis is concerned with improving our knowledge of damage to the built environment associated with the latter, lava flows. The best source of data for understanding potential damage from lava flows is empirical impact data collected during past events; however, buildings or infrastructure in contact with lava have commonly been assumed to be completely destroyed. This means that tools used to forecast impact (such as fragility functions relating hazard intensity to damage probability) are binary: in contact with lava results in complete destruction, not in contact with lava results in no damage. Few building-level lava flow damage assessments exist, resulting in a lack of understanding of how structures interact with lava, or the frequency of these events. Studies that have evaluated individual structures present examples of structures exhibiting resistance to lava flows, with structures damaged but not destroyed. In this thesis, I argue that the impacts from lava flows are not always binary and develop methods that assess non-binary lava flow damage to structures, assess recent eruption impacts, and use the results to provide tools that help forecast future lava flow damage. To start, I compiled the first dataset of past lava flow impact events to the built environment, to quantify event frequency, and to assess temporal and spatial trends of events. Next, I developed the first damage-state schema for structures impacted by lava flows, and apply this using ground surveys and aerial imagery to quantify the range of structure damage from the 2018 LERZ lava flows of Kīlauea volcano, Hawai‘i. Finally, for those events with available building-level hazard and damage data: 2021 Cumbre Vieja lava flows, La Palma, Spain, and 2014-2015 Fogo lava flows, Cabo Verde, I conducted damage assessments to provide damage states for over 9,207 structures. Along with the 2018 Kīlauea assessment, I used this compiled dataset to assess relationships between lava flow thickness as a proxy for hazard intensity and damage severity, presenting the first empirically-derived and non-linear fragility functions for lava flow structure damage, for different structure types. The findings show a recent increase in recorded impact events, and that these events occur more frequently than previously thought, with almost three impact events, to population centres, per decade in the past 100 years. With the collected empirical building-level data, I found all structures were destroyed at >6 m lava thickness, while there was a range of damage severity at <6 m lava thickness, dependent on structure type. This suggests that structure typology is important, with concrete and cylindrical structures showing resistance to thin flows. I show that damage extended beyond the flow channel in breakout flows and overflows, and extended up 600 m away from the lava flow margin due to the fire spread at Kīlauea, extending potential impact beyond the modelled channel footprint. Out of those structures impacted across the three case studies, 6% are classified as damaged (and not destroyed), showing that the binary impact assumption is an oversimplification. This thesis provides empirical data, and related tools and collection guidelines that enhance our understanding of building vulnerability to lava flows and support more robust forecasting of building damage from lava flows.