| Abstrak/Abstract |
Tephra from large explosive eruptions can cause damage to buildings over wide geographical areas, creating a variety of issues for post-eruption recovery. This means that evaluating the extent and nature of likely building damage from future eruptions is an important aspect of volcanic risk assessment. However, our ability to make accurate assessments is currently limited by poor characterisation of how buildings perform under varying tephra loads. In this study we use remotely assessed building damage to generate new building vulnerability models, estimate the total number of buildings damaged and their repair costs, using the Kelud 2014 eruption as a case study. A total of 1154 buildings affected by falls 1-10 cm thick were assessed, with 790 showing signs that they sustained damage in the time between pre- and post-eruption satellite image acquisitions. Only 27 of the buildings surveyed appear to have experienced severe roof or building collapse. Damage was more commonly characterised by collapse of roof overhangs and verandas or damage that required roof cladding replacement.
To estimate tephra loads received by each building we used Tephra2 inversion and interpolation of hand-contoured isopachs on the same set of deposit measurements. Combining loads from both methods with our damage assessment, we develop the first set of fragility curves that consider damage severities lower than severe roof collapse. These curves were applied across two separate building exposure inventories to estimate the total number of buildings damaged during the 2014 eruption. The four possible combinations of hazard and exposure produced ranges from 37,000 – 87,000 buildings, due to large discrepancies between the two exposure datasets and small (<1>20,000 buildings are likely to collapse in 25% of future VEI5 eruptions with repair costs > $650 million USD. This study highlights that damage and loss estimates are subject to large variability as uncertainty is propagated through different stages of the assessment. This suggests the need for standardised, comprehensive approaches to be developed and adopted in the future.
|
