Revisiting fire following earthquake modelling for Wellington City

(Inc. GST)
(Ex. GST)
Write a Review

Scheele F, Lukovic B, Horspool NA. 2019. Revisiting fire following earthquake modelling for Wellington City. Lower Hutt (NZ): GNS Science. 27 p. (GNS Science report; 2019/24). doi:10.21420/QYQ1-RE79.

Fires are a common secondary hazard following earthquakes and, on rare occasions, can develop into major events with severe consequences. Wellington City has many characteristics that make it susceptible to fire following earthquake (FFE), including the potential for conflagrations and significant losses (e.g. property, infrastructure and casualties).
Through the ‘It’s Our Fault’ research programme, we are revisiting FFE modelling for Wellington City, building on the previous models developed by Cousins et al. (2002). We have developed a relative FFE risk map, which identifies areas that may warrant increased attention. These are locations where both ignition and fire spread are more likely; however, FFE could occur anywhere in Wellington City.
Through a literature review of globally available models, we have identified and applied an ignition model that enables the probability of ignition to be estimated for individual buildings, which is essential for fire spread modelling.
The fire spread model now includes both wind speed and direction, which is a significant improvement on the previous static model (Cousins et al. 2002). Stochastic modelling utilising both the spread and ignition models and varying input parameters can produce loss and impact results across many potential scenarios. One hundred simulation runs of a Wellington Fault magnitude 7.5 earthquake scenario were performed for this study, which provide a preliminary indication of the distribution of loss and impacts for FFE in Wellington City. The average number of ignitions per scenario was 34.
The stochastic modelling results show a lognormal distribution of losses. The median scenario loss is $1.49 billion (ranging between $0.55 billion and $17.65 billion at 95% confidence); however, nearly 20% of scenarios had estimated losses exceeding $4 billion. A median loss for a single burn zone (the extent of fire spread following an ignition) of $24 million (ranging between $0.5 million and $670 million at 95% confidence) indicates the potential benefit of suppressing a single ignition.
Examination of the relationship between wind speed and loss shows a strong correlation.
For wind speeds over 25 km/h, an increase of 10 km/h in wind speed is associated with ~$4 billion of increased loss. This indicates that by considering the weather conditions following an earthquake event, a rapid assessment of the threat from FFE could be undertaken.
Future work will involve refinement and validation of the modelling and processing a greater number of simulation runs to produce stable results. The FFE modelling is informed by frequent feedback from key stakeholders, and future efforts will continue to be directed towards improving the understanding of FFE risk in Wellington City and supporting risk reduction initiatives. The refined model will be integrated into the RiskScape 2.0 software.