Scientists have simulated how Auckland could recover from a major volcanic eruption within the city - costing industries billions of dollars and transforming entire residential areas.
Auckland straddles a large volcanic field home to more than 50 volcanic centres, posing an ever-present hazard to the city's 1.6m-strong population.
While its existing volcanoes are thought unlikely to erupt again, the Auckland Volcanic Field is young and potentially active - with the potential to bring fast-moving surges of hot rock and gas, and widespread ashfall.
Recent research has suggested signs of an imminent eruption could occur with anywhere between five and 15 days' warning, forcing the evacuation of perhaps more than 400,000 people from their homes.
In a new study, researchers used sophisticated modelling tools to explore the social and economic fall-out of a mock eruption in South Auckland.
They also looked at how a disaster might drive changes in local land use and infrastructure under different scenarios over 20 years – and how they compared with status quo projections.
Their simulated event, taking place in 2020, began with weeks of earthquakes and gas emissions around Māngere, before a steam-driven eruption burst from a freshly-opened vent amid the suburb, launching a pyroclastic surge and a 3km-high ash column.
More eruptions followed over the next few weeks - some launching rocks, fountaining magma, lava flows and more ash and gas - with activity fully ceasing after three months.
The eruptions caused an initial economic shock and forced around 80,000 people to evacuate the region.
Under a "quick recovery" scenario, about 30,200 people were able to return within two months, with their homes repaired enough to be able to live in again.
After about 15 months, water and electricity networks, which provided services to areas inside and outside of the eruption impact area, were fully restored.
The researchers estimated the impact on Auckland's industrial sector in the first year would be around $5.2b, along with $4.1b for the commercial sector.
Over a 10-year period, however, the commercial sector faced a $25.5b toll, compared with $12b for industries - something owing to a lengthy impact on operating surpluses due to new capital having to be chanelled into restoration.
Their modelling found some sectors would fare worse than others.
Local finance and insurance sectors would take a $560m hit over two years, for instance, while rental, hiring and real estate services could suffer a $586m loss.
To analyse what the disaster meant for land use and population growth, the researchers split Auckland into three broad blocks - the "eruption zone" around the Māngere site and zones north and south of it.
If no disaster happened, around 84,700 people would be living within the eruption zone by 2040, compared with about 68,200 today.
Under the modelled eruption scenario, however, there'd be around 53,000 residents in the area 20 years after the event if recovery from the disaster was fast - or just 24,700 if it was slow.
There were also big implications for areas outside it: the "south zone" would contain between 30 and 32 per cent of Auckland's population by 2040, compared with 26 per cent with no disaster, while the north would contain around 67 or 68 per cent, compared with 70 per cent normally.
"The results also illustrate the possibility of dramatic changes in urban form development, particularly around where industrial and commercial centres would develop," said study lead author Robert Cardwell, a Market Economics senior analyst and a University of Auckland doctoral candidate.
That could bring "significant implications" for Auckland's urban form well outside of the immediate eruption impact area, he said.
"For example, our simulations demonstrate how taking longer to remediate the area directly impacted by the volcanic eruption can result in more urban development on the urban-rural boundaries than might have otherwise occurred."
Within the eruption zone, only between 60ha and 120ha of industrial land would remain by 2040, compared with 340ha otherwise.
The modelling also showed a disaster would eventually result in much more of this type of land to the south by 2040 - and slightly less to the north.
Study co-author and University of Auckland volcanologist, Professor Jan Lindsay, said the modelling was only the first demonstration of its kind, and came with many assumptions.
"A different vent location might have led to quite different results," she said.
"It should therefore be seen as a first cut, from which such modelling capabilities can be developed further to make them more meaningful and robust."
Nonetheless, the authors said, such modelling could help planners take a longer view of natural disasters like eruptions.
"The models can help planners understand and evaluate the relative impacts of and relationships between things like the initial evacuation and people leaving the region," said co-author Associate Professor Liam Wotherspoon, of the University of Auckland.
"There are also factors like reconstruction costs and land rehabilitation, and the long term impact on the attractiveness of Auckland to foreign investment and high-skilled workers.
"Some of the next stages of development would be to develop a better understanding of the factors driving these outcomes."
Another author, Market Economics director Garry McDonald, said the tools used in the study could also enable alternative recovery strategies to be tested.
"We don't have to wait until a volcanic eruption occurs to get an idea of how long-term land use and economic development could play out for each of the possible recovery strategies we could implement."
The study, carried out as part of the collaborative Resilience to Nature's Challenges National Science Challenge, has been published in the Journal of Volcanology and Geothermal Research.
• To learn more about research into city's eruption risk, visit the Determining Volcanic Risk in Auckland (DEVORA) project's website.
The death of the 19-year-old has been ruled as accidental.