Scientists have moved closer to creating a realistic picture of what types of eruptions could unfold around New Zealand's biggest city.
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 and its billions of dollars of property and infrastructure.
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 studies have 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.
Now scientists working in concert with the Determining Volcanic Risk in Auckland (Devora) project have taken a much deeper look at the hazard, by combining two approaches that together enable specific risks to be matched to locations.
It's allowed them to more clearly see how areas north of the airport are more at risk of phreatomagmatic eruptions – not too dissimilar to that of Whakaari/White Island's deadly blow in December.
The central city, meanwhile, could be exposed to mainly magmatic effects, like lava fountaining from the ground and lava flows.
It came after Devora researchers put together a series of eight hypothetical yet realistic eruption scenarios around the field.
Those included phreatomagmatic eruptions at Auckland Airport, a magmatic eruption at Mt Eden, and phreatomagmatic events that transitioned to magmatic ones at Otahuhu, Mangere Bridge, Birkenhead and Rangitoto Island.
There was also the reverse scenario of a magmatic eruption transitioning to a phreatomagmatic-style bang at Waitematā Port.
In the Rangitoto Channel, a "Surtseyan" eruption - where seawater floods into an open vent - could see jets of ejected material fired several hundred metres above sea level.
"Each of the scenarios was assigned a specific vent location, so that specific hazards could be modelled according to the topography and environmental setting of the area, and so that consequent impacts could be assessed," University of Auckland volcanologist Professor Jan Lindsay said.
While those scenarios painted a general picture of Auckland's risk, they still only represented eight possible locations across the whole area – and the actual likelihood of any of those eruptions happening at those spots still remained unknown.
"The Devora scenarios got us thinking: wouldn't it be good to critically examine those locations against the proposed scenarios, to make sure the size and style of eruption proposed actually made sense in that location?" Lindsay said.
"And not only that, are there other locations in the AVF where that particular type of scenario is likely to occur as well?
"We thought that would be good to know, because then you could flip the problem and say: 'okay, out of these eight scenarios, which would be the most likely to occur at, say, the airport? Or in the CBD?"
To get their answers, the research team effectively developed a fresh method for assessing hazard spatially in volcanic fields, by assigning relative probabilities, conditional on location, to each of the eruption scenarios.
That involved examining a mix of factors at each site, such as how high the vent sat above sea level, its distance to the harbour and known faults, and what comprised local geology.
"We combined all this in a probability tree for the scenarios, and then produced probability maps of the AVF that show the spatial probability of occurrence of each scenario across the entire AVF," said study co-author Susanna Jenkins, an Assistant Professor at Nanyang Technological University's Earth Observatory of Singapore.
"We found that the locations of two of the Devora scenarios were bang on – in that their scenario matched the most likely scenario to occur at the proposed location."
The team determined the most likely scenario in the flatlands to the north and east of Auckland Airport would be a pure phreatomagmatic eruption, triggered by rising magma interacting with surface or groundwater.
"For the elevated region immediately to the south of the CBD, we were able to say that the most likely scenario is a magmatic eruption – or one that involves fire fountaining of lava and generation of a scoria cone and lava flows," Jenkins said.
Another co-author, Massey University's Professor Mark Bebbington, said the new study could lead to more work in two main areas.
One was expanding the range of possible scenarios to approach a fully probabilistic assessment, rather than the "pseudo-probabilistic" evaluation the team made.
"The second area for further work is the scenario probabilities could be combined with numerical hazard models and exposure and vulnerability information, such as population, infrastructure and cultural assets, to better determine the risk from a possible AVF eruption," Bebbington said.
"Once individual hazard footprints are calculated from each scenario at each location and weighted by the probability of that scenario, an overall hazard map can then be constructed by weighting those results according to a model for vent susceptibility.
"We think this information, which would include spatial dependence of hazard impacts, could be really useful for planners."
He also saw potential to apply the team's new method to other volcanic zones around New Zealand and overseas.
The new study has been published in the Journal of Volcanology and Geothermal Research.