Decades of data has revealed how the bed of New Zealand's largest lake has been constantly rising and falling – something scientists attribute to moving magma within the enormous volcano beneath it.
A just-published study, based on trove of measurements spanning 42 years, has found areas north of Lake Taupō have sunk by about 14cm over that time – while the lakebed near Horomatangi Reefs has been uplifted by around 16cm.
But scientists stress Taupō is no closer to blowing again.
The huge, hidden caldera at the centre of the North Island – regarded as the world's most frequently active supervolcano system - has caused some of the largest eruptions in our planet's history.
Lake Taupō essentially fills the hole left by one of those monster blows - the Oruanui eruption, around 25,400 years ago, which spewed more than 1100 cubic km of pumice and ash into the atmosphere that travelled as far as Antarctica.
Taupō's most recent major episode - a devastating event 1800 years ago - fired out more than 120 cubic km of pumice and ash and obliterated the surrounding landscape.
With that hazard in mind, in the 1970s, an experimental project overseen by volcano geologist George Grindley, of the former NZ Geological Survey, was set up to measure vertical ground movements at fixed points around the lake.
Beginning in 1979, surveyor Peter Otway and colleagues developed a novel surveying technique that used a small portable gauge, hung from 22 fixed points around the shore and from a specially-constructed platform on Horomatangi Reef to measure changes in lake levels.
By correcting for the current level of the lake itself, any changes in ground elevation at that point could be measured within eight millimetres of accuracy - and four surveys have been carried out each year since.
This scheme, now part of the wider GeoNet monitoring system, happened to be established just in time to capture changes that coincided with a swarm of quakes around the Kaiapo Fault in 1983.
"I simply couldn't believe my luck in being in the right place at the right time to catch the very small movement before any earthquakes had begun," said Otway, now retired from GNS Science.
"When a lady in Kinloch rang the office to say she had been nearly sloshed out of her bath by an earthquake - soon followed by a flood of earthquakes reports - I was delighted to see that the lake levelling surveys had apparently made the first detection in New Zealand of pre-earthquake deformation."
Yet, rather than just foreshadowing the rupture of the Kaiapo fault – which would have been a hugely exciting prospect for earthquake prediction - it soon became clear the 50mm of deformation observed was instead tied to volcanic activity beneath the lake.
That was made all the more certain when the survey data revealed more uplift under Rangatira Point, just southwest of Taupō.
Over the next few decades, the network recorded further flurries of movement – showing clear patterns of rising and sinking across areas of the lake floor.
"We found that the ground surface at Taupō behaves in two distinct ways," explained study co-author Dr Finn Illsley-Kemp, a volcano seismologist at Victoria University.
"North of the lake, in the Kinloch area, we find that the ground is subsiding with a total of 140mm in 42 years.
"We believe this is caused by the Taupō rift, which causes the major faults in this area and is acting to stretch the crust apart."
The second pattern was seen within the caldera in the northeast of the lake, near Horomatangi Reefs.
In 2019, an analysis of more than 7000 recorded quakes helped them pinpoint a hidden magma chamber that was at least 20 per cent molten, and spanning roughly 250 cubic km in volume.
"[Near the reefs] the ground is rising, and does so in short spurts, we believe this is the result of magma being injected into the magma reservoir that we highlighted in our study of unrest in 2019," Illsley-Kemp said.
The data from the lake levelling survey was unique, he said, as it showed how the ground was shifting on a timescale much longer than we've had modern GPS measurements.
"I was surprised that the area above the magma reservoir has been uplifting near constantly over the last 42 years," he said.
"This raises the question, how long has this been happening? When will it stop? And what does this tell us about the potential for future eruptions?"
Unlike in 1970s, Illsley-Kemp pointed out that scientists could draw on real-time monitoring of ground deformation through GPS sensors.
"These show really interesting patterns and we have recently found evidence that the volcano may respond to external events, such as the Kaikōura earthquake," he said.
"But there are still so many things we don't understand about Taupō."
Illsley-Kemp and fellow scientists at Victoria, GNS Science and Massey University are gradually building a clearer picture of the volcano through the five-year, $8.2m MBIE-funded ECLIPSE project.
For Otway, the long dataset, along with satellite and seismic monitoring, reflected the fact Taupō was indeed an active caldera volcano.
"However, the scheme also suggests this type of activity has not increased during the last four decades, or possibly for much longer," Otway added.
"We would expect the inflation to accelerate rapidly before the next eruption - so it is essential that the deformation continue to be closely monitored."