The world's largest mud volcano, which left 13,000 families homeless, is likely to continue erupting for another 26 years, researchers have estimated.
It first erupted back in May 2006, and - at its peak - was spewing 180,000 cubic metres of mud a day, equivalent to 50 Olympic-sized swimming pools.
The volcano, in East Java, Indonesia, has buried homes, schools and farmlands over seven square kilometres.
The findings have been published in the Journal of the Geological Society.
This is the first reliable estimate on how long Lusi (derived from the Indonesian word for mud, lumpur, and the place where the eruption occurred, Sidoarjo) would continue erupting.
Co-author Richard Davies, a geologist from Durham University's Department of Earth Sciences, said there were a number of factors that meant it was not possible to produce an estimate until now.
"We did a provisional estimate in 2008, but we have significantly improved the methodology," he told BBC News. "Also, for two or three years there was a lot of debate about what caused it.
"I think there were a lot of people involved in working what happened - so now we have seen the natural progression from wanting to know what caused it to wanting to know how long it will last."
The cause of the eruption was a subject of a prolonged scientific debate, with opinions divided on whether it was the result of a drilling well in the area, or a magnitude 6.3 earthquake that shook the area a few days earlier.
In 2008, during a conference in South Africa, supporters of both hypotheses presented their arguments before a panel of independent experts.
The debate was chaired by Edinburgh University's Professor John Underhill, who was also a top level football referee. The majority of experts, 42 out of 74, favoured the drilling explanation.
Professor Davies supported the drilling hypothesis: "There is a lot of evidence now that shows it was caused by drilling - there was a blowout that was not controlled."
In 2008, the company that was drilling in the area and blamed for triggering the volcano agreed to pay compensation to the 50,000 displaced people. However, it did not say the drilling activities were the sole cause of the volcano.
Professor Davies added that the team now had confidence to publish an estimate because they also had four year's worth of data on how much material was coming out of Lusi, allowing them to calibrate the computational model developed by co-author and fellow Durham University researcher Simon Mathias.
He explained that the volcano was driven by carbonated water in an aquifer, estimated to be about 2.5-3.5km beneath the surface, that travelled up through the drilled borehole and passed through a layer of muddy material (known as the Upper Kalibeng) before erupting on the surface via a 50-metre central vent.
The inputs for the model of what was happening underground came from two wells; one located on the site of the central vent, and another located about 6km away, where there was a natural, smaller mud volcano.
"This gave us the pressure in the actual aquifer," explained Professor Davies.
"Then we considered factors such as the permeability and porosity of the rock to estimate how long would it take for fluid pressure to decline until no more fluid would come out of the vent.
"It is a fairly standard methodology, but it has never really been applied to mud volcanoes before."
He added that the team used a lot of different probabilities and combined them, creating a huge number of outputs (called realisations).
"Out of the 10,000 realisations, we actually rejected a large number because we wanted to fit the actual rate which the mud had come out of the volcano to fit those data points," he said.
They ended up with 381 realisations, which allowed them to reach the estimate of 26 years.
But Professor Davies added: "Within the modelling, there is a 10% chance that it will last for more than 100 years, and there is a 90% chance that it will last longer than 10 years."
The team worked on the assumption that the aquifer was not being recharged from another source, so there was no way for pressure to return once it had been discharged.
"If there is recharge taking place then the 26-year estimate will be too short," he conceded. "Basically, does Lusi have any surprises left for us."