POZZUOLLI, Italy—Tragedy struck at the Solfatara volcano crater north of Naples a few weeks ago. An 11-year-old boy climbed over a low, wooden fence, ventured onto the chalky moonscape, and fell into an open fissure. His parents frantically tried to pull him out, just as the hollow floor of the crater crumbled, sending them all to their deaths in a gaseous pit of boiling gray mud, as the family’s seven-year-old boy watched in horror. The accident was a freak occurrence, responsible for the only recorded deaths on this crater in centuries. But those hot gases hold an ominous story: Solfatara is part of the massive Phlegrean Fields, a threatening supervolcano experts agree could begin erupting anytime.

Phlegrean Fields—a network of two dozen craters and other volcanic structures—is especially worrisome because the nested craters stretch under the Gulf of Naples, right alongside the greater Naples metropolitan area, home to more than three million people. Much like Yellowstone in Wyoming and Long Valley in California, Phlegrean Fields is classified as a supervolcano that, when (not if) it next erupts, could cause a catastrophic global event. Its eruption 200,000 years ago—the largest ever in what is now Europe—darkened skies worldwide, causing a terrible volcanic winter. An eruption 40,000 years ago may have wiped out Neanderthals, according to a 2010 report that is hotly debated. Phlegrean Fields last blew in 1538, an eight-day eruption that created the 440-foot Monte Nuovo volcano on the coast.

Even if Phlegrean Fields takes another 500 years to erupt, it offers a unique research opportunity because of its particular style of unrest. Prior to and after the 1538 eruption the ground in this highly populated area experienced very slow, vertical oscillations known as bradyseism, which do not exist to the same extent in any other caldera in the world. In the 1970s these movements became strong again—the ground rose and fell 10 to 15 centimeters (about four to six inches) a year. Volcanologists were concerned the complex might be reaching another eruptive phase. The fear became worse after bradyseism caused the ground to lift an incredible 110 centimeters (43.3 inches) in 1982, and the movements then stopped almost entirely in 1983. But Phlegrean Fields did not erupt.

In 2016 the volcano was upgraded from green to yellow on a scale that continues up to orange and then red, which indicates an imminent eruption. Scientists last declared it “red” in 1983 after roads, houses and a hospital were destroyed by the sudden 110-centimeter rise. More than 30,000 people had to be evacuated. Whether or not the odd, current silence means a threat is imminent is open to scientific interpretation, just as changes at other volcanoes worldwide are. The latest studies provide some intriguing suggestions, and offer insights about possible eruptions elsewhere on Earth.

Part of determining just when Solfatara or another portion of Phlegrean Fields might erupt is understanding what is under the crust. That has been difficult to ascertain with certainty because most of the crater is covered by a one- to two-kilometer-thick caprock. Several recent studies paint different pictures of just what lies below this protective lid. Luca De Siena, an expert in geophysics and volcanology at the University of Aberdeen in Scotland, thinks there are reservoirs of static, thick magma about a kilometer long and five to 10 meters thick that last migrated between ancient cavities around the 1980s. Tiziana Vanorio, director of Stanford University’s Rock Physics Laboratory, instead thinks the caldera’s cavities have less structure and hold hot, fluid liquids and gases that influence the stress and strain on the caprock.

Both of the Italian experts grew up on the Phlegrean Fields caldera and are happy to credit each other’s research, even as they disagree with it. If the next eruption is explosive like in the past, rather than effusive (which would be relatively gentler), Vanorio, who was evacuated after the 1982 unrest, says what is under the caprock really will not matter for people in cities such as nearby Pozzuoli. “Whether the eruption involves viscous magma containing sufficient gas dissolved under pressure or hot mudflows and ash-laden gas clouds, the town of Pozzuoli and its surroundings would not be subjected to such different risk exposures,” she says. In other words, events resulting from conditions predicted by either hypothesis would be equally catastrophic.

Phlegrean Fields offers an intriguing research opportunity because companies drilled almost a dozen geothermal wells there prior to the 1982 fright. No longer used, the wells now allow extensive monitoring, provide direct data about underground characteristics such as temperature and allow researchers to drill cores into the deep interior. “Having cores is like having a view through a window to a landscape that has been long-covered,” Vanorio says. Techniques used in Phlegrean Fields are already being applied at other major volcanoes, including Mount Saint Helens in the U.S., Deception Island in Antarctica, Mount Teide in the Canary Islands and Mount Aso in Japan.

And yet, the debate about what exactly is under the caprock of this supervolcano stems from lack of access beyond the geothermal wells. In 2010 funding was made available for a deep-drill project that would have penetrated all the way through the caprock to see what was below. But the Neapolitan mayor, backed by a handful of experts, stopped the plan based on fears that drilling might cause an eruption. By the time a new city administration green-lighted drilling again several years later, the funds had been sent elsewhere. Some researchers are trying to revive the project.

Unfazed, De Siena and a group of experts published a study in August that explores the dramatic unrest in the 1980s and sets out new methods to measure and analyze even the subtlest movements and sounds. He hopes that new techniques will take some of the guesswork out of modeling eruptions. One approach would be to monitor underground water, underground rocks, and oil and gas fields together, providing unified data that go beyond studying such factors independently.

All of this could make it safer for populations that live on volcanoes around the world. “Unlike earthquake prediction, scientists can actually do something about volcanoes,” Vanorio says. “Though each volcano is unique and may change its eruptive behavior at any time, there are trends in precursors. That helps scientists get better at predicting eruptions.”

Despite new tools and renewed enthusiasm for useful results, scientists still cannot say how soon Phlegrean Fields might erupt. “Are we in danger?” De Siena asks. “The answer is always yes. There are more than three million people there. Any eruption of any sort is going to be dangerous.”