On Jan. 17, 2001, Hutchinson, Kan., awoke to an apocalypse.

Gas that had silently collected inside a downtown appliance store ignited, reducing two buildings to tinder carcasses and shattering windows for blocks.

Three miles away, a geyser of gas shot out of the earth, sending mud and rocks 30 feet into the air. Elsewhere, the ground popped open like the rotten hull of a boat, spraying brown briny water or catching fire.

The next morning, just when the earth seemed to recover its temper, a new plume of gas and water shot through the floor of a mobile home, killing two people. Hundreds of other Hutchinson residents were evacuated from their homes, many for months.

The mysterious disaster claimed national headlines, but there was little public discussion of the fact that it was caused by problems with underground injection wells.

Among a small community of geologists and regulators, however, the explosions in Hutchinson — which ranked among the worst injection-related accidents in history — exposed fundamental risks of underground leakage and prompted fresh doubts about the geological science of injection itself.

Geologists in Hutchinson determined that the eruptions had sprung from an underground gas storage field seven miles away. For years, a local utility had injected natural gas between 600 and 900 feet down into old salt caverns, storing it in a rock layer believed to be airtight so that it could later be pumped back out and sold. The gas had leaked out and migrated miles into abandoned injection wells once used to mine salt, then shot to the surface.

“It was an unusual event,” said Bill Bryson, a member of the Kansas Geological Survey and a former head of the Kansas Corporation Commission’s oil and gas conservation division. “Nobody really had a feeling that if there was a leak, it would travel seven miles and hit wells that were unknown.”

Though regulated under different laws than waste injection wells, gas storage wells operate under similar principles and assumptions: that deeply buried layers of rock will prevent injected substances from leaking into water supplies or back to the surface.

In this case the injected material had done everything that scientists usually describe as impossible: It migrated over a large distance, travelled upward through rock, reached the open air and then blew up.

The case, described as “a continuing series of geologic surprises and unexpected complexities” by the Kansas Geological Survey, flummoxed some of the leading injection experts in the world.

Perhaps more troubling was that some of the officials assumed to be most knowledgeable about injection wells and the risks of underground storage seemed oblivious to the conditions that led to the accident.

“The existence of those widespread formations and old salt-solution wells was unknown to the operators of the storage facility, the Kansas State Geologic Survey, city personnel, and its inhabitants,” noted a 2006 paper authored by Sally Benson, a leading geoscientist at Lawrence Berkeley Lab’s earth sciences division, and others. “It is still not clear how long the leakage occurred.”

Bryson agrees that officials should have known more about the number of abandoned wells in the area, but he says that otherwise Kansas’ regulations worked as intended.

The cause of the accident was identified because workers were diligently monitoring pressure changes in the gas injection well, as they are required to do. Once in a while, accidents are going to happen, he said.

“How far do you go to make sure that nothing will ever happen?” he said. “Lets face it: Something is going to go wrong… states have to be trusted enough to let us deal with that.”