Near the bottom of the Moab Khotsong gold mine in South Africa, scientists aim to drill to the site of a recent earthquake. 


Deep in a South African gold mine, scientists drill for the heart of an earthquake

It’s no easy feat to drill into the faults that cause earthquakes. Intercepting such active ruptures, which are buried kilometers beneath the surface, requires specialized equipment, skilled crews, and a lot of money and time. There are no shortcuts. Well, maybe there’s one: an express elevator in a South African gold mine that runs 3 kilometers straight into Earth.

This week, scientists were set to begin drilling a 750-meter-deep hole from the bottom of the Moab Khotsong mine, located some 180 kilometers southwest of Johannesburg, South Africa. Three years ago, a magnitude-5.5 earthquake erupted from a previously unknown fault beneath the mine. Aftershocks continue. The $1.6 million project, if successful, could reveal the internal workings of what may be a “fresh” fault: one that has experienced its first earthquake. Unsullied by previous quakes, the fault could yield new insights into how quakes unfold—and how humans can inadvertently trigger them.

The international project, dubbed Drilling into Seismogenic Zones (DSeis) and backed by the International Continental Scientific Drilling Program, is the brainchild of several Japanese researchers who have been working in South Africa since the fall of apartheid in 1991. Other efforts to drill into active faults, such as the San Andreas in California or the Alpine in New Zealand, have cost tens of millions of dollars and yielded one or two cores. In South Africa’s mines, the researchers saw a quick and relatively cheap way to get a detailed look at the fracturing rock that generates earthquakes. “In Japan, we couldn’t easily access the seismic source,” says Yasuo Yabe, a seismologist at Tohoku University in Sendai, Japan, fresh from a trip down the mine. “Here people are going in and out daily.”

DSeis was already planning drilling projects in several other mines when the magnitude-5.5 Orkney earthquake struck beneath Moab Khotsong on 5 August 2014, killing one person. The event kicked the project into high gear. It was the country’s largest quake in nearly 50 years. South Africa’s gold mines experience frequent small quakes triggered by their explosive excavations. But this one was far stronger and deeper; unlike most induced quakes it was not on the same level as the mine. The fault ran perpendicular to other known faults. And its rocks slid past each other horizontally—a strike-slip motion—rather than vertically, as in most smaller quakes.

These mysteries worried the mine companies, who need to plunge deeper for more gold and don’t want seismic surprises. They were receptive to a pitch from a longtime collaborator, Hiroshi Ogasawara, a seismologist at Ritsumeikan University in Kusatsu, Japan, to add the Orkney fault to the list of DSeis targets.

Most campaigns to drill into earthquake zones have targeted faults like the San Andreas, which has seen thousands or millions of earthquakes. But there’s good reason to believe, given its quiet tectonic environment, that the Orkney earthquake came from a fault that has ruptured just once or, at most, a few times, says DSeis collaborator William Ellsworth, a geophysicist at Stanford University in Palo Alto, California. Such a fault would lack the “rock flour” that builds up in active faults as they grind away and makes it hard to discern how an individual earthquake divides its energy into seismic waves, heat, strain, and pulverizing rock. “We have a tremendous opportunity to get a look at what is essentially an ordinary earthquake,” Ellsworth says.

The mine is a challenging work environment. Moab Khotsong claims to be home to the world’s deepest mine shaft: a 3000–meter descent at some 15 meters per second in a shaking cage, with only the arcs of head lamps piercing the darkness. At the drilling level, 95 floors down, it’s a 10-minute ride in a rail carriage to the drill site. It can be eerily quiet, and the air carries an acrid scent of burnt rock and ammonia from recent dynamiting. “It’s a smell that’d be associated with Hades,” says Tullis Onstott, a geomicrobiologist at Princeton University who has also joined the project.

Onstott’s goal is to learn whether earthquakes can favor the microbes that live in deep rock. In an experiment he’s been trying to do for a decade, ever since it failed at a San Andreas drill site, he will install an automated sampling system, triggered by seismicity, in the borehole at the fault. Past work has shown that an earthquake can release a pulse of hydrogen gas, which might have been trapped in the rock or generated by chemical reactions caused by the fracturing. Onstott hopes, as more aftershocks strike, to discover a cascade of microbial populations feeding off this chemical energy. If so, similar faults could become targets for a search for life on Mars, which might influence the selection of a landing site for NASA’s Mars 2020 rover mission, Onstott says.

Mine operators might also get what they are seeking: a clearer picture of how human activity can trigger quakes. In the United States, the underground disposal of wastewater from oil and gas drilling is known to trigger quakes by boosting pressure in the rock pores. Pore pressure is unlikely to be a factor in South Africa, says Ze’ev Reches, a geophysicist at the University of Oklahoma in Norman. But he notes that the Orkney quake, like some of the largest induced quakes in Oklahoma, occurred deep in basement rock along a previously unmapped fault. (Reches is planning to drill into a fault in Oklahoma, but he’ll have to go 4.5 kilometers down.)

Plenty can still go wrong for DSeis. In May, operators canceled a drill site at another mine that the project planned to use, citing an increased risk of falling rocks. But Ogasawara says they have plenty of candidate earthquake faults for their campaign, including small, shallow ones accessible from nearby mines, which may allow them to target the very heart of an earthquake: the hypocenter, the exact place where a fault first starts to give way. “That is what we want to see,” Ogasawara says. “We’re expecting to show what a hypocenter looks like.” 

Corrections, 1 June, 1:42 p.m.: This story has been updated to correct several errors: The Orkney quake killed a resident of a nearby town, not a mine worker. In May, the project was ruled out from using another mine due to falling rocks, but the mine stayed open. And Yasuo Yabe is a seismologist, not geologist.