For some seismologists, the deadly tremor that hit Nepal on Saturday came as little surprise.
"It's been well known in the seismological world that the Himalayan system is a very dangerous system," Tom Heaton, a seismologist at the California Institute of Technology, told VICE News. "If you asked seismologists around the world where is a particularly hazardous place, that would be one of them."
But knowing that some areas are more likely to have quakes doesn't mean seismologists can predict them. Earthquake forecasting is still a vague and imprecise science. And pinpointing who will get one — when — is barely a glimmer in seismologists' eyes.
"We cannot currently predict earthquakes," Emily Brodsky, a seismologist at University of California, Santa Cruz, told VICE News. "It does seem reasonable to look for patterns in past events and try to figure out if they might recur, but in general, we don't have enough of a history to be fully confident of our success."
'There's a shortage of instrumentation.'
Collecting greater amounts of data on past earthquakes might help scientists better predict when a quake is coming next, Brodsky said, but history alone can't solve the problem.
An earthquake occurs when the energy that has built up along a fault line — over years, decades, or hundreds of years — suddenly bursts forth in potentially destructive seismic waves.
The Earth's plates are constantly shifting about very slowly, which means some of that energy is always seeping out. But in the absence of accurate information about how much energy is lost through that slow motion, seismologists don't know how much is building up — and when it might reach a point that it triggers a major convulsion.
"Very often the missing piece in our understanding of the system is the slow motion," Brodsky told VICE News. "There's a shortage of instrumentation."
But one thing scientists do know is that earthquakes tend to strike in clusters. The risk of aftershocks lingers for days, weeks, and sometimes years after a quake. And, they know that an aftershock can be greater in magnitude than the original tremor.
Statistical modeling shows, for example, that one out of every 20 earthquakes is actually a foreshock to an even bigger one.
Scientists remain unsure why one earthquake might become a magnitude 4 versus a magnitude 8 and why a tremor might last thirty seconds rather than a minute, like the one that struck Nepal on Saturday. And even if they did that still wouldn't help predict when an earthquake might strike.
"That's pretty nonspecific," Tom Heaton, a seismologist at the California Institute of Technology, told VICE News. "When we get in the middle of a crisis we can only tell people, 'Oh yeah, we're having earthquakes.' There's always some chance something bigger will happen."
Still, there are hints that scientists might one day be able to improve their prediction skills. In March 2014, instruments placed on the floor of the Pacific Ocean, off the coast of northern Chile, began detecting moderate seismic activity. For two weeks, the tremors grew in magnitude and moved north along the fault at a rate of several miles per day. On April 1, a magnitude 8 quake hit Chile, killing six people, and displacing tens of thousands.
"That was exciting to a lot of people as a genuinely predictable earthquake," Brodsky told VICE News. "There's sort of this open question of, did we see it because the instrumentation was better, or was it different in some way?"
Japan and Canada are similarly monitoring seismic activity. Instruments off the coast of Japan allow scientists to observe seismic activity along the fault that triggered the 2011 earthquake and tsunami. Canada is monitoring a 600-mile fault that runs south from Vancouver down to Cape Mendocino in northern California.
The US, though, lacks offshore instruments like those of Japan and Canada, despite the fact that seismologists expect the fault off the Pacific coast could soon produce a strong earthquake.
Japan has something else the United States doesn't: an early warning system, designed to give individuals seconds or minutes of notice before a quake hits. Modern communication signals move more quickly than seismic waves from an earthquake. So when Japan's instruments pick up a major event, authorities can alert the public, giving them time to take cover or stop if they're in a moving vehicle.
The US Geological Survey has been working to implement an earthquake alert system since 2006. But low funding has held up progress, Heaton says, and it still hasn't moved beyond the demonstration phase.
Meanwhile, there's little evidence that earthquake forecasting could become anything akin to your weekly weather report.
"The more we study them, the harder they look to predict," Heaton told VICE News. "It is disappointing. It would be oh so nice if we could actually give people a warning weeks ahead of time, but it doesn't appear that we've got a strategy to do that."
Follow Laura Dattaro on Twitter: @ldattaro