TSUNAMI are terrible things. And part of their terror lies in their unpredictability. Even when a submarine earthquake that may cause one is detected, the information that is needed to determine whether a giant wave has actually been created takes time to gather. That is time unavailable for the evacuation of coastlines at risk. Contrariwise, issuing a warning when no subsequent wave arrives provokes cynicism and a tendency to ignore future evacuation calls.

Such tsunami-warning systems as do exist rely on seismometers to detect earthquakes, and tide gauges and special buoys to track a wave’s passage. That is reliable, but can often be too late to get people away from threatened coastlines. What these warning systems cannot do reliably is predict immediately whether a given earthquake will cause a tsunami. And that, in the view of some seismologists, is a scandal. For, as the annual meeting of the American Association for the Advancement of Science learned from Gerald Bawden of NASA, Paul Huang of America’s National Tsunami Warning Centre, Tim Melbourne of Central Washington University, and Meghan Miller of UNAVCO, a geoscience research consortium, the tools for accurate tsunami prediction already exist. All that needs to happen is to connect them up.

The nub of the problem is that it is hard to distinguish immediately whether a submarine earthquake is powerful enough to cause a tsunami. Big quakes (those above about magnitude 7.3) involve slippage along many kilometres of a fault. That means their energy is not radiating from a point. A single seismometer therefore has difficulty distinguishing between a quake of magnitude 7.3, 8.3 or even 9.3 (about as large as they get). The logarithmic nature of the earthquake-magnitude scale, though, means the third of these is 1,000 times more powerful than the first. And the more powerful the shock, the more likely it is that a dangerous tsunami will result. Once the seismic waves from an earthquake have reached enough seismometers, the distinction becomes clear. But near any given quake there are rarely enough seismometers around.

Except, as the panellists pointed out, there are. America’s satellite-based Global Positioning System and subsequent similar efforts from other countries (known collectively as GNSS, the Global Navigation Satellite System) have permitted the creation in many places of networks of sensors that measure, within millimetres, local distortions of Earth’s crust. The main reason for doing this is to understand the build-up of earthquake-causing strain in the crust, so such monitors are most abundant where tremors are commonest. And, if a tremor does happen, monitors nearby will be shaken by it.

There are, by the panellists’ estimates, about 17,000 such monitoring devices around the world. Of those, around 2,300 make their data available instantly. If these instant monitors’ signals could all be gathered together and run through suitable software, the true nature of a big submarine earthquake would be apparent almost at once, and appropriate warnings could be issued.

At the moment, two regional projects are testing this idea. One, READI, on the Pacific coast of America, is under the aegis of NASA. The other, GEONET, in Japan, is organised by that country’s land-mapping agency. The hope is that, if these local ventures work, other countries will join in and a global network can be created over the next decade.

Really clever use of the GNSS, moreover, might be able to do even better than this, by tracking a tsunami as it travels. Though the most visible consequence of a tsunami is a wave in the ocean, it also creates one in the atmosphere. This affects the arrival time of GNSS radio waves in a way that, with enough ground-based detectors, would permit the passage of the wave to be followed. And these detectors, too, will soon be commonplace. For many years, smartphones have contained GNSS receivers, so a phone’s apps can use location information. The latest phones have equipment so sensitive that it could, in principle, detect a passing tsunami in the atmosphere. All this would require is for someone to write a suitable app, and for enough phone users to download it.