DW-TV: Japan is a society that has played a pioneering role in a wide number of technologies -- including of course earthquake warning systems. But even then, people have just five seconds after the warning comes to brace themselves.

Cutting straight to the chase, thousands of detection stations around the world, international networks, decades of research -- why can't we predict earthquakes yet?

Danijel Schorlemmer: Well, even though we have thousands of stations, we have a big problem. We only measure the signals on the earth’s surface. Unlike in meteorology, where you can measure all the values you’re interested in, like humidity, wind-speed and so on in 3-D, we only see the earth’s surface. We cannot make measurements in the earth, which would be very important to understand what’s going on. And we’re also lacking a precursor phenomenon, a signal that would tell us that an earthquake is imminent. And there have been a lot of these precursors suggested, even animal behavior. But none of them has been shown that it works and that we can use it to forecast or predict earthquakes.

Was there really no sign or hint of a magnitude quake of this size?

Well there’s always a possibility, a small chance that after a large earthquake there will come an even larger earthquake. And now, in hindsight, you can say that the magnitude seven two days before the big destructive earthquake was this foreshock. But if you were to evacuate people every time you have a big earthquake, or sort of moderate sized earthquake, because there will be a bigger one, after the fifth time, nobody will follow you anymore.

If we can't predict earthquakes, is there something else we could do to save lives?

Yes, we can implement early warning systems. One is implemented in Japan, as you mentioned. It gives people some seconds of warning to get into a safe position or out of the house. We can install these warning systems in many places. Another thing is simply hazard estimates. We try to compute what can happen. What do we have to be prepared for. And then design building codes and other mitigation measures to save lives.

You're working on a project with Japanese colleagues on understanding earthquake predictability. Can you tell us more about that?

Yes, we have a so-called earthquake forecast experiment going on where different models are being tested to show what their forecasting performance is. But it has to be understood that this is not predicting particular large earthquakes, it’s a statistical estimate. Where do we have to expect earthquakes? And the better we are in this the better hazard assessments we can make.

What about a system that’s going to be reliable? That can reliably predict an earthquake? Are we going to see that in our lifetime?

I fear I have to say that I don’t think we will see this in our lifetimes.

Why? Why are you so pessimistic?

Because, as I said in my answer to the first question, we’re still lacking understanding of the basic measurements. But I think we can do a great deal in saving lives by having better hazard assessments.

(Interview: Ben Fajzullin)