Why can't we predict earthquakes?

The prospect of finding any more survivors from the 7.8-magnitude earthquake that struck Nepal is getting increasingly bleak. Nepal lies in a region of South Asia that's known to be prone to earthquakes. Seismologists knew the country was vulnerable to quakes, but they couldn't say exactly when one would hit. But how come? Why, after decades of earthquake research and monitoring, are scientists still not able to predict earthquakes?

Brian Tucker, seismologist and president and founder of GeoHazards International, explains the complex science of earthquakes and why we still can't predict them.

What happens, briefly, with the tectonic plates during an earthquake?    
Basically, the plates move either past each other, under each other, or over each other, but they are accommodating motion relative to the two plates.    

So these are massive parts of the Earth and an enormous amount of force. Why can't we predict exactly when those plates are going to shift?    
Well, imagine picking up a big matzo and bending it to break it in half. You can't predict just where the cracker will break. You might want to break it in half, but it may break anywhere. And similarly, we can't predict just where and when the fault will rupture. The two plates are stuck together and the stickiness varies along the fault. When the plates are being pressed, you don't know which part of the plate is the weakest, and therefore which will break. 

But you know that the possibility is there, and we know that certain areas are more at risk for earthquakes than others. Nepal is one of those. So was there seismic activity in and around Nepal before this earthquake hit?     
There was, but it was normal. It was like rain showers in British Columbia. There have always been earthquakes [in Nepal], so there was no reason to expect that at this time there would be such a large earthquake. There was a big earthquake back in 1934 and in 1833 so we knew that big earthquakes have occurred and will occur. But we just didn't know exactly when. 

But we do have an idea of the duration between giant quakes, so were you surprised to see an earthquake of this magnitude when we did? 
There was absolutely no surprise at all. We know that there is a sort of return period. But these are only averages, it's not like a Swiss clock. And that's why we've been working there for the last twenty years.    

In March of last year, moderate seismic activity was detected off the coast of northern Chile, and those tremors continued. Then there was a magnitude 8.0 earthquake that hit the country, so doesn't that suggest that science might be getting closer to predicting earthquakes?            
There are occasions when there's a whole burst of activity of small earthquakes and then they're followed by a large earthquake, but it's not always the case. The most famous case occurred in 1975 in China when volunteer observers spread out all over the country looking for any weird, anomalous behavior from ground shaking, to snakes. For example, they saw snakes coming out of the ground in winter, all concentrated in one particular area. The government was so convinced that this was going to be an earthquake that they ordered people out of their houses in the middle of the winter. Lo and behold, that was followed by a large earthquake and thousands of lives were saved because of it. But then, the next year in Tangshan, China, there was another earthquake that was not preceded by anomalous activity and some report 400,000 lives were lost in that earthquake. So, we still do not have any means of predicting earthquakes.

As sensors become more sophisticated, and data becomes more abundant, do you think that seismologists will ever be able to accurately predict earthquakes?
I don't think it'll happen in my lifetime. I feel that the best use of our time is to go to places where we know earthquakes will occur and try to prepare the communities rather than more effort trying to predict them.

You've been to Nepal. When you spoke to people about the possibility of earthquakes, did you feel that they wanted to hear what you had to tell them?     
No. They wanted to listen, although, they were hearing a different message than what I was saying. For example, when I described that we would expect 40,000 people to die in a repeat of the 1934 event, one man told me, 'OK I know what we need to do to prepare. We need to collect wood.' And I said 'Wood? How's that going to save anybody?' And he said 'No, no, we will need wood to cremate people.'  He totally missed that what we wanted to do was to reduce the deaths, not to prepare how to cremate all of the people who would die.