# Why aren't seismic stations installed very deep underground so as to pre-warn from earthquakes?

The velocity of p-waves emanating from earthquakes is in the range of 5-8 km/s (link)--let's assume it is 5 km/s. The earthquake depth is up to hundreds of kms deep underground (link)--let's assume it is 100 km.

That said, if a seismic station is installed at a depth of 50 km, and there are many of them in any given metropolitan area, then we can have a warning that is tens of seconds before the earthquake reaches the surface.

While I realize that drilling down to 50 kn is no easy task, I would have imagined that saving human life is well worth the efforts. Why hasn't this been done so far? Is it that such a short notice (10s of seconds) isn't worth it?

• I can't really give a good answer, but from my understanding: (1) Saying it's "not an easy task" is quite an understatement. (2) You would need to have a dense network of these stations, many of them under water. Good luck. – yo' Jan 21 '16 at 21:20
• en.wikipedia.org/wiki/Earthscope - only two boreholes and extreme hassle of persuading other ... savants at Congress and the OMB. – Deer Hunter Jan 21 '16 at 22:52
• But aren't the most severe earthquakes (in terms of human impact) usually the shallow ones? And you're cheating by assuming the slowest possible wave propagation, instead of either the average or fastest. – David Richerby Jan 22 '16 at 0:52
• @DavidRicherby you are right--all what I wanted here is just to have an idea. – student1 Jan 22 '16 at 1:33
• @SimonW that's a big misconception people have on the structure of the earth. The rock in the crust and the mantle are overwhelmingly solid. There is no magma ocean. Melting is a rare localised phenomenon. Oceanic crust is about 10 km thick, continental crust can be up to 100 km thick. Below the crust is the mantle which is also solid. – Gimelist Jan 22 '16 at 20:28

The simple answer is that you can't drill to 50 km depth.

The deepest holes ever drilled were to a little more than 12 km, one is named the Kola Superdeep Borehole in Russia, which was a scientific drilling project. The very few others were oil exploration boreholes.

Drilling that deep is extremely expensive and hard. If you go and ask anyone who ever worked on a drill rig, drilling the second 100 metres is always harder than the first 100 metres. And we're talking about kilometres here! There are several problems with drilling that deep. It's extremely hot down there, and the drilling equipment just breaks and stops working. You also need to pump cooling water in and pump out the stuff you're drilling and it gets harder with depth.

This is simply not feasible. Now let's say that you did somehow manage to drill a hole to that depth. How would you put monitoring equipment inside? That equipment has to sustain heat and pressure and still keep working, while being able to transmit whatever it's reading back to the surface. This is not going to happen, not at 50 or 10 km depth.

Another problem is that not all earthquakes are that deep. Some earthquakes originate near the surface, or just several km deep. Having a monitoring station down there isn't going to help. The 2011 Tohoku earthquake (the one that triggered the tsunami at Fukushima) was only 30 km deep. Same thing for the 2004 Indian Ocean earthquake.

• To add to this answer, not only does the monitoring equipment have to work. It has to work extremely reliably: there's no point if you can't tell distinguish whether an earthquake is happening or whether the equipment is malfunctioning or broken. – Hugh Jan 23 '16 at 6:59

10 seconds is a worthwhile warning time. The Japanese use it to switch off compressors on gas lines (or open release valves on them), emergency break bullet trains, and in particular shut down nuclear power plants, all in anticipation of a power loss moments later. In the case of Japan, however, the warning time doesn't come from having deep sensors, but from having sensors everywhere -- and getting the advance warning because earthquake waves don't just travel upward, but along the earth surface for dozens of kilometers before they become too weak to cause much destruction.

The issue with drilling down to 50 km is simply that we have neither the technology nor the financial resources. The Germans drilled down to slightly more than 10 km, but it cost a billion dollars. The issues are that (i) it gets incredibly warm at these depths, (ii) you need to rotate a drill at the end of a 10 km long pipe that is so hot that it would rather deform than transmit this torque, (iii) there is tremendous pressure on the sides of the well from the surrounding rock, causing it to push inward (and in the process jamming the drill string). I don't think we'll ever have technology that can drill to a depth of 50 km.

• +1 on specific details of securing gas lines, Japanese trains, and cost of drilling. I wonder if the OP accepted my answer too soon. – Deer Hunter Jan 22 '16 at 8:00
• One billion US$for 10 km? Do you have a source for that price? I'm aware it's expensive, but so expensive? – gerrit Jan 22 '16 at 10:47 • @DeerHunter I am not actually sure which answer to accept since all of them present new perspective supported by some evidence and I am by no means expert or even knowledgeable so I might just accept the one with the most +1s. – student1 Jan 22 '16 at 19:20 • @gerrit German Continental Deep Drilling Program considering today prices: ~0.48 billion$US – klanomath Jan 23 '16 at 14:43
• Oh, my bad -- my memory was off by a factor of 2. But the order of magnitude was certainly correct. – Wolfgang Bangerth Jan 24 '16 at 2:56

I can see two major benefits of having a 10 second warning time:

• shutting down gas lines to prevent uncontrollable ruptures and fires after the tremor when comms channels are either destroyed or overloaded;
• giving people inside buildings time to hide in relatively more survivable spots before the building collapses. Duck and cover, reenacted against earthquakes.

With this in mind, it's still a question of cost-benefit analysis and politics. Deep boreholes are costly; it took almost \$200M to set up EarthScope with two boreholes at San Andreas, the deeper one going down 1.67 miles. Evacuation benefits really start to accrue when you can get people outside rubble islands, and it isn't 10 seconds, more like 90 seconds upwards.

• Do NOT hide in doorways in a quake. You will be hurt either by the wildly swinging door, or by people fleeing through it. The phrase to remember is "Drop, cover, and hold on" - that is, get down low so you aren't thrown off your feet, banging your head open; get under cover such as a table; and hold on to prevent yourself being flung about, or flung out from cover. – Dewi Morgan Jan 24 '16 at 9:25

I think you've answered your own question when you ask if 10 seconds is too short notice and is it worth it.

A question for you, what can anyone do in 10 seconds and how far can anyone travel in 10 seconds?

The only way to possibly save lives when an earthquake is about to occur for people to move away from the earthquake site to a safe area. As you state earthquakes travel at 5 to 8 kilometres per second. Most people might be able to walk 5 to 8 kilometres per hour.

People are not going to be able to get far in 10 seconds. Most people won't even be able to get out of the building they would be in.

There is nothing to be gained by having 10 seconds of warning prior to a major natural disastrous event.

• Shutting down gas lines isn't worth it? There's also the possibility of hiding within the building (near door frames etc). – Deer Hunter Jan 21 '16 at 22:47
• @ Deer_Hunter: Can gas lines be shut down & safely cleared of gas in 10 seconds? – Fred Jan 21 '16 at 22:53
• I don't know that, and it's a good question! (for Engineering SE, perhaps?) – Deer Hunter Jan 21 '16 at 23:04
• I'd say that a closed line is better than an open line when it ruptures, and even if it wasn't exactly clear or actually closed, a small amount of gas from a closed, or soon-to-be closed line is probably less explosive. 10 seconds could be enough time to avoid catastrophe or make sure you're clear of any chandeliers. It might also be long enough for someone to get CLOSER to a safer spot than otherwise. Not to mention, getting outside of a building in an earthquake is probably the least safe option. – coblr Jan 22 '16 at 0:40
• @DeerHunter: Do NOT hide in doorways in a quake. You will be hurt either by the wildly swinging door, or by people fleeing through it. The phrase to remember is "Drop, cover, and hold on" - that is, get down low so you aren't thrown off your feet, banging your head open; get under cover such as a table; and hold on to prevent yourself being flung about, or flung out from cover. – Dewi Morgan Jan 24 '16 at 9:25

The cost of digging very deep boreholes is generally prohibitive and unnecessary given the accuracy that detectors can be built with modern technology and the time advantages of deep mine detectors is minimal given the velocity of wave propagation through a solid. The Tevatron Particle Accelerator in the US was able to detect earth quakes from the other side of the world during its operation.

Japan has started to put vital services such as Water, Power and Gas in very very deep utility tunnels ( 1km+ deep ) as they are only minimally affected at the deeper depth by Earthquakes, the seismic shift is greatly reduced ( About 1-3cm ) and oscillations occur at a different frequency. In most of the great earthquakes to hit Japan most people have been killed not by building collapse but by incidents caused by ruptured utilities and the aftermath of having no services ie access to clean water.

These deep service tunnels are also populated with seismic detectors as they are insulated from surface vibrations and provide a stable baseline for earthquake detection and monitoring.

This massive change in city planning demonstrates that advanced warning alone is not enough to protect a city and its inhabitants.

• Hi, and welcome to stackexchange. As this is a Q&A site, rather than a discussion forum, all answers should be direct answers to the question,and not to previous responses. While the first two paragraphs of your answer are interesting, I don't think they actually answer the question; your third paragraph does that. Please consider editing? – Semidiurnal Simon Jan 23 '16 at 11:14