What is an earthquake?
Perhaps the question should have been; when does an earthquake start? For me as a physicist what we normally perceive as an earthquake is just the boring business end of a rather fascinating process, because the earthquake actually starts far back in time.
What we normally see and feel as an earthquake is nothing else than rock breaking. During a small earthquake it is not a larger piece than what you might find in your back yard, and when it is a large earthquake it can be a breakage running for hundreds of kilometers in a fault-line. To simplify it we say that the amount of energy released is the same as the area of the rock breaking in a fault line. There are other factors of course, but this is not the time for that.
Back to when an earthquake starts. To answer that well we have to understand that there are 3 types of earthquakes (that we need to mind today, there are more types of course).
This kind of earthquake is driven by the movement of the continental plates; the motion is rather slow from a human standpoint. The speed of the continental plates is normally ranging from 1 to 5 centimeters per year. This type comes in 3 common subtypes; rifting (Mid Atlantic Rift) is when two plates are being pulled apart, clearly visible as the MAR slowly pulls Iceland apart. The second is subduction earthquakes as one continent plate is pushed down under another. Japan is a good example of this. Then we have shearing rifts, which is when two continental plates are sliding past each other.
In general the rifting earthquakes are the weakest, followed by the medium of shearing rift quakes, and then the subduction earthquakes as the strongest (Japan and Chile). The reason for this being the relative crust thickness, a rift zone has to thin crust to create the largest earthquakes, a shearing quake only affects crustal parts that are rubbing together, and the old crusts that are involved in subduction earthquakes has had the time to grow thick enough.
Math is simple. If the fracture is ten km long in a rift zone with ten km thick average crust, then a full faulting would affect 100 square kilometer. The same for 50 km thick subduction earthquake would be 500 square kilometer. Also we have the fact that the old rock in a subduction zone is generally more brittle and have a longer faulting.
A magmatectonic earthquake is when magma pushes upwards and releases pent up energy that already exists in a rift, shear or subduction zone. Here the movement of magma works as a trigger mechanism, but is rarely the cause of the earthquake. It works more like the small stone that topples the wagon.
These are caused by inflow of magma into dykes or magma chambers. The pressure of the magma pushes the rock until it flexes and finally breaks. I will come back to this one.
Age of earthquakes
Here comes what I find to be the interesting part of earthquakes. The age and the processes that make some places able to contain more energy (the last part of which won’t be covered here). I am going to use MAR earthquakes as an example here. As the MAR is spreading with an average of 5cm a year it starts to stretch an area of newly formed crust, our part of the crust is when it starts to stretch 10km thick (just an example). It is rather ductile, so it can take quite a bit of stretching, how much of it we know fairly well.
How do we know it? Well, as an earthquake happens it makes a crack, these cracks normally ranges from 1 to 10 meters or sometimes even more. Our earthquake will give a 5 meter crack. Simple mathematics would give that it would take about a hundred years for our earthquake to form. Problem is just that normally only half of the energy pent up is released in one event. So, it would be more like two hundred years since our earthquake got under way.
The size of our quake will now be determined by the area affected, and effected via the length and depth of the faulting line.
I have here made enormous simplifications, so anybody who wants to explain more in detail in a comment is most welcome to do so.
A magmatectonic earthquakes age is harder to determine, it can be a small amount of magma that hits an old pre-earthquake fault line that is close to faulting. Or it can be a lot of magma hitting a juvenile fault line. The effect will be pretty much the same. With one difference, the first one will most likely not cause an eruption; the second is likely to do so. And, if you have a mature fault line hit by a lot of magma, well that would be your basic Laki eruption.
Magmatic earthquakes are normally fast, anything from a few years, to a few days. In the case of Theistareykjarbunga it took about a year from inflation started to the first earthquake, at El Hierro it took weeks. This is normally governed by the size of the magma chamber, which in Theistareykjarbungas case is huge, compared to El Hierros chambers.
Earthquakes and Volcanoes
Magmatic earthquakes are caused solely by pressure from magma and the gases released by the magma. The exception is of course rift zone volcanoes like Krafla. It can either be pressure from below as new magma comes up through the mantle, it can be dyke intrusions, feeder tubes opening up, or inflation of magma chambers, and of course as the vent is opening up as the last stage of an eruption.
Here comes the thing, size does matter. A small 1M earthquake is cracking your average garden stone (well, a large one). A 3M quake is producing a fault that is roughly 100 times larger, and releasing 1000 more times the energy counted in destructive force. I am not going into the math of that here, it is fairly complex. But if someone wants to do it in a comment, please feel free.
So it takes 100 1M earthquakes to fault as large an area as one 3M earthquake, and it takes a thousand of them to crush as much stone (this would be the important number when talking about magma chamber formation). You should think mining blasts here. It is one thing to crack a fault line the length of the detonation on the video below, it is something completely else to blow all that rock to smithereens.
So please, if you see one earthquake of 1M strength, it is not important. If you see a hundred it might be interesting as a sign of something. But to actually have any significance at all, you need a lot more than that. Before a normal eruption you have hundreds or thousands of earthquakes ranging from M1 to M3 or more. And the 1Ms are not even important really. Only exception to this rule that I know of is of course Hekla. That one just needs a few below 1M, and then one around 2M, and then it goes off. But that is the known exception.
And let me reiterate this, I am not a geologist, nor a geophysicist, I am just a normal physicist trying to explain something really hard, and by doing so I have simplified things a lot. If you find that it is too simplified, or simplified so much that it is useless, feel free to compliment and explain further in the comment field. I will only appreciate it.
And once again, small earthquakes are really small, don’t be fooled.
And now over to a huge detonation courtesy of New Boliden Mining AB, I just like big detonations. Sorry, it is probably genetic among Swedes, blame Nobel. 1100 ton explosives used, 2Kton destructive nuclear device equivalent. And now to the real candy, this a 4M earthquake equivalent.