Why on earth am I using the term ”debunked” about a piece of equipment that is used by scientists across the globe? The answer is quite simple, it is when laymen start interpreting them that they are overused, or used in ways they were never intended. And this causes a plethora of faulty reasoning and wild theories.
But, first we need to explain how they work and what they do measure before we get to a few interesting things that they do measure that has nothing to do with eruptions or earthquakes as such.
Types of strainmeters
The first strainmeter was built in 1932 by the great seismologist and inventor Benioff. It consisted of an iron pipe that was attached inside a borehole with a strain-gauge at the top. As such it was intended to measure changes in distance up and down, or in other words if the mountain was expanding up and down. The values given was increased strain if the steel pipe was stretched or decreased strain if the pipe was relaxing.
The only problem was that this did not work that well. Benioff had forgotten that metal tend to expand and contract due to temperature differences, so if the mountain heated up even minutely it showed as a strain decrease, and vice versa if it cooled down. But, the theory behind the instrument was sound.
So, the steel pipe was after a while replaced by pre-stressed quartz crystal rods and finally useful data started to emerge. It was now possible to accurately measure very long frequency waves and record increase in strain over months or even years. The big problem was that it was only measuring what happens in a single direction.
The second type of strainmeter works like the common laser range finder. This version uses a laser beam that is reflected back to measure changes in distance over hundreds of meters. It is either placed in long boreholes or inside old mines. Otherwise it works in the same way as the quartz rod strainmeter and has the same drawback in that it only measures one direction of movement.
The third and most useful member of the strainmeter family is the volumetric strainmeter, or as it is also known, the dilatometer. It is basically a metal tube filled with silicone oil that is topped with a pressure gauge. The tube is inserted into a borehole of equal diameter and as the mountain flexes and bends it will measure how the volume changes over time. At times the pressure becomes too large and a special valve will automatically open resetting the pressure to a baseline value.
This type measures changes in all directions and will give you very exact readings, but there is of course a drawback. You will only know the sum of changes, but you will never know the direction of the changes. Think about it for a while and you will understand why.
What is measured?
A strainmeter will only pick up changes in the rock that is surrounding it. It will not pick up changes that are even relatively nearby. Basically, it is the surrounding rocks motion that is measured and motion on the other side of a small fault may not be picked up at all. Therefore, largescale tectonic motion is not possible to pick up and should be left to GPS-trajectory analysis.
Now quite a few of you will be protesting and saying that you can pick up both earthtides and distant large earthquakes on strainmeters. And the answer is that yes, you can pick those up. But what you are seeing is the effect on the local rock by the distant cause.
In other words, a large earthquake is the cause, the low frequency waveforms are the mode of transportation of energy that ultimately has the effect that it causes deformation of the local rock around the strainmeter. So, unlike a seismometer a strainmeter does not measure the original cause at a distance, it measures the local effect in the surrounding bedrock of the distal event. This definition is important, and the lack of understanding of this leads people astray when interpreting strainmeters.
But let us take a closer look at one of the most important strainmeter networks on the planet and discuss its uses and limitations.
The Hekla Strainmeter Network
The HSN consists of 7 stations, Búrfell, Geldingaá, Hekla, Hella, Saurbaer, Stórólfsvoll and Ytri Skógar. Geldingaá and Saurbaer was taken offline in 2013 and are now kept as backups for the general Hekla Network. It is not known if Icelandic Met Office are still using them internally.
If I am correct Hella and Stórólfsvoll are quartz strainmeters and Búrfell, Hekla and Ytri Skógar are volumetric strainmeters.
So, what do they measure? First of all, they measure strain increase caused by slowly accumulating magma in the Hekla volcano itself. The problem here is that Hekla fissure itself is spreading at the same speed as magma enters the system, so you will not really know if you are measuring magma pressure increase or effects of the spreading rift on the local bedrock.
To glean any data in regards of this you would need a complete record spanning years if not decades. And only the Icelandic Met Office has those records. So, for an amateur sitting at home watching the Hekla strainmeters they will not tell anything about the state of the volcano itself between eruptions.
The strainmeters also measure changes ranging from the Hekla fissure over to the next large fissure. To the east that would be the Vatnafjöll Volcanic Fissure and to the west it would be the closest fissure(s) in the Southern Icelandic Seismic Zone. Every other volcanic signal would be filtered away by the fissures and a pesky law of nature that states that for every time you double the distance the signal strength decreases fourfold.
It also measures strain changes in the nearest Faultline out from Hekla.
So, we have several different signals affecting the network at the same time and no way of knowing in which direction the signals are emanating from. And remember that there have been large earthquakes on both sides of Hekla, so the seismic effects of the tectonic motion is far larger than any volcanic signal in the years in between eruptions.
Now you may be thinking that they are pretty useless. After all they only record that strain is increasing or decreasing in an area over years or even decades and that we will never get to know where the changes are. Well, that is only partially correct.
The great use of a strainmeters are twofold, one is to give a longterm baseline and the other is for when the shit hits the fan. It is as a large nearby earthquake happens or a large eruption happens that the strainmeter comes into play and gives the most beautiful data of what is going on during the event. Let us take a closer look.
During the eruption of Hekla in the year 2000 the strainmeters gave the best instrumental data. The eruption yielded very little data in the form of GPS-trajectory changes and earthquake data. Instead it was the strainmeters that gave results that are still being analysed and reinterpreted.
One thing that became clear was that something was “off” with the Búrfell strainmeter. Where the others gave a rapid and large strain increase Búrfell gave off an even larger and equally rapid strain release.
In the end, it was apparent that the anomaly was caused by an old semi-solid magma reservoir under the preholocene volcano of Búrfell and that the magma chamber did both relax during onset of Hekla eruption and that it did amplify the inverted signal.
This data was later used to further correct the earthquake depths recorded around Hekla during the eruption making it possible to incrementally refine the location of the magma reservoir under Hekla. It was previously believed that Hekla had a discretely located magma chamber like most Icelandic central volcanoes, instead the reinterpreted data gave at hand that Hekla has a magma reservoir open at depth to the mantle and that is running all along the fissure underlying Hekla and that it is rising to a depth of about 1km below the base of the volcano.
Another result given is that Hekla does not erupt due to pressure increasing above the threshold of what the overburden can constrain, instead Hekla erupts due to tectonic forces pulling the fissure apart.
This last part explains the extremely rapid onset of eruption. During onset of the 2000 eruption Hekla opened up a conduit 4 km deep and 5 km long in a period that lasted 32 minutes, during the same time magma rose the same distance. The entire event took about 60 minutes, but it is during those 32 minutes that the main onset happens.
Cool strainmeter stuff
I have so far said that strainmeters are only measuring changes in the immediately surrounding bedrock or showing distal signals affecting that bedrock. So, what distal effects can be seen?
The first and most common effect are earthtides. These are not the same as oceanic tidal waves, instead these are the very minute movements in the mantle that is caused by tidal stress from the moon.
I always chuckle a bit about this because all the people stating that the moon causes earthquakes use the oceanic tides when they try to prove their erroneous theory. Instead they should be using earthtides since they move the crust minutely. They would still be wrong, but they would at least try to use the correct power. For those who missed the debunking of the moon as a cause for earthquakes I recommend reading GeoLurkings Tour de Force piece about it where he tried to prove the hypothesis that the moon is causing earthquakes by using all of Iceland’s recorded earthquakes without finding any proof.
Earthtides are best followed on the Búrfell strainmeter where they are visible as regular swings up and down on the plot.
The other thing that you can watch on the HSN are distal large earthquakes. What you will see are the low-frequency waves that has travelled inside the crust around the planet. On the image you can see the energy transferred from the M7.8 Solomon Islands earthquake.
I once again repeat, what you are seeing is not the event itself, instead what you see is energy moving around the globe that is affecting the bedrock close to the strainmeter.
Another thing you can see is a bit more surprising, and that is wind. As storms pass over the top of the strainmeter the rapidly shifting wind will both buffet the bedrock and cause rapid pressure changes that affect the pressure gauge. The effect will cause the plot to show a thickened curve.
Many people think that strainmeters are useful to predict eruptions and earthquakes, this mistake was even common among geologists in the seventies when they emplaced strainmeters to predict earthquakes along the San Andreas fault in California.
So far, the only result is that there has been a recorded strain increase running for years between eruptions of Hekla, otherwise no other useful predictive data has been given by any strainmeter on earth. They have though given lots of useful data about events as they happen.
Currently there is a lot of talk outside of the scientific community about the usefulness of the Hekla Strainmeter Network as a predictive tool for Bárdarbunga earthquakes. The original hypothesis was that a concerted strain raise was a signal that a larger seismic episode was coming at Bárdarbunga. This did not pan out against statistics. Instead an ad hoc extension of the hypothesis was added to save it. The new hypothesis was that a concerted strain increase or a concerted strain release was a sign of an impending seismic episode. After this deus ex machine intervention the theory could on the surface of things predict 90 percent of all seismic episodes at Bárdarbunga. Theory proven? Quite not.
The first hurdle here is that the strainmeters at Hekla does not measure the bedrock around Bárdarbunga. They are not even on the same plate shard, in fact they are several fissures over from Bárdarbunga. In fact, they are not recording any direct event at Bárdarbunga since they are on a completely different stress regimen.
So, can we save the hypothesis that you can predict earthquakes at Bárdarbunga via the HSN if we see it as distal energy transferal? Let us test that. After a large earthquake we would be seeing a marked strain drop that should occur immediately after the event. We have never seen this happening after a Bárdarbunga earthquake, not even after the largest of them. Why would we see a strain drop? Well, an earthquake is the release of pent up strain so by pure laws of physics we should see it if the strain increase measured was relevant to the earthquake.
Also, if we could predict anything it would look pretty much like what happens just prior to and during a Hekla eruption, and that is that we would get an inverse signal from the Búrfell strainmeter due to the magma anomaly. This is actually chucking an enormous spanner in the hypothesis. Now, have we seen an inverse signal before a Bárdarbunga event? No, we have not. Nor will we since the strainmeters can’t record things from Bárdarbunga to begin with.
Now some of you are saying that I just admitted that the amended hypothesis is accurately predicting 90 percent of the seismic episodes. Yes I did, but not in the way you think. The fact is that 90 percent of the time the strainmeters will move up and down together due to the common earthtide, so you would naturally find 90 percent of the seismic events during a concerted movement anyway. And here the hypothesis moves into the realm of myths and as such it is debunked utterly.
Or in other words, the entire idea is unscientific balderdash without any hope of ever being saved.
The Icelandic Met Office and other agencies around the world are giving the volcano amateur community access to loads of wonderful advanced tools to watch and to speculate about. It is quite possible to get at least a helpful understanding as an amateur of what a volcano is up to by looking at seismometers and GPS-signals prior to an eruption. But the strainmeters are a completely different set of beasts altogether since they are good at measuring data during the absolute onset of eruption and during an eruption, but are useless in between. They are as such in the realm of mathematical interpretation by geophysicists, and even we have a hard time getting it right.
They may though give relevant data for Hekla itself as a predictive tool, but only if used in conjunction with other types of instruments like seismometers and GPS-stations.