Two earthquake swarms during the night.

During the night there where two earthquake swarms in Iceland. One of this earthquake swarms was man made.

The larger earthquake swarm was in Hengill volcano and was man made. It happens and the hydrothermal plant pumps down cold water into the bedrock and changes the pressure and strain in the bedrock and nearby sediments that make up this area. The earthquakes only happens when cold water is being pumped down into the ground. When they are not pumping cold water into the ground, no earthquakes happens. Largest earthquakes in this man made earthquake swarm where ML2.5 and ML2.2 according to automatic measurements of Iceland Met Office. But most of the earthquakes where less then ML1.0 in size.

The man made earthquake swarm in Hengill volcano. Copyright of this picture belongs to Iceland Met Office.

The second earthquake swarm that did happen during the night was in Katla volcano. This earthquake swarm is no surprise to people how have been watching Katla volcano activity since July. This earthquake swarm in Katla volcano was however small during the night. But there did more earthquake activity in Katla volcano around 18:00 UTC yesterday. But overall it seems that Katla volcano has slowed down a bit compared to last few days.

Earthquakes in Katla volcano during the past 48 hours. Copyright of this picture belongs to Iceland Met Office.

For the moment it is rather quiet in Iceland, as there is currently not that much activity in Katla volcano. But that can change without any warning at all. As Iceland is highly active in both volcanoes and earthquakes.

Icelandic news about the earthquake activity in Hengill volcano.

Jarðskjálftar í boði Orkuveitunnar (Rú, Icelandic)
Jarðskjálfti á Hellisheiði (Ví, Icelandic)

268 Replies to “Two earthquake swarms during the night.”

    1. I like how they put a heart under Eyja in the wiki version.

      By the way, I believe the cryptodome shown is Godabunga but what is that eastern lava dome?

      1. The cryptodome is Godabunga, and they have nicely enough not connected it to either Eyjafjallajökull, nor Katla. This is actually good since no one really knows what volcano it belongs to, if either of them.
        The east lava-dome is I think a protuded plugg of lava that has been pushed up at some time or another. Ie, remnants of an old eruption. At least that is the explanation I remember for it. But I might be mistaken.

    1. Wind speed 13,4 M on average 12.1

      It’s wind even if we see 2 quakes yesterday at 9:58 & 14:56

  1. One question of the effects of the wind:
    Several times when something is changing in tremor graphs it has been explained to be the effect of the wind. At the moment average wind is 12,1 m/s with over 20 m/s gusts in Heklubyggð, Jon’s seismograph has gone bezerk, but I do not see any significant change in tremor graphs…
    Does the direction of the wind matter or hoe strong gusts there is or what?

    1. The SIL seismometers are considerable better wind protected then my geophone.

      But even so, wind is going effect the SIL seismometers. Along with ocean waves and other noise from the nearby environment.

    2. Wind affects mostly higher frequencies, i.e. mostly blue but also green lines. Red line is affected by deep ocean waves. Surface (ocean) waves affect coastal stations only (approx. less than 50 km from waterline), I guess in all bands.

      1. Yeah, I thought it was something like that 😛 Is there supposed to be even worse weather in Iceland next upcoming days?

  2. AUST GPS station looks like it’s moving south west again.. Deflation?
    The station is close to the clusters of earthquakes in the north of caldera which we saw in the last weeks.. Could the quakes be the result of deflation of magma instead of inflation?

    The other GPS station don’t seem to be affected that much.. Only a little movement to the south what could mean inflation.. Magma shifting to the south of Katla from the northern part?

    1. If it is shifting it is probably moving towards the 1755 fissure.

      But Austmannsbunga has been doing it’s dance for months now going all possible directions, but the general trend has always been steadily upwards. And sofar there seems to be no down-movement to speak about really. We need at least a week of data to start talking about deflation.
      My interpretation is that there is no deflation at Austmannsbunga, but that there might be inflation over at the 1755 fissure too.

      1. Is there a map with all the fissures in Katla? I know it is somewhere but can’t find it..


  3. Extract from Ruv today about Katla (translation by Google)

    “Still TREMBLES Katla

    Nearly twenty earthquakes were detected near Goðabungu and Hábungu Mýrdalsjökull in the night, the largest of approximately 2.7 meters under the weather service. Source of these earthquakes were in the middle of the Katla caldera, measurements of turbulence in the area at night did not indicate that volcanic eruptions are looming. Two major earthquakes in the night were for kilometers, according to the Meteorological Office.” Here is the link:

    1. downgraded

      12.09.2011 10:28:39 63.647 -19.325 0.1 km 0.7 99.0 3.8 km WNW of Goðabunga

  4. @Lurking, Raving and Jack et al:
    Regarding Bragin et al, as for all papers like this it is hard to take math like this and make it into something usefull. Trust me on this one.
    I did a lot of work with the Kalman conjecture and pseudo-cyclicity back in the late nineties.
    The reason for your plots turning out ugly Lurking is the eternal noise versus significance problem. There are two short-cuts you can take here. Since the Raving method presented earlier with binning is a simplification we can continue on this road. Nota bene, non-simplification is not possible due to it becoming a computational NP-problem. Ie, there is not enough computational power available on the planet for it to be done with simplification.
    First short-cut you can do is to remove all quakes that arbitrarily can be said to be noise, ie, remove all quakes that are to week (you eat hamburgers, you do not measure them as quakes), shallow (ice-quakes, construction blasts and so forth). That would clean it up a bit. Then you can change the scale, plotting like this can give a scale that is hard to comprehend. Think here of a painting of Jackson Pollock. If you stick your nose to the picture you just see garble, as you back out you start to see shapes and patterns, and when at the right distance you get the picture (understanding it is something else…).
    You would probably need to do both.
    Problem is just that the pattern sought after might be in the removed stuff. So I would do a plot on the noise to.

    Thank you for your concern. I went to New York over the weekend for a bit of R’n’R.

    1. Actually I liked Raving’s proposal very much as it will reveal if the system is quasi-periodic (in the sense of the chaos theory). If it proves quasi-periodic, however, it does not give us any direct methods to predict anything.

      If this does not reveal anything (physically) meaningful, then it must be concluded there’s no periodicity or quasi-periodicity whatsoever there…

      1. Yes, but what I meant was that the noise-threshold can be to high for us to say anything in any way right now. And that by eliminating a bit of noise we can see, or not see, if there is a quasi-periodicity in there.

        Jack, isn’t the wonderfull thing with chaos-theory that we cannot predict anything, but we always believe that we can do it?

      2. Well, the fact with chaos theory is that the data from a chaotic system will always look noisy, but the attractor will not IF there is any quasi-periodicity in there. If there is no clear attractor (no closed loops or surfaces), there is no quasi-periodicity.

        The beauty of the “trick” lies in finding a regular, quasi-periodic orbit (called attractor) in a noisy-looking data, thus revealing the true nature of the phenomenon itself (essentially non-linear systems). So, the system is behaving perfectly deterministically, it is just a little picky on the conditions due to the inherent non-linearity of the governing equations…

      3. Well. To be honest I got confused half way through the steps.

        I got the daily quake counts for the SISZ, came up with a way of stepping though the differences in the count at steadily incrementing offsets, and got one mondo set of data that seemed, visually, to have an error in it.

        In my original “pseudo-cycle” sheet, I had a large spreadsheet with the cells extracting the info. You could see where the cell equation was solved. Non solutions were left as blank. This gave gave a visual indication in the form of diagonal traces of numbers as the parameters for the cell moved up and down the spreadsheet. I expected something like that.

        What I got were the diagonal traces, but also an artifact that went horizontally down the row columns. Checking on the cells to make sure they had the right procedure showed that they did, so I continued the process.

        Plotting this behemoth of a spreadsheet with a Z-table gave me a nice green plot with some artifacts that represented the diagonal traces I had seen in the numeric form. This plot took 5 minutes to make as the two programs duked it out. It was a bit above 300,000 data points, each with and x,y,z. Then the graphing program had to munch through the generation of triangles for the surface plot. My computer groaned, spit out the plot, and I sat there confused about what I had. My gut feel was that it was garbage.

        So, I ran a simple xyyy plot that I linked to earlier, made a comment, and went to bed.

        The part that confused me was that after working through that, per Ravings instructions I was supposed to repeat what I had just done? A completely different set of offset calculations?

        If so, this may computationally be beyond what I can muster with this machine… and it’s a Quad core Phenom II and 4G of ram.

        (No, I am not going to put together a distributed processing system with my wife’s machine.. she has a hard enough time with XP)

      4. Kind of what I meant that it quickly turns into a computational NP-problem. Imagine than that you do it as fractional integers instead of binning and we are into hell…
        Try to de-noise it in some way and you should probably have removed enough to be down to a tenth of the datapoints or less, then you could probably run it against the second aggregate.
        When we set the pattern-recognition system for the hydrophone I found that the principles for pattern-recognition was the easiest part (although not simple math), finding a suitable threshold between de-noising and computational power was another ball-game alltogether.
        Weak instances are normaly found in abundance around stronger instances, so if one remove the weakest the stronger stand out a bit more. But… sometimes the weaker instances may be what you are looking for.

      5. Lurking: Do not be afraid if the result looks a bit odd. Those phase plots are a bit odd-looking by their nature!

  5. Completely offtopic, but I was strollin’ through some GPS plots and ended up at Krysuvik, which, as we all know, shows clear signs of inflation. But then it came to me that I’ve always understood that the Reykjanes volcanoes do not have a central magma chamber, instead they have a system of magma-filled sills. Are sills actually capable of producing this much inflation? With a magma chamber, yes, but sills? A magma chamber is one concentrated space filled with magma acting like a balloon. Sills are more spread out and therefor the general trend up during a magma injection should be less.

    1. Ahem, the phrase magma-chamber might be the most ill-coined one in volcanology. It quite simply does not exist as such in the form of a hollow chamber. It is even in its simplest form a behemoth of tubes, sills, cavities and dykes.
      Krisuvik is believed to be a fissure volcano, but with a magma-reservoir between two and four kilometres in the form of dykes and sills. And as those whidens as they fill with magma they can cause considerable (for the area) uplift.
      Eyjafjallajökull for instance was believed to have a “chamber” before it erupted, but when the eruption occured it was quickly discovered that she did not have one.
      Boris Behncke is the leading authority on hating the word magma-chamber and one of his articles lead me to coin the term “behnckian tubinged magma-reservoir”. Look at his pictures of Etna and you start to understand what I am talking about.

      1. When thinking with common sense about a magma chamber it feels already not logic.. such a large hole in the ground.. it has to collapse.. Since things tend to go to a low energy state which is basically a filled hole because gravity wants it to be filled :p
        On the other hand, how do caldera’s form then? Is the entire tube system blown out then? and what remains is a hole in the ground?

      2. There are 4 versions of caldera-forming events that I know of, which is not to say that it cannot be more of them 😉
        1. Explosive caldera formation Krakatau-style. The reservoir gets so damaged during eruption that it caves in. For Krakatau the ocean then fell into the reservoir and Ka-Boom!
        2. Rapid subsidence caving in style (Yellowstone), where the mass of ejecta is so large in short time that the reservoir falls in on itself. With or without explosion as it falls in on itself.
        3. Toba style, whereas the volcano just blows off the top of the reservoir. For Iceland, that would be Askja for you. If the large caldera in Askja happened in the same way as the 1875AD and the 8900BC VEI-5s we have our basic VEI-7.
        4. Slow subsidance through blocking. As the reservoir cools down the volume diminishes and the various roofy parts start to fall down in an ultra-sized sinkhole.

      3. Many steam based power plants have a “fire box” with a set of water filled tubes in it. The idea is to present as much surface area to the flame as possible. Like a pot of water on the stove, the goal is to get the water up to a certain temperature.

        Structurally different, but they do the same thing, and the heat transfer equations apply equally well.

        Think of a ball of cotton. It occupies a contiguous area and can be described as having height-width-depth. Upon closer examination you notice that it’s not a solid object, but is made up of a lot of little interwoven and connected fibers. Those fibers would be the equivalent of the interconnected tubes in a magma “chamber”

      4. Sander

        “…Is the entire tube system blown out then?…”

        Collected magma is not a complete mass of liquid. Geologists (real ones) tend to think in terms of percentage of partial melt. Not all of the material in the magma is in a liquid state. Included in the magma are dissolved gases.

        When a volcano erupts, the first thing that happens is that some of this material (under very high pressure) has found it’s way to near the surface. Once the surrounding rock experiences enough force to fail, it does. This releases pressure on the magma and the gas in solution finds itself in a lower pressure environment. At this point the magma that has had the most radical drop in pressure will have it’s gas to start coming out of solution and collecting in bubbles, these bubbles expand as more gas is collected in them and they start to rise. If the process happens fast enough, things cascade into a violent evacuation of the area where the magma had collected.

        All this happens on very small time scales.

        With that much force and the pressure waves traveling up and down through the system, yes, it can finally break the tubing structure and evacuate the whole thing en mass.

        Add an influx of water into the event and things can get nasty fast.

        Each volcano is different, but these are the basics as I know them.

        (I’m also not a Geologist, so I accept the possibility that some of what I state could be wrong)

  6. Somebody has been out doing field-work at Krisuvik.
    GVP has updated the eruption list a bit.
    If what they now are saying is anywhere close to the truth an eruption at Krisuvik would be disruptive.
    Before the consensus here seemed to be that if an eruption occured it would be most likely a hawaiian style eruption, with a remote possibility that it could be explosive if it was erupting under Lake Kleifarvatn.
    But now the list states that the first two known eruption are large-scale lava lake events. After that we have a series of nine eruptions, and seven of those have been explosive. They where not massively explosive, but since it is not that far from Reykjavik I guess it could be a bother and two if/when Krisuvik erupts.
    Fun thing, this list has been redone times now in one year.
    I guess it is someone doing field work on it now.

  7. Not seeing marked wind noise on multiple stations or a warning at IMO as yet, so are we sure it is particularly windy at the moment? I know Katia remnants are due to brush Scotland and N.England today, but am unconvinced that the systems path is taking it over Iceland. If it isn’t massively windy, Grimsfjall tremor looks interesting.

    1. Looking at the IMO site probable winds of between 8 – 12 m/s – I am afraid as its not in miles per hour I’m uncertain to the strength of that, but am guessing at @ anywhere between 30 – 40mph, so, quite strong but I am not sure enough to account for this tremor pattern across all wavelengths.

    2. I was going to say that looking at the webcams, especially for Geysir (, the wind seems to have dropped a bit. However, when I went back to get the link above, the wind is still gusting (spray from the geyser is being carried a few metres by the wind).

      Tremor graphs indicate that the earthquake map should be showing at least small recent earthquakes. Maybe IMO’s seismology department is having a team meeting this afternoon?

  8. jon do you have an icelandic account i can deposit to i don’t have pay pal or credit cards

    1. There seems to be more water flowing at Jokulsarlon compared to yesterday (See link: IMO’s hydrology maps in that area not showing more flow in the nearby rivers that they do measure (assuming that I have got the right area & not got confused by Icelandic naming.)

      Is this near to where you are seeing the harmonic tremors?

      1. The webcam is attached to a bridge over the outflow of the lake. One can often see clearly that the normal high tide brings the water from the sea into the lake. And at low tide the water flows back out. The larger icebergs then accumulate before the runoff. On the beach behind the bridge (in the back of the webcam) often lie the remains of icebergs.

    1. It does look quiet – this is a screen grab from IMO
      Allot of activity in the early hours (3-4am) now nothing at all but there is nothing since 11am. 2-4 HZ shows movement at ALF & GOD but what this means I have no idea. I think the Katla is taking a breath waiting for her weekend burps (or when Jon says it’s quiet) 😀

  9. If Katla is a bully she will wait until we have given up before she thinks about erupting.
    The lull before the storm will be a long wait in my opinion.
    A full, properly transcribed version of Katlas folklore in either American English or Original English is needed, Google is no good in this case. I bet there are Icelandic speaking English bankers. 😆 Help…

  10. Jokulsarlon cam NOW – what’s the orange cloud centre background ?rising smoke? or is it the sun setting – more likely

  11. If it is a bright light, it is either an Icelander in a 4×4, in a helicopter, or an aeroplane, even just up there with a torch.
    Islanders are the same no matter which island they come from. They do odd things, or it appears so to outsiders. They get to places that are deadly to outsiders with ease.
    I have lived a long time on 3 different islands, Iceland is no different.
    Folklore works on islands better than anywhere else on Earth. It is harder to tell lies on an island after all.

    1. Have a look centre background above the dark cloud line, looks orange probably thin cloud with setting sun, but changes a little, may be poor optics and atmospheric distortion, but its a long way off, not lights

      1. Jokulsarlon does seem to have smoke rising in the background. I thought it was pink to orange, now it looks like 4 puffs of smoke breaking up.

      2. This is NOT smoke, just clouds. There is no way that it could be smoke or something. Don’t let sunset light fool you!

      3. Note I said smoke ie particulate matter from incomplete combustion of carbonaceous material, nothing else!

  12. Someone had too much coffee to drink and is shaking the seismometer! 😀
    2.4 11 Sep 10:30:38 74.8 2.5 km SE of Litla Kaffistofan

    1. Sorry the link above is corrupted. But if you Google Jökulsárlón, you will find the correct link.

Comments are closed.