More harmonic tremors in Katla volcano

It seems that Katla volcano is not as quiet it seems to be. For the moment the harmonic tremor seems to be small and only properly observable on one SIL station due to how weak it is at current time. So far two harmonic spikes have happened. This is most likely a hydrothermal system going crazy under the glacier close to Lágu Hvolar SIL station, rather then magma on the move (or at least that is what I think).

Harmonic tremor at Lágu Hvolar SIL station. The harmonic tremor can be seen at the end of this tremor plot. Copyright of this picture belongs to Iceland Met Office.

There has not been a lot of earthquake activity in Katla volcano during the last 24 hours or so. But that might not mean anything, as eruption can and have started without any warning at all (example: Grímsvötn volcano May 2011). So for now, it is more wait and see and more waiting it seems.

90 Replies to “More harmonic tremors in Katla volcano”

  1. would like to know views re: non-eruptive outgassing and possibility of water vapour release. Please. (This is not a scaremongering, suspended atmospheric vapour question)

  2. This is mainly for Carl and Jack@Finland.

    I’ve been playing with the FFT (Fourier Transform) and the various sets/subset of Icelandic quakes over the last year or so, and found nothing that really sticks out that strikes me as really that interesting… just stuff that is somewhat dubious.

    A couple of months ago, Willis Eschenbach, guest post writer over at WUWT and who is fully into all things data much more that I am, wrote a post about Periodicity Analysis. In this interesting article (to me anyway) he outlined the fundamentals of the nuts and bolts contained in and provided a segment of the R program that he used in his examples.

    I got as far as loading the R program interpreter and managed to get distracted by work and other more pressing things.

    REVISITING this method today, I finally figured out how to fundamentally do the transform in Excel, and applied it to the data that I have been ruminating on for quite some time… the Icelandic Quake list.

    Now… this is NOT a Fourier transform. It is a self determining extraction of “psuedo-cycles.” It does not rely on a collection of sinusoidal components to make up the base signal. All it does is to identify periodicity in the Icelandic quake set.

    My method to get the data into a manageable form, was to first bin the quakes into 3 day bins. I then ran the period counts and looked for integer bin for each increment, averaging each count and saving it for later plotting. I ran this through 738 counts. This comprises about 2214 days, or 6 years.

    This is the plot

    I have labeled prominent points. The ones that really stick out are likely the Bárðarbunga and Grímsvötn systems. The other peaks I don’t really have much of a clue about. But you two might, so I figured I would share.

    For the casual observer, this doesn’t prove anything. It’s just an interesting way of looking at quake data.

    The original post at WUWT that sort of lead me down this path.

    1. Note: That “six years” is a six year window. Any periods greater than that would not have been detected. This window was applied over the entire quake collection (1995 to 2011)

      1. There is no date-line. This just identifies time periods in the data.

        A few days ago, Boston Volcano Heads pieced together a long graphic of the areas around Katla, Eyj, Goda etc.

        That long running graph, (actually, one for the whole of Iceland) would be how to key this into a specific start date… but it may not (likely) fit any specific area. This is just a characteristic of the data.

    2. A more interesting application might be detection of harmonic tremors, which I believe currently uses a FFT filter. Of course you’d need to get your hands on some raw digitized geophone output, you could make an interesting plot showing predominant frequency by colour (say red – blue for 0.5hz to 4hz) and amplitude on the y axis.

      1. Being as a quake is made up of sinusoidal waveform, yes, that would work just fine. In fact, that is one way that raw seismic signals are processed. (It may even be in Jons software package).

        Seisgram2K has that feature if you can get your hands on the raw seed files for the seismic gear in Iceland, but all I know of is one station in Reykjavik that is part of the Iris BUD system.

        Failing that it’s a no-go.

    3. Came back from a trip to Russia only to see the most interesting volcano week went already! :o(

      Lurking, Let me look into it more closely… I’ve heard both of R-language and periodicity analysis before, but had no time to dig into details by then.

    4. It is very hard to interpret it in any way… Surprise is how small the Eyja event is.

      Regarding the 2 large spikes 3,5 yars ago. One of them should be the 2008 swarm-storm of Reykjaness Ridge, bunches of 4s, not many know about them though, I would guess that is the highest of the peaks. The third highest could be the 2008 Sprungur-Quake. The second largest I think might be TFZ, if I remember correctly that one had a heck of hickup around that time.
      Tectonically speaking 2008 was a whale of a year so I am not so shure that things are magmatically related for that year.
      Caveat, I am by far not shure… I am guesstimating wildly here!

      1. I’m sorry, that is not a “time in the past” sort of thing or any specific events.

        Like an FFT, this is a measure of the strength of the periods contained within the quake count for Iceland.

        But, I emphasis that this is not an FFT. An FFT yields the size of the various sine waves that that you need to generate the original signal. All this does it point to periods that are in the signal… no matter what makes them. (square/triangle/pulse/sine/complex whatever, even a combination of something… even the periodic bumping of plates)

        I also guess I need to apologize for the plotting software, it likes to connect the points in anything plotted. A more realistic plot would be just the points. This is not a continuous function.

        Now.. a klutzed up explanation of it. The linked references will give you better detail, but I’m gonna take a stab at it here.

        Once the data is in a regular ordered time sequence, done here with binning the quake count. (number of quakes every three days) or taking a monthly measurement, you number the individual measurements/data.

        sample/measurement # 1
        sample/measurement # 2
        sample/measurement # 3
        sample/measurement # 4

        If you divide the sample # by 2, half of the sample #’s will be integers. Filtering for that, and then finding the average value gives you a value for “2”.

        Repeating this across the whole set (3, 4, 5, 6, 7, 8 …) yields the curve in my original post.

        All it indicates is that there is some periodic component in the original data and that that component has a relative strength when compared to the other values.

        It doesn’t tell you what it is, just that there is something in there.

        Since I was using bins of known size (3 days), I can convert the period counts back into something we are more familiar with.

        Again, this is not a “time into the past” or “specific event” sort of plot, just that something of that length is in the quake data. That’s why I think that the ≈3.3 and ≈3.5 year signal equates to Grimsvotn and Bardabunga, thats in the ballpark of their average repeat time.

        A better plot/version of this plot would be to extend the size of the window past seven to nine years so that we could detect longer period events. Once the window size gets larger that one half of the length of the sample data, I have a feeling that it’s usefulness starts to degrade. This is new to me so I can not really say for sure.

    5. That is an interesting plot. I wonder if the periodicity remains roughly the same if you omit, say, the 1st or the last quarter of data. This might help to tell if some of the peaks are purely coincidental.

      To get some sense of the significance of the peaks, one could analyse synthetic data with the same total number of quakes as well, e.g. fully random and data with pre-set periodicity – or some combination of these.

      1. Actually, that would be a pretty good test. I can generate Gaussian noise using the mean and sd of the original data.

        If I remember correctly, Jack (I think) had recommended that method when dealing with peak significance in FFTs.

      2. The really cool part about this method, is that the data itself spits out the information.

        I plan on filtering the events down to the area from Hekla to the triple point and running this against that data to look for periodicities there.

        That should partially nail down the proposed movement of stress forces there… and possibly add to Carl’s ammunition belt.

    1. Wulffmorgenthaler usually has some dark jokes but they are funny. 🙂 This one made me giggle. 😉

  3. How about the earthquake swarm in Northern Italy? More than twenty at the same coordinates!

  4. Jon: What are the resolution and noise figurs for your helicorders in terms of g (mean acceleration on Earth due to gravity, 9,8 m/s^2)? Are they in the milli-g or micro-g range?

  5. @Irpsit AND alan_c:

    You guys have answers in the old post, I didn’t notice that Jón hade put a new post untill I had answered.
    Both answers are technical and directed directly as an answer to questions from you guys about technology or safety.
    alan_c, please do not hesitate if you wish a more detailed explanation.

  6. Krafla is still active.
    I have tried like a maniac to find out if they are doing something new and the Krafla Geothermal Station, but as far as I know they do not, and the test-drilling at Theystareykir seems to be over for the the season.
    So the harmonic tremoring, is probably not man-made. As far as I know it is though not magmatic. Probably just some sort of hydrothermodynamic event happening.
    I would though like to caveat this, it looks suspiciously periodical for being nature only playing around.



    1. Noticed the acceleration in the periodicity? First approx. 12 hours, then 10 hours, then 8 hours, … Something is cooking there. Is it Mother Nature or some human bastard?

      1. Thars da rub…

        I cannot get my head around if it is human-made or nature. I tried to call Landsmannavirkjun earlier this morning to ask is they are trying to run water into the deep well that they have drilled into the magma-reservoir (Worlds first magma sample!!!).
        Because they might be trying to push water in incremental steps into that one. But, as far as I know that drill-hole was south of the inverse graben and should be noticeable on the SILs south of it, not north of it. It is still the only thing I can imagine giving results like that and still be man-made.

        The other explanations are A) some naturally occuring hydrothermal event in increasing periodicity, or B) rapid magmatic filling of the system north of the Graben. I have a problem with B:
        If it was magmatic it should give quakes north of the Graben since that is not a dead zone. South of the Graben is the Krafla ductile zone that only gives micro-quakes when filling.
        As far as I know the krafla-fires was south of the graben, but I might be wrong on this.
        As usual, time will tell.

      2. It would be the second magma sample. In 1977 magma came out of a geothermal steam well during the eruption of Kafla on Sept 7th, 1977. A fracture is thought to have broken the well casing. The well seems to emitted small lava bombs.

      3. P.S. my source was a book I have just borrowed from my local library: “Volcanoes” Academic version, third edition, by Robert Decker and Barbara Decker, W H Freeman and Company, 1998.

  7. Good morning!

    If there is a geothermal activity there around this regular movements ain’t them like a geyser like pattern?
    Maybe a new born?? 😀

  8. I have a small question regarding the earthquake map. Some quakes have a black outlining, other have grey. Does that mean anything?

  9. @Norway: Black circle: Confirmed and checked by IMO.

    BTW: She’s alive!

    09.09.2011 08:50:27 63.636 -19.174 1.1 km 2.3 85.77 3.8 km E of Goðabunga

  10. @ Carl & Jon
    Thanks for the seismic workings info – my insight into seismics is very dated and limited mainly to borehole airgun data responses and a little vibroseis surface analysis, from the early 80’s (I’m now a petrologist with Norfolk CC looking at Glacial Drift – boulder clay and sand&gravel all very much the same!!). I’m afraid accelerometry is too new for me Carl, so if you could – just a little info please, I’d be v grateful.
    OT but tropical storm Maria look to be heading splat towards the Windward Isles then SE USA coast(ish)

    1. An accelerometer is a solid state digital component that measures the acceleration in a vibration or waveform in (normally) a binary form. They are cheaper, more exact and more reliable than the old analogue MMC seismometers.
      It measures the power of the acceleration in the form of Gravity and then the duration of the shift in movement. It gives as you understand very exact measurments when dealing with low frequency waves. It is also very easy to transfom into whatever power-unit you wish, joules, watts, erg, or “mega-fippel” (mega-fippel, any valid energy-unit, one mega-fippel is of course 1 million times larger than the ordinary fippel).

  11. The earthquake swarm close to Hellisheiðarvirkjun is due to the fact that they are pumping down cold water into the bed rock. That creates earthquakes. From what I can tell, I am rather sure that this earthquake swarm is man made.

    It seems that Katla volcano has now started a new earthquake swarm.

      1. okay, thanks.. that could perhaps be 😉

        but.. sorry for the dumb question now: .. isn´t it a little dangerous to drill holes inside a vulcano, which could probably errupt? – just to enjoy some geothermal heat?


  12. Out of interest, what constitutes a swarm? Is it a set number of earthquakes of at least a certain size over a specific time frame?

  13. Since yesterday morning I have been unable to access tremor, strain and GPS measurements on the IMO site. If I click on articles, volcanos, reports, individual EQ maps etc there is no problem I am straight on!

    Has anyone else experienced the same glitch? Anyone any ideas why this is happening?

    1. Please read…
      PFXD01_EGRR – Received at 1054 on 8 September 2011 – (PNG, 85 kB)
      Some folks? I know it’s all a bit misty on top.

      1. What exactly are you pointing to? That it’s a test? That was clear to me – that’s why i said they’re preparing. 🙂

        What do you mean with “some folks?”?

      2. Tests happen all the time, equipment needs checking.
        There was a point made about sensationalists earlier.

  14. The possibility that the Katla caldera gets bigger by eruption and small inwards flank collapses, which then get covered by glacier and eroded away to the sea by glacial action and run, was brought up in discussion with interested friends today over lunch. The caldera could have grown like this slowly?

      1. That graphic is showing well the confusion of the Katla caldera, inwards collapse is possible with that diagram.

      2. Calderas under glaciers are often larger from erosion. Think about it, the floods after an eruption take a lot of material. These floods heavily erode things both outside and inside the volcano. After the eruption is over the glacier pushes back in, and grinds the ground up more. Also there is steam explosions that help to break up the ground. No vei 7 eruption needed to explain katla caldera.

      3. We were thinking also that when ice is melted and the inwards edge of the caldera is unsupported by the glacier, it would tip or fall in. To be eroded away…

  15. They need to do some work on their automated system – how can they register EQs of 2.5 and 2.3 and then realise they’re only <0.5 – that's an energy difference of a multiple of 1000 (assuming they're using Mw and not Richter)

    1. so i hope that the “might be” erruption is not such big as 1815 (the year without summer) !

  16. There seems to be a new earthquake swarm in Katla volcano. So far it is quite small in size. Both regarding number of earthquakes and the size of those earthquakes.

    1. I noticed this Jon. It’s Friday Night remember ! Katla is getting ready for the weekend again.
      By the way ! Does the IMO weather forecaster know something we have missed? Ashes expected in the south according to Giggle translation! 🙂
      “Forecaster’s remarks
      A drifting ash is expected in the southernmost part.”

      The UK has a weather warning out now for Monday. Severe winds may cause structural damage and travel disruption from the remains of Hurricane Katia!

      1. This is wind blown volcano ash from Eyjafjallajökull volcano and Grímsfjall volcano. It is going to remain a issue in Iceland for years to come. As it takes time for the largest part of it to get buried down into the ground by nature.

      2. Oh Goodness Jon! It must be a real nuisance especially for farmers and crop growers.
        I didn’t realise this fact. I am sorry if I seemed flippant. I didn’t understand.

      3. Hurricane Katia is still passing to the east of North America, it’s not the storm that is about to hit the UK! Please be aware of that. The same was for storm Irene, it only passed by Iceland, but it did not affect at all the UK of which some people spoke of.

  17. Just a small (maybe stupid) question…

    In the event of an enormous eruption in Katla, could we witness another caldera formation if the eruption is violent enough? What I mean is: Could the caldera-bottom/magma-chamber-roof give away again to form even deeper and maybe bigger caldera?

    just speculating

  18. to my uneducated eyes the quakes on the map these past few days seem to be outlining a circle.

    (photoshopped image courtesy of Icelandic Meteorological Office)

    a similar circular pattern shows up in this map Jon posted on an earlier blog post..

    (Copyright of this picture belongs to Iceland Met Office.)

    what do you think? someone’s probably pointed this out earlier, I know. I’m probably way off, but I see patterns in things. A bit too much at times, as Renato will testify. lol

  19. Drifting ash definitively present in the air. I can see it to the south this thick brownish haze, but it did not reach here due to the northerly wind. I can’t feel any ash in my eyes yet.

    The weather is pretty cold, went down to -2ºC at night, and only up to 7ºC during day.

    And best of all, there was just now the first impact of the solar flare a few days ago, so we will get big, bright and fast-moving northern lights tonight!!!

    1. I’m sorry Irpsit but Katia is heading right for the UK for late Sunday and Monday. We’ve an “amber” warning out for north England and South Scotland. Please do check out the Met Office Bracknell charts:
      I’m rather fed up about it as my wife flies back in to Heathrow right in the middle of it all. Hope the pilots are OK landing in a stiff gusty cross-wind!

      1. Planes often fly in hurricane / tropical storm conditions so your wife should be fine. If the wind is in the right direction, she might even arrive here early. If she has a tendancy towards air-sickness, maybe she should consider taking the tablets / remedy with her.

        And yes, I have seen the weather warning too. Also saw the projected path, surprised they didn’t think was going to come further south. Suppose they hope the Gulf Stream will direct it that way.

  20. Update from Iceland!!!

    I went outside and I saw something crazy!!!

    Katla is not to my east anymore!
    She left to Reykjavík!

    A friend called me and told me that Katia was seen walking down Laugavegur (Reykjavik main street) getting already drunk at bars and it’s only 6pm!!!

    What is going to happen tomorrow, OMG!

Comments are closed.