GPS changes in Grímsfjall due to ice on GPS antenna

I did just speak with geologist at Icelandic Met Office and the he told that the recent changes in GPS measurements in Grímsfjöll where not real. But they where created by ice. But now the ice has been cleared of the GPS antenna and the data is now correct.

But the real movement is to the south and Grímsfjall has continued to do that. So Grímsfjall is continues to inflate at relative fast rate and prepares for a eruption as magma flows into the Grímsfjall magma chamber.

68 Replies to “GPS changes in Grímsfjall due to ice on GPS antenna”

  1. Wait a minute…. that annual signal might be some poor soul that has to trudge out every winter and scrape the ice off?

  2. @ Jon – What is your take on the tremor at SNB today? It seems only to be the 1-2Hz band which is affected. Is it harmonic tremors due to magmatic activity?

    1. @Daniel_Swe, I am unclear on what that actually is. I didn’t show up on my geophone. But if it is something interesting, we might get a repeat on that signal soon.

  3. Jón: What I don’t understand is if the recent anomalies observed from GPS measurements are still reflecting a different inflation pattern from before, or if they just came back to its former trends.

    1. @Renato Rio, When the ice is gone I would expect the GPS station return to normal trend. That trend being measuring the inflation at Grímsfjall volcano.

  4. @Renato Rio

    What should be making you curious is how GPS, which relies on the timing of signals and not the amplitude, is so adversely affected by ice to generate an annual bump.

    Personally, I’m not buying it. Ice on the antenna is going to affect how many satellites area available for a fix. Thicker ice means fewer satellites can penetrate the ice and therefore the fix quality would go down. This should show up as an increase in the error rate (wider vertical bars), not the actual central point. The central point is the average of the fixes.

    You can see this in the latest GFUM plot.

    So… back to the original question, why the climb in winter? Antenna ice doesn’t cut it.

    1. @Lurking, The climb during the winter would be inflation in Grímsfjall volcano. But there are other factors that go into this. That being snow and how much of it is coming down during the winter and how much the summer melt was and so on.

  5. BTW… I found my coffee.

    LORAN C, TACAN, OMEGA, and GPS are all based on timing signals. Sure, the old OMEGA system used a measurement of the signal amplitudes in one of it’s early incarnations but they did away with that later. All the amplitude determines is if a receiver set can get a good enough signal to determine the timing from it. As mentioned in my last post, amplitude doesn’t determine the fix, just the timing.

  6. @Lurking glad u found your coffee and ice doesn’t sound plausible to me…perhaps it was an excuse for a first hand observation.

  7. Well, i’m glad Jon could talk to an official smart person…………… it’s nice to have an on site person. And ice can do some amazing things so i’d have to believe them.
    Poor person that had to go out and scrape the ice off………. we are scrapping the ice off the car windshields in the mornings now……….. Any days past Sept 15th without snow are extras……….. That’s the earliest i’ve seen it snow here and stay. (it doesn’t count until it stays 😉 ) And since it will be here until April; i’m not anxious to have it come. Heavy frost tho and the temp says it’s 27 F.

  8. @Lurking: I’m sure you must have a point there. But cannot discuss it any further for sheer ignorance. I don’t clearly understand what do you mean by “The central point is the average of the fixes.”
    @Mots Fo: Not an ice expert here, being from the tropics. But I do think ice is a tricky thing. Just think about the way water volumes increase when solidifying. Just figure that in a wider, glacial, scale. Don’t know what to say, but, “keep yourself warmed, Mots Fo”. Send you a warm nod from Rio (it’s nice and cool right now. 19º C). Best!RenatoRio

  9. @Lurking and everyone,

    I don’t think the ice had much influence on the reception of the GPS signals. Rather, I could imagine that the ice buildup on the antenna caused the system to be dislocated by lever action and thus create the extra few centimeters of ‘inflation’. I don’t know what kind of set-up they are using up on that glacier, but such an explanation would make more sense to me….

  10. It’s not at all that hard to believe the explanation provided (ice on antenna). Lurking has it partly right that GPS positioning uses the timing of signals to calculate the position.

    To put it more sharply, I’d say GPS positioning uses the time differences of several radio signals sharing an identical base time reference (GPS master clock, i.e. intenational UTC time from a bunch of atomic clocks). The are typically 5-8 satellites available for evaluating the relative differences for each pair of signals. None of them is held as “local reference”, but from the time difference values the position cal be calculated for each base time value.

    Well, what’s up with ice then? The optical index of refraction for ice is about 1,3 but for radio frequencies the index of refraction varies with frequency, ranging typically from 1,4 to nearly 2,0. The laws of electrodynamics state, that the propagation velocity of an electromagnetic field in a medium is that in vacuum (i.e. speed of light) divided by the index of refraction at the given frequency.

    Well, this enables us to estimate the thickness of ice needed to cause the apparent inflation of 60 mm! So, light (RF wave) travels 60 mm in air (or in vacuum) 0,2 nanoseconds. In 60 mm of ice (assuming index of refraction 1,667) light travels in 0,33 nanoseconds. The difference is 0,13 nanoseconds, or 39 mm in air, or 23 mm in ice!

    So, an ice cover of approx 20-25 mm on top of Grimsfjäll GPS antenna is enough to explain the observed apparent inflation in the direction UP. Since it caused also a 40 mm shift to north, I expect to have seen (if I were there to clear it) something like 10-15 mm of ice on the northern face of the GPS antenna…

    The only case, when ice accumulation does not affect the results of positioning is when the ice cover is even in thickness AND spherical in shape.

    1. This is only due to the fact, that the antenna or GPS software does not know if it is covered with ice or not.

      1. @Jack: Thank you very much for the explanation / calculations. It seems very clear to me now. Still we need somebody to go to the spot and tell us what’s really happening.

      2. Jack @ Finland says: that the antenna or GPS software does not know if it is covered with ice or not.

        In “Ice” Land? Lol. (Freezing rain ?)

        Would have thought that the effects of ice/snow build up and and design features to to manage it would be built into the instrumentation. Live and learn, etc.

  11. Anyone care to venture and educated guess on the probably VEI if this does erupt and whether it would likely be subglaciel or a large phreatic eruption to the surface?

    Also, is anyone expecting that there would be a high frequency of EQ’s moving upward as an early warning for Grimsvotn’ or a large EQ? I am wondering how much warning this will give if it does have an eruption.

    1. Ron: As Jón has posted earlier, it would take a big quake swarm short before the eruption as a sign of its approach. As for the VEI – well I don’t think this would be much different from the other Grimsvötn’s periodic subglacial eruptions beneath Vatnajökull. So I hope, for they can get dangerous if a larger fissure occurs and the lake under the glacier meets lava.

  12. The only problem with Jacks explanation is that all signals have to pass through the ice. (assuming even coating) The relative differences are not going to be specific to one or two satellites since the entire constellation is in relative motion. I can see the apparent delay causing a perceived decrease in elevation due to the additional lag in time. Not an increase.

    1. Lurking: I’m not saying I agree with Jack’s explanation I only said that he gave me the tools to better understand the problem we are talking about. I’m sure you know exactly what you mean and I’m eager to follow the discussion, from the back. Many thanks.

  13. Following are some 3D images of SE Iceland, with GPS site locations marked and caldera rims outlined, using Google Earth to obtain a 3-D perspective (except as noted). I especially like the 3D views to get a feel for where the heck the GPS monitors are compared to the topology of the volcano itself. And seeing the outline of the calderas in 3D gives a much more vivid and memorable image of volcano.

    The GPS locations are exact, as specified by the GPS installers. The caldera are taken from MET maps, with the image copy carefully positioned in Google Earth as an overlay, then using the Google Earth path tool to trace the caldera rim. As a last step, the caldera was matched to the terrain using high zoom and adjusting the points making up the outline of the caldera in order to get as best as possible positioning.

    Since 3D perspective is used, the distance scale varies with location, so the only picture where distance can be fairly accurately determined for any locations on the earth surface shown is the one specified as “Overhead“, since that is actually a 2D representation, like an ordinary map.

    Eyjafjallajokull & Katla viewed from the north in 3D, showing GPS sites for AUST, ENTA,, GOLA, FIM2, BAS2 & STE2.

    Eyjafjallajokull & Katla volcanoes of Iceland, showing 10 GPS monitoring locations. In 3D, w/ calderas outlined. …GPS sites shown: HAMR, DAGF, SVBH, THEY, FIM2, SKOG, SOHO, GOLA, ENTA, AUST, HVOL; and more at Hekla.

    Overhead view of SE Iceland= Eyja, Katla, Hekla, etc showing all calderas, scaled distances, & GPS sites. (GPS sites shown: HAMR, DAGF, SVBH, THEY, FIM2, SKOG, STE2, BAS2, SOHO, GOLA, ENTA, AUST, HVOL, SNAE, with Hekla sites in background.)

    Wide view of Iceland from Eyja & Katla volcanoes to the north coast: showing GPS locales, claderas, & glaciers.

    Wide view of SE Iceland showing many volcanoes, calderas, glaciers and GPS monitors for ground movements.

    Eyjafjallajokull & Katla volcanoes of Iceland, showing north-side GPS locations. In 3D, w/ calderas outlined.

    Eyjafjallajokull & Katla volcanoes of Iceland, showing 10 GPS monitoring locations. In 3D, w/ calderas outlined.

    Hekla volcano, Iceland, in 3D, with 6 GPS monitors shown, to detect ground uplift & movement (before eruption).

    Torfajokull Volcano, Iceland in 3-D with caldera outlined. Eljafjallajokull, & Katla also shown.

    Tinfjallajokull Volcano, Iceland in 3-D, with caldera outlined.

      1. Are there enough stations as to make a 3d for Vatnajökull and its many volcanoes? Tks.

      2. Unfortunately, for Vatnajokull most interesting volcanoes there are way under the ice.

        And the GPS units have to be cemented to solid ground, so they are all off the glacier, except for GFUM, for which a little area of rock was found for its feet.

        This shows all the GPS on the island, including Vatna area:

      3. Thanks a lot @ William. That’s great! But I can’t open this links:

        Eyjafjallajokull & Katla volcanoes of Iceland, showing 10 GPS monitoring locations. In 3D, w/ calderas outlined. …

        Overhead view of SE Iceland= Eyja, Katla, Hekla, etc showing all calderas, scaled distances, & GPS sites

        Wide view of SE Iceland showing many volcanoes, calderas, glaciers and GPS monitors for ground movements

    1. Three of the above inexplicably went missing at twitpic, after definitely being there. The have been added back, but at new urls, which are in a post below.

  14. Monday
    11.10.2010 19:58:02 64.398 -17.280 3.0 km 2.1 90.01 0.8 km SSW of Grímsfjall

  15. @Renato

    saw that one just too. wonder what will happen next, will there be a sign of this in the tremor graph or not.

    You’ve created a wonderful Blog, just the right thing for an iceland&volcano-fan like mine.

  16. Well, part of being honest with yourself is being honest when you are wrong.

    Intuitively, Jack’s explanation was wrong to me. But I now think he has it right.

    Water has a dielectric constant of about 80.4 at a temperature of 68°F (via, Ice is around 4.15 @ 1 MHz / 3.2 @ 3 GHz (again, via This means that you will have to have quite a bit of ice to be equivalent to water.

    So… I ran an experiment.

    I took a “SmartBlue” bluetooth GPS module (for use with laptops and smartphones) and set it up in the back yard at about 0.45 meters (foot and a half) above the ground. I ran the unit for about 730 samples and saved the data. I then took a ziplock bag and placed 1 cm of water over the unit, and did another 730 samples. Without water, the unit averaged 44.568 meters. With water, it averaged 59.9306… a 15.3626 increase in elevation. (at 3 cm of water, I lost all signals and could not even connect to the unit)

    Bear in mind that this is a consumer grade terrestrial GPS unit. The elevation reading is not it’s strong point.

    So… here is a plot of the data.

    1. I didn’t throw the http:// part on rfcafe… (didn’t want a link to it), yet the forum software did and now placed my post in a wait que?


      1. Nah, just stubborn.

        I’m a technician by training, not an engineer. In my worldview, technicians bust their arse to make it work, engineers bust their arse to figure out why it broke.

        With that in mind, I am always a bit conflicted when someone refers to me as a field engineer… which technically is my job description.

      1. It has been revised. Here it is:
        11.10.2010 19:58:02 64.398 -17.280 3.0 km 2.1 90.01 0.8 km SSW of Grímsfjall

  17. Diana, thanks for letting me know that 3 of those pictures went missing. I uploaded them again, but they have new urls …

    Wide view of SE Iceland showing many volcanoes, calderas, glaciers and GPS monitors for ground movements.

    Overhead view of SE Iceland:= Eyja, Katla, Hekla, etc showing all calderas, scaled distances, & GPS sites.

    Eyjafjallajokull & Katla volcanoes of Iceland, showing 10 GPS monitoring locations. In 3D, w/ calderas outlined.

    The links for the other 7 pics (posted up above) continue to work ok.

    I think the disappearance of the 3 pics might be a conspiracy of Google and the Borg to prevent disclosure of the proof, shown in the pics, that the Borg Cube is in sector 001 (Earth!) …!! …. (See the proof and comments here.) …. Oh no! That also implies that twitpic has also been assimulated by the Borg. This is scary!

      1. @ Jón: You were faster with your answer. I forgot to register again. I really learn a lot in your fantastic Blog! 🙂

      2. Of course I meant “Login”. It’s a little bit late in Germany…! 😉 Too late to follow all the water postings. Borgs are easier to understand at this time. 🙂

  18. Lurking, … amazing detective work.

    But if a cm of water can have that effect on the GPS’s displayed elevation, then what about when it rains? … and transmission through the atmosphere saturated with water? Why isn’t that seen in the GPS plots?

  19. It probably is… to a point. There is also a very real possibility that all I saw was the vagaries of satellite signals, the trailing end of the no water trace is suspiciously close to the beginning of the 1 cm water trace… almost as if you could fit them together. Seeing as that one sample set was started within a few minutes of the other, it’s possible. Also note that the set with the water over the antenna had a lot more variation in the signal. I think that’s important. As for atmospherics not having the same effect, well, this was a solid mass of water. Rain/snow/sleet/clouds would offer a more diffuse attenuation or dielectric effect.

    While revealing that I may have been totally wrong in my initial assumptions, it doesn’t indicate just how much of an effect that it really is. In order to have half a chance of doing that, many more test runs would have to be done, and the use of a GPS unit that is designed to accurately measure altitude would be a boon. Probably at least to the level of survey equipment. This is just a consumer grade unit made to get you from point A to point B.

  20. Well folks its all about D-GPS, meaning differential GPS. To get it down to the MM level it has to sample at that bandwidth on an oscilloscope if you wanted to use that as a measuring device. California uses lasers, and D-GPS. Thats where a local fixed transmitter is used as the base. It also has about 14 feet of error for every give or take 80 miles. The instruments are working just fine with or without ice because the degradation of the sampling is taken as a high, low and in between. The Goldilocks info. The PRF (pulse repetition frequency) is 4-6 pulses per second. That means that even if one satellite is snafu’d and giving slightly erroneous data transfer, the other 6-8 up 14 in places will pick up the slack.

    Ice would degradate the signal from all of the satellites if the receiver head was covered depending on thickness. Maybe not a lot depending on the density (super cold ice) or nearly water ice. Aircraft using point in space computers that are constantly updating are not as a rule hazarded by this either as the signal is very strong and the clouds they may be flying through roughly uniform in density. But for landings, the best they can do right now is 14 feet from centerline. If the three landings that Flight Check aircraft perform are uniformly 14 feet, they adjust the points in space for GPS approaches right or left and vertically, and most of them will also use a fixed transmitter (DME/NDB-Distance measuring equipment which is provided by the TACAN portion of a beacon, or a non-direction beacon) to use as the final fix. In Iceland the distance to that NDB or even a normal ILS to pick up the slack could be large. We dont flight check instruments on the sides of volcanoes… they are kinda stuck on their own for what is right/wrong. Generally speaking though if you depot one, they take the GPS out and test it and if its good, its likely good for a year or more before calibration is required.

    But forget all this truss…. The question is whether this mother is inflating… It is.

  21. And yet another one:
    12.10.2010 02:08:51 64.496 -17.423 1.1 km 2.7 90.03 12.4 km NW of Grímsfjall

  22. Jón: What wee those spikes at your helicorders from 21:10 to 23:40 h and at 5:10?

  23. @Renato Rio, That was human noise. That is traffic of some type. I don’t know what it was doing there at this time of night, as the area is remote and just has summer houses.

    1. I was wondering if I had a helicorder with such readings whilst I was asleep I would get scared… 😉

    1. @Jon: Well, a lot of small earthquakes… You surely know, what it may mean?!

      1. @Jack @ Finland, This is high season for small earthquakes in Katla / Mýrdalsjökull so it might not mean anything. It just earthquakes and so far a lot of nothing is happening at the moment.

  24. Volcano monitoring is serious business…sometimes humor makes it easier…

    Mom and Dad Volcano:
    What did the dad volcano say to the mom volcano?
    Do you lava me like I lava you?

    A man goes into a restaurant, sits down and starts reading the menu. The menu says:
    Broiled Accountant $5.95 per plate
    Fried Engineer $7.95 per plate
    Toasted Teacher $7.95 per plate
    Grilled Geologist $25.95 per plate
    The man calls a waiter over and asks “Hey, why does the Grilled Geologist cost so much more?”
    The waiter says, ” Are you kidding? Do you know how hard it is to clean one of them!?”

  25. While looking at GPS data last night… something occurred to me about the coordinated swings in this data… you know, where all stations make a simultaneous move.

    The clue is “miðað við REYK”

    Roughly… if I understand it right, ‘in reference to Reykjavík’ Remember that this is all relative positional data.

  26. And……

    Reykavik is the base station for their D-GPS and the Icelandics map show it in relation to the fixed facility of REY. I guess it makes sense, probably too expensive to tag up a multi-axis GPS system and then have them reference through the system.

    Elevation too would have a tremendous effect on the strength. The water signal thing from above doesnt really flip my skirt too much but its damned good thinking. This was always developed for putting a bomb in the barrel. Hence the term pickle one off…Pickle barrel.

    Subs use INS and submarine buoy pings for positioning that come out. Eg. SOSUS can talk to them if they want it to by sending a pair or more of undewater signals.

    But as to the ice. Ice is different from water, the signals would be attenuated without a doubt, but the fact is that the signals could have been attenuated and then…suddenly they got their chipper and deicer cans out and its still up by what 40 MM?

    I leave it to the scientists here about whether its going to blow its head off and how bad. Aviation though is my rice bowl. They are using an aviation application and the equipment again, is working fine. If there was only a wind of 5 mph, sublimation would take a 1 inch ice coat off in a couple of hours on the leading edge of the receiver and it would get those signals. I am quite impressed with the brain pan of this little august group though…Lotta thinking going on. For me, I think they should go do their site surveys for when this motor scooter gets started so they can place them according. Laki and this thing as I read it gin out Hydroflouric acid in high levels.

    Them dummies in the EU might fly through weak ash clouds, but not acid. All of this junk mostly crossing the pond now is fly by wire and wires and air data systems will not tolerate acid for very long.

    It will blow and they wont go, then they’ll parse and put a few up. Then they’ll cross and say everything is fine..Right up until the second they get the first MAYDAY call.

    I can hardly wait. Remember when we only had to think about terrorists on aircraft?

Comments are closed.