Green Gabbro

If you like apocalypse porn, you have probably been following the current earthquake swarm at Yellowstone. While there is no reason to believe that this is part of the lead-up to a giant caldera-forming eruption that will wipe out most of North America, or indeed to any eruption at all, there is also no reason to let that spoil your fun.

If you’re watching the recent earthquakes page, take the reported earthquake depths with an especially big piece of salt. Earthquake depth is difficult to accurately determine, especially with real-time automated processing techniques. Think about it: You’ve got seismometers scattered all over the surface, with lots of variation in x and y but no variation in z.

So even if earthquakes are listed at, say, 0.2 or 13 or 37 km, that doesn’t mean there is actually a “chimney” of seismic activity that penetrates the Yellowstone magma chamber or reaches up to the surface. According to USGS scientist Jacob Lowenstern, the most accurately located earthquakes so far have all been between 3 and 5 km depth.

You can also see more or less real-time data from Yellowstone’s continuous GPS network (though this is not a candidate for obsessive reloading – because most surface deformations of geological interest are smaller than errors in an instantaneous GPS measurement, the data is not useful until it’s been time-averaged). If there is something interesting (inflation or deflation) happening to Yellowstone’s magma chamber, or a gas-pressurized reservoir, we would expect to eventually see it as a slight change in surface topography.

Comments

  1. #1 bigTom
    January 2, 2009

    Earthquake depth can be pinned down pretty well, if one looks at the waveforms, not just the arrival times. Reflections off the surface, and the mantle crust boundary have a large impact on the waveforms. However, you are probably right, in that these depths you are referring to are probably generated from only arrival times.

  2. #2 Maria Brumm
    January 2, 2009

    IIRC, they start doing full waveform modeling and moment tensor solutions at magnitude 4; largest event in the swarm so far has been 3.9.

    Or maybe I am confused, and that’s just how Berkeley does it. Anyway, even with a full waveform your locations are still only as good as your earth structure model. I would expect any serious scientific work on the swarm to begin with double difference relocation of all the catalog events.

  3. #3 Greg Laden
    January 2, 2009

    Who knows. Since we’ve never seen the run up to a mega super caldera gigundaruption, perhaps they start with a series of tiny earthquakes in the middle of the eruption zone and then suddenly…..

    Oh, probably not.

  4. #4 christopher guerra
    January 4, 2009

    “they start with a series of tiny earthquakes in the middle of the eruption zone and then suddenly…..”

    THAT’S what she said!

    *crickets chirping*

  5. #5 Pyre
    January 5, 2009

    “Earthquake depth is difficult to accurately determine, especially with real-time automated processing techniques. Think about it: You’ve got seismometers scattered all over the surface, with lots of variation in x and y but no variation in z.”

    Variation in z, helpful as it might be, isn’t necessary.

    Visual parallax determinations aren’t the best analogy, but your two eyes (or telescopes taking photos six months apart) actually determine depth best when separated along a line at right angles to the target’s line-of-sight. Likewise the Very Large Array of radio telescopes (along a plane at right angles…).

    But (you object) a single eye affords 2D location, a line or direction to the target, which a single seismograph cannot?

    Fair enough. As few as three seismographs in a roughly equilateral triangle, however, can together give that sort of information. Add another set of seismographs for another “eye”, and you’ve got parallax. Add still more seismographs, get better “vision”. They may all be in the same plane, but they can still measure distance at right angles to that plane.

    The problem is that the ground is “foggy” with varying materials/faults/boundaries, affecting propagation rates and cluttering noise (e.g. the echoes or reflections bigTom mentions). The “picture” often needs enhancing.

    But as a geometric issue, a sensor array needn’t have a “variation in z” in order to measure a target’s distance along z.

  6. #6 Rob R.
    January 7, 2009

    Hello Ms. Brumm,

    You can also see more or less real-time data from Yellowstone’s continuous GPS network

    [...]

    If there is something interesting (inflation or deflation) happening to Yellowstone’s magma chamber, or a gas-pressurized reservoir, we would expect to eventually see it as a slight change in surface topography.

    I’ve been following along with the recent happenings at Yellowstone (that is, as best as I can as a layman) but haven’t seen that site before. Could you explain (or link to) what I’m seeing there and what a “change in surface topography” might look like? Thanks in advance.

    Regards.