12 March 2011

Why earthquakes and eruptions are rarely linked

Posted by Jessica Ball

USGS Earthquake Location Map for the 11 March Honshu earthquake

The geoblogosphere – and the rest of the news – have been buzzing with information and discussion about the recent M8.9 earthquake in Japan. Despite being a country that is relatively well-prepared for events like these, even Japan couldn’t withstand the power of such a quake and the resulting tsunami, and they will need help. Please consider donating to a relief organization such as the Red Cross, Doctors Without Borders, or Save the Children.

ResearchBlogging.orgErik Klemetti over at the Eruptions Blog mentioned the new eruption at Karangetang volcano in Indonesia, which occurred just after the earthquake in Japan – and made the point that the two events probably aren’t linked (since the volcano was already showing signs of unrest). This is likely the case in most situations where eruptions occur in close timing with earthquakes – so the answer to the question of “Did this earthquake cause an eruption?” is, probably not. But that still leaves this question: Can earthquakes trigger eruptions? This is the main subject of a paper I read recently: Manga, M. and Brodsky, E., 2006, Seismic Triggering of Earthquakes in the Far Field: Volcanoes and Geysers.

The answer is, predictably, complicated.  Studies have indicated that  eruptions do occur in the vicinity of large earthquakes more often than would be expected by chance – at least on the short term (within 5 days). But this only appears to be true for a very small percentage of eruptions, and only with the largest earthquakes (which implies that the stress changes caused by earthquakes are small compared to the stress changes needed to initiate eruptions). This works out to <<1% of eruptions.

Manga and Brodsky, assuming that very large earthquakes could trigger eruptions, set out to describe how this could occur. They divide triggers into two categories: mechanisms which increase the overpressure within the magma (something like shaking a bottle of soda until it explodes), and mechanisms that could trigger eruptions from outside the magma chamber. (There are a few suggestions made in the paper that I’m leaving out, because the authors make a point of saying that their effects are not strong enough to make them plausible eruption triggers – the criteria is that a mechanism must produce at least a 10 MPa overpressure).

From within the magma,

  • Bubble nucleation: Creating new bubbles by increasing the nucleation rate through pressure differences caused by the passage of seismic waves. This mainly depends on the melt being supersaturated with gas to begin with; if seismic waves cause new bubbles to form, large gas pressures compared to the ambient (magma) pressure may cause bubbles to grow quickly and explosively.
  • Falling roofs: Aggregates of crystals (or crystal “mushes”) formed near the roofs of magma chambers could be shaken loose by seismic waves; the sinking crystals could initiate convection in the chamber, bringing hotter magma to the top. This magma could vesiculate and expand as it rises, creating overpressures and possibly initiating an eruption. This mechanism depends on the actual existence of crystal “mushes” (hard to observe directly), as well as the velocity of their descent being great enough that they initiate overturn on a short timescale.

From outside the magma chamber,

  • Surface unloading: If a distant earthquake was strong enough to shake loose an already weak part of a volcanic edifice and cause a landslide, it could effectively “uncork” a pressurized magma chamber and initiate an eruption. (This is essentially what happened in the 1980 eruption of Mount St. Helens, although the earthquake that accompanied that eruption was not distant or particularly large.) This mechanism has never been observed in relation to distant earthquakes, but it’s plausible if the volcano is weak enough.
  • Fatigue and crack growth: Repeated or fluctuating stresses are known to cause material fatigue and cracking before the normal failure stress is reached. This could be somewhat related to the previous mechanism (in that large cracks could then cause surface unloading), or it could create rifts that magma could exploit to form dikes and initiate rift eruptions (like what’s happening at Pu’u O’o right now). The authors note that this would require very high pore pressures in something like a volcanic hydrothermal system, and even then earthquakes could trigger eruptions via this mechanism only 0.01% of the time.

What is the upshot of all this hypothesizing? Basically, it’s that very large earthquakes could trigger eruptions, but we don’t yet know exactly how – and that it will require a great deal of careful monitoring and study to possibly catch these mechanisms in action. Because we can’t yet say for sure whether these mechanisms actually occur, it’s silly to immediately start speculating about a link between the recent Japan earthquake and any eruptions that are occurring right now. For the moment, it’s important to remember that any natural hazard requires careful monitoring all on its own – whether or not there seems to be an upswing in the occurrence of other natural events nearby. This means that things like earthquake activity, tsunamis, and volcanic eruptions should be paid attention to all the time – not just when it seems like more of them are happening.

And don’t even think about trying to link something like the phase of the moon and natural disasters – if an earthquake usually can’t trigger an eruption, there’s even less chance  a full moon will, no matter how close it is to the Earth.


Manga, M., & Brodsky, E. (2006). SEISMIC TRIGGERING OF ERUPTIONS IN THE FAR FIELD: Volcanoes and Geysers Annual Review of Earth and Planetary Sciences, 34 (1), 263-291 DOI: 10.1146/annurev.earth.34.031405.125125