After discussing how supermoon was not going to kill us all, Donald-the-linguist and Jon-the-astrophysicist became so intrigued by the relationship between tides and earthquakes that I had to pull together a literature review on the topic for them.
A few points to get out of the way:
- Tides are caused by the gravitational pull of the moon and sun on the earth. For more detail: 1 2 3 4
- Earthquakes are caused by tectonic plates getting stuck while moving past each other, until overcoming a threshold and slipping violently. For more detail: 1 2 3
- The reviewed literature is non-exhaustive: I am heavily biased to literature written in English, journals that are either open-access or accessible either my university’s online credentials, and this is a casual review outside my usual area of expertise.
Now, on to a review of the last 50 years of science on the interaction between astronomical tides and earthquakes:
Tides pull on water, but they also pull the ground. The wavelength is so long and amplitude is so small that we don’t notice it, but dirt and rock does get pulled by the tides (called a “solid earth tide”). Logically, it makes sense that if a fault is stuck and has just-barely-below threshold energy build-up, that getting yanked around by tides would tip the balance and trigger an earthquake. Alas, it is not so simple!
Earthquakes on the moon are closely tied to tides (Lammlein 1977). However, the moon has much smaller tectonic stresses, and much larger tidal stresses, so this doesn’t necessarily parallel the situation here on Earth. The quest to make a link between tides & earthquakes on Earth falls into a few categories:
- Apply tidal strain to faults. Mathematically account for all the components of tidal stress, and determine if it’s enough to trigger an earthquake. This technique is limited by real life always being more complicated than models, especially since experiments are restricted to downsized laboratory models.
- Look for periodic earthquake activity. If the earthquake cycles match tidal cycles (12 hours for high-low tides, or on longer monthly cycles), then tides may be the cause of the cycle. However, correlation is not causation, so even if the timing matches, two cycles could be totally unrelated physically.
- Use statistics to find patterns, then match patterns to tidal cycles. If more earthquakes happen during a particular point in the tidal cycle than any other, then tides may be causing earthquakes. Again, correlation is not causation, and “statistical significance” is a far weaker statement than we colloquially use the word “significant.”
In a massive overview of studies performed until the mid-90s, Emter (1997) finds almost a direct split between researchers who found a connection between earthquakes and tides, and those who didn’t. Even worse, authors are inconsistent as to what relationship between tidal stress and the faults is necessary to produce earthquakes — minimum tidal stress, maximum tidal stress, changing tidal stress, or when tidal stress lines up with faults — and even the types of faults theoretically affected by tidal stress and magnitude of earthquakes produced.
Statistics are tricky, and it’s very easy to accidentally lie to oneself by choosing the wrong comparisons or datasets. For example, Knopoff (1964) accidentally correlated timing of perigee (closest point in the moon’s orbit, related to the moon’s month) to phase of the moon (also related to the moon’s month) instead of either to earthquakes. Thus, when evaluating the reliability of any paper on earthquake-tide relations, it is important to carefully consider the meaning of the statistical analysis.
Similarly, dividing data into subsets can cause all sorts of headaches that have more to do with mathematics than science. Statistics requires large datasets to say anything both reliable and interesting; subsets break data into smaller groups which may no longer be large enough to analyze reasonably. This is nicely illustrated by Heaton, whose 1975 data suggested a link between tides and earthquakes, but a more rigorous analysis on a larger dataset in 1982 by the same author failed to support the link. The same problem pops up when dropping data into temporal bins, grouping events into hours-long blocks in an effort to simplify analysis (for example, Klein 1976). The same thing can happen for data selected during particular time periods: Shlien (1972) used a dataset of seismic events at the equator (which experience maximum tidal stress) categorized by tectonic setting, yet found nothing significant beyond a cluster in one location at one year that did not hold true for other locations the same year, or the same location other years. An analysis performed using that one subset of data would have totally different results than the data from all years with no plausible justification for what made tides that year different from any other.
Once a correlation is identified, researchers must provide physically plausible connections between earthquakes and tidal stress. Research reviewed by Emter (1997) found patterns between earthquakes and odd aspects of the moon’s cycle that lacked in physical justification for influencing earthquakes, while failing to find a connection to other, far more physically dominating aspects of the moon’s cycle. Several researchers (Heaton 1975, Klein 1976, Tsuruoka 1995, among many others) have attempted a more physically rigorous approach, attempting to correlate tidal stress over time to failure mechanisms. Most commonly this involves either tidal stress oriented to enhance tectonic stress (which requires knowing the exact fault orientation when performing statistical analysis), or ocean loading enhancing stress (which only applies to coastal subduction zone earthquakes).
Laboratory testing simulating tectonic stress and tidal stress (Beeler 2003) suggests that the connection between earthquakes and tides is so small, it would require dataset of at least 13,000 earthquake events to detect the connection with any statistical significance. Using a very large dataset (13,042 events) on a well-known fault plane, Vidale and Agnew (1998) found a 2% increase in earthquake frequency when tidal stress promotes failure along normal faults, yet openly declare the difference is not statistically significant and may be due entirely to random chance. With a similarly impressive dataset (9,350 events), Tanaka et al. (2002) also found no strong correlations, but the little correlations they did find were only for reverse faults. In a truly massive dataset (27,000 events), Cochran et al. (2004) claim yet another relationship, determining that tidal phase is closely correlated with shallow thrust earthquakes. With so much inconsistency between studies, it is difficult to have confidence in their conclusions.
Besides earthquakes, tectonic plates can also move by low-frequency (as in, “the waves are low and rumbly”, not “it happens rarely”) aseismic events called tremors. Some evidence suggests that tremors near volcanoes may be correlated to tidal stress (Kasahara 2002, Richardson 2008). However, tremors won’t kill you, so they fall well outside the science I want to talk about today.
Emter (1997) says it best in his delightfully snarky conclusions:
But there exists also a tendency for negative results [linking earthquakes to tidal cycles] with increasing quality of the data sets and statistical methods. […] It may be a reality, as Burton (1986)states in a critical review, that the coherency between earth tides and global and local seismicity is usually within the noise level of any analysis.
I’m sure the future will hold a lot more papers on the topic with bigger and better datasets, more detailed fault and tide models, and clearer physical mechanisms for failure, but until then: a clear link between earthquakes and tides is a myth unsupported by reality.
Beeler, N. (2003). Why earthquakes correlate weakly with the solid Earth tides: Effects of periodic stress on the rate and probability of earthquake occurrence Journal of Geophysical Research, 108 (B8) DOI: 10.1029/2001JB001518
Burton, P. (1986). Geophysics: Is there coherence between Earth tides and earthquakes? Nature, 321 (6066), 115-115 DOI: 10.1038/321115a0
Cochran, E. (2004). Earth Tides Can Trigger Shallow Thrust Fault Earthquakes Science, 306 (5699), 1164-1166 DOI: 10.1126/science.1103961
Emter, D. (1997). Tidal triggering of earthquakes and volcanic events Tidal Phenomena, 293-309 : 10.1007/BFb0011468
Heaton, T. (1975). Tidal Triggering of Earthquakes Geophysical Journal International, 43 (2), 307-326 DOI: 10.1111/j.1365-246X.1975.tb00637.x
Heaton, T. H (1982). Tidal triggering of earthquakes Bulletin of the Seismological Society of America, 72 (6A), 2181-2200 article
Kasahara, J. (2002). GEOPHYSICS: Tides, Earthquakes, and Volcanoes Science, 297 (5580), 348-349 DOI: 10.1126/science.1074601
Klein, F. (1976). Earthquake Swarms and the Semidiurnal Solid Earth Tide Geophysical Journal International, 45 (2), 245-295 DOI: 10.1111/j.1365-246X.1976.tb00326.x
Knopoff, L. (1964). Earth tides as a triggering mechanism for earthquakes Bulletin of the Seismological Society of America, 64 (6A), 1865-1870 article
Lammlein, D. (1977). Lunar seismicity and tectonics Physics of The Earth and Planetary Interiors, 14 (3), 224-273 DOI: 10.1016/0031-9201(77)90175-3
Richardson, E., & Marone, C. (2008). GEOPHYSICS: What Triggers Tremor? Science, 319 (5860), 166-167 DOI: 10.1126/science.1152877
Shlien, S. (1972). Earthquake-Tide Correlation Geophysical Journal International, 28 (1), 27-34 DOI: 10.1111/j.1365-246X.1972.tb06108.x
Tanaka, S. (2002). Evidence for tidal triggering of earthquakes as revealed from statistical analysis of global data Journal of Geophysical Research, 107 (B10) DOI: 10.1029/2001JB001577
Tsuruoka, H., Ohtake, M., & Sato, H. (1995). Statistical test of the tidal triggering of earthquakes: contribution of the ocean tide loading effect Geophysical Journal International, 122 (1), 183-194 DOI: 10.1111/j.1365-246X.1995.tb03546.x
Vidale, J., Agnew, D., Johnston, M., & Oppenheimer, D. (1998). Absence of earthquake correlation with Earth tides: An indication of high preseismic fault stress rate Journal of Geophysical Research, 103 (B10), 24567-24572 DOI: 10.1029/98JB00594
Edited 9 April 2011 to format citations in accordance with Research Blogging guidelines.
Edited 19 February 2014 to reformat citations after the long, slow death of my reference-footnote plugin.