Several times, I’ve seen statements like “The Great SDMB Fault has had major quakes every 100,000 years. It’s been 125,000 years since the last major quake, so it is 25,000 years overdue.” I’ve also seen this applied to volcanoes and the like.
Is there any meaning or power, predictive or otherwise, to such statements?
IANA geo-anybody, but just from the math side of things the answer is: Not much.
From a strictly statistical pespective you’d need quite a few repetitions of the event and accurate measurements of their dates to establish intervals with any confidence.
For example, if we had, say, 20 events accurately dated with each 100,000+/-500 years apart, we could make a statistically plausible prediction about the next one and therefore about the overdue-ness if one doesn’t happen on schedule.
But what if we have just 3 events in our history at 100,000 +/- 25000 years ago, 225,000 +/- 40000 years ago, and 275,000 +/- 40,000 years ago? The average is still about 100,000 years apart, but the noise in the signal is so large that even *if *the underlying geophysics are periodic, we can’t say it’s “overdue” until a lot more than 100K years after the last one.
And we’d be real hard-pressed to make a statistical argument that the geophysics even *are *periodic with data like that. A geophysicist may have other reasons to beleive it’s periodic, but this data isn’t good enough to show that.
The Yellowstone supervolcano is an example of something where we have like 2 or 3 noisy data points, yet the popular press has us becoming officially overdue & moreso by the day. Bollocks.
geologist here but IANA specialist in volcanology and seismology. yes, those terms usually apply to earthquake and volcano monitoring.
for volcanoes, an active volcano is one that has erupted as recently as 10,000 years ago (smithsonian standard.) standards may differ among different countries. gone are old terms like “dormant” and “extinct” volcanoes. in countries like indonesia (country with the most vocanoes) and the philippines, “active” may be one that blew out 500 years ago or more recently. hawaii has one that has been erupting since the 1980s.
for the above definitions, one could say an erution is overdue when the time a new eruption episode is supposed to happen has come and gone. for this, you have to have at least two documented eruptions thus giving you a rough estimate of time between episodes.
now, if a volcano (active or inactive) is showing signs of restiveness, then geologists spring into action (like in dante’s peak) and start monitoring. they may estimate the time it will erupt (months, or days.) there are many types of eruption but the one most people know is the real spectacular full-blown eruption of magma coming out in glorius fashion. if the estimated time lapses, they may again call it overdue (using a different time scale) or say the probability of eruption has gone down.
It’s somewhat related to the Gambler’s Fallacy.
If a roulette wheel goes for, say, 100 spins without ever hitting the number “17,” which ought, by expectations, come up every 38 spins on the average, then the wheel is “overdue” for a 17. But that doesn’t mean that the number 17 is actually any more likely to come up. It remains at the same stable 1-chance-in-38 (American style wheel – do European style wheels exist any longer?)
On the other hand, some events, like earthquakes, are not “independent” instances, but gather increased probability over time as stresses build with the continued movement of tectonic plates. Thus, California is not only “overdue” for a big earthquake, but the odds are (slightly!) higher as each year goes by without a stress-relieving quake.
(I suggest we buy the state a drink…)
Trinopus
Usually the context it is used in is trying to scare people. I don’t think it’s often used accurately. If we could accurately predict events, and one was late, it would apply.
Earthquakes happen when two adjacent moving plates of the earth’s crust suddenly let go and slip. On earth-sized scales of hundreds or thousands of miles, even rock is not rock-solid. It can stretch and squeeze a bit, and build up pressure. However, that built up pressure eventually will release.
Scientists have several ways to measure this probability. They can use old surveys to determine how much the two crust plates have moved. They measure minor earthquakes in the area; they look at how much movement there has been in plates at either end of the “stuck” zone. Sonic conduction can tell them the geological characteristics of the area. The historical “one every 10,000 years” is also valid. (For example, they can tell the last giant Seattle quake, about 1400AD IIRC, from the tree-rings of fallen trees buried in the mud).
Going by how much stress has built up, and how likely that sort of geology is to release based on the amount of stress, is a good indicator of whether a quake is likely soon, and how big it could be if it does let go. The Asia tsunami IIRC resulted in about a 9-foot slide of the two plates.