Hacker News Reader

Comment by graderjs | Hacker News Reader

Largest quake in at least a year in Japan. Also strangely almost exactly 12 months since a 7.0 in the very same area (20 March 2021): https://earthquake.usgs.gov/earthquakes/eventpage/us7000dl6y...

View on HN

The infamous 2011 earthquake was also in March. Is there any plausible reason for earthquakes to occur in the same region around the same time of year over and over again, or is it just coincidence? I don't think Japan has any large glaciers that could melt in the spring and shift the center of mass of the plate underneath, for example.

My guess would be that ice melting in the caps triggers a release tension cascade in several parts of the crust, but is just a feeling. Less weight in the poles could make the planet tend to go more "sphaerical".

Just my own stupid theory. I'm not a geologist and could be wrong by a mile.

In any case since several years ago I have noticed a sort of end of the winter pattern: Eyjafjallajökull (2010), Fukushima (2011), Villarica Chile (2015), Fagradalsfjall (2021), New Zealand earthquakes (2022)... all in March.

And then is a waterfall later travelling towards equator: Volcan de Fuego Guatemala (may 2012). Feels a little like a tennis play with eruption waves bouncing, but maybe I'm just cherry picking facts together.

Interesting theory, but wouldn't the ice only melt at one pole cap at a time?

Good point. Dunno. Antarctic and Arctic are very different in terms of Geology. First is a continent with mountains. This means that has a thick crust. North Pole is submerged with some Islands around. Crust here is more thin and should be more prone to move when you remove weight or apply forces in any other direction.

Maybe thermal expansion, since March means warming after winter.

20 meters under ground there is already no temperature difference between summer and winter.

Great question. 1/12 x 1/12 x 1/12 -- odds seem kind of low...but could possibly be related to ocean currents, seasonal flow systems probably significantly shift the center of mass of large bodies of water, maybe related to temperature.

Probabilities and statistics don't work like that. Between 2011 and now there's been loads of earthquakes in Japan, in most months, many stronger than 7.0.

Heh :) That's totally incorrect. Do a search:

https://earthquake.usgs.gov/earthquakes/map/?extent=-32.6208...

That searches magnitude 7 and above for the last 24 months in a rectangular region that includes Japan and Taiwan. There's only been three, one was yesterday and two roughly a year ago in February and March.

So your intuition said arrogantly about earthquakes in Japan is incorrect, waay incorrect. Also I don't know if I got the probabilities totally right (but it seems fair for the probability of those other three earthquakes being in March, versus other months), but you said this as if you knew what you were talking about. But you don't. Seems like I'm the one here who knows what he's talking about, not you, sorry. Why not do a little research before you make your bold and arrogant claims trying to contradict someone else next time?

To get any meaningful results, we'll need a decent sample size. Magnitude 7.0+ quakes in Japan are a fairly rare phenomenon, occurring 1-2 times per year. If we pull out a somewhat larger data set of 136 such quakes in the last 100 years (https://www.volcanodiscovery.com/earthquakes/japan/largest.h...), we'll notice that the quakes are fairly evenly distributed between the months.

The mean is (of course) 8.33%, and the standard deviation 1.91%. All months fall within 2 standard deviations from the mean (March is highest at 11.62%, followed by Feb at 11.06%, Jun 10.51% and Nov 9.76%). So, nothing particularly interesting to see here. Also, if we do the same "three-month window" analysis you did, the winner is February (with 30.68%, somewhat higher than the expected 25%).

Wow, I just checked out that website and I can't believe you did that analysis! That's pretty awesome. You somehow got the data out of that site, and analyzed it.

But I did a quick sanity check on your numbers (searching on that site quakes from March 18 1922 to March 18 2022) and there are actually 153 quakes 7+ (not 136). If you share the raw data it will be more convincing!

Unless you do I'm going to take my 10 year sample from 2011 with a 68.75% of chance of quake in March (+- 1 month) as gospel :)

So pretty hard to predict and essentially random, meaning predicting when a given quake will occur should be, extremely challenging, essentially predicting a random variable, which would be pretty incredibly amazing. Good to see.

As a further update, consider this search:

https://earthquake.usgs.gov/earthquakes/map/?extent=16.80454...

All quakes 7 and above from 2011-01-01 until today. Some stats:

Total quakes - 16

Number of quakes in March - 6

Number of quakes in April - 3

Number of quakes in February - 2

Months with 1 quake - 5, 7, 8, 10, 12

Probability of quake in March - 37.5%

Probability of quake in March (+- 1 month) - 68.75%

Interestingly the two biggest earthquakes in my memory in Mexico City happened in the same day. Sept 19 1985 and 2017

Every year in Sep 19 there is a city wide simulacro. And then in 2017 an earthquake happened for real.

Maybe something to do with orbital forces / sea currents...?

Reply on news.ycombinator.com