Earthquakes Dataplot - Earth becoming Unquiet?

Out of curiosity, some time ago I downloaded from USGS their 'NEIC' earthquakes catalog containing all earthquakes worldwide since 1973, from here. The plot of that data is below - the curve is smoothed by using a running median over 90 days to make it better understandable.



(You can click the image for full size)



What is apparent from looking at this plot is that until 1979 or so there relatively few earthquakes, about 10 or so per day and there was almost no growth in the number of earthquakes per day. At about that time (1979) the curve changes and the number of earthquakes/day increases, but also, the roughness of the curve increases, meaning that there is greater variance between periods of quiet and periods with more earthquake activity. This lasts until about 1992, when the growth in number of earthquakes increases as well as the bumpiness of the curve, until the plateau ends abruptly about 2008, when the average number of earthquakes per day decreases sharply and appears to stabilize at a level around 37 earthquakes/day in mid-2010.


It woule be interesting to know what causes these discernible changes in seismic behavior of all of earth. The sharp decrease in activity in 2008/2009 occurs in the middle of the minimum between solar cycles and the seismic activity generally appears to not have anything to do with solar cycles. It appears as if there is an increase of stress on earth building up for about 30-35 years, stress which is then released or diminished very fast, over 1-2 years.


Whatever the plot means, it look weird. I dont even know if it is appropriate to interpret anything into it - to suppose that earth as a whole can be interpreted to have "one" overall seismic behavior. If this is the case and if this behavior is from inside earth, then what could it be? If what causes this is outside earth, equally, what could it be? It would probably not be from interaction with the known planets - as this would probably show some kind of cyclic behavior. OTOH, there is nothing known about the presence of anything in the solar system able to have this effect, just the usual secretive behavior of TPTB.


I first discovered this some time ago and thought that there was error in the data because I was combining data from different sources. I now found the time to redo this with the USGS catalog linked above and it gives the same result.
BTW, the big Japan earthquake of March/2011 is the third peak from the right.

Hello name,
Interresting plot. It is not clear how the seismograph's network and data analysis developed in the last thirty years, that would be an interresting factor to consider.
On an another side, using a running median is sometimes tricky when your series are not stationnary. I would preferably use a running mean in that case.
I am not sure if the earthquake rate is really the best proxy. For instance, if you have two 3mag earthquakes one day and two 6mag earthquakes the following day, you will have the same number (2) for both days. Maybe a better count would be to estimate a cumulative energy. And even then, it would be interresting to see how to separate a major earthquake from its replica.
In your graph, the Japanease earthquake is 1, but it shows as a peak because of an exceptional number of replica. Not sure it happens for all major events though.
Just a FWIW remarks, I hope you continue, maybe there are interresting things indeed :)

mkrnhr said:

Interresting plot. It is not clear how the seismograph's network and data analysis developed in the last thirty years, that would be an interresting factor to consider.

Click to expand...

Yes, and I think that would be why the available data before 1973 is scant. Still, that does not explain the steep decline in total num of registered earthquakes seen in the graph.


[quote author=mkrnhr]
On an another side, using a running median is sometimes tricky when your series are not stationnary. I would preferably use a running mean in that case.

[/quote]
Thanks for that. I spoke with somebody else today and he too told me to use average. Avg works out the bumps and valleys better - see graph below.
[quote author=mkrnhr]
I am not sure if the earthquake rate is really the best proxy. For instance, if you have two 3mag earthquakes one day and two 6mag earthquakes the following day, you will have the same number (2) for both days. Maybe a better count would be to estimate a cumulative energy. And even then, it would be interresting to see how to separate a major earthquake from its replica.

[/quote]
Probably not. Through this your suggestion I have today found out that calculating energy output of earthquakes is not trivial and will involve some work of looking up data. I Wrote USGS to ask what magnitude they publish (there are several) and received a bureaucratic non-answer telling me only that they use "preferredly Mw but select others according to convenience".
[quote author=mkrnhr]
In your graph, the Japanease earthquake is 1, but it shows as a peak because of an exceptional number of replica. Not sure it happens for all major events though.
Just a FWIW remarks, I hope you continue, maybe there are interresting things indeed :)
[/quote]
Actually the japanese earthquake is 1 bump but about 2 weeks of days with over 100 earthquakes, the two first days of the incident having over 500 eq's each.
Here is an interesting page about the big japanese eq of last year. The videos should be explaatory.

Hi name,
Something seems to have happenned indeed around 2010. It is probably related to solar activity.
For instance, it is at that time that the Ap index ("a measure of the disturbances of the Earth's magnetic field brought about by the solar wind hitting the Earth") reached a minimum : _https://wattsupwiththat.files.wordpress.com/2012/07/image.png

Approximate formula to link magnitudes to energy can be searched in this paper (I don't have access though)
_http://www.agu.org/pubs/crossref/1995/95JB01969.shtml
However, from _http://earthquake.usgs.gov/learn/topics/measure.php , it is possible I guess to have an approximate estimation of the released energy from Mw by using

E~10^(1.5(Mw+10.7))x1.6e5 in erg

Even if it is just a rough approximation, it may give some additional insights to the graphs. At least, averaging energies can be more easily interpreted than averaging magnitudes (logarithmic scales) or number of occurrences IMHO.

Edit: verify the equation if you want to implement it, I could have made a mistake.

Thank for your input, mkrnhr.
[quote author=mkrnhr]

Something seems to have happenned indeed around 2010. It is probably related to solar activity.
For instance, it is at that time that the Ap index ("a measure of the disturbances of the Earth's magnetic field brought about by the solar wind hitting the Earth") reached a minimum : _https://wattsupwiththat.files.wordpress.com/2012/07/image.png
[/quote]

Solar activity is another intriguing area, of which I understand about nothing. Just so much: earthquakes appear to not be correlated to sunspots. And 2010 was when the (long) minimum between solar cycles ended and the new cycle started.


[quote author=mkrnhr]

Approximate formula to link magnitudes to energy can be searched in this paper (I don't have access though)
_http://www.agu.org/pubs/crossref/1995/95JB01969.shtml
However, from _http://earthquake.usgs.gov/learn/topics/measure.php , it is possible I guess to have an approximate estimation of the released energy from Mw by using

E~10^(1.5(Mw+10.7))x1.6e5 in erg

Even if it is just a rough approximation, it may give some additional insights to the graphs. At least, averaging energies can be more easily interpreted than averaging magnitudes (logarithmic scales) or number of occurrences IMHO.

Edit: verify the equation if you want to implement it, I could have made a mistake.
[/quote]


Here is the moving average of the energy output of all earthquakes above M=1 in Terajoules, in logarithmic scale:

To my eye there is a short cycle of about 3-4 years there, which lasts until about 1989/1990. From there on, the behavior in the curve changes, and since 2004 (tsunami) there have been several big events in fast succession - so far I remember, the last earthquake comparable to the 4 we have in about 6 years was the one in Chile in 1961. I noticed that I had completely forgotten the earthquake of last April - the MSM just mentioned it in short reports, concluded that nothing was amiss since there had been no tsunami, and moved on.


Here is a (linear) plot which shows the 4 big events of the past years compared to what is "normal" (more or less a flat line)


So, the plot of the earthquake energy output over the years allows confirms that earth is becoming unquiet, but I still don't know why. Perhaps looking at the solar weather data has something useful for this. That's it for now.

I looked at the SOHO website and I find it "interesting" that they offer their data in a java application which precludes mass downloads, and apparently no direct download of their data. I found out that the data about space weather, apparently with a "prediction" value but without the ap/Ap data - which is an "official" calculation done by some Uni in Germany, can be had from a USAF website. That I also find "interesting": Why is there apparently no civilian agency offering that kind of data?

Anyway, I'll be back in about 10 days and look at that data.

Hello,
This second plot shows an other story: less earthquake but more energetic?
As for the short cycle I don't know, filtering a random signal often gives that impression.
One other point to take into account also is that the series extents only from the 70's. There have been major earthquakes in the 50's-60's and in the 20's ( http://earthquake.usgs.gov/earthquakes/world/10_largest_world.php ) as in the 2000's, maybe denoting a cycle that is larger than the actual sample? It could be interresting to compare the actual situation with those surges in seismic activity over a century as well IMHO.

Edit: link

Interesting that the graphs which show the rise in earthquakes over the past 100 years closely match the graphs that shows oil consumption for the past 100 years. They both start to rise at the same time and both show a large spike during the 1970's.

Do an image search using the keywords "oil consumption 100 years" and "earthquakes 100 years"

Welcome to the forum MarcClintDion.

Actually, seeing as this is your first post on the forum, we would appreciate it if you would post a brief intro about yourself in the Newbies section, telling us how you found this forum, how long you've been reading it and/or the SOTT page, whether or not you've read any of Laura's books yet, etc.