Why do earthquakes mostly happen outside of daylight hours?

Loma Prieta is the one exception I can think of. Apart from that, it seems that earthquakes generally happen when it’s dark in whatever part of the world they happen in. Just a coincidence? Or do fault lines respond to a lack of activity on the surface?

The 1923 Kanto Earthquake (Tokyo) struck at two minutes to midday.

I had never heard that this was the case.

I just called up the USGS list of recent earthquakes of magnitude 5 and above.

Anyone care to work those out in local time? :slight_smile:

All of the earthquake times in Wikipedia are cited in UTC, which I’m sure is scientifically rigorous and all that, but local time would be useful too. I expect there’s some anal reason why this information wouldn’t be acceptable.

I don’t care enough to work out all the times, but even of the first three, two are during daytime: around 5 PM and 3 PM of local time.

The biggest earthquake that I’ve personally experienced was at 10:30 am local time.

The only reason why the local position of the Sun could affect earthquakes would be tidal forces, and they follow a cycle roughly 12 hours long – a bit more than 12 hours, because they are mainly affected by the position of the Moon. Perhaps the question should be about the timing of earthquakes relative to where the Moon is in the sky.

I think the OP is experiencing confirmation bias.

The Good Friday quake in Alaska was 8.4 on the Richter scale. It happened at 5:36 pm local time on 3/27/64.

The 5.9 Whittier Narrows earthquake was on 10/1/87 at 7:42 am local time.

The magnitude 7.7 Kashmir quake was 8:50 local time on 10/8/2005.

The magnitude 7.9 Gujarat quake was the morning of 1/26/2001 at 3:16 UTC, which is broad daylight locally.

Here is a handy calculator to see what part of the earth is in daylight for any date and Universal Time.

You don’t need to work it out precisely (since time zones are frequently arbitrary anyway); you can just use the longitude to normalise the times, since plus and minus 180 roughly corresponds to a 12 hour shift from UTC, so you can just apply a correction value.

I’m not really sure there are enough samples in that list to be statistically significant, but there certainly doesn’t appear to be any correlation between local time and frequency of earthquakes:

Ah, histograms…

Living in one of the most earthquake-prone countries on Earth, I certainly don’t get the impression that they happen more often at night. For instance, take a look at Japan’s Meteorological Agency’s list of recent perceptible earthquakes. The page is in Japanese, but you can see the time on the left of each row. It looks fairly evenly distributed.

As a matter of fact, I did the calculations and I get an average time of 11:17 with a standard deviation of 7 hours and 7 minutes. I also ran a simple histogram on the data and counted the number of events that occurred between 18:00 and 6:00 (night) and between 6:00 and 18:00 (day). As it turns out, there were 50 nighttime quakes and 50 daytime quakes.

Essentially, everything points towards a very even distribution.

Well, I suppose I’m glad I’m wrong!

Here is a histogram based on the Centennial catalog of large earthquakes, with times converted to approximate local solar time (the same method that Dead Badger used):


                                  #
                                  #
 #        #  #  #     #  #     #  #     #              #              #
 #  #  #  #  #  #     #  #     #  #     #     #  #  #  #  #  #  #     #
 #  #  #  #  #  #     #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23
13030 events; each # represents 25 events.

Whoa!!! Look at that statistically significant drop at 6 am!!! There MUST be a reason!!! :smiley:

I wonder if the reason it might seem like there are more earthquakes at night than at day is because it’s easier to miss a quake if you’re up and about than if you’re lying down. For instance, both my mom and I completely missed this earthquake, which took place while we were walking around downtown San Francisco in the middle of rush hour (my dad and sister, who were much closer to the epicenter, told us about it when we got home; THEY had felt it), while I have been jolted awake by much smaller quakes. Someone in my elementary school told me she didn’t feel the Loma Prieta quake because she had been playing soccer at the time. (And okay, we lived 100 miles from Loma Prieta, but that’s still somewhat unbelievable - Loma Prieta was a **big ** shake.)

All the earthquakes I’ve been in, all minor tremors of no discernible effect, were during the day. But then, if they were when I was asleep I probably wouldn’t even wake, so maybe there were some.

Those who record such things don’t want to get out of bed to write it down at 6AM. Seismologists like to sleep late.

I felt that Illinois-based quake yesterday morning here in the Cincinnati area.
It was pretty freaky, since it was the first one I’d ever experienced and the fact that the Midwest doesn’t see many of them.

I suppose I just popped my earthquake cherry!

I think the posts in between have shown that the OP is not true, but I offer a totally spurious explanation of why it might be;

Because rocks are afraid of the dark. Earthquakes happen at night because the rocks are shaking in fear. :smiley:

A lot of big quakes, at least in CA seem to happen in the early morning:

Ft Tejon Quake 1857= 8:20 AM
1872 Lone Pine Owens Valley= 2;35 AM

SF 1906 =5:12 A.M.

Kern County 1952= 4:52 AM
1971 San Fernando earthquake 6:00: a.m
Landers Earthquake magnitude 7.3 1992 4:57AM
Northridge earthquake 1994 at 4:31 AM

Weird huh?

Only one that wasn’t= Loma Prieta earthquake, 1989 at 5:04 p.m.

Note that the USGS sez "http://earthquake.usgs.gov/learning/faq.php?categoryID=6&faqID=17
*Q: Are there more earthquakes in the morning/in the evening/at a certain time of the month?

A: Earthquakes are equally as likely to occur at any time of the day or month or year. The factors that vary between the time of the day, month, or year do not affect the forces in the earth that cause earthquakes*

:eek: It’s true! Binning the California events (actually everything between 32°N and 42°N, and 113°E to 125°E) in four-hour groups, including the early-morning hours 03:00-06:59:


    #
    #
    #
    #
    #
    #
    #
    #
    #
    #
    #
    #  #  #
    #  #  #  #
    #  #  #  #
    #  #  #  #
 #  #  #  #  #  #
 #  #  #  #  #  #
 #  #  #  #  #  #
 #  #  #  #  #  #
 #  #  #  #  #  #
 #  #  #  #  #  #
 #  #  #  #  #  #
23 03 07 11 15 19
02 06 10 14 18 22
68 events.

I admit that I chose the time groups to make the peak as large as I could :D; binned only by hour it looks like


                                              #
                #  #                          #
          #  #  #  #                          #
          #  #  #  #     #        #           #
       #  #  #  #  #  #  #        #        #  #
       #  #  #  #  #  #  #  #  #  #  #  #  #  #        #     #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #     #  #  #  #  #  #
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23

which isn’t quite so impressive.

With a larger and different data set (this one, using magnitudes 4+), the peak is slightly later (6-8am) and quite a bit smaller:



                      #
                   #  #
                   #  #
                   #  #
                   #  #  #
                   #  #  #  #  #  #           #  #
             #  #  #  #  #  #  #  #  #        #  #
 #     #  #  #  #  #  #  #  #  #  #  #  #  #  #  #     #           #  #
 #     #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #     #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23
2500 events; each # represents 10 events.

Probably not statistically significant, but surprising to me that it’s still visible.

Here’s one more histogram, this one using the California historical data (1769-1974) from the same page:


       #                          #           #  #        #     #  #
 #  #  #  #  #  #  #  #  #     #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
 #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #  #
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23
25750 events; each # represents 100 events.

…no more early-morning peak.