Yearly Earthquake Update

I am still occasionally running into claims going around on the internet that the number of earthquakes has been increasing in recent years (which is then often used as “evidence” for some doomsday scenario). However, as I have been reporting in my yearly earthquake updates, there is no truth to these claims. In fact, the past year (2020) had some of the lowest number of earthquakes in the past 20 years.

The plot below shows the total number of (global) earthquakes of magnitude 5 or higher (M5+) in each year from 1970 to 2020. All data is obtained from the USGS. This plot shows that there was indeed an increase during the first decade of this millennium, but after 2011 it seems to have been “business as usual” again.

One thing that stands out, though, is the unusually high number of earthquakes during the years 2007, 2010, and 2011. What caused these “anomalies”?

One possibility could be that those years had one or more very strong earthquakes. Every earthquake causes multiple (smaller) aftershocks. The larger the magnitude of an earthquake, the more aftershocks it will cause. So perhaps there is a correlation between the magnitude of the strongest earthquake and the total number of earthquakes in a given year.

The next image shows a scatter plot of the total number of earthquakes (vertical axis) against the magnitude of the strongest earthquake (horizontal axis) in a given year, for the past 25 years (1996-2020). There does indeed seem to be some correlation. Years with a larger “strongest” earthquake (more to the right) also tend to have a larger total number of earthquakes (more to the top).

One striking outlier seems to be the year 2004, which actually had the strongest earthquake (M9.1) of all years, but only an average total number of earthquakes. However, this can be (at least partially) explained by the fact that this M9.1 earthquake happened on December 26th of that year, just a few days before the end of the year. Since aftershocks can easily continue for a week or longer after the initial quake, part of the aftershocks of this 2004 earthquake will actually have been included in the total for 2005.

Leaving out this outlier, and performing a linear regression on the remaining 24 years results in the diagonal line shown in the plot. This linear regression has an R2 value of around 0.45, indicating that about 45% of the total variation in the yearly number of earthquakes can be explained by the magnitude of the strongest earthquake in each year. Not bad, for just a single statistic to explain almost half of this variation! Furthermore, the p-value of this regression is 0.0003, suggesting that it is extremely unlikely that the observed correlation is just a coincidence. If you are not familiar with linear regression, and what these numbers mean, you can read a simple “Maths in a minute” explanation.

I assume this correlation is known among earthquake researchers, but I have never seen it mentioned or shown explicitly anywhere (although I admit I am not familiar with the earthquake literature).

Update: After inquiring with the USGS, they kindly responded pointing to some basic information about forecasting aftershocks.