A Recent Increase in Earth Activity? The Data Provides Little Support.

The internet is buzzing with claims that there has been a significant recent increase in seismic, volcanic, and atmospheric activity on our planet, and there seems to be a specific emphasis on the years 2012 and beyond. Suggested causes for such an increase include global warming having an effect on these earth activities, deliberate geo-engineering, changes in solar activity and/or cosmic radiation, and an as yet unknown or undisclosed disturbance within our solar system. However, a simple analysis of relevant and publicly available data provides surprisingly little support for such claims.  [Last update: 29 Dec 2015]

In this post, I show the results of analyzing three publicly available data sets: earthquakes, volcanic eruptions, and tropical storms. First, I will indicate exactly which data sets were used and where they were obtained from. Next, for each of these data sets simple monthly or yearly summary statistics and trends were calculated, which were then plotted over the given time frame. These resulting plots are presented here, showing that there is little empirical and statistical support for claims of a significant recent increase in seismic, volcanic, or atmospheric activity. Finally, several possible reasons for the existence of such claims, or the lack of evidence for them, are discussed.

Earthquakes
Starting with seismic activity, a dataset of global earthquakes of magnitude 5 (M5.0) or larger over the past 20 years was downloaded from the website of the US Geological Survey (as a .csv file). A short R script was then written to extract these magnitudes on a month-by-month basis (from January 1996 to December 2015), and to count the number of earthquakes of a certain minimum magnitude (M5.0, M6.0, and M7.0, respectively) for each of these months. The monthly counts were then plotted over time, as shown in Fig. 1.

Fig.1: The monthly number of global earthquakes of minimum magnitude M5.0 (red line), M6.0 (green line), and M7.0 (blue line) from January 1996 to December 2015. Click the pic for a full-size view.
M5
M6M7Fig. 2: The counts and trends (as a 12-month moving average) for each of the three magnitude categories (M5+, M6+, and M7+). Click each pic for a full-size view.

The most striking feature in this graph is the large spike of just over 700 M5.0 or larger earthquakes. This is due to the M9.0 earthquake off the eastern coast of Japan in March 2011, which apparently generated many large aftershocks. In fact, there seems to be an increasing trend in the number of M5.0 or larger earthquakes (red line) leading up to this major event. However, after that there is actually a decreasing trend! Could this M9.0 earthquake have released much of the pressure that was apparently building up, causing an actual decrease in the total number of earthquakes in its wake? Also, the initial increasing trend up to March 2011 is not apparent in the M6.0 or larger earthquakes, and seems to be mostly confined to the M5-range.

To see these trends (or absence thereof) more clearly, a 12-month moving average was calculated for each of the three categories (M5+, M6+, and M7+). For the M5+ earthquakes, the large March 2011 event was left out for calculating this trend, assuming it was a singular event and not part of the actual trend (although it might have caused the change in direction of the trend!). The three graphs in Fig. 2 show the counts and their trends on appropriate scales for each magnitude category separately. Click on each of the graphs to view them in full size.

The M5+ graph (top) clearly shows the increasing trend up to the March 2011 event, after which the trend abruptly changes direction (going down) and then levels off. However, as the other two graphs show, any clear trend (whether increasing or decreasing) is absent for the M6+ and M7+ earthquakes. Finally, yearly counts were also calculated for each category, which shows the same results: at first an increase in M5+ earthquakes up to 2011, and then a sharp drop, but no such trends for the M6+ and M7+ categories (graphs not shown here, but available upon request).

Volcanic eruptions
Continuing with volcanic activity, a data set of global volcanic eruptions was downloaded from the website of the Smithsonian Institution (as a .xls file). Data for the past 40 years (1976 to 2015) was used, and another short R script was written to calculate the total number of volcanic eruptions for each of these years. The yearly counts were then plotted over time, as shown in Fig. 3.

Fig. 3: The yearly number of global volcanic eruptions from 1976 to 2015. Click the pic for a full-size view.

No clear increasing trend can be seen in this graph. In fact, the numbers from 2009 onwards are actually lower than in the immediately preceding period. Note, however, that the data for the year 2015 is incomplete, as the last reported eruption for that year is from mid-August (reason unknown). However, extrapolating from these first 7.5 months (with 27 eruptions up to then) to a full year (taking (12/7.5)*27=43), still remains well within the previous-40-year range. So, there is no significant or recent increase.

To show this statistically, a linear regression was performed (leaving out 2015, as the data for that year is incomplete), shown as the dotted blue line. There appears to be a slight increasing slope, but this is not statistically significant. In fact, the t-statistic for the estimate of the slope is equal to 1.879, which is below the significance level of 2. Furthermore, the p-value for testing the null hypothesis that the slope of the regression is zero (i.e., that there is no dependence of the number of volcanic eruptions on time), is equal to 0.0681. This is clearly larger than the standard significance level of 0.05, so there is no statistical justification for rejecting the null hypothesis. In other words, the slope cannot be assumed to be significantly different from zero.

Tropical storms
Ending with atmospheric activity, data on tropical storms in the Atlantic and eastern Pacific was obtained from the website of the National Hurricane Center. Monthly hurricane reports for the years 1999 to 2015 (i.e., the past 17 years) are available under the Tropical Cyclone Monthly Summary Archive section. The final report of each year contains a full summary of all storms for that particular year, including maximum sustained winds (in MPH). Storms with maximum sustained winds of at least 40MPH are included in these reports, which are separated between the north Atlantic (including the Caribbean Sea and the Gulf of Mexico) and the eastern north Pacific (everything east of 140 west longitude). The data from these reports was compiled (by hand) into a spreadsheet, combined over the two geographical areas, and an R script was used to calculate yearly storm counts, and average and maximum strengths (in terms of wind speed) for each year. The results were then plotted over time, as shown in Fig. 4.

Fig. 4: The yearly number of tropical storms (red line), average strength (green line), and maximum strength (black line) from 1999 to 2015. Click the pic for a full-size view.

As the graph shows, there is no increasing trend in either the total number of storms or their average strength (in terms of maximum sustained winds in MPH). The year 2005 was exceptional with the largest number of storms, and 2015 saw the strongest storm on record: hurricane Patricia in the eastern Pacific, with sustained winds up to 200MPH (we will have to wait and see if this is a freak accident or perhaps the start of a trend). It would be interesting to have data further back in the past to have a more complete picture.

To show statistically the lack of a trend in the number of storms, a linear regression was again performed, shown as the dotted blue line. The t-statistic for the estimate of the slope is equal to 0.961, which is well below the significance level of 2. The p-value for testing the null hypothesis that the slope of the regression is zero is equal to 0.3517. So, there is even less statistical justification to reject the null hypothesis for the number of tropical storms as there was for the number of volcanic eruptions (i.e., there is an even larger p-value, well above 0.05).

Discussion
The above analysis of relevant and publicly available data does not provide much support for claims of a significant recent increase in seismic, volcanic, and atmospheric activity on our planet. There was indeed a significant increase in M5+ earthquakes in the build-up to March 2011, but this was followed by a clear decrease since then. Furthermore, any clear trend (in either direction) is completely absent in the number of M6+ and M7+ earthquakes, volcanic eruptions and tropical storms. Anyone with internet access and some basic knowledge of the statistical package R (or even a simple spreadsheet program) can perform the same analyses as I have described here, and judge the results for themselves.

So, how come these claims exist in the first place, or, if they are somehow true, why is there an apparent lack of evidence? I can imagine several possible reasons for this.

  • Unreliable or incomplete data: I would assume that publicly available data from the USGS, Smithsonian, and NHC are generally reliable. However, given that these are all three US-based and US government-funded organizations, there could be a bias in the data. If this is the case, it should be possible to detect this from a comparison with similar data from alternative sources. In case of discrepancies, it would be worth repeating the above analyses with such alternative data.
  • Lack of longer-term data: One could argue that to see real trends, we need data even further back in time. However, given that the majority of “doomsday claims” seem to specifically emphasize the past few years, in particular 2012 and beyond, data for the past several decades should still clearly show the claimed recent increase. Furthermore, older data might not be reliable or complete enough. For example, on the USGS website it is clearly stated that there were no machine-readable earthquake records available for the years pre-1980. This older data literally had to be scanned in from paper documents. That certainly brings up the question how reliable and complete such older data is (in other words, possible increasing trends over longer time scales may simply be an artefact of missing data).
  • Other data: Of course the data analyzed here does not tell the whole story, but it seemed the most obvious place to start. If time permits I would like to find and analyze additional relevant data, and also look at certain aspects in more detail.
  • Increase in population size: The world’s population has increased exponentially over the past few decades. As a consequence, urban areas have grown rapidly, both in size and number. Even if the frequency of natural disasters like earthquakes, volcanic eruptions, and hurricanes has remained the same, the average number of affected people per event will inevitably have increased, perhaps giving the illusions that things have become worse, or more frequent.
  • Increase in news reporting: On a similar note, news reporting has increased drastically as well over the past years or decades. Anytime anything happens anywhere, we hear or read about it these days. However, this increase in the number and extent of news reports on natural incidents does not necessary reflect an actual increase in such events, but may also simply give the illusion of an increased frequency.

Disclaimer
I am an independent scientist, not currently employed by any institution or organization. I have no particular stake in either proving or disproving certain claims about a significant recent increase in seismic, volcanic, and atmospheric activity. In fact, the reason I performed the analyses described here is that I actually expected to see evidence for such claims, presented in an unbiased and transparent manner. However, the data I used and analyzed here provides surprisingly little support for such claims. If anyone has (or can suggest) publicly available and reliable alternative data and/or results that do clearly show a significant recent increasing trend, I would be interested in hearing about it.