Ronni Grapenthin - Notes

twitter: @rngrp
New Mexico Tech
Dept. of Earth &
Environmental Science
801 Leroy Place
Socorro, NM-87801

Earthquakes vs. Volcanoes: Answers from Big Data

Published: 2013/05/19

During the 2013 SSA meeting in Salt Lake City Utah I ended up at a bar with a few friends and we debated whether earthquakes or volcanoes were seen as a greater threat by society. This question was somewhat motivated by recent funding trends, and other reasons I won't go into here. Also, there's a lot of hackling with my friends in the geosciences about who's working on the more interesting stuff, but I won't include "glacier" or "climate change" here, because that's rather depressing: try the exercise I go through below including these terms if you wish. So, let's consider the question from society's point of view - What does have a greater impact: earthquakes or volcanoes?

While there is enough evidence for very impactful volcanic eruptions in stratigraphic records (and some in historic records), I believe that people are generally more worried about earthquakes - many large cities are built on fault zones, which experience occasional small earthquakes. Even when exposed to today's fast-paced media, events like the 2011 Tohoku earthquake or the 2004 Sumatra earthquake are still in the back to the mind of people. Volcanic activity, however, was rather moderate lately - the big news maker was clearly Eyjafjallajökull in Iceland in 2010. What else do you remember? I worked on the 2011 Grimsvötn eruption, also Iceland, and the 2009 Redoubt eruption (pdf) in Alaska. Have you heard of Kluchevskoy Volcano in Kamchatka, Bezymianny, Tolbachik? Pavlof in Alaska is currently erupting. While all of these events have some impact on air traffic; planes are rerouted, or flights are canceled, nobody was killed during these eruptions (as far as I know). So news drama may have been rather slim. Here, we are at the fundamental difference between volcanoes and earthquakes: scale! Earthquakes can affect large regions over a very short time period, whereas even large eruptions take more time to evolve (and our understanding of volcanoes and hence our monitoring techniques has become a lot better which results in fewer fatalities).

But - what is it that people are really interested in? One obvious pointer to societal interest is how often people search for these terms. So I checked with Google Trends. There are issues with this (total search volume would be nice), I am aware. Let's stay on the safe side and not over interpret, and instead have fun and look for the big picture here:

Click here, if the above figure didn't show.

This figure shows Google searches for 'volcano' (blue) and 'earthquake' (red). I limit this to the US as they dominate the world wide market in terms of English search terms. As far as I understand, the curves are normalized for the highest occurrence of one search term (Here, 'earthquake' in March 2011, which is the Tohoku event).

Those of us into volcanoes know that there were quite a few volcanic eruptions in Alaska, Kamchatka, well world wide during the time from 2004-now (check: Global Volcanism Program). As I mentioned above, some of these events had rather an impact. See little blip in 2009? My guess is that that's Redoubt. Eyjafjallajökull had significant impact on Europe and in Germany therefore "volcano" ranks a bit higher than "earthquakes" (again Japan):

Click here, if the above figure didn't show.

Unless you include the German terms for earthquake and volcano: "Erdbeben", and "Vulkan", respectively:

Click here, if the above figure didn't show.

The largest impact on Americans made the Japan earthquake (Mw 9.0); closely followed by the Virginia earthquake (Mw 5.8), which interesting, as it didn't really cause much damage but hit a seismically rather quiet area at a magnitude that can be felt.

What can we learn from this? When nothing's going on the average U.S. citizen cares as much about earthquakes as about volcanoes: almost not at all (although, when you zoom in, earthquakes seem to win). Generally people seem to be more interested in earthquakes than in volcanoes, though. Don't forget that St. Helens did stuff in 2004 (until 2008) that caught attention yet the blip is hardly visible, and that's much closer to population than anything in Alaska. Overall, the closer the earthquake, the more interest. However, this seems to inversely scale with magnitude, which makes sense. Interest seems to peak when something happens in an area where it was not expected (Virginia).

After I emailed the above to my friends, an interesting new question came up: What percentage of the volcano searches can be attributed to children doing science fair projects? Volcano school projects probably outnumber earthquakes 100 to 1. Well, here's Big Data:

Click here, if the above figures didn't show.

These are searches for "volcano science fair" (blue) and "earthquake science fair" (red); again limited to the U.S. There is not much to interpret here, volcanoes clearly win, which brings up an interesting question: If kids are more interested in volcanoes than earthquakes compared to the general population - what triggers the change in interest? Clearly there is lots of interest in science fair volcanoes in Texas and the east coast - so one guess would be that people grow up, and start worrying about their property. But that's just me guessing.

Now, what are the blips in earthquake science fair that come up? Considering that all of these are searched in California, I would propose the following events (based on distance vs. magnitude vs. recentness):

For the volanoes, the 2004-08 St. Helens eruption seems duly noted (although Washington State does not show up in the regional interest plots which is somewhat suspicious). The seasonal pattern in volcano searches reflects the cyclicality of science fairs well and may provides an estimate of general interest vs. peak interest during the Mt. St. Helens event.

If you don't believe there's seasonal interest in science fairs look at this plot:

Click here, if the above figure didn't show.

This shows searches for "volcano" in blue and "science fair" in red. The seasonality in science fair searches is obvious and explains a fraction of the seasonal effects in volcano searches. What's more interesting is the general decline in science fair searches - considering that the plots are thought to be normalized by "peak search interest" there should not be any effect due to "more people on the Internet" or anything like that. Does it really reflect a general decline in interest in science fairs? While I judged from 2008-2012, I have no idea of the total numbers of submitted projects, but this article in the NY Times suggests that numbers are in fact declining.

After several emails went back and forth, one of my friends put all of this into perspective:

Click here, if the above figure didn't show.

Anybody want to learn about correlation vs. causation? That uptake in cat searches in late 2010 and the Japan earthquake qualify!

To put a summary below this: There are interesting trends in U.S. search volumes which can be linked to individual geologic events. Overall the interest in phenomena such as volcanoes and earthquakes seems rather low, unless something happens which results in peaks of searches (this may explain why searches for "glacier" and "climate change" are so much lower than volcano/earthquake). The interest then seems to scale by distance and magnitude, and surprise (Mt. St. Helens, Virginia earthquake), whereas small magnitude earthquakes that are close can outsearch large events with big impact that are far away; so distance clearly dominates. The fact that there are more earthquakes in more populated areas as well as their suddeness may in fact explain why they are more popular in searches. Apart from the cyclicality in the science fairs and their impact on volcano searches, none of what I wrote should be much of a surprise; it's just great to get it in "numbers."

ronni <at> gi <dot> alaska <dot> edu | Created: 2013/05/19 | Last modified: December 17 2014 01:56.