Aurorasaurus Study Shows Auroral Mapping with Twitter

By Nathan Case

An aurora, yellowish at the bottom, fades to red at the top across a starry night background.

A red aurora as seen from Paradox Valley, Colorado on March 17, 2015 at 3:23am (Photo credit: Derick Wilson)

As our users will know, we love Twitter here at Aurorasaurus. In addition to using the social media service ourselves (check out the Twitter list of Aurorasaurus Team Members here), we present aurora-related tweets for our users to verify them as aurora sightings (a process called crowdsourcing.) We then use these “verified tweets” in just the same way as the other observations reported via our website or apps, for example, by comparing their location with the modeled auroral ovals.

In this blog post, we wanted to highlight a recent academic publication we wrote about using Twitter as a tool to map the aurora. For those with access to AGU journals, the study can be found here. In the study, we looked at how Twitter can provide a picture of when, and from where, an aurora can be seen. As you’ll probably already know, people love tweeting about when they see the aurora (and we love reading those tweets too!). When people do tweet about their sightings, they often include their location (either directly through GPS embedding or in the text of their tweet) and other useful info (such as a photo).

The study showed that the number of aurora-related tweets correlated well with several different “geomagnetic indices”. These common, but somewhat confusing, indices are proxies for auroral activity. This means that they are not true measures of auroral activity but as these indices increase so does auroral activity. This showed that as the level of auroral activity increased, and an aurora was visible to more people over a wider area, the number of aurora-related tweets increased.

Additionally, “peaks” in the number of aurora-related tweets were found to frequently coincide with particularly strong auroral displays.

Sounds fairly obvious right? Well, perhaps it is, but it had never been shown before! Similar relationships had been found with many other phenomena (such as earthquakes and influenza outbreaks) but this was the first time a study had looked at using Twitter as a tool for mapping the aurora.

So, remember, if you see an aurora then please let us know either via our website or mobile apps, or post about your sighting on Twitter. If you do post on Twitter, always be sure to post promptly (don’t wait until the next day) and to include your location (by turning on Tweet Location or being very clear in the message of your tweet).

The study has received coverage on LiveScience and the Smithsonian Institute.  We are excited to help demonstrate one of the many ways citizen-science data can be used for scientific purposes.


Nathan is a space weather scientist based at University of Lancaster, United Kingdom, formerly at NASA’s Goddard Space Flight Center. As a past member of the Aurorasaurus team, he helped merge traditional forms of space weather data with citizen science observations to improve the nowcasting ability of the auroral oval models – though you may also see the occasional blog post or Facebook update from him too!

Share this post by clicking an icon below!
Share on Google+Share on TumblrShare on FacebookEmail this to someoneShare on RedditTweet about this on Twitter

Leave a Reply

Your email address will not be published. Required fields are marked *