Twitter Based Earthquake Detection System Puts Behind Sensor Based Systems in Detection Time

The US Geological Survey (USGS) has created an earthquake detection system based on Twitter alerts. The system was being developed as a pet project by a student, and the USGS funded it later with funds from the American Recovery and Reinvestment Act. Clearly, the USGS saw a good future prospect in the project, and the funding is beginning to prove fruitful.

During the recent earthquakes in Philippines, the USGS’ Twitter Earthquake Detection (TED) system was able to give early warnings, much before any of the sensor-based systems in place. This is groundbreaking, as sensor-based systems take anything between 2 to 20 minutes to detect an earthquake, but the TED system is almost instant. The official page for TED describes it as,

@USGSted  (USGS Tweet Earthquake Dispatch) distributes alerts for earthquakes worldwide with magnitudes of 5.5 and above. We may modify this criteria in the future to tweet alerts for more earthquakes of potential interest. @USGSted earthquake tweets contain a magnitude descriptor, location, origin time, and a link to the USGS webpage with the most recent information about the event.

In the recent Philippines earthquake, the TED system detected tweets and the location of the earthquake in just one minutes and seven seconds. Systems like TED are good for augmenting traditional earthquake detection systems based on sensors. However, they also suffer the risk of being gamed by an overwhelming amount of tweets crying wolf.

This reminds me of an XKCD comic.


You can follow the TED twitter account for latest earthquake warnings. Also read how Twitter is being used for emergency calls in Japan.

Published by

Chinmoy Kanjilal

Chinmoy Kanjilal is a FOSS enthusiast and evangelist. He is passionate about Android. Security exploits turn him on and he loves to tinker with computer networks. He rants occasionally at You can connect with him on Twitter @ckandroid.