Paper No. 4
Presentation Time: 8:45 AM
HOW TWITTER CAN IMPROVE EARTHQUAKE RESPONSE
EARLE, Paul S., National Earthquake Information Center, USGS, PO Box 25046, MS-966, Denver, CO 80225, GUY, Michelle, USGS, National Earthquake Information Center, MS 966, Box 25046, DFC, Denver, 80225 and HORVATH, Scott R., USGS, 12201 Sunrise Valley Dr, Mail Stop 119, Reston, VA 20192, pearle@usgs.gov
The U.S. Geological Survey (USGS) is investigating how the social networking site Twitter can augment its earthquake response products and the delivery of hazard information. The goal of this project is to gather near real-time, earthquake-related messages (tweets) and provide geo-located earthquake detections and rough maps of the corresponding felt areas. Twitter and other social Internet technologies are providing the general public with anecdotal earthquake hazard information before scientific information has been published from authoritative sources. People local to an event often publish information within seconds of an earthquake via these technologies. In contrast, depending on the location of the earthquake, scientific alerts take between 2 to 20 minutes.
In regions with active Twitter users, the frequency of “earthquake” tweets often rises above the background level within seconds after a felt earthquake. Examining these tweets shows it is possible (in some cases) to roughly estimate the felt area of an earthquake using the geolocated Twitter responses. The tweets also provide (very) short first-impression narratives from people who experienced the shaking.
The main drawback of this approach is the lack of quantitative information it produces such as epicenter, magnitude, and strong-motion recordings. Without quantitative data, prioritization of response measures, including building and infrastructure inspection, are not possible. The main advantage of Twitter is speed, especially in sparsely instrumented areas. A Twitter-based system could potentially provide a quick notification that a possible event has occurred, and that seismographically derived information will follow.