Federal researchers are poking around the Internet to see if tweets can beat seismographs in reporting earthquakes..
"There are going to be times when you see quake tweets before you get the scientific data from the seismograph networks," said USGS seismologist Paul Earle. "They are actually doing this right now."
The USGS coordinates a continent-wide system of motion acceleration detectors, sesimographs, communications lines and scientific experts that can report a quake's epicenter, magnitude and danger within 2-3 minutes of a quake's occurrence.
Twitter quake reports start to come in within 60 seconds, Earle said.
"We're just starting to look into how well a system based on social networks would work," Earle said.
Scientists have started using a Twitter pipeline to monitor and count the use of the words quake, earthquake and terramoto -- the Spanish word for earthquake -- on the Web.
An existing system called "Did You Feel It?" asks people visiting the USGS website to answer questions about how strong an earthquake was felt, and has provided valuable and accurate information to seismologists about stricken areas in rapid time after quakes.
Mining Twitter and other networks will not deliver "really good, actionable information like you would need to decide on shutting down the nuclear power plant.
"But it will give us a bulletin that something has happened."
Of course, that only can happen if the quake has not severed the tens of thousands of routers, servers, cell sites, fiber relays and other components that bounce tweets around the globe when the system is working.
"That's another big problem, we would not depend on Twitter and the Internet always being up," Earle said. "When Michael Jackson died, Twitter slowed down nearly to a stop."
Danny Sullivan, a Web columnist who operates the Search Engine Land service on the Internet, recently analyzed the Chinese quake, interviewed USGS officials, and concluded that Twitter beat the USGS by two minutes. But he said the verification and scientific accuracy of the USGS data, coming 120 seconds behind unverified reports, was reassuring.
Twitter users can learn more about the project by joining @usgsted.