The US Geological Survey (USGS) is tapping into wildly popular microblogging service Twitter to quickly pinpoint where and how brutally earthquakes hit.

The US agency that tracks and reports on all earthquakes in this country and major temblors around the world said terse text "tweets" can provide instant assessments of what is felt when the ground shudders.

The USGS has launched a Twitter Earthquake Detection (TED) project that continuously gathers tweets containing the word "earthquake" and its equivalent in other languages then meshes results in online maps of seismic activity.

"Many people use Twitter to tell others what is going on around them," USGS scientist Paul Earle said Tuesday in an audio interview at the agency's website.

"So after an earthquake they often rapidly report that an earthquake has occurred and describe the experience."

Twitter users in areas shaken by earthquakes typically begin "tweeting" in seconds while it can take as long as 20 minutes for data from sensors to be analyzed, according to USGS scientist Michelle Guy.

"The basic difference is speed versus accuracy," Guy said in the website interview.

"Analyzing the tweets provides an early indication of what people experience before the quantitative information becomes available."

Software filters out misleading references such as "earthquake brownie" desserts or the board game "Quake."

Text messages regarding bona fide temblors and available information about locations of the people that sent them are merged with data from USGS sensors and mapped, according to Earle.

"By collecting these tweets, which are open for anyone to search and analyze, it is possible to obtain firsthand accounts, albeit short, of what people experience during an earthquake," Earle said.

Earthquake-related tweets with vague or missing location information are skipped, because the system is intended to pinpoint where shaking is felt and how strongly.

Twitter, which allows users to pepper one another with 140-character-or-less messages, has grown rapidly in popularity since it was launched in August 2006 and claims millions of users.