Social media can be a lifesaver in a disaster, new research demonstrates. It can assist relief teams responding to catastrophes by providing a measure both for the intensity of damage, and tracking where help is most required.

“Natural disasters are costly,” a paper in Science Advances notes. “They are costly in terms of property, political stability, and lives lost. Unfortunately, as a result of climate change, natural disasters, such as hurricanes, floods, and tornadoes, are also likely to become more common, more intense, and subsequently more costly in the future. Developing rapid response tools that are designed to aid in adapting to these forthcoming changes is critical.”

Information technology can make a huge difference, however. The number of lives lost in cyclones, for example, has plunged since satellites have provided warnings of where and when they will strike. But it is much easier to get information out from a central source to the general community than back in the other direction.

Experience has shown many ways social media can facilitate disaster relief, for example by spreading safety advice. “More recently, researchers have begun using social media platforms to derive information about disaster events themselves,” the authors noted. “For instance, the number of photographs uploaded to Flickr was shown to correlate strongly with physical variables that characterize natural disasters (atmospheric pressure during Hurricane Sandy).”

Dr. Yury Kryvasheyeu of National ICT Australia, and his coauthors, tracked the number of tweets about Hurricane Sandy per person in New Jersey and New York and compared this with damage estimates from the Federal Emergency Management Agency (FEMA). The found 52.55 million Sandy-related tweets from 13.75 million users. Even when limiting the study to geocoded tweets they still had 9.7 million to work with, a more than adequate database.

“The number of messages slowly increases with a strong peak on the day of hurricane landfall, followed by a gradual decline in the tweet activity level,” the authors report. “Geographically, the trend is similar almost everywhere, but the magnitude of the normalized response changes depending on the proximity to the hurricane, determined through the shortest distance to the path of the hurricane.”

Heat maps created using Sandy tweets proved a better match with damage sustained than FEMA's estimates at the time.

^{A comparison of the Twitter activity around Hurricane Sandy and associated damage for New Jersey (above) and parts of New York. Kryvasheyeu et al.}

Analysis for the paper was conducted years after the event, but could be done in real time by relief agencies assessing where they are most needed. The authors demonstrated that differentiating between original content and retweets can provide an even more sophisticated picture.

“The method for the assessment of the damage distribution proposed here offers a range of advantages to complement traditional alternatives (modeling, postdisaster surveying, and collection of data from multiple institutions): the advantages of fine spatial resolution, speed, low cost, and simplicity,” the authors conclude. An additional advantage is that it is much easier to crunch the numbers of millions of tweets, than it is to photograph disaster-struck regions from the air to find the epicenter of need.

Tracking of social media is proving useful to science in more benign circumstances as well, such as being used to track aurora.

^{Twitter could have other uses in disasters as well, but XKCD is skeptical. Randall Munroe.}