Meteorologists Did "Robot Journalism" Before It Was Cool

Something that shook-up the media more than the media getting shaken-up in Monday morning's 4.4 earthquake in Los Angeles is the revelation that the Los Angeles Times has a set of code written to automatically post an article when the United States Geological Survey (USGS) reports that an earthquake occurred in California.

Monday's temblor was the latest in a long series of automatic earthquake reports generated by the L.A. Times, all of which end with the following disclaimer:

This information comes from the USGS Earthquake Notification Service and this post was created by an algorithm written by the author.

Almost every popular news website (Slate, BBC, The Guardian, Huffington Post, and even io9) wrote a story about the seemingly-groundbreaking practice in journalism, which uses the term "robot" to make it sound like a futuristic Smart House-type of technology.

The practice, however, is nothing new. Writes Will Oremus from the above-linked Slate article, the L.A. Times' "Quakebot" isn't the first of its kind at the paper:

Quakebot isn't the first bot of its kind at the LAT. Schwencke and his colleagues on the paper's data team modeled it on a similar bot that generates automatic reports about homicides in the paper's coverage area. Again, it's up to the humans to decide which ones merit further reporting.

The practice of using an automated computer algorithm to quickly distribute basic, helpful information to the public isn't limited to earthquakes or homicides. Meteorologists have used this technology for years to quickly distribute life-saving severe weather warnings from the National Weather Service to the public in as little time possible.

The most obvious of these automated warning dissemination systems is the Emergency Alert System that broadcasts tests and warnings to television and radio users across the country. Started as a way for the President of the United States to quickly address the American public in the event of a Soviet attack in the 1950s, CONELRAD gave way to the Emergency Broadcast System, which eventually evolved into the Emergency Alert System that we know today as the annoying screen that interrupts programming on television and emits a screeching tone during a test or severe weather event.

Take, for instance, a tornado warning. When a meteorologist at your local National Weather Service (NWS) office issues a tornado warning for your county, the computers at the NWS immediately broadcast out the warning along with a coded series of tones (the annoying screeching you hear) to help receivers read the warning. When your local cable or satellite provider receives the coded tones along with the warning, the Emergency Alert System equipment reads the tones and deciphers what the message is and how to process it to relay the information to the public.

When everything works the way it's supposed to, the Emergency Alert System equipment at your local cable company is expected to read the tones, convert them into a readable and audible alert, and send it to your television set to preempt your programming and give you the emergency information.

This system, while often drawing ire from people watching a crucial episode of Wheel of Fortune, is able to save lives by activating within seconds of the issuance of a severe weather warning by the NWS.

Another way that meteorologists are able to automate the relay of important weather information to the public is through services that post severe weather warnings to social media the minute they're issued.

WRAL-TV meteorologist Nate Johnson did an excellent job covering both the pros and cons of the automation of severe weather alerts on social media last November. Many popular meteorologists are subscribed to services that automatically publish alerts like severe thunderstorm or tornado warnings through their Facebook and Twitter feeds, allowing their tens of thousands of followers to receive this information almost immediately if they weren't paying attention to other sources at the time.

As noted by Oremus in his Slate article, this robo-info is up to humans to interpret and provide the all-important value added aspect, and this is especially true for meteorology. Just as using raw weather model output needs a human touch to have any relevance or accuracy, pure emergency weather information needs someone to help make sense of it when much of the public still doesn't know the difference between, say, a "tornado watch" and a "tornado warning." (The former means conditions are favorable for tornadoes, while the latter means that a tornado may be imminent).

Relaying important weather information to the public using automated means is a great way to make them aware of immediate dangers, but the added value of having a meteorologist explain on television or through a quick Facebook post exactly what's going on helps the public further evaluate the danger to themselves and those around them.

Whether or not more news agencies start to write code based this week's obsession with "robot journalism," you'll be glad to know that you've unknowingly benefited from this type of automation in meteorology for years.

[Screenshot and YouTube video by Stefan Cassadee]