[BBI45]

That's the alert I was referring to, the pink and blue one. I'm assuming those are automated at NHK from external data? All the examples I've seen seem to suggest if an alert occurs in a pre-recorded programme the system will just throw it up and then NHK crash out of the programme once somebody in the studio is ready and then its just becomes a rolling news network for x hours.

The system is just generally fascinating, I still don't quite know how it all works or how the "chatter" gets to fire up TVs and what not, but... its quite cool really

Now that I'm back from my new year's day trip to Emley Moor (the things you do when you're bored during new year ), I can do my usual spiel about the system. There are two main systems that have been in use today: The Earthquake Early Warning (EEW) system, and the Emergency Warning Broadcast System (EWBS).

The Earthquake Early Warning system has been around in its current form since 2007, and is run by the Japan Meteorological Agency and their network of 4,000+ seismometers throughout the country. Data is constantly being fed from these seismometers to JMA offices throughout the country. Because of the science of how earthquake energy dissipates (which even I can't quite explain), there will be multiple waves of shaking caused. The seismometers pick up the faster and weaker waves and quickly make predictions on how strong the quake could be and where it could pose a major hazard. Once there's enough data for the system to be reasonably confident that a strong earthquake has occurred, warning data is automatically pushed out to be used as necessary. There is very little human involvement, as there is a very short space of time to get warnings out (as you'll see in the clips below, the worst impact can sometimes hit areas closest to the epicentre before the warning arrives). These warnings are used in a variety of ways, such as on the rail network to allow trains to stop, hospitals to allow surgery and scans to be halted and patients to be evacuated.

Bringing it back to media, once the warning data is pushed by the JMA, the EEW system kicks in. For TV and radio, this will take in the data, see if the warning is applicable to their broadcast area (NHK covers all warnings by law), generates the relevant graphics (TV) or audio message (radio), and pushes them out over the top of the primary output. These warning and their graphics/messages can expand as the JMA gets more data in, and once again, don't require human involvement to broadcast.

As you've said, once stations have had time to acknowledge the system, they will get a journalist on air if needed. NHK will always put some sort of update out, and always have a newsreader based directly outside a flash studio who can go from being at their desk to being on air in around 30 seconds. Other broadcasters will have people on standby who can go on air, but will generally use text updates at the top of the screen unless the earthquake is likely to justify rolling coverage.

Moving onto the EWBS, this system covers a wide variety of emergencies. From typhoons to tsunamis, and meteorites to North Korean missiles! OK, maybe not meteorites, but if something is likely to pose a threat to a population group, it's covered by the EWBS in some form or another. When it comes to tsunamis, the JMA will take the near finalised data from any earthquakes, take date from previous disasters, and assess if a tsunami is a viable risk. Once again, this is generally done by computers, but will generally be checked by JMA staff before being published.

If a warning is issued, once again, graphics are generated (or a script for radio) and all info is put out on air. This can once again be done automatically, but if an earthquake is big enough to have potentially generated a tsunami, then there will normally be people already on air who will take the lead instead. After this has happened, emergency broadcasting will continue with tidal data from JMA being used to keep track of the situation until all warnings have been downgraded to advisories or cancelled completely.

As for the "chatter", it's just data being carried in audio form like the Emergency Alert System in the U.S. I'm fairly sure there are now ways they can achieve the same effect of switching on/over TV sets without the screeching noise (given it's been around longer than I've been on this planet). But it seems to be one of those things where it gives the newsreader a few moments to process all the extra info, and the general public instantly recognise the sound and its meaning, so the screech survives.

youtu.be 
youtu.be 
youtu.be 

There are now better videos being uploaded, and there are three things I'd like to note:

1: In the first clip, you can hear the presenter's distress when the tsunami warnings are issued (6:40) and height forecast increased (16:20). Normally, the presenters are expected to stay calm and not sound alarmed. However, in this situation, the epicentre was extremely close to shore meaning that evacuation time is even more limited (on this occasion, a 1.2 meter tsunami was recorded within 12 minutes of the strongest tremor).

2: In the second clip, you can see how NHK handled it in comparison to Nippon TV (NTV), TV Asahi (EX), TBS (TBS), TV Tokyo (TX), and Fuji TV (CX). It nicely demonstrates my point about how NHK will cover every warning issued, whilst the others will only cover it once it's either relevant to their broadcast area, or is clearly becoming a major story. You can also see the text updates being used by other broadcasters whilst they assess the situation and decide to break into normal programming.

3: In the final clip, you can see that as the preliminary tremor is being reported on, the largest earthquake hits and because of how close the epicentre is and the short delay in getting the data from the source to the JMA to media outlets, the warning is too late to be of any use. Unfortunately, this is one of the pitfalls of the system and there isn't any workaround.