2 March 2010

A hectic weekend at the Pacific Tsunami Warning Center

Posted by Michael McFadden

Brian Shiro at his PTWC office (Courtesy of Brian Shiro).

Brian Shiro at his PTWC office (Courtesy of Brian Shiro).

Brian Shiro is a geophysicist who has been working at the Pacific Tsunami Warning Center (PTWC) for over four years. On a normal work day, his tasks go from checking that the warning system works to maintaining the center’s website and working on expanding Hawaii’s seismic network. But last weekend, after a 8.8 magnitude earthquake hit Chile and a Pacific-wide tsunami alert was issued for the first time since 1964. Brian spent about 40 hours working (almost) non-stop at PTWC. In total, the center issued 27 tsunami-related messages over the course of the event; the first one on Friday evening (Hawaii time) and the last one, calling the warning off, near midnight Saturday night. Brian took some time yesterday to tell us about his experience during the emergency. For two nights and a day, he was fielding calls from the media and public, and pushing tsunami scientific and safety information out on the Web via emails and social media sites, such as on Twitter.

Q: In your blog, you describe how on Friday night you went to the office in your flip-flops to fix a problem with the website, unaware of what was about to transpire…

A: Yes, I had gone home a couple of hours earlier from work to spend the evening with my family. When I initially got the call back to the office to fix something on the website, I didn’t at first expect the tsunami warning to be as major as it was. But as soon as the magnitude of the threat was obvious, I ended up staying there and helping. The whole ordeal for me was from about 9:30 PM Friday night to 4 PM Sunday afternoon, because after staying there to help on Friday night and Saturday, my “real” shift started early in the morning the next day.

Q: Were you working all the time?

A: I took a small break on Saturday afternoon and tried to take a nap, but that was the time when the tsunami was arriving in Hawaii, so I stayed glued to the news coverage like everyone else. I was on standby duty for the evening during the latter hours of the warning. Being on standby is like being a fireman for tsunamis: you can be at the center and watch TV, take a nap, etc., but you have to have the pager turned on and be ready. We are usually on standby two days a week. But this time, I ended up working the whole time.

PTWC operation room (Photo courtesy of NOAA)

PTWC operations room (Photo courtesy of NOAA)

Q: You wrote that during this time you were answering emails, handling media requests, and also monitoring social media between media calls; why did you think it was important to pay attention to what was happening on Twitter?

A; Well, to begin with, if I hadn’t been monitoring the emails, I don’t know who would have. We had a lot of emails coming in, and no spare people to answer them. Also, we were all sharing the load with the phone calls. About the social media: I am the youngest person in here. I am very interested in social media right now, so I’m going to be looking at it regardless. It’s partly a generational thing. But I also believe social media has become such an important means of communication right now; it’s a good way to have a connection with people, even if you’ve never met them. It makes sharing news more participatory.

I was [tweeting] unofficially, since we don’t have official Twitter accounts in the center. So I was just speaking for myself; I was trying to say all I could, whenever I could, but I had to be careful only to disclose information we had already made public. Sometimes I was trying to correct misinformation like when a rumor was spreading that a 35 foot wave was going to reach Hawaii.

Q: What kind of information requests did you get from social media and from good old-fashioned calls and emails?

A: We still get the most requests by phone, and email is also a big way people contact us. Since we don’t have an official presence on social media, people are not going to go there to look for us. But some journalists saw my tweets, and I got some interview requests through that.

We get the most phone calls from journalists, with all types of information requests. At some point, every local media outlet had its van parked here and was broadcasting live from PTWC. Besides the media calls, we got calls from concerned people, who were trying to find out what to do. For example, there was a man who called saying that his daughter was in the Galapagos doing research when the Chilean earthquake struck, and he didn’t know how to reach her. He asked if we had heard anything about it, and I told him that yes, that people in Galapagos had been evacuated.

Once the attention was on Hawaii, we got a lot of phone calls from local people asking what to do; “shall I leave my house, how high should I go?” We were trying to tell people not to panic, telling them “as long as you don’t go into the water, you’ll be OK.” We were trying to dispel some of the hype that was going on out there, but at the same time trying to take it seriously. We knew Hilo and Kahului bays were going to get the highest waves, and were most concerned about them.

Q: How was the evacuation? Were Hawaiians well-prepared for a tsunami emergency?

A: It was a big success story: there were no major gridlocks on roads as some people feared. Hawaii had tsunami evacuations in 1986 and 1994, and both times there was gridlock on the roads, which means that if the tsunami had hit, the cars along the coast could have been washed to the sea. So we try to tell people not to get in the car and just to walk to higher ground.

Also, hotels and resort areas were very well-organized, knocking on doors and telling people what to do: either to go to higher floors or to evacuate, depending on the hotel’s policy. The police worked on closing some roads, and there were buses busing people out in some areas.

On the other hand, some people evacuated when they didn’t have to, because they were outside the danger zone. We still have some work to do getting information out there.

But overall, this event served as an excellent end-to-end test of Hawaii’s tsunami warning system: from the earthquake detection and tsunami inundation models that we do to the organization of the evacuation by civil defense authorities.

Q: What are Hawaiians saying about the decision to evacuate, despite the fact that in the end there was a tsunami no-show?

A: The reaction has been overwhelmingly positive: we’ve got many emails and phone calls, as well as people coming up to us in the street, recognizing us from TV, and saying “thank you for your good work”. That’s the good news.

We’ve also gotten some criticism from people saying the warning shouldn’t have been hyped as much as it was. Of course, that was somewhat out of our control. We never said there would be a huge tsunami but that there was a potential for danger. We kept telling people that most places would see 1 meter (3.3 feet) but that Hilo and Kahului could see up to 2 or 3 meters (6.5 to 9.8 feet). It ended up being about half or so of what we expected… which, in this business, is actually pretty close.

Part of the reason for the evacuation was that the tsunami warning system is set up conservatively, so that you can make sure you never miss a warning when it’s needed. There is always a danger of crying wolf, but we hope people realize we’d rather err on the side of caution to keep them safe than hedge our bets that an event might not be dangerous. This means you might overwarn, but that’s preferable to the alternative of underwarning.

Q: (Radio host) Rush Limbaugh went even further and suggested that scientists are manipulative.

A: There seems to be this theme going on with some people, and it’s very hard to refute because they have their mind made up that the government is withholding information from them. But this is not true: all the information we had, we put out there in the hourly warning messages we issued. If anyone wants to know what we said, they can look at them. The message can get spun different ways when it gets picked up by the media; we do our best to try to re-address it, but we have no total control on that.

Also, Limbaugh’s comment on why we weren’t able to watch the tsunami with a satellite illustrates his ignorance about the state of the art: we can’t see tsunamis from space, at least not in a reliable, operational way that the tsunami warning system can rely upon. We might in the future, but not today. Our measurements of the tsunamis are only in the ocean, we gauge information as a tsunami passes by an observational buoy or coastal gauge.

Brian Shiro at the 2009 Fall Meeting (Photo by MJ Viñas).

Brian Shiro at the 2009 Fall Meeting (Photo by Peter Weiss).

Q: Why is it difficult to forecast exactly how high a tsunami is going to be?

A: Tsunami forecasting is a very active area of research, and it’s improving rapidly. Especially since the 2004 Indian Ocean tsunami, because there has been more public awareness and more computing power. But still, there are so many assumptions that go into a forecast model. All we have to work with is that an earthquake happened in this location, and it was this big. Sometimes you can also say how the fault moved once you have the focal mechanism. So you put the information into a model and make assumptions: if the earthquake moved this way, how does it translate into seafloor motion? Then you have your equations of motion of the water that happened from this excitation, and these predictions are good in deep ocean. But when you get into shallower water, you enter a non-linear regime and you have to use different assumptions in your model. It is still difficult to predict the exact runup at specific location will experience due to uncertainties in the model, computing limitations, and lack of high resolution data availability.

Also, the period of the tsunami is important too. A tsunami is actually a series of waves and the period is the amount of time between each wave. We were concerned about Hilo Bay because it’s notorious for amplifying tsunami waves: it’s just the right shape and size for tsunamis to fit perfectly inside the bay. The energy of the tsunami can’t dissipate in there, and when the next wave of the tsunami comes, its energy adds to the previous one, and the waves grow higher.

Q: So what needs to be done to improve tsunami forecasting?

A: There’s a lot of room for improvement. For starters, there need to be more ways to get the information out, so that there isn’t so much confusion when the prediction gets filtered through the media. For example, with social media we would have a way to control the message a bit better.

Also, this tsunami has taught us some valuable lessons in how we can improve the modeling. And the modeling has become so important in our decision-making that such improvements are critical. Thanks to refinements that are already in the works for tsunami forecast models, this time we could have excluded some areas from the warnings. But for this event we didn’t exercise that, we just said that this was going to be large enough that everybody is at risk.

As computing power gets better, the more complicated models that can depict in detail what’s going on near the shorelines are going to get better and better. Another big improvement will be having more data: one thing that delays us is having to wait for data to come in. Even when tsunamis travel fast, it takes them time to cross the ocean and reach detectors; if we had more detectors, we’d get the data faster and could have more accurate predictions..

I am optimistic that we will improve our ability to observe and forecast tsunamis in the future. Now we do the best we can with the information we have, which is actually pretty good.

Maria-José Viñas, AGU science writer