Five years ago, on a visit to the Pacific Tsunami Museum in Hilo, Hawaii, I played an interesting simulation game: setting off an undersea earthquake and deciding whether or not to issue a tsunami warning to the many countries in and around the Pacific.

The volunteer-run museum, based in ‘the tsunami capital of the world’, engages visitors on the science, history and sociology of tsunamis. The exhibits are mostly mechanical or use basic electronic displays, but the messages are carefully thought out.

The game allowed me to imagine being Director of the Pacific Tsunami Warning Centre (PTWC), a US government scientific facility in Ewa Beach, Hawaii, where geophysicists monitor seismic activity round the clock. When the magnitude exceeds 7.5, its epicentre is located and a tsunami watch is set up. Then, combining the seismic, sea level and historical data, PTWC decides if it should be upped to a warning.

The museum game allows players to choose one of three locations where an earthquake happens — Alaska, Chile or Japan — and also decide on its magnitude from 6.0 to 8.5 on the Richter Scale.

This is an instance where scientists must quickly process large volumes of information and add their own judgement to the mix. With rapid onset hazards like tsunamis, every second counts. Delays or inaction can be costly — but false alarms don’t come cheap either.

I played the game thrice, and erring on the side of caution, issued a local (Hawaiian) evacuation every time. If it were for real, that would have caused chaos and cost the islanders a lot of money.

In fact, those who make decisions on tsunami alerts or warnings have to take many factors into account – including safety, economic impact and even political fall-out.

PTWC is only an expert facility that recommends action: countries covered by its technical advisories are left to make their own national decisions. Depending on local circumstances and considerations, they may act on it immediately – or decide to wait and see.

After playing the simulation game, I can better appreciate the predicament government officials who shoulder this responsibility. They walk a tight rope, balancing short-term public safety and long term public trust in the entire early warning system.

“Tsunami prediction is an inexact art practised in conditions of imperfect information and time pressure,” says Dr Rohan Samarajiva, head of the regional think tank LIRNEasia and a former telecom regulator in Sri Lanka. “In the Pacific Basin, which has had the most experience with tsunamis, 75 per cent of all warnings are false.  But this causes little harm because the false warnings do not get through to the general population for the most part.”

Indian Ocean tsunamis

In contrast, countries in the Indian Ocean have less than a decade’s experience in dealing with tsunamis. There may have been killer waves in historical times, but modern memories begin with the devastating tsunami of 26 December 2004.

On that fateful day, PTWC detected and warned about it, but most countries in South and Southeast Asia lacked national decision making capacity and public warning systems to act on it quickly and resolutely. As a result, than 250,000 lives were lost.

Those memories no doubt played a big role on 11 April 2012, following a powerful undersea earthquake, government officials in many Indian Ocean rim countries agonised over the right response. To warn or not to warn — that was the question.

The 8.6 magnitude quake occurred at 8.38 UTC (14:08 Sri Lanka Time), 440 km southwest of Banda Aceh in Indonesia and 33 km beneath the ocean floor. That was relatively close to the location from where the December 2004 tsunami originated.

PTWC issued its first information bulletin six minutes after the 4/11 quake. It introduced an Indian Ocean-wide Tsunami Watch, recommending a state of readiness to act, covering 28 countries and territories.

Over the next few hours, they updated their assessment, but didn’t escalate the tsunami watch to a tsunami warning, as only minor tsunamis were generated during the aftermath. At 12:36 UCT (18:06 SL Time), they called off the tsunami watch.

This time around, the Indian Ocean had its own tsunami early warning system, set up under the UN Intergovernmental Oceanographic Commission (IOC) following the 2004 tragedy.

Anchored around three early warning providers – operated by Australia, India and Indonesia – the Indian Ocean system combines data from underwater probes, orbiting global positioning system satellites and floating buoys to better detect a coming tidal wave.

On April 11, individual Indian Ocean countries reacted differently. Indonesia issued a warning five minutes after the quake. India’s was in eight minutes, and Australia’s, in 10. Several others issued warnings; some followed it up with coastal evacuation.

According to news reports, Thai authorities shut down the Phuket international airport, and evacuated hotel guests in the coastal resort area to the hills behind. In southern India, meanwhile, the port of Chennai closed down for a few hours. These were among the locations badly hit in 2004.

Sri Lanka’s official warning and coastal evacuation order came around 15:30 SL Time, issued by the Department of Meteorology and released to the media and public by the Disaster Management Centre (DMC).

If a trans-oceanic tsunami was indeed generated, it would have reached Sri Lanka’s east coast (Trincomalee) in just over two hours from the quake. So the window to act was tight – but long enough for coastal evacuation.

But there was considerable chaos just before and after the official warning. The undersea quake itself was felt almost instantaneously in many parts of Sri Lanka as a mild tremours, physically alerting people about the Indian Ocean’s rumble. Within minutes, at least two dozen radio and TV channels were in ‘Breaking News’ mode. The online social media platforms like Facebook and Twitter also exploded with citizen updates, opinions and emotions.

In the absence of a steady flow of official information, radio and TV kept repeating whatever they found online. Not all of it was drawn from credible or trusted sources. That, combined with inexperience of some announcers, contributed to public confusion.

“An ignorant public at a time chaos is not a good thing,” says journalist Amantha Perera, who reports from Sri Lanka for global media outlets including TIME Magazine. “There was also massive confusion: the government said it was a tsunami warning and got people out of the coast, but there was no follow-up information coming from official channels, which really led to chaos.”

Individual agencies acting in isolation belied the absence of a coordinated response strategy. Buses and train services were stopped on the coastal lines. Electricity supply was shut down in certain coastal areas. Public and private offices were closed early, with thousands of workers suddenly asked go home.

Dr Samarajiva sees these as “a disorganised random set of responses uninformed either by realistic assessments of the risk (exemplified by the model predicting arrival times) or by definitive guidance from the government.”

In a post mortem, he adds: “When it came to issuance of warnings, evacuation orders, etc., the government earned a failing grade. Not enough authoritative direction was provided in time.” (full text at: http://tiny.cc/RS411).

Fear and Panic

Journalist Amantha Perera felt that most mass decisions that day were motivated by fear. “Media, government and others did not do anything to keep the people clam and informed – telling them the situation was under control…it appeared no one was in charge, at least when it came to  keeping the public informed after the warning.”

He adds: “There was also no firm authoritative figure who came out and assured the people. [In contrast] we saw what the Indonesian president did: when it became very clear that a tsunami was unlikely, he was the one who assured the country.”

We also found the limits of early warning technology. Post-quake overloading of telecom networks was predictable. More worrying was how the DMC’s arrangement for cell broadcasting – a method that can send text messages simultaneously to a large number of mobile phone users in a given area – failed. Two weeks later, that has yet to be explained.

At least 10 coastal warning towers – built after 2004 with foreign aid – also didn’t work. This is being investigated and remedial action has been promised.

Thorkild Aarup, Head of the Tsunami Unit of the UN Intergovernmental Oceanographic Commission (IOC), has acknowledged gaps in the Indian Ocean system.

Speaking within hours of the April 11 incident, he said: “But the Indian Ocean is much better prepared than it was in 2004. The tsunami early warning systems are like the atmospheric systems used by meteorologists which are constantly being improved by new technology. The same is true for the tsunami warning systems”.

In the end, the problems have less to do with specific technologies of monitoring and data gathering, and more rooted in the human systems of decision making and crisis response. The biggest challenge, for national authorities, is how to make the best possible decisions on the run, under immense pressure, and often with incomplete information.

Here too, the Indian Ocean can learn a few things from the Pacific experience. Since the system was set up in 1947, it has never missed warning of a damaging tsunami — but there have been a number of very expensive evacuations that turned out to be unnecessary.

“These precautions are needed to ensure public safety, but scientists are working to minimise unnecessary warnings without ever missing a hazardous event,” explains one panel at the Pacific Tsunami Museum.

Easier said than done! As Dr Samarajiva says: “Disaster risk-reduction professionals know that false warnings are an artefact of the inexact art of predicting the onset of hazards: but the general public does not.  If they are subject to too many false warnings, they will not respond even to true warnings.”

Too much of a good thing?

So was the tsunami warning and coastal evacuation on April 11 justified? This needs careful, dispassionate analysis in the coming weeks.

“Better safe than sorry” might work the first few times, but let us remember the cry-wolf syndrome. False alarms and evacuation orders can reduce public trust and cooperation over time.

Public behaviour – in both good times and bad — is a composite phenomenon made up of millions of individual citizens making private decisions in their self interest. While many heed their ‘herd instinct’ during emergencies, some can — and do — refuse due to their own reasoning. People can’t be saved at gun point.

A case in point is what happened in southern Bangladesh in November 2007 as cyclone Sidr approached. A false tsunami alert and evacuation two months earlier (on 13 September 2007) had led thousands of Bangladeshis to ignore cyclone early warnings. As villagers and officials later admitted, this accounted for many of the over 1,000 lives that were lost to the cyclone.

On the whole, Bangladesh is an outstanding success story in community based early warning systems – it saves thousands of lives from cyclones that regularly hit the deltaic country. But as the Sidr experience showed, too much of a good thing can be harmful.

The rapid spread of information and communication technologies (ICTs) introduces a new dimension to emergencies. The multiplicity of information sources, channels and access devices is certainly better than their absence. But they also make it harder to achieve a coherent and coordinated response. Controlled release of information is no longer an option for any government.

During those tensed hours on April 11, it is likely that many of the 15 to 20 per cent of Lankans with web access went online to look up and/or share information. We saw the power of social media: in a spontaneous collaboration, several regular Twitter users (tweeps) stayed active throughout the period.

Tweets not only updated on what was happening in Colombo and other coastal areas, but also relayed latest news from established wire news services (such as AP and Reuters) and mainstream media in Sri Lanka. When some radio broadcasts were creating panic, tweets pointed out that at first, it was a tsunami watch — not a warning.

In contrast, the official websites of the DMC  and the Met Department had no updates for at least 90 minutes after the quake.

As I noted in another recent essay, we can’t expect state agencies to become twitter-happy overnight (although timely updates of their websites would be a good idea). At a minimum, they must realise the info landscape is now transformed.

“If we don’t get the language and communication right, greater use of (social) media can actually aggravate confusion and chaos,” cautions Sanjana Hattotuwa, one of several Lankans who covered April 11 crisis on twitter, from @groundviews.

For a start, everyone needs to discern a tsunami WATCH (stand-by for more) from a WARNING (take action). Many – including some journalists – still don’t appreciate the difference. It is even more confusing when hastily translated into local languages.

A simple and language-neutral colour code system can help. As I’ve been saying, why not adopt the well known hierarchy of green–amber–red, already well known in traffic lights?

Yes, we have come a long way since 2004. But we still face many challenges, now of a different kind. A little learning can be dangerous. Patchy awareness – combined with fear and rumours — can easily trigger panic.

Disaster early warnings are pure public goods. But in our modern information societies, each member of the public must decide what is good for them. Public trust is the lubricant that will move the wheels of law and order as well as public safety in the right direction.

Science writer Nalaka Gunawardene has been covering disasters as a journalist and film maker for over 20 years. In 2007, he co-edited Communicating Disasters: An Asia Pacific Resource Book