Making it out alive

When a major tsunami arrives, the efficiency of our retreat will be decisive. There’s debate over the best way to prepare, reports Charlie Mitchell.

The timer starts. Red dots on the map start moving. The tsunami is on its way. The dots spill out on to the streets, swarming to high ground. They get stuck in congestion, slowing them down; they choose other routes, looking for the path of least resistance.

When they reach high ground, past the blue line, the dots turn green. They are safe. But many stay red.

This is the product of a sophisticated computer model, which tries to predict how a rapid tsunami evacuation would unfold in a densely populated area.

On a given weekday in Wellington city, about 85,000 people live and work in a tsunami evacuation zone. That zone mainly includes land reclaimed after the 1855 Wairarapa earthquake. It is low-lying and has little natural defence against a tsunami.

In the event of a large earthquake in the Hikurangi subduction zone, a tsunami could reach Wellington city in about 30 minutes. Thus, a problem: how do 85,000 people leave the evacuation zone within 30 minutes?

In the model, each person is represented by a red dot. The dots are programmed to take the most efficient route to high ground, unless the volume of dots in a given space is too high (congestion), in which case they take a different route.

‘‘The evacuation process is very complicated, and our model is a simplification,’’ says Dr William Power, a tsunami modeller at GNS Science who co-authored the research. ‘‘The main thing we hoped to find with it are some of the bottlenecks where people could be slowed down due to congestion.’’

When the simulation begins, and the dots begin their hikoi, they strike a problem.

For most people, the quickest route is via The Terrace, an elevated road arcing towards the Beehive. But getting from Lambton Quay to The Terrace requires moving through alleyways, which start as wide roads, but end in steep, narrow stairways. Many of the dots get here and stop: the congestion is too high.

After 30 minutes, the simulation ends, and the dots stop moving. Around 65,000 green dots have made it out of the evacuation zone; about 20,000 stay red. Although it’s a novel way to simulate how humans behave, the modelling has significant limitations.

It doesn’t account for people who drive or those who choose not to evacuate at all. Many people wouldn’t take the most logically efficient route. Collapsed buildings and debris might block roads, and make evacuating harder. It doesn’t specifically account for wheelchair users or people with limited mobility.

What it does show is how a city’s design, mixed with human behaviour, can be a barrier to saving lives.

‘‘On the whole, an evacuation is likely to work better if most people in the CBD use the main roads rather than trying to use these alleys and staircases,’’ Power says. ‘‘A sufficiently timely evacuation will always depend on people getting out of buildings quickly, and maintaining a good flow of movement along the main roads.

‘‘These are factors that can be improved significantly through good and effective education and signage.’’

War gaming

While the physical science of subduction zone earthquakes is quite well understood, how real-life New Zealanders will behave in the precious minutes before the waves arrive is a mystery.

How individuals behave is an important determiner of whether someone survives a tsunami. In the 2011 Japan tsunami, research shows early evacuation was strongly linked to whether someone survived. Such decisions were often influenced by previous participation in tsunami drills.

‘‘If you look at the Japan experience, around 95% of people were able to evacuate safely. So even though there were still a large number of deaths, that in itself was good news, which is probably down to the fact they do a lot of drills and a lot of practices and people knew what to do,’’ says Dr Julia Becker, an associate professor with Massey University’s Joint Centre for Disaster Research, who studies natural hazards from a social science perspective.

‘‘Drawing on that experience, making sure people are aware of tsunami hazards, know what they need to do, and practising that, is really important.’’

Local-source tsunamis are a devilishly difficult hazard to prepare for. They are likely to happen in the immediate aftermath of an earthquake, one that may have caused injuries or infrastructure damage. They happen so quickly that no official warning can be given. They require us, as individuals, to have a plan, which we must execute quickly in a period of utter chaos.

For authorities and experts, preparing the public for this specific threat requires a type of psychological war gaming. It’s not enough to assume we will behave ideally, even if we know what that ideal looks like. They have to navigate our biases, our preconceptions and our mental constructs of the world. It’s an imperfect science, which often changes.

After a tsunami devastated Indonesia in 2004, many parts of New Zealand bolstered their preparedness by installing tsunami sirens. Hundreds of sirens were set up to blare when a tsunami warning was issued.

Some communities refused, including Wellington. The reasoning is complicated – and contested – and shows the complex psychology of disaster preparedness. The logic is that in the event of a near-source tsunami, there would not be enough time to activate a siren. Any coastal resident waiting for the sound as their cue to evacuate would be in danger.

Sometimes, sirens go off accidentally, causing needless chaos and conditioning people not to trust them. They can trigger people not in an evacuation zone to leave, adding to congestion. It might not even be clear to some people what the sound is trying to communicate.

These all proved to be problems during the Japanese tsunami. Despite a sophisticated early warning system, some researchers argue it did more harm than good.

After Japan, both GNS Science and the National Emergency Management Agency (Nema) recommended against installing sirens. Because disaster management is managed regionally, some communities have taken heed, while others have not.

It speaks to a broader problem. Unlike some other hazards, New Zealand has limited experience with large tsunamis. It can be difficult to know which measures will work, and which will have unintended consequences, until it is too late.

In 2011, Wellington city trialled an innovation in tsunami preparedness: blue lines painted on roads,

indicating where a tsunami safe zone begins. They have since been rolled out in other regions, and even overseas.

They are a gentle, psychological nudge that tsunamis pose a threat, and give people an inkling of where they should go when they evacuate. Many experts support them. But could there be unexpected consequences?

‘‘For example, would people tend to evacuate only to that blue line, and would they stop there?’’ says Dr Lauren Vinnell, a researcher at Massey University’s Joint Centre for Disaster Research who studies natural hazard preparation. ‘‘Ideally, we’d like them to go further, so we don’t end

up with a bottleneck at these points. With things like that, it’s hard to know until we have a large event and see how they work in practice.’’

In recent years, public messaging about tsunami evacuations has centred on the phrase ‘‘Long or Strong, Get Gone’’, intended to remind people to evacuate immediately in the event of a long or strong earthquake.

It has been effective: in a recent survey, 90% of people knew that was what they needed to do if they were near the coast after a quake.

But there’s a difference between being aware of something, and understanding it.

‘‘Do people understand that it’s

long or strong, not long and strong?’’ Vinnell says. ‘‘And do they know what we mean by that?’’

Long or strong

A few decades ago, many New Zealanders had little direct experience with earthquakes. They had lived in a relatively quiet seismic period, where stories of the devastation in Napier in 1931 were relics of a distant past.

That changed with the Canterbury earthquake sequence, followed by the North Canterbury/ Kaiko¯ ura earthquake. Over half a decade, more than one million New Zealanders had direct experience with earthquakes.

That is not so for tsunamis. We remain in a quiet period, which will almost certainly influence our response. In 1960, the largest earthquake ever recorded – a magnitude 9.5 – struck off the coast of Chile, sending waves hurtling across the Pacific Ocean towards New Zealand.

At the time, there were no tsunami warning systems, even for distant quakes. Large waves hit the eastern coastline in the night; noone died, but there was considerable damage to infrastructure.

It shocked authorities into action. Just three days later, an aftershock led to a mass evacuation; many thousands of people were evacuated to high ground, largely under the direction of the police.

Some people behaved differently. A photo published at the time shows a line of people – including children – sitting on the Napier wharf, waiting for the tsunami to arrive.

Some communities were particularly vigilant, while others did nothing. In the days afterwards, The Press reported ‘‘glaring’’ deficiencies.

It would almost certainly be different in the modern era. The experiences of Indonesia in 2004 and Japan in 2011 were widely publicised; more New Zealanders than ever are, at least in theory, alert to the threat of tsunami. We do, however, have reason to worry.

Shortly after midnight on November 14, 2016, a long, severe earthquake rattled much of the country. At first, there was no reason to fear a tsunami. The earthquake’s epicentre was about 30km inland, in North Canterbury.

At 12.40am authorities issued an advisory noting there was no tsunami threat. No-one could have known the earthquake was among the most complex recorded, rupturing a network of at least 21 fault lines, including several that went out to sea.

Shortly after the advisory was issued, GNS Science warned that a tidal gauge off Kaiko¯ ura had recorded a sudden drop in the tide – a precursor to a tsunami. A tsunami warning went out to all residents of the North Island’s south coast at 1am, which 25 minutes later was expanded to the eastern coasts of both islands. In the dead of night, many people left their homes for high ground.

To researchers, the community response that evening was a rare glimpse into the future. The earthquake was strong, with shaking that lasted more than a minute: At the time, there was no way of knowing it hadn’t triggered a local-source tsunami capable of devastating coastal areas within minutes.

How well did people do? Researchers surveyed people in Petone, at the top of Wellington harbour, and nearby Eastbourne.

The results, researchers later noted, were ‘‘alarming’’. In a perfect world, residents would have known in advance whether they lived in a tsunami evacuation zone. If the answer was yes, they would have evacuated immediately, once the shaking stopped.

They would not have waited for official guidance. They would not have looked on Facebook for more information. They would not have

collected supplies or waited for friends or family so they could evacuate together. If they were not gone within 10 minutes, they were too slow.

The surveys showed that twothirds of respondents did not evacuate within 10 minutes. While most did evacuate at some point, the most common reason for doing so was advice from Nema, which came an hour after the earthquake. Only 11% of people specifically cited the earthquake – the only warning you would get for a localsource tsunami – as the reason they evacuated. Even among those who evacuated, the majority took too long.

‘‘There’s a reasonable number of people who understand they have to evacuate, and perhaps do evacuate, but they’re a little bit slow about it,’’ says Becker, from Massey. ‘‘It’s important to speed up some of that, especially for the local events where you might only have 10 or 15 minutes’ warning before a tsunami.’’

For many of us, our experience of tsunamis is through events like the Kaiko¯ ura earthquake; a tsunami is deemed possible, but does not arrive. When it keeps happening, we become conditioned to expect this next time, too.

March 5, 2021

Weeks after that earthquake, Nema rolled out its ‘‘Long or Strong, Get Gone’’ campaign. It was followed by the emergency mobile alert system, which commandeers mobile phones to produce a sirenlike sound in an emergency.

Last month, the first nationallevel tsunami hazard map was released. Nema has opened a Monitoring, Alerting and Reporting Centre which is staffed 24/7. To what extent they have helped remains to be seen. Significant changes are likely to take time.

When a local-source earthquake struck off the coast of Gisborne on March 5, 2021, it produced another scenario to study.

Researchers again deployed surveys in the aftermath.

The results were mixed – again, only about a third of people evacuated immediately – but because the earthquake was experienced as shorter and weaker, it might represent an improvement.

For the most part, authorities found the response that day encouraging. While there were individual cases of people doing the wrong thing, the issues were not deemed systemic.

‘‘By my estimation, it was the largest mass evacuation of New Zealanders in modern history, and it was done under conditions that were sudden onset and at large scale,’’ says Roger Ball, Nema’s national operations manager, who was the agency’s acting director that day.

‘‘On the whole, this was the community doing what we asked them to do, which was keep an eye out for the natural signs, look after themselves, look after their neighbours, implement their plans . . . It was tremendous to see.’’

Tsunami responses can be difficult because they often depend on every person doing the right thing, in the absence of any advice from authorities. The March 5 research surveys found a possible upside to this. ‘‘We’re relying on individuals doing the right thing, but when we looked at, for example, March 5, about one-third of them evacuated because friends, family, or neighbours evacuated,’’ says Vinnell, from Massey, who led the research.

‘‘We know that in emergency situations like that, which are uncertain or a bit scary, people do look to those around them to see what they’re doing and sort of follow their behaviour. So while it is individual, if we can get enough people in a community or a neighbourhood evacuating, then that’ll help prompt others who might not have otherwise evacuated to do so when they see their neighbours going.’’