Interim report: Tsunami activity in Northland following Hunga Tonga-Hunga Ha'apai volcanic eruption

This report focuses primarily on the tsunami advisory of 15 January 2022 following an eruption at Hunga Tonga-Hunga Ha’apai volcano in Tonga, and the impacts at Tutukaka Marina in particular.

Background/Tuhinga

New Zealand’s procedures to receive, assess and disseminate tsunami notifications at the national level are set out in the National Tsunami Advisory and Warning Plan, which is available on the National Emergency Management Agency (NEMA) website at: civildefence.govt.nz

The diagram below is taken from a NEMA guide produced to help explain the respective roles and responsibilities and processes to the media (please note this is the process for regional and distant source tsunami – it differs for local-source tsunami [those generated close to New Zealand’s coastline]).

Of note is the explanation in Box 4 of the different advisories and warnings that can be issued. The general intent of a National Warning: Land and Marine Threat and a National Advisory: Tsunami Activity are summarised below the diagram (wording also from the same NEMA document):

Our process for regional and distant-sources tsunami

‘NATIONAL WARNING: LAND AND MARINE THREAT

Land and marine threats are as serious as our warnings get. A land and marine threat means that tsunami waves are coming, and will reach inland. A Land and Marine Threat will usually mean evacuation for at least some areas.

Always remember that the first tsunami waves may not be the largest.

‘NATIONAL ADVISORY: TSUNAMI ACTIVITY

Beach and marine threats are unlikely to require any evacuations (besides getting people off beaches). Strong and unusual currents and unpredictable surges are obviously dangerous for people in or on beaches and coastal waters, but people on dry land don’t need to worry about waves reaching them.

It’s extremely important that we get the strong currents and surges information out there, but generally this type of threat should be positioned as ‘low risk’ for anyone not actually in the water or at the beach.’

The majority of the content of these is templated, with details such as locations added as applicable.

Tsunami warnings and advisories issued by NEMA are sent directly to the media (at the same time and via the same mechanism as they are sent to CDEM Groups and other stakeholders) and also posted to NEMA’s social media channels and website.

Nationally and also within Northland, a Warning of a threat to land i.e. an expectation of inundation is the threshold for the use of alerting platforms such as Emergency Mobile Alerts, and tsunami sirens where these are in place.

Advisories for strong and unusual currents and unpredictable surges can be issued relatively frequently – up to four or five times a year in some years and often through the hours of darkness. While some commenters have pointed out their awareness that the siren sound is a signal to ‘seek further information’, past experience has however shown that a significant proportion of the population  will start to evacuate as soon as they hear a tsunami siren (or anything that resembles one).

The national process allows for local decision-making – however, this is generally in the implementation of the GNS assessment/NEMA advisory or warning at local level, rather than taking actions that are (or are perceived as) inconsistent. Given that media and the public have full and immediate visibility of the national warnings/advisories (and information is shared between friends in different regions), even variations with a sound logical basis can lead to reduced understanding and/or credibility.

The National Tsunami Advisory and Warning Plan outlines the thresholds for assessment of earthquakes for tsunami risk. Not all earthquakes are assessed (following the 15 January event, there is an expectation from some within Northland that this will be the case). Since 15 January, there has already been one further earthquake assessed (M6.4 in the Kermadecs region on 29 January), the outcome of which was ‘no tsunami threat to New Zealand’.

Sequence of events on Saturday 15 January

GNS assessment process

GNS Science is the mandated and official scientific assessor of earthquake/volcano/potential tsunami for New Zealand. GNS has published an article which gives an insight into its experience on this occasion, which can be found at the link below (and has since been followed by other articles about the eruption in general). The italicised content below is a compilation from two of these articles:

Hunga Tonga-Hunga Ha'apai eruption - information about the tsunami at: www.geonet.org.nz

Watch: A GNS Science animation of the event at: www.gns.cri.nz

‘A very large, explosive eruption occurred at Hunga Tonga-Hunga Ha’apai volcano in Tonga on 15 January 2022. The eruption was very energetic, generating a tall eruption cloud, explosions, shockwaves, audible booming and it created tsunami waves that travelled around the Pacific.

‘The huge eruption was the largest in the current eruptive episode at Hunga Tonga-Hunga Ha’apai that spans back to 2009. Back then, all you could see of the volcano was two small, elongated  islands poking about 100 metres out of the ocean. Hunga Tonga-Hunga Ha’apai volcano is located 65 kilometres north of Tonga’s capital Nuku‘alofa.

‘During 2014 and 2015 eruptions, a cone built at the volcano, bridging the gap between these two smaller islands to create one larger island.

‘An earlier eruption on Friday 14 January destroyed this central cone, flooding the vent with seawater. This added to the instability of the volcano, and the addition of large amounts of water may have contributed to the Saturday 15 January eruption being more explosive. The eruption exploded through the ocean, triggering a tsunami, with waves that headed for coastlines right around the Pacific Ocean. All that remains now of the volcano above water are two much smaller islands again.

‘The eruption was much more violent than scientists had expected, given the volcano’s run of smaller eruptions in recent decades. The atmospheric shockwave travelled around the globe several  times and was picked up on air pressure sensors as far away as Iceland - it continues to circle the globe. Audible booming could be heard from New Zealand to the south and Alaska to the north. This was due to the low-frequency bass-like booms produced during the eruption that can travel thousands of kilometres away from the source. This eruption now holds the world record for being ‘heard’ so far from the volcano.

‘The eruption was rare in that it caused tsunami waves large enough to impact thousands of kilometres away from the volcano. There are several mechanisms scientists believe could have contributed to generating tsunami waves: the huge explosion through the ocean, the shockwave that pushed out as it travelled, the effect of the collapsing eruption column onto the ocean, and a  possible caldera collapse of the volcano itself underwater. We haven’t seen a volcanic-source tsunami like this since Krakatau, Indonesia in 1883.

‘As soon as tsunami activity was reported following the eruption, the GNS Science team swung into action to assess the tsunami threat to New Zealand shores. The first sign of unusual activity came when the New Zealand DART buoy closest to Tonga triggered at 5:48pm. Over time, as data was recorded at more of our DART buoys and coastal tsunami gauges, we were able to build a picture of what threat there might be to New Zealand.

‘Assessing this tsunami was different from assessing earthquake-source tsunamis that occur more frequently. Where our experts usually use tools like modelling and forecasting, for this volcano-source event they relied on real time observations from our DART buoy network. The DART network was invaluable in our ability to make assessments about when and where the tsunami might arrive as well as its amplitude.

‘Cyclone Cody was also a complicating factor. The storm caused heightened swells around the east coast of the North Island and made our data more difficult to decipher. Storm energy shows up in recordings with a much shorter period than tsunami waves and it can make picking out the signal of the tsunami very difficult.

[Conversely, some of the criticism in Northland has centred around the view that the visibility of the eruption from space, the fact that the sonic boom was heard extensively in New Zealand, low atmospheric pressure and sea levels associated with Cyclone Cody, and the phase of the moon made it ‘obvious’ that a tsunami warning should have been issued].

‘Based on our assessments, the National Emergency Management Agency (NEMA) issued a National Advisory: Tsunami Activity for the east and north coast of the North Island and the Chatham  Islands around 8:30pm on Saturday. This was extended to the West Coast of the South Island around 9:40am on Sunday based on further observations from the network. They warned people to  expect strong and unusual currents and unpredictable surges at the shore’. – GNS Science. 

Northland CDEM Group summary of actions

Northland CDEM team members had already been monitoring the effects of Cyclone Cody, and were in communication with both NEMA and GNS Science during the assessment process which  followed the volcanic eruption.

Shortly after 8pm, an evacuation of three homes at risk from the Waiharara fire was also advised by Fire and Emergency NZ (the potential for this had been comprehensively planned in advance with the residents and communicated to all agencies. The Northland CDEM involvement in this was managed by one of the team’s Far North members, so did not reduce the overall capacity of the team).

In accordance with Northland CDEM’s SOP, when NEMA issued the advisory of strong and unusual currents and unpredictable surges at 8.14pm, it was emailed directly to stakeholders (e.g. emergency services, councils, port and marina facilities, lifelines utilities and the Welfare Coordination Group, as well as Community Response Group co-ordinators around the region) and was posted on the Northland CDEM Group Facebook page, also referencing/reiterating the existing forecasts of potentially hazardous rip currents, sea surges and coastal inundation created by Cyclone Cody, and the likelihood that the latter would continue beyond the duration of the tsunami advisory.

Having received the advisory forwarded by Northland CDEM, the Tutukaka Marina manager returned from home to the marina and was working through the marina’s own protocols for this type of  event in the leadup to the most damaging surges.

Some of the media coverage resulting from the NEMA advisory was shared to coastal community Facebook pages One of the Tutukaka boat owners who has been extensively interviewed by the media has commented that he was aware of the volcanic eruption/tsunami from media reports but was awaiting a siren or cellphone alert.

The option also exists within Northland CDEM’s SOP to send notifications via the Hazard app. This has a lower threshold than the Emergency Mobile Alerts platform but also reaches fewer people in Northland (this is simply a reflection of the fact that it requires people to download a free app). This option has been used for tsunami advisories in the past, typically when there has been a concern about large numbers of people out (or heading out) on to the water. Given the time of day (after 8pm), the existing sea conditions and the publicity that had been given to the Cyclone Cody forecasts, the decision was made not to do this.

Progressively from 9.29pm, Northland CDEM team members received reports from campgrounds in the Far North that had been evacuated as a precaution after elevated sea levels/wave heights  were observed.

Although some of these subsequently proved to have been less severe than was thought at the time, the collective view of the Northland CDEM team was that a more accurate picture was necessary and the following actions were taken: 

• A number of CDEM Community Response Group co-ordinators on the east coast of Northland were contacted to collect information
• The GNS tidal gauges were monitored via the GNS website
• The head of the NRC Hydrology team was contacted to gather hydrographic information at the Whangaroa and Ōpua tidal sites and then all other east coast tidal sites. These gauges are not
  intended for the purpose of tsunami assessment and the readings were affected by both the set reporting intervals and the swells generated by Cyclone Cody (which at that stage were measured at 6m+ at the Ninepin [northern entrance to the Bay of Islands]; however the rise-falls recorded by the tidal gauges at Whangaroa and Ōpua were in the order of 0.5m)
• Emergency services (Police and Fire and Emergency) were contacted to gather intelligence
• The Duty NEMA officer was contacted for further GNS updates
• We fed our observations back to NEMA and GNS Science during the evening and updated the following morning.

In the meantime at Tutukaka, surges at a level similar to previous events had been taking place at the entrance to the marina but escalated from 9.34pm, with the damage following very quickly from this point. Northland CDEM was advised of damage from 9.49pm. Surges continued at damaging levels until approximately 11pm and were evident (at gradually-diminishing levels) for a number of days.

The damage to vessels and to the marina structures has been well documented. Emergency services (Fire and Emergency, and Police) worked with the marina manager, community members and  local businesses to deal with the immediate priorities (evacuating liveaboards, securing loose boats and damaged parts of the marina as best possible, establishing cordons). Most of these  organisations were back again the following day, along with NRC Maritime staff from both Whangārei and Ōpua, CDEM staff and contractors engaged by the marina trust.

There were no reported injuries. However, the marina manager pulled a woman out of the water just after the initial surge. She was shaken but uninjured.

No liveaboard boats were sunk and no people were on board any of the boats that went underwater. All liveaboards at Tutukaka Marina have been back living on their respective boats (some in  different berths).

The surges did not overtop the breakwater at the outer (harbour) end of the marina, nor the seawall at the inshore (Village Green) end. By the technical definition, it remained a marine event and in  that regard, consistent with the GNS assessment.

What has subsequently emerged

The combination of storm surge, high sea levels and a volcanic eruption made the scientific assessment very difficult. Tsunami generated by volcanic eruptions have different characteristics to those generated by earthquakes (which represent 80% of tsunami), there is a relative scarcity of modelling and this is now acknowledged and being urgently addressed by the international community.

Observed effects on the west coast of both the North and South Islands were greater than anticipated in the original advisory, which was extended to include the west coast of the South Island at  9.40am on Sunday 16 January and cancelled at 7.06pm on the same day.

This event and the impacts at Tutukaka in particular are being studied by experts from a number of organisations. On the very limited number of occasions that media have sought input from parties with a perspective to offer other than Northland CDEM, the views expressed publicly to date have been in alignment with those of GNS Science. One of these experts, Jose Borrero of eCoast Marine Consulting and Research, has previously carried out detailed modelling of Northland ports and harbours (including Tutukaka) in a project co-ordinated through the Northland Regional Harbourmaster and Northland CDEM and funded through the NEMA Resilience Fund. He has commented that wave heights were no higher than in the 5 March 2021 event but the currents were much stronger, and has also suggested that there may be an additional offshore influence at Tutukaka contributing to the surges experienced there.

Around Northland, there were more people camping in low-lying coastal locations or fishing than anticipated, given the Cyclone Cody forecasts and the sea conditions at the time. Additional people took to the surf on Sunday in the hope of experiencing the effects from the tsunami surges. However a significant number have expressed a view that although they knowingly disregarded the  cyclone Cody forecasts and conditions, they would have acted differently had they been aware of the tsunami advisory.

A number of the effects observed elsewhere in the Northland region can be largely attributed to the storm surge generated by Cyclone Cody (see graph at end of this section). However two stand out (in addition to the campgrounds which were evacuated):

• The flounder fishers on the Hokianga Harbour
• People on board boats in Whangaroa Harbour in particular, which experienced violent swirling and surges to an extent that caused them to be concerned whether their anchor gear would fail  under the load. While these effects are broadly covered under ‘strong and unusual currents and unpredictable surges’ the risk of injury on this occasion exceeds the previous effects associated  with such an advisory.

In combination with the Far North campgrounds, some of these locations are (or were, due to conditions on 15 January) outside of cell coverage and the existing tsunami siren network.

A proportion of the public reaction relates to/contains references to COVID and/or Government decisions and is indicative of general frustration at how the past two years have unfolded.

Maximum wave heights recorded by the NRC wave buoy (located near the Ninepin in the Bay of Islands). Time period is from Monday 10 January to Monday 17 to show the increase in wave heights associated with Cyclone Cody. The ‘spike’ indicates the approximate time of the Hunga Tonga-Hunga Ha’apai eruption:

What could have been done differently

Had there been any indication – either in the advice from GNS Science, or from observed activity anywhere in Northland, or from past experience – of what was to ensue within Tutukaka Marina, a pre-emptive, local evacuation of those on board boats could have been undertaken to mitigate the risk of injury. This could have been achieved by similar (but sooner and under less stress) means to those that were ultimately used: the marina manager and others carried out the equivalent of a ‘doorknock’. The liveaboard community at Tutukaka is – intentionally - in a relatively compact section  of the marina. The use of sirens or Emergency Mobile Alerts would not have been necessary for this to be carried out (even if it was possible to restrict these to the affected area) and could have led to confusion and over-evacuation.

(Of relevance at this point is the comparison with the Far North campgrounds which undertook self-initiated precautionary evacuations at about the same time of the evening).

Components which gave way under load within Tutukaka Marina ranged from deck cleats (attachment points) on boats, lines (ropes), both wood and polystyrene/concrete parts of floating marina  fingers, and piles into the seabed. Given this, standard precautions such as securing boats with additional lines would not have altered the outcome in the instances where for example, cleats were torn from the boat, or boats went along with all or part of the structure to which they were attached.

Swell conditions at Tutukaka at the time (estimated by the marina manager as 4m at the entrance to Tutukaka Harbour), the onset of darkness and the relatively short amount of time available  meant relocating boats to deep water would have come with significant risks on this occasion.

In the years since 2010, Northland CDEM has carried out three evacuations in response to tsunami warnings using the siren network, for which one (5 March 2021) Emergency Mobile Alerts platform was also available to be used. The platform was launched in 2017 for ‘highest-priority alerts for risks that:

• Have occurred, are ongoing or where their probability of occurrence is greater than 50%
• Have a significant threat to life, health and property
• Where response actions should be taken immediately.’

An advisory for strong and unusual currents and unpredictable surges would normally fall below the threshold set by NEMA for use of an EMA. Conversely, the risk to people on boats at Tutukaka  from what unfolded would have done so (although as noted elsewhere, evacuation of that number of people from a defined area and with assistance already present or available, would not have necessitated an EMA). However the fact remains that an EMA has become the default expectation from some at least and its absence shaped their view of the risk and their consequent decisions/actions.

From feedback since the event, it has also become clear that people in other locations (including those which experienced minimal or no effects) have come to regard EMA as a source of information even when response actions are not required on their part. This represents a fundamental shift from how Northland CDEM has previously approached public alerting, which can be summarised as: ‘if there’s something you need to do, we’ll let you know but otherwise allow you to continue with life’.

Comparisons have also been drawn with the use of EMAs for COVID-19 Alert Level changes, often accompanied by suggestions that COVID had distracted the focus of the emergency management system from other hazards.

Both NEMA and GNS Science are well aware of the depth of feeling in Northland, with a significant amount of feedback having been posted on NEMA’s social media channels or emailed directly.

Comments posted on GNS’ platforms have been fewer and less critical, with a number still focusing blame on NEMA and/or CDEM.

What could be done differently in future

Local arrangements could be put in place that recognise Tutukaka Marina is more vulnerable to potentially-damaging surges than other locations, and that there is a desire on the part of the marina  community (and potentially those in the businesses and accommodation providers immediately surrounding the marina) to evacuate at lower thresholds. Any changes proposed by/advice  commissioned by the marina trust would need to form a part of this.

Neither of the present options for expected tsunami effects – either an advisory of strong and unusual currents and unpredictable surges, or a threat to land – are a particularly good fit for the risk to those living (either permanently or on shorter overnight/holiday cruises) on the water. One key means of reaching boaties (marine VHF broadcasts) is less effective through the hours of darkness; tsunami sirens are not the answer either, leaving EMAs which will also be received by a larger number of people not subject to the same risk and not intentionally targeted by the warning.

An option is to seek the creation of a new advisory of ‘strong and unusual currents and unpredictable surges which may in some locations be enough to present a risk of injury to those on boats’.

Similarly additional means of reaching those locations in the Far Far North in particular where cellphone coverage is not complete and there are significant gaps between tsunami sirens will be explored (communications to iwi via DoC radio having been the key link on the night).

Separately, a parallel may be drawn with the current relationship between CDEM Groups/regional councils and MetService, where the issue of a Severe Weather Warning is normally accompanied by direct communication with the duty forecaster, in which any indications that impacts may exceed the amounts stated in the forecast, specific local concerns or other variables can be conveyed.

Conclusions

A key determinant of the public perception of this event is the severity of the damage that took place within Tutukaka Marina, relative to what was anticipated by the advisory. As an event largely confined to a facility with a previous history of elevated levels of surge activity– albeit not to the extent experienced on this occasion – this indicates that any solutions put in place should also be localised.

Even foreknowledge of the damage that took place within Tutukaka Marina would not have been a trigger for activation of the siren network. However a local evacuation of the marina and surrounding buildings would have mitigated the risk to people, and this option and its implications should be worked through directly with community.

Comments from the public that they would rather receive EMAs and be evacuated on a precautionary basis are at odds with what was expressed following the 5 March 2021 warnings and evacuations, and the relative frequency of advisories for strong and unusual currents and unpredictable surges.

The desire on the part of some members of the public to see Emergency Mobile Alerts used on an ‘information and updates’ basis (rather than an alert to immediate actions required) will be problematic. Although the advisory on this occasion was issued at a time that would be publicly acceptable, this is often not the case and even continuing to use EMAs as this event unfolded would have brought a reaction from the public. The sound associated with the alerts was designed to be intrusive and does have an effect on recipients. In most situations, it is extremely difficult to confine the alert to those directly impacted (because it will deliver to all users of compatible phones within the coverage area of selected cell towers).

Northland CDEM will pursue the possibility of a new advisory for strong and unusual currents and unpredictable surges with additional risks through the national Tsunami Working Group, and will also explore potential localised approaches with the Tutukaka Marina and surrounding community once the marina trust has had the opportunity to arrive at some conclusions on its own future  plans.

A residual risk remains of further activity at Hunga Tonga-Hunga Ha’apai (both eruptions and landslides of parts of the remaining volcano into the sea).