Future Quakes: Communicating Long-Term Earthquake Forecasts
- Sally Potter
- Oct 29, 2024
- 7 min read
Updated: Oct 30, 2024
Effectively communicating forecasts for prolonged events can be really challenging.
Our brand-new published research investigated New Zealander’s experiences in communicating aftershock forecasts following the Canterbury Earthquake Sequence that began in 2010. The findings have helped inform how scientific agencies, such as GNS Science in New Zealand and the U.S. Geological Survey, as well as emergency managers, can let people know what the likelihoods and outcomes are from continuing earthquakes. If you are someone that might need to communicate about prolonged hazardous events in future, including long-term volcanic unrest or eruptions, this post is for you.
I’ve described my experiences in relation to the 2010-11 Canterbury Earthquakes here. You can see more on the impacts of these main earthquakes in our open access paper - Potter et al. (2015). One of the more significant aftershocks in the sequence was a M5.7 earthquake that occurred under Christchurch on 14 February 2016 – five years almost to the day after the most impactful aftershock in the sequence. Dubbed the ‘Valentine’s Day’ earthquake, this aftershock caused further damage, rock falls and power outages, but thankfully no reported injuries or fatalities.
GNS Science has been communicating the probabilities of aftershocks since the main event, and in fact, still are 14 years on. Co-author Matt Gerstenberger explains aftershock forecasting for the Canterbury Earthquake Sequence in the video below, and how there can be spikes of activity over time:
The 1-year likelihood of at least one M5.0-5.9 earthquake at the time of Valentine’s Day event was 50%, with 0 – 3 expected.
At home in Wellington with a five-week-old baby, my memories of that time are hazy – we were more focused on survival from sleep-deprivation. But our research team, led by Anne Wein from U.S. Geological Survey and Julia Becker from Massey University NZ, had already been studying the communication of the earthquake sequence by then (e.g., Becker et al., 2015), and kicked into action.
Conducting workshops with Christchurch residents and disaster researchers, emergency managers, and health officials, the team explored information needs and uses, and how they had evolved throughout the earthquake sequence.
Here is what we found. I’ve cut straight to the key implications for people communicating long-term forecasts – check out the publication if you want to read more about the data and reasoning behind this advice.
What to consider and communicate throughout a long-term earthquake sequence
If you are using this information to decide what to communicate and when for a major earthquake sequence, please note that you need to find out from your official local scientific advisers what the right context is for your situation. For example, the location (including depth) and magnitude of the mainshock may make it unlikely for any aftershocks to be felt. Or it may be that there is a reasonable chance of a significant aftershock occurring that is larger than the mainshock. The advice below is based on the context and our experiences of the Canterbury Earthquake Sequence.
Before a big earthquake occurs:
Instead of talking about a future large earthquake (singular), talk about it as a sequence of earthquakes, to plant that seed right from the start. This helps to avoid people being surprised later on, when large or impactful aftershocks occur.
Include aftershocks and their forecasts in readiness events and exercises. An example of discussion points to do this is in the HayWired Scenario Toolkit. Also, ShakeOut earthquake drills could include information about aftershocks.
Involve staff in readiness activities and exercises to ensure they know about procedures and plans for when an earthquake occurs.
Risk perceptions are usually low and may relate to more distant, potentially highly impactful hazards.
In the days after a big mainshock earthquake:
Communicate at the start of the sequence about how the number of earthquakes generally goes down over time, but that you can get spikes of activity that reinvigorates the sequence.
Risk perceptions understandably tend to relate to the impacts that occurred in the mainshock. They may be quite low if the initial mainshock didn’t cause too much damage, or high if the consequences were significant.
Forecast communication can use a variety of formats, such as through text, plots and graphics, and tables (explaining how to read them, if they are technical). Scenarios are another good option. Forecasts may be provided in relation to baseline risks. The forecasts should also include the uncertainties by describing the ranges in the (numerical) probability.
When using scenarios, three tends to be a good number to use, including a worse-case scenario. The likelihood of each of the scenarios should include the numerical probability if this is available and use the related qualitative/verbal terms (e.g. ‘unlikely’) alongside it. Including potential impacts can be helpful for decision-making, and ideally it would be accompanied by preparedness/action advice. Don’t forget to include the time period that the forecast relates to. Sharing the range in the number of expected earthquakes (rather than just the probability) has also been found to be effective. I’ll aim to write another more detailed post in future on how to communicate probabilities!
Omori's Law shows how the number of earthquakes goes down over time. Ideally this plot would have little upward spikes to show how the sequence can be reinvigorated by larger aftershocks. Image from GeoNet.
During periods of quiet after a mainshock or between aftershocks:
Risk perceptions tend to drop off as the time between damaging aftershocks increases.
Continue to communicate that aftershocks can still happen and give related information about hazards (such as unstable cliffs). Balance this with the need for people to get on with their lives, but don’t shy away from that proactive hazard, impact, and action/preparedness messaging. How frequently this message needs to go out depends on context. In the Valentine’s Day case where no large aftershocks had occurred for a while, and the annual probability of a M5.0-5.9 event was 50%, it was thought that around every 6 months would have been good. The message should be pushed out to get it in front of busy people with a wide range of responsibilities (including policy makers and central government). The best agency to do this may be the scientific monitoring agency in coordination with other agencies.
Scientific agencies producing the forecasts can update the information to show retrospectively what happened in comparison to the forecast. This can help to build trust in the forecast and agency over time.
Remind staff about earthquake activation procedures and plans. New staff in particular will need to know these.
Continue education with the public and preparedness campaigns, from earthquake mechanisms for the interested public, to securing tall and heavy furniture. New people may have moved to the area that have not done these actions.
Immediately after aftershocks:
Rapidly gain situational awareness to know what the consequences were from the aftershock. Other agencies may take their lead from whether there is emergency management response, and so wider and indirect impacts of the aftershock should be also considered in the decision to stand up the Emergency Operations Centre. Even if there were not many physical impacts, a communication response may be beneficial.
Proactively reiterate that having aftershocks or ‘spikes’ of activity is normal – people can find this reassuring and helps to make sense of what has happened.
Put out educational messaging about earthquake preparedness and response actions while people are engaged.
Scientific agencies could link the earthquake information web page to the aftershock forecasts to help people find them and see the longer-term context more easily.
In the days after aftershocks:
Risk perceptions can be very high and decision-makers risk averse for quite some time if aftershocks are damaging. Perceptions relate to the safety of the built environment in which people are spending their time – people feel safer once vulnerable buildings are demolished and stronger buildings have been constructed in their place.
Even if the physical impacts of the shaking were not significant, be prepared to communicate to two audiences – 1) those that are ‘earthquake seasoned’, and 2) those that are inexperienced with earthquakes. Both groups would benefit from receiving information about coping, taking safety precautions, and finding psychological support.
1) The ‘earthquake seasoned’ group will include those who have suffered emotionally and mentally from the stress of the ongoing earthquakes, and people who were experiencing the aftershocks from a different context such as having moved to a different part of the city or region since experiencing the mainshock. Some experienced people may feel that they don’t need as much information about earthquakes and what to expect than they did at the start of the sequence, whereas others may feel they are quite knowledgeable about earthquakes and desire more (e.g. technical) information, such as which fault the latest earthquake was on.
2) The 'inexperienced' group includes tourists, young children, and new residents, including those who may have moved to the area to help with reconstruction and recovery. This group may benefit from more empathetic messages.
Wrapping up
In all communications, the messaging needs to be coordinated between agencies. Use multiple authorities, channels, and complimentary messages with hazard information coupled with behavioural and psychosocial advice. Be transparent, open and honest in communications, as trust is a key factor in hazardous events that last a long time.
GeoNet shows that currently (at the time of writing) the likelihood of at least one M5.0-5.9 earthquake in the Canterbury area within the next year (calculated on 1 February 2024) is 29% (unlikely), with 0-2 earthquakes of this magnitude expected. As an example, they state:
“With every month that passes without a major aftershock, probabilities will continue falling. However, if another large aftershock occurs it can re-energise the system and spark a resurgence of earthquake activity for a month or so; this was seen with both the February and June 2011 magnitude 6.3 earthquakes.”
Nice work, GeoNet.
For more detailed information, you can check out the publication, especially Table 4. This research was led by Anne Wein from the U.S. Geological Survey, with co-authors Sara K. McBride (U.S. Geological Survey), Julia Becker (Massey University, NZ), Annemarie Christopherson (GNS Science, NZ), Emma H. Doyle (Massey University, NZ), Matthew Gerstenberger (GNS Science, NZ), and me, Sally Potter (at GNS Science during the period of conducting the research).
Publication details:
A. M. Wein, S. K. McBride, J. S. Becker, A. Christophersen, E. E. H. Doyle, M. C. Gerstenberger, S. H. Potter (2024). Long-term communication of aftershock forecasts: The Canterbury earthquake sequence in New Zealand. International Journal of Disaster Risk Reduction, Volume 114, 104878, ISSN 2212-4209, https://doi.org/10.1016/j.ijdrr.2024.104878.
Comentarios