Haiti is still rebuilding from the massive earthquake that occurred 11 years ago and is dealing with the consequences of the president’s assassination in July. As a result, the island nation is insufficiently prepared for the magnitude 7.2 earthquake that struck its western region on August 14. At the time of publication, the official death toll was hovering around 2,000, although media reports listed hundreds of deaths in the first few days after the earthquake.
But according to a scientific model of the US Geological Survey, the true number of deaths may be at least 5 to 50 times this number. Its tool is called the Rapid Assessment of Global Earthquake Response (PAGER), which automatically combines earthquake information with demographic and other data in the affected area to simulate the scale of a possible disaster, including death toll and economic impact.
PagerIt was launched in 2007 and does not predict the exact number of deaths. Instead, it estimates the probability that the number is within a certain range. For the recent earthquake in Haiti, PAGER gives a 35% probability of 10,000 to 100,000 deaths, and a 32% probability that the death toll will exceed 100,000.
This earthquake may not be as deadly as the earthquake that hit Haiti in 2010. PAGER correctly predicted more than 100,000 deaths. The tool also estimates that there is a 66% chance that the damage caused by this year’s earthquake will be between 100 million and 10 billion U.S. dollars. In contrast, the damage caused by the 2010 earthquake was between 7 billion and 14 billion U.S. dollars.
Early and accurate estimates of deaths and losses can greatly affect the scale of the response. News organizations tend to report the official death toll, that is, the number of dead bodies found so far. In a country like Haiti, with poor infrastructure and limited access to remote areas, it may take months to determine the actual number of deaths in this way. This has caused problems for international rescuers, humanitarian aid organizations, and other countries, all of whom need to assess the extent of the loss and determine appropriate response measures.
Scientific american Speaking with David Wald, a seismologist from the U.S. Geological Survey, who helped develop PAGER and other seismic modeling tools.
[An edited transcript of the interview follows.]
Why is the official death toll in Haiti so much lower than that of PAGER?
There are many things that the model predicts that you can’t see. What you don’t see are all the small towns and villages and remote buildings, some of which are located on these hillsides, which may not only be affected by earthquakes, but also by landslides. It will take a long time to realize all these areas, especially because of the challenges facing access and civil infrastructure and governance.So those [fatality] As we have always expected, the numbers will grow over time. We can say with certainty that the road will have problems due to landslides-because the road passes through steep terrain, where there is strong vibration.
What data does PAGER use to simulate the death toll?
It is impossible to obtain the exact number of deaths. There are too many uncertainties and too many unknowns. But to make a prediction, three factors need to be considered: the vibration, the population exposed to each level of vibration, and the vulnerability of that population (based on the building) [people] exist.
We know that for large earthquakes, the pattern of shaking can be very complicated, so we try to capture this situation as much as possible with a tool called ShakeMap. If you have thousands of sites, you will know that there is shaking everywhere, this will be a very constrained map.
In California, you have hundreds of radio stations. If you go to Haiti, there are not many seismic stations there and there is additional uncertainty. Until we help Haiti equip more instruments, it will always become more uncertain. The second factor is people exposed to these different levels of shock. You can use the population grid and calculate the population exposed at all different shock intensity levels. The last issue is its impact. In China or Haiti, buildings are very fragile. For the same vibration level, you may experience many times more deaths than in places with stricter building codes such as California, New Zealand, or Japan.
As far as Haiti is concerned, it has a large population, strong vibrations and very fragile buildings. This leads you to conclude that we will issue what we call a red alert on PAGER, and we may have a death toll of 1,000 or more.
Why is there so much uncertainty?
You don’t have many musical instruments. Haiti has an earthquake network that can locate earthquakes. But we cannot use these instruments for ShakeMap. They must be special instruments called strong motion instruments. There are about six or seven in Haiti, and we only have data for two of them.
Another source of uncertainty in our calculations is trying to figure out [the location and shape of] mistake. It also depends on the data and the quality of the data and the complexity of the earthquake. In this case, a very complex and challenging fault must be identified. Sometimes these things take months to really resolve. But we always try our best to complete it in the first few hours and days.
For modeling, the big problem is: we don’t even know how many people died in 2010. It is estimated to be between 100,000 and 320,000. Most countries have good reports. If the number of deaths is three, four or five, you can consider it very accurate. But once you fall into these truly huge losses, such as in Haiti, the truth is uncertain.
In this business, we can never be precise, but we can make a difference. For Haiti, it is in a red alert state anyway, so this is an international response.
When an earthquake occurs, how does PAGER record the earthquake?
First, the National Earthquake Information Center [determines] Magnitude and epicenter. It runs 24/7 to send information and trigger ShakeMap, which will use the available seismic data to generate vibration maps. Once ShakeMap is generated, PAGER will run. It needs to shake and cover the population. Using the country’s model, it can estimate the number of deaths. It will also send ShakeMap to all parts of the world to estimate vibrations and damaged systems.
Most earthquakes occur in green areas because they will occur in oceans or sparsely populated areas.Even level 8.0 is often in the subduction zone [collisions between tectonic plates in which one sinks beneath the other] Offshore, does not affect anything. One of the most important things about PAGER is to say that nothing happened, not what happened.
However, if it is an orange or red alert for Haiti, then it will page us and we can see the model generation. We will review the results before sending them. We sat on it for about 10 or 20 minutes while looking at other incoming information, such as better magnitude and better location. We believe this is a good starting point.
News reports use official figures, which may be far lower than actual figures. Does this mean that people will be less concerned about disasters?
In the media, there were some earthquakes that occurred at night, and no one paid attention. Then this will be another news cycle, Afghanistan and COVID and many other things, so the media is paying attention to these.
We said, “This will be worse than the initial report. The situation will be worse.” Based on some of these uncertainties, we may have overestimated the total loss. For the media, dealing with uncertain numbers is a challenge. This is a challenge for institutions, financial institutions and urban search and rescue teams. But we are pushing them in a direction that is “worse than what we have seen so far.”
Different PAGER users have different timelines. City search and rescue should start within a few hours, so you can use these uncertain numbers. Whether to provide aid worth 1 billion U.S. dollars can wait for a while. After a few days, things stabilized, and your answer was more stringent than initially estimated. So you can treat the model as an uncertain estimate, along with what is happening on the ground, and weigh these appropriately.