Are earthquake "early warning systems" to be trusted?

The opinion : NO and YES, ... with the right use of the alert, many things can be done. Unfortunately, the vast majority of people are doing exactly the opposite of what they should do in such an event.

After every major earthquake there is hype for "earthquake early warning systems" like smartphone apps, computer tools, etc.

Let's explain first what happens when an earthquake is triggered and probably explaining it, will give already part of the answer!
- Due to the extreme strength of 2 moving plates or sections of plates,  the crust will rupture at the weakest point and an earthquake is generated.  The rupture will lead to 2 main waves which will be disseminated from the rupture point or area.
a) the P-wave or Primary wave which is used by the alert systems to trigger the alert. Depending on the Magnitude and the composition of the ground layers the P wave generally travels somewhere between 4 to 8 km/second.
b) the S-wave travels approx. at 60% of the P-waves (let's call it not so fast) and is responsible for the start of the shaking.
When a massive shallow earthquake hits below land, the shaking will be almost simultaneously felt together with the alert sound of an early warning app or tool, as there are usually only a few seconds in between both waves at short distances.
The area at risk for damage and injuries in 95% of the earthquakes is limited to tens of km, not to hundreds of km (with the exception of soft soil zones and large M7.5 events).  Only national seismology agencies who have installed multiple instruments all over the country are able to trigger an alert. Measuring the alert and redistributing an alert signal will at least need some seconds. The area at risk is already shaking heavily at that time. Drop, Cover and Hold (the generally accepted preparedness standard) is the only recommended way to protect you from injuries. Further away from the epicenter (from 50 to several hundred km) the time in between the alert and the time the shaking starts can be used to take the Drop, Cover and Hold position or to follow the approved evacuation methods.

EAS can be of serious help in the case of Tsunamis and if the signal is triggered by National seismology agency. In most cases however, authorities have installed professional tsunami alert sirens along the coast.

The video below has been made by the Huffington Post, a USA on-line news website. Dr. Thomas Heaton directs the Earthquake Engineering Research Laboratory at CalTech who tells the audience that also the US is working on it, but some hurdles have to be taken. Please be aware that there is a short advertisement at the start.

 The conclusion :
- early alert systems give a false feeling of security near the epicenter
- early alert signals should be provided by government or local authority specialized agencies
- early alert systems have to be used to take the Drop, Cover and Hold position until the shaking is over.
- early alert systems may never be used to evacuate buildings, unless the technology is there to invent alerts which can predict an exact time that the secondary wave would reach your location.
- early alert systems are being used and should be used even more in automatic "safety switches" like : trains (automatic stops), elevators (mandatory stop at next level blocking doors open), etc.
- early alert smartphone systems can be useful in smartphones when the Alert message is combined with your GPS location and if a voice can speak loudly  "earthquake - weak shaking expected at your location". All this without having to manipulate your phone!
- early alert systems can be beneficial in case of tsunami danger and for shutting down hospitals, lifelines or switching power supplies to retain essential services
- early alert systems DO NOT predict earthquakes

Early Warning Systems in Mexico City, Japan and cities where they are a long distance from the major fault zones but built on soft soils give useful data for citizens to prepare. However, other early alert systems should only be used for authorities.

We have found a very good article which was written in the The article describes what currently is tested in the USA and what an alert should have to do, how much time it allows etc.

Remarks based on our extensive experience following up what happens in real life during an earthquake :
- computers are normally not connected with backup power or batteries. We experience in 97% of all destructive earthquakes that power is lost within seconds
- smartphones are great BUT most Mobile network sending masts are NOT battery powered and will go off-line instantly when there is a power failure
- in case of battery powered masts (like in New Zealand) people are almost instantly run into an overheated network, disabling the great apps on your smartphone. A simple solution is repeated over an over again by Block voice calls and send a Text message that "Due to an earthquake, only Text messages will be allowed until further notice".


  1. John Vidale says

    A number of details are not quite right in this article - refer to the interview with Heaton or Berkeley's Richard Allen for a more nuanced discussion. The basic conclusions that the usefulness of earthquake early warning can be oversold, and will take a lot of work to implement many capabilities are correct, however.

    • Armand Vervaeck says

      Thanks John for commenting. We invite you to be more specific. We gave described, just like Heaton, that these EWS can be very beneficial, but only if properly used and if the software and hardware is overcoming a lot of repeated anomalies.

  2. I think this article raises important issues about the limitations of an EEW system, but at the same time we should look to Japan for how they communicate this to the general public. Their EEW education (including that from the Japan Meteorological Agency) makes it very clear that little to no warning will be provided to locate at or immediately near the epicenter. I think many countries planning to adopt an EEW system should similarly be honest to their citizens about the limitations to avoid a false sense of security. I still believe that areas somewhat far from the epicenter can still benefit; the Marina district in San Francisco was quite far away from the Loma Prieta epicenter and sustained heavy damage; one commercial vendor here states they could provide about 11-12 seconds of warning should another Loma Prieta earthquake occur at the same place and magnitude.

    To go along with how to alert people, I completely agree that the alert should be provided as an instruction rather than a mechanical tone or siren. We don't want people to try to interpret the sound - instead it should clearly state in a voice "Earthquake alert, moderate shaking detected, immediately drop and take cover until shaking stops". The current alerting sounds of buzzers and such are far too confusing for most people, particularly when the alert is heard infrequently. The biggest benefit of an EEW system is for automated processes, specifically stopping elevators to prevent entrapment, since this is a fairly common situation after a major earthquake and provides a heavy burden to fire departments needing to be called in.

    Finally, I think we should not discount the benefit of having on-site detection, which some places in the United States utilize (as no other options exist for them). Onsite alerting has benefits in that it can be built to be redundant from area wide data or power failure and may provide protection even when the earthquake is near or close to the location. A few systems in Japan utilize a hybird combination/internet connected EEW alerting system.

    • Thanks for the insightful comment Mark. A must read for everyone!
      Extremely true. I think that there is also much benefit in EEW system for essential services, soft soil site further away from the general major occurring earthquake zones (Mexico City, sometimes Japan, San Francisco etc.). Authorities just have to be realistic!

  3. Gerardo Lazos says

    In the case of Mexico City, the Sistema de Alerta Sísmica (Seismic Alert System) has proven useful for the general population, as we have been building up a seismic awareness culture through the years. Mexico City is located 300 km away from the closes epicentral areas, which provides us with an advantage early warning of up to one minute before the seismic waves impact the city. That allows for employees to head for the safety areas within skyscrapers or to evacuate the smaller ones. Certainly, one such system cannot save everyone, but can do the job for many and that is absolute gain for us and everyone seeks to be warned one minute before impact. But then, that is Mexico City.

    Not all cities are lucky enough to be located as far from the epicentral areas. Then an early warning system would not be as effective, but then a not so early warning is better than no warning at all. Think of this as a fire alarm without the sprinklers. It might not put off the fire, but it will tell you when to take action. The same principle is true for the seismic early warning systems.

    Some of your readers might wonder why Mexico City requires such a Seismic Alert System when the epicentral areas are so far away. The city was built upon a lake in the time of the Aztecs. Therefore, the muddy soild conditions in the city amplify the seismic waves and so they feel as intensely as in the epicenter, although differently, because here we get to feel a roking motion as if we were on a boat. But the rolling can get pretty nasty as it happened as recently as March 20, when many people had the chance to evacuate in a timely fashion thanks to the Seismic Alert System.

    • Armand Vervaeck says

      We agree on the specifics of Mexico City and are happy that the last earthquake did not create any damage