Tsunamis are one of the most destructive and feared natural phenomena for coastal populations. In a matter of minutes, these giant tsunami waves can wipe out entire communities, cause tens of thousands of deaths, and result in incalculable economic damage. In recent decades, the world has advanced in its ability to detect tsunamis through early warning systems. However, the reliability of these technologies against tsunamis is still far from optimal.
A large part of current systems is based on the interpretation of seismic data, which involves a high margin of error: many times, false tsunami alarms are triggered, and in other cases, the alerts arrive too late. In response to this problem, a group of scientists at Cardiff University has developed GREAT v1.0 (Global Real-time Early Assessment of Tsunamis), a system that combines acoustic analysis and artificial intelligence algorithms to detect tsunamis almost in real time.
This innovation, already undergoing operational testing in Europe, could redefine global standards in the prevention of natural disasters like tsunamis. In this article from ITD Consulting, we explore how GREAT v1.0 works, why it represents a technological leap, what its current strengths and limitations are, and how its impact is projected in the future of oceanic risk management, such as tsunamis.
Traditional Warning Systems: Strengths and Limitations
Since the 1940s, the countries most prone to tsunamis — such as Japan, Indonesia, or the United States — have developed warning systems based mainly on seismic data. The logic is simple: if an earthquake occurs under the ocean and exceeds a certain magnitude, it is likely to generate a tsunami. The system issues a preliminary tsunami alert while other factors are analyzed, such as the earthquake’s depth and the deformation of the seabed. Although these mechanisms have been useful, they present several serious problems:
- Frequent false alarms: According to studies cited by Cardiff University, approximately 75% of tsunami alerts that led to evacuations since the 1950s turned out to be false. This has significant economic and social consequences.
- Delays in confirmation: Traditional systems need time to verify the threat through buoys or other sensors, which limits the reaction capacity in the face of a tsunami.
- Incomplete assessments: Not all undersea earthquakes generate tsunamis, and those that do depend on multiple variables, such as the direction of tectonic plate movement.
One of the most impactful examples of the consequences of false alarms was the evacuation of Honolulu in 1986. Although the tsunami never arrived, the economic losses exceeded 30 million dollars. This type of event has caused a gradual loss of trust in warning systems, endangering the public’s response in real emergencies.
The 2004 Tsunami: A Catalyst for Innovation
The global turning point in awareness of the need to improve these systems came in 2004, when a 9.1-magnitude earthquake off the coast of Sumatra triggered one of the deadliest tsunamis in history. More than 230,000 people died in 14 countries, and it is estimated that the damages exceeded 15 billion dollars due to the tsunami.
After this catastrophe, a global network coordinated by the United Nations was created to improve tsunami detection and alerting. DART buoys (Deep-ocean Assessment and Reporting of Tsunamis) were installed, monitoring centers were strengthened, and international data exchange was promoted. However, even with these advances, most systems still rely on seismic data, with all the limitations already mentioned.
The Birth of GREAT v1.0: An Acoustic and Intelligent Approach
In this context of searching for more precise solutions to be prepared for tsunamis, GREAT v1.0 emerged — a completely different approach. Instead of focusing on vibrations in the Earth’s crust, the system analyzes acoustic waves generated by the deformation of the ocean floor to predict tsunamis.
When an undersea earthquake occurs, the energy released travels not only through the earth (as with traditional seismic waves), but also through the water in the form of pressure waves or T-waves. These waves travel faster than the tsunami waves themselves, so they can be detected before the disaster manifests at the surface.
Main Features of GREAT v1.0
- Speed: The system can complete the full analysis in a matter of seconds using a conventional computer to issue a tsunami alert.
- Accuracy: Thanks to artificial intelligence, it can distinguish between horizontal seismic movements (less dangerous) and vertical ones (more likely to generate tsunamis).
- Modularity: It can be integrated into existing monitoring systems and scaled to more powerful platforms.
- Reliability: It reduces the number of false tsunami alarms, one of the main historical problems in this field.
How Does GREAT v1.0 Work?
The tsunami detection system is structured in several stages:
- Acoustic data capture: Hydrophones, installed at great depths, are used to capture the waves generated by water compression during an earthquake. These signals travel at more than 1,400 meters per second.
- Mathematical processing: The software calculates the location of the epicenter, depth, and submarine topography to project potential wave trajectories.
- AI analysis: Using machine learning techniques, the system classifies the type of tectonic movement (horizontal or vertical) and estimates the earthquake’s magnitude based on acoustic patterns.
- Impact simulation: If a risk is detected, the wave amplitude and arrival time to various coastal regions are estimated.
The process can be carried out in a few seconds, which provides a greater response window than current systems, improving tsunami prediction and timely evacuation. Furthermore, by not relying exclusively on seismic or surface physical data, GREAT is capable of avoiding many false tsunami alarms.
Scientific Validation: Case Studies
To validate its effectiveness, GREAT v1.0 was tested with data from several historical tsunamis. The results were promising:
Sumatra (2004): The system analyzed acoustic signals detected over 3,000 km from the epicenter. It successfully identified the regions at highest tsunami risk, such as Sri Lanka and Madagascar.
Tohoku (2011): The system’s estimates matched the records from DART buoys, especially in areas close to the tsunami’s epicenter.
Alaska (2018): This was a false tsunami alarm triggered by a conventional system. GREAT confirmed in less than 30 seconds that there was no risk, which could have prevented unnecessary evacuations.
Tateyama (2009): Although the event was low in magnitude, the system provided consistent results at nearby stations.
These validations suggest that the system can operate accurately in both extreme events and more subtle situations, allowing timely warnings about the occurrence of tsunamis and the areas they primarily affect.
Current Implementation and Limitations
Since June 2024, GREAT v1.0 has been in an operational phase at the Tsunami Warning Center of the Instituto Português do Mar e da Atmosfera (IPMA). This real-world trial allows the system to be evaluated under operational conditions, including possible technical failures, transmission delays, and geographical variations.
Despite its advances, GREAT v1.0 faces some important limitations that affect its performance and reach. First, the current coverage of hydrophones is insufficient to ensure effective global tsunami surveillance. Only six stations, managed by the Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO), are available for this type of analysis, and of those, only four are actively used for acoustic tsunami detection.
Each of these stations can monitor an area with a radius of approximately 1,000 kilometers, leaving large regions without direct coverage. To achieve truly global surveillance capacity for tsunami prediction, it is estimated that at least 30 strategically distributed stations across the oceans would be necessary.
Another significant challenge is the limited database used to train the artificial intelligence model. Although GREAT has been developed and validated with significant historical events, the quantity and variety of data are not enough to perfect its accuracy and adaptability across different scenarios. Expanding this dataset with satellite information, more recent oceanic records, and various types of seismic events would significantly enhance the system’s predictive capability and response to real tsunami situations.
Finally, the current version of GREAT presents restrictions regarding the phenomena it can detect. Its design focuses mainly on tsunamis generated by underwater earthquakes, while other major causes — such as submarine landslides or volcanic eruptions — are not yet included in its algorithms. For the system to be more comprehensive and useful in a wider variety of situations, future updates are expected to incorporate the ability to detect and assess these non-seismic phenomena, thus expanding the scope of protection for coastal communities.
Global Impact and Future Perspectives
The implementation of a technology like GREAT v1.0 has the potential to significantly transform global standards for tsunami protection. One of its most notable benefits is the reduction of false tsunami alarms — a historical problem that has led to unnecessary evacuations and high costs for communities and governments. By decreasing the number of erroneous tsunami alerts, GREAT not only helps to avoid economic losses but also helps to minimize the emotional and social toll caused by frequent evacuations that turn out to be unnecessary.
This improvement in the accuracy of tsunami alerts has a direct impact on public trust. When alerts are more accurate, the population begins to take evacuation instructions more seriously, increasing their willingness to act quickly in the face of a real threat. Credibility in the tsunami warning system is essential for safety protocols to be followed effectively, and GREAT’s technology could be key to restoring and strengthening that trust — something that traditional systems have largely lost due to their high false-alarm rate.
Additionally, the technological accessibility of GREAT v1.0 is a crucial point for its global adoption. Unlike other systems that require complex infrastructures or supercomputers, this software can operate on conventional equipment, making it easier to implement in developing countries or regions with limited resources. This democratization of access to advanced early warning technologies could save countless lives in areas where current systems are not viable due to cost or lack of infrastructure.
Finally, GREAT does not aim to replace existing systems but to complement them. Its ability to integrate with traditional seismic sensors, oceanographic buoys, and satellite monitoring creates a more robust and multidimensional ocean surveillance ecosystem. This complementarity expands data sources and improves the speed and accuracy of tsunami detection and evaluation, enabling faster and more reliable emergency responses.
Toward Global Coverage
For GREAT to have a truly global impact, the acoustic sensor network must be expanded. International collaboration will be key at this point, especially with entities like the CTBTO, oceanographic organizations, and coastal governments. There is also the possibility of incorporating mobile sensors or warning systems attached to underwater drones.
A future development path for GREAT will include the detection of tsunamis generated by other non-seismic mechanisms. For example, submarine landslides or volcanic eruptions can generate large tsunamis, as occurred in Tonga in 2022. This will require new system modules as well as additional training for the AI algorithms with different types of acoustic events.
The development and implementation of technologies like GREAT v1.0 align with several of the United Nations’ Sustainable Development Goals (SDGs). In particular:
SDG 11: Sustainable Cities and Communities – By improving resilience to disasters, especially in densely populated coastal areas.
SDG 13: Climate Action – By providing technological tools for climate change adaptation, which can increase the frequency or impact of extreme ocean events.
SDG 17: Partnerships for the Goals – By promoting scientific and technological cooperation between countries and institutions.
Moreover, the Cardiff team’s approach also responds to the recommendations of the Sendai Framework for Disaster Risk Reduction 2015–2030, promoted by the United Nations. This framework emphasizes strengthening early warning and risk understanding as fundamental pillars to reduce global vulnerability.
The emergence of GREAT v1.0 represents a radical shift in how the global tsunami risk is understood and anticipated. Its innovative approach — combining ocean acoustic analysis with artificial intelligence — enables alerts that are faster, more accurate, and more reliable than traditional systems, so that people can be safer when tsunamis occur.
This advancement not only has the potential to save thousands of lives in the future but also contributes to a global culture of prevention, sustainability, and scientific cooperation. Although it still faces technical challenges — such as the limited sensor network and the need for more diverse training data — the path it has opened is clear: a smarter, faster, and more human warning system is possible.
In a world increasingly exposed to the extremes of climate change and uncontrolled coastal urbanization, technologies like GREAT v1.0 are not just scientific innovations — they are vital tools for building a safer future. If you want to learn more about the innovations brought by AI, such as tsunami detection, write to us at [email protected]. We have a tech team ready to help modernize your systems.
EN 
ES