The result of natural calamities like earthquakes, underwater landslides, volcanoes and other causes, Tsunamis are among the deadliest, with the capability of widespread destruction across tens of thousands of miles. For the past 100 years, 58 Tsunamis were reported, which wiped out 260,000 lives or an average of 4,600 per disaster, as per the data of the UN in November 2022. For example, the Tsunami that struck the Indian Ocean in 2004 resulted in around 230,000 deaths and casualties as the mammoth walls of displaced water struck Indonesia, India, Sri Lanka, Thailand, Somalia, Myanmar, Maldives, Tanzania, Malaysia, Bangladesh, Seychelles, Yemen, South Africa and Kenya. They can be incredibly destructive as they cause huge loss of life and will eradicate infrastructure.

Need for early warning systems

Here, early warnings are challenging and have limitations as these kinds of disasters depend on the underwater earthquake’s features that predominantly trigger them. Earthquakes can result in Tsunamis if a large amount of water gets displaced. The warning systems are useful yet has imitations. The Tsunamis form and travel due to a compound interplay of factors. This is why few destroy the whole region while others might only raise the water level a few feet when they reach the shore.

Hence determining the type and risk of earthquakes is significant. According to the Physics of Fluids by American Institute of Physics Publishing, the researchers of the University of California, Los Angeles and Cardiff University in the UK have developed an early warning system that combines state-of-art acoustic technology and Artificial Intelligence for an immediate classification of earthquakes and observes its potential Tsunami risk.

Co-author Bernabe Gomez says, “Tectonic events with a strong vertical slip element are more likely to raise or lower the water column compared to horizontal slip elements”. “Thus, knowing the slip type at the early stages of the assessment can reduce false alarms and enhance the reliability of the warning systems through independent cross-validation”, he added.

Hydrophones to measure the acoustic radiation

Relying on seismographs and buoys attached by bottom-pressure sensors is not always the right method, as all earthquakes do not result in Tsunamis, and the buoys can detect Tsunamis only when one passes them, leaving less time to react and evacuate. In such cases, time is of the essence. So, the researchers put forwards the technique of measuring acoustic radiation- the sound produced by the earthquake that carries information about the tectonic events and travels considerably faster than the Tsunami waves. The underwater microphones, named hydrophones, record the acoustic waves and monitor the tectonic activity in real time.

Co-author Usama Kadri says, “Acoustic radiation travels through the water column much faster than Tsunami waves. It carries information about the originating source, and its pressure field can be recorded at distant locations, even thousands of kilometres away from the sources. The derivation of analytical solutions for the pressure field is a key factor in the real-time analysis”.

(The four different past earthquake scenarios associated with Tsunami events. The red and yellow rectangles denote the projected earthquake dimensions, locations and orientations retrieved by the proposed model for acoustic radiation. a) Matayai, Samoa: Sept 29, 2009; b) Bonin Islands, Japan region: Dec 21, 2010; c) Kushiro, Japan: March 14, 2012; d) off the east coast of Honshu, Japan: Oct 25, 2013. Credit: Bernabe Gomez and Usama Kadri)

How AI helps in classifying the magnitude of waves

The mathematical model the Cardiff team developed triangulates the earthquake data from the hydrophones, and then the AI algorithms classify its slip type and magnitude. The important properties such as the length, width, duration, the uplift speed are calculated to dictate the estimated size of the Tsunami.

The authors tested the model with data available in the hydrophone, detected it almost immediately, and successfully defined the earthquake parameters with low computational demand. They are reportedly enhancing the model by factoring in more data to increase the accuracy of Tsunami characterisation. This project of predicting the potential Tsunami risk is part of a larger project to improve the early warning system.

• Photo by Silas Baisch on Unsplash