The group of researchers from University of Cambridge in the UK and Boston University in the US has claimed that they have developed an Artificial Intelligence (AI) system to successfully predict the timing and magnitude of earthquakes in advance.
Learn & spot the Seismic Zones in India (PDF)
Research Paper
The study, which is published in the Geophysical Research Letters journal, could be the most advanced technique till date for detecting earthquakes.
The study describes that tiny sounds from rocks can give a clue to scientists to predict earthquakes one week in advance.
According to researchers, this AI breakthrough will help to be prepared for natural disasters and potentially save millions of lives. They have identified a hidden signal leading up to earthquakes, and used this fingerprint to train a machine learning algorithm to predict future earthquakes.
Before Earthquakes?
Prof. Colin Humphreys , a materials scientist at Cambridge University, said: What happens before an earthquake is that rocks emit noise because one grain of rock is rubbing against another grain of rock. It’s a little like a squeaky door.
He further said:
There’s always a little rumbling in the Earth, but the team found they could train a computer to recognize changes in that rumbling that occur when an earthquake is imminent—about a week away.
This study pave the way to predict earthquakes through Artificial Intelligence (AI) mechanism.
READ MORE US scientists develop new tool to improve Indian monsoon forecast
AI Approach
The researchers studied the interactions among earthquakes, precursor quakes and faults to develop a method which can help predict earthquakes well in time.
For this, a lab-based system has been used by researchers that mimics real earthquakes.
In it, they used machine learning techniques to analyse acoustic signals which comes from the disturbance as it moved and looked for patterns. And then, record the seismic signals and sounds that are emitted.
Next What?
Researchers used steel blocks to closely reproduce the physical forces involved in a real earthquake, and then record the seismic signals and sounds that are emitted.
After that, machine learning was used to find the relationship between the acoustic signal coming from the fault and the degree of its failure.
The machine learning algorithm was able to identify a particular pattern in the sound, previously thought to be nothing more than noise, which occurs long before an earthquake.
Sound Pattern
Researchers said: The characteristics of this sound pattern can be used to give a precise estimate of the stress on the fault and to estimate the time remaining before failure, which gets more and more precise as failure approaches.
Prof. Humphreys said:
This is the first time that machine learning has been used to analyse acoustic data to predict when an earthquake will occur, long before it does so that plenty of warning time can be given and lives are saved.
