45 Years Ago, My Country Predicted a 7.3 Magnitude Earthquake 20 Days in Advance

Using machine learning technologies such as deep learning and neural networks to analyze and study earthquake issues can reveal the value of some data that people cannot see, and achieve higher accuracy in the prediction of aftershocks and micro-earthquakes.

Earthquake early warning: a race against death to save yourself

 1. Earthquake early warning for the Yibin earthquake and the Hualien earthquake 

This Yibin earthquake early warning is not the first successful early warning case.

At 13:01 on April 18 this year, a sudden 6.7 magnitude earthquakeThis was the largest earthquake to hit Taiwan in the past 20 years, and the tremors were felt in many areas near the Taiwan Strait.

Before the earthquake was felt, New Taipei City received a warning message 20 seconds in advance.

The message received in advance in the video is the same as the 61-second advance warning in the Sichuan earthquake.Earthquake Early WarningIt is estimated that10 seconds of early warning can reduce casualties in 39%; If there had been an early warning when the Wenchuan earthquake occurred, the death toll might have been reduced 20,000 to 30,000.

 2. Earthquake early warning is of great significance, but it is not the same as prediction 

Earthquake early warning means:After an earthquake, useThe principle that electromagnetic waves travel faster than seismic waves.This will allow early warnings to be issued to areas not at the epicenter of the earthquake.

For individuals, receiving early warning information can enable them to take timely protection and escape measures.

For other institutions, hospitals can avoid mistakes in ongoing surgeries, trains and high-speed trains can brake in time to avoid derailment, and elevators can be controlled in time to avoid people being trapped...

Specifically, earthquake early warning isGain time difference.

When an earthquake occurs, the detection instrument deep underground detects the longitudinal wave and transmits it to the computer, and calculates the magnitude, intensity, epicenter and other general information immediately through the time it takes for the P wave to reach the surface and the speed difference between the P wave and the S wave. The relevant departments can send an alarm to the ground more than 10 seconds before the transverse wave reaches the ground.

Areas other than the epicenter can receive early warning 20 to 50 seconds before the earthquake waves arrive.

Earthquake prediction: How many days in advance should we prepare for an earthquake?

However, earthquake early warning is not the same as prediction. Although earthquake early warning is relatively mature, it is still difficult to predict.

Earthquake prediction meansBefore an earthquake occurs, the time, location and magnitude of the earthquake can be accurately calculated, and countermeasures can be arranged in advance.

  1. Forecast 20 days in advance,

  The Haicheng Earthquake that was warned on that day  

The only accurate prediction and successful forecast in history wasHaicheng Earthquake. For the 7.3 magnitude earthquake that occurred on February 4, 1975, due to the advance prediction of scientists and the issuance of short-term forecasts, the death toll in the region was 1,328, accounting for 1.3% of the total population. 0.02% .

January 16, 197520 days before earthquake, the National Earthquake Bureau proposed "The Liaodong Peninsula, especially southern Liaoning, may be pregnant with a large earthquake, with a high probability of magnitude 6."Short-term forecasts were made, and then earthquake forecasting was fully implemented.

At 10:30 a.m. on February 4, the provincial government issued a telephone broadcast to the entire province, announcing an impending earthquake warning. Factories stopped production, rallies were cancelled, and rescue teams were on standby. It was winter, and the government ordered "people to leave their homes and animals to leave their pens."

At 7:36 that evening, a strong earthquake of magnitude 7.3 occurred in Haicheng.

Based on population density and general earthquake casualty rates, if no defenses were taken, the Haicheng earthquake would have caused more than 50,000 deaths.

Although this prediction was later proven to be scientific, it still shocked the world at the time.For the first time in human history, an earthquake was successfully predicted.It even attracted foreign scientists to come and learn.

But just when scientists thought they had solved the problem of earthquake prediction, the Tangshan earthquake hit their theory to the bottom. There is no truly effective prediction method for complex earthquake problems, and the advance prediction of the Haicheng earthquake has never been repeated.

 2. Prediction success is often accidental   

Although the Haicheng earthquake forecast was a combination of confusion, empirical analysis, intuitive judgment, and good luck, it was the first time in practice thatNo major earthquake forecast has ever failedattempt.

Because the causes of earthquakes are complex and data is scarce, we still cannot accurately predict earthquakes to this day.It is common for events to happen that were predicted but did not occur, and for events to happen that were not predicted. And phenomena such as the rumored earthquake precursors have gradually been proven by many examples to not apply to all earthquakes.

However, it is gratifying that although the problem of earthquake prediction has not been solved, scientists have begun to try in recent years.Use machine learning techniques such as deep learning and neural networks to analyze and study earthquake problems.Good prediction results have been achieved in forecasts such as aftershocks and microearthquakes.

Aftershock prediction: the only way to avoid further harm

Earthquakes are often not isolated events. Aftershocks are frequent after major earthquakes, and large aftershocks can cause serious secondary injuries. A typical example is the May 12 Wenchuan earthquake, which saw tens of thousands of aftershocks after the main shock, posing a great threat to rescue work.

Therefore, the detection of aftershocks is also an important part of earthquake prediction.Harvard University is working with machine learning experts from Google to try to use deep learning to predict where aftershocks will occur.

Their research achieved a breakthrough, and the final results were published in Nature in August 2018.

Their database containsInformation about at least 199 major earthquakes that have occurred around the world.They eventually obtained the optimal aftershock location prediction model. Although the system still needs to be improved, it means that this is a step forward in this direction.

The research also had an unexpected benefit: it helped the team identify the physical quantities involved in earthquakes, which is very important for earthquake research.

Meade, one of the team members, once explained: "Traditional seismologists are more like pathologists, they study what happens after catastrophic earthquake events. We don't want to do that, we want to be more like epidemiologists,We want to understand the triggers and causes of these events."
 

It is expected that in the future, machine learning can uncover the mysteries behind earthquakes and reduce the damage they cause.

Science and Technology: Fighting against nature, the fun is endless

Machine learning techniques can be used to preserve simulated records of past earthquakes in large quantities. As the media that records this data gradually degrades, seismologists are racing to preserve this valuable information.

For a long time, some scholars believed that it was impossible to completely predict earthquakes. But from the current research results, perhaps predicting earthquakes is no longer "impossible".

Natural disasters are uncontrollable. We can only hope that we can use the power of modern technology to reduce the damage caused by disasters.