How Artificial Intelligence (AI) Can Help Predict Earthquake
A scientific report said that scientists are currently undergoing a test with a newly made artificial neural network (ANN) to study the spatial relationships between more than 130,000 main earthquakes and their aftershocks.
The reporter said that, though it is impossible for the seismologists to predict where and when an earthquake will happen, they do know much about the aftershocks.
Susan Hough, a geophysicist of the U.S. Geological Survey in Pasadena, Calif., said that the spatial cluster and decay over time of the aftershocks is a long-time knowledge.
In 1992, lots of interest came that forced the necessary bodies to find ways of locating exactly where an aftershock might occur.
The series of history gathered regarding the earthquake and its aftershock has made scientists gather enough data which can be incorporated into the artificial intelligence technology.
The ability to know how and when an earthquake will occur is not easy, sometimes it may involve calculations, past records, and geographical local, and environmental conditions, in most cases it may be practically impossible to determine it.
A source said, seismologists care more about the aftershock because that impacts most on the people in the areas affected than the earthquake itself.
“But fault orientations in the subsurface can be as complicated as a three-dimensional crazy quilt, and stress can push on the faults from many different directions at once.
Imagine a book sitting on a table: Shear stress pushes the book sideways, and might cause it to slide to the left or right.
Normal stress pushes downward on the book, perpendicular to the table so that it wouldn’t budge. Such a thorny computational problem may be tailor-made for a neural network” according to Hough.
According to a test conducted by an Earthquake scientist Phoebe Devries of Harvard University and colleagues, including a Cambridge, Mass-based team from Google AI.
An artificial intelligence (AI) was fed with data on more than 130,000 mainshock-aftershock pairs which included the location and magnitude of the earthquake, and to also measure the changes in stress on the faults coming from it.
The AI was expected to use the data and determine how likely an aftershock will occur in a given place, and then possibly pinpoint the location of the aftershock using data from another 30,000 mainshock-aftershock pairs.
The test result revealed that the AI was able to predict the aftershock locations much better than the Coulomb failure criterion founded by the researchers.
There is believe that the AI was able to achieve it positive result by strongly correlating with other measures of stress change, such as the maximum amount of change in shear stress on a fault.
However, the study focused just on static stresses, which are permanent shifts in stress due to a quake.
However, aftershocks may also be triggered by a more ephemeral source of stress known as dynamic stress caused by the quake’s rumbling through the ground.
There are trending questions regarding the efficiency of this invention and how fast it can help to predict the aftershock of an earthquake.
According to Lucy Jones of Caltech based in Los Angeles “using a neural network to study the aftershock problem is a really nice, efficient approach”.
Improving the system to have a rapid response that will be the expectation of the scientists is another challenge that some belief could be achieved.
Work is ongoing to see the invention improved upon and put into useful purpose in tackling the issue of earthquakes and their aftershock problems.
Philip is a graduate of Mechanical engineering and an NDT inspector with vast practical knowledge in other engineering fields, and software.
He loves to write and share information relating to engineering and technology fields, science and environmental issues, and Technical posts. His posts are based on personal ideas, researched knowledge, and discovery, from engineering, science & investment fields, etc.
You can submit your article for free review and publication by using the “PUBLISH YOUR ARTICLE” page at the MENU Buttons.
If you love this post please share it with your friends using the social media buttons provided.