AI Could Help Predict Natural Disasters, Here is HowEnvironment
Artificial intelligence (AI) and machine learning (ML) may soon help to alleviate major global environmental crises, as researchers and scientists are exploring how powerful algorithms and machines with "learning" abilities could tackle some of the greatest challenges humanity has ever faced.
Whether fighting climate change, helping biodiversity, or cleaning oceans and air, AI could be applied to assist humans in better understanding what is happening on the planet.
Recently, disaster resilience has been an increasingly popular area where AI technology could be harnessed to know how, when, and where natural disasters occur.
Through various methods, AI systems could help humans monitor levels, mitigate risks, reduce impacts, and save lives.
AI-based systems can scan images obtained from satellites orbiting the earth, and look for any changes that could help to predict the occurrence of a natural disaster like landslides, floods, volcanic eruptions, and tsunamis.
The Japanese government announced plans to launch a new satellite-based system for disaster prediction in 2020. The system would "accurately forecast the locations of disasters such as floods and landslides, thus enabling local governments to issue an early warning of disasters, including timely evacuation orders to residents," according to Japan News.
In India, Google is developing an AI platform that could warn users about impending floods through Google Maps and Google Search. Google's engineering Vice-President Yossi Matias explained that the AI-based system would make the predictions through a combination of machine learning, rainfall records, and flood simulations.
A combination of satellite images and machine learning could also offer a more effective technique to predict cyclones and hurricanes.
For example, when Hurricane Harvey hit southern Texas in 2017, NASA and Development Seed tracked the hurricane's intensity and path using satellite images and machine learning, which "proved to be six times better than the usual techniques, as the hurricane can be tracked every hour instead of every six hours with the traditional methods," according to this Forbes article.
Volcanologists are also combining satellite measurements of ground movements with machine learning to monitor and predict volcanic eruptions more accurately.
Early Detection of Earthquakes and Aftershocks
Deep learning systems could help analyze seismic data, which contains the magnitude and patterns of previous earthquakes. The analysis could enable the early detection of future earthquakes and aftershocks.
Scientists from Google and Harvard are working on an AI system that can predict the aftershocks of an earthquake. They have studied more than 131,000 earthquakes and aftershocks to build a neural network, and then tested the network on 30,000 events.
At the end of the experience, the AI-system predicted the aftershock locations more precisely than traditional methods.
Another innovative method to predict earthquakes came out of Oklahoma, where AI was used to amplify the sensitivity of the state's earthquake detectors to identify them sooner and prepare for them. In this method, convolutional neural networks are trained to ignore the regular geological rumblings, and identify the early signs of earthquakes and cause for alarm.
AI and ML systems could potentially save millions of lives when it comes to managing, early-detecting, and predicting disasters. However, as for any innovation that comes out of this emerging technology, the AI-based systems still need to go through multiple reliability tests to prove their readiness for holistic and real-life implementations.