A critical part of any disaster response is knowing where to direct the survivors for the greatest degree of safety. With the Fukushima Daiichi disaster – and the unreliable tools at the time – still looming in people’s recent memories, a team at the University of Tokyo Institute of Industrial Science has developed an AI capable of assisting in responses to nuclear disasters by predicting the dispersion patterns of radioactive material. Using machine learning applied to weather forecasts, the AI can draw projections for windows of over 30 hours.
“Our new tool was first trained using years of weather-related data to predict where radioactivity would be distributed if it were released from a particular point. In subsequent testing, it could predict the direction of dispersion with at least 85% accuracy, with this rising to 95 per cent in winter when there are more predictable weather patterns,” said lead author Takao Yoshikane in a statement.
“The accuracy of this approach did not decrease when predicting over 30 hours into the future is extremely important in disaster scenarios. This gives authorities time to arrange evacuation plans in the most badly affected areas, and to issue guidance to people in specific areas about avoiding eating fresh produce and taking potassium iodide, which can limit the absorption of ingested radioactive isotopes by the body,” said Yoshikane.
The report can be found in full at Nature.com.
Sources:
https://www.nature.com/articles/s41598-018-27955-4
