Engineers at the Massachusetts Institute of Technology have developed an AI tool called Extreme Event Aware, or η-learning, that can predict extreme weather events without relying on historical disaster records. According to Techub News, the system can generate maps of statistically plausible extreme weather events that have not previously occurred in a given region, while also estimating their duration, intensity, and geographic reach.
The method was described by researcher Kai Chang and Professor Themis Sapsis in a paper published in Nature Communications on August 20. The researchers said the algorithm combines point-based statistical data with spatial maps and learns statistical patterns from the relationship between low-resolution and high-resolution maps. That design allows it to generate spatial patterns for extreme weather that go beyond the range of the training data.
In testing, the tool was applied to precipitation data from the continental United States to generate maps for plausible once-in-a-century storms. Techub News said the system could simulate a reasonable storm scenario in which rainfall in New York City reaches 300 millimeters, compared with a historical record of 200 millimeters. The report said such outputs could serve as reference material for contingency planning in urban planning, insurance, and grid operations.
Engineers at the Massachusetts Institute of Technology have developed an AI tool called Extreme Event Aware, or η-learning, that can predict extreme weather events without using historical disaster data.
According to Techub News, the tool can generate maps of statistically plausible extreme weather events that have never been recorded in a specific region, and it can estimate the duration, intensity, and affected area of those events.
Method described in an August 20 Nature Communications paper
Researcher Kai Chang and Professor Themis Sapsis described the approach in a paper published in Nature Communications on August 20. The report said the algorithm combines point-based statistical data with spatial maps. It learns statistical regularities from the relationship between low-resolution and high-resolution maps, which allows it to produce spatial patterns for extreme weather beyond the range covered by the training data.
Testing focused on precipitation across the continental United States
In testing, the tool generated maps of plausible once-in-a-century storms using precipitation data from the continental United States. As one example, the report said it could simulate a reasonable storm scenario in which rainfall in New York City reaches 300 millimeters, versus a historical peak of 200 millimeters.
Techub News said the output could be used as a reference for contingency planning in urban planning, insurance, and power grid operations. The cited source was Nature Communications.
This article was originally published by Bit.Fan. For more cryptocurrency news and market insights, visit www.bit.fan. Disclaimer:
The market information, project data, and third-party content displayed on this platform are for industry information sharing only and do not constitute any form of investment advice or return commitment.
Cryptocurrency trading carries high risks. Users should fully assess their risk tolerance and make independent decisions. All profits, losses, and legal responsibilities are borne by the users themselves.
