The Science Dev portal is open for submissions! Sign up →

Predicting the unpredictable: New algorithm improves extreme weather forecasts

MIT researchers develop a method to refine climate models and predict the frequency of extreme weather events with greater accuracy.

Predicting the unpredictable: New algorithm improves extreme weather forecasts

Researchers at MIT have developed a groundbreaking AI algorithm that enhances the accuracy of extreme weather predictions.

For decades, scientists have relied on complex climate models to predict future weather patterns and assess the risk of extreme events. However, these models often lack the resolution necessary to provide accurate forecasts for specific locations, limiting their effectiveness in guiding local preparedness efforts.

A team of researchers at MIT, led by Professor Themistoklis Sapsis, has developed a novel approach that leverages the power of machine learning to enhance the accuracy of these models. Their method focuses on “correcting” the output of existing climate models by integrating data-driven insights and refining their predictions.

Traditional climate models simulate large-scale weather patterns, such as temperature, humidity, and precipitation, across a global grid. Due to computational limitations, these models often average out features over large areas, resulting in a loss of crucial details about local weather phenomena.

“These models still do not resolve very important processes like clouds or storms, which occur over smaller scales”, explains Professor Sapsis.

To address this challenge, the MIT team developed a machine learning algorithm that learns the complex relationships between various weather variables from historical data. This algorithm then uses these learned associations to refine the output of existing climate models, bringing them closer to real-world observations.

“What we’re doing is trying to correct dynamics, as in how an extreme weather feature, such as the windspeeds during a Hurricane Sandy event, will look like in the coarse model, versus in reality,” says Sapsis.

The researchers tested their approach on the Energy Exascale Earth System Model (E3SM), a widely used climate model. By applying their machine learning algorithm, they were able to significantly improve the model’s accuracy in predicting extreme weather events, such as tropical cyclones, in specific locations.

We now have a coarse model that can get you the right frequency of events, for the present climate,

Sapsis states. “Once we correct the dynamics, this is a relevant correction, even when you have a different average global temperature, and it can be used for understanding how forest fires, flooding events, and heat waves will look in a future climate.”

This breakthrough has significant implications for climate change adaptation and disaster preparedness. By providing more accurate and localized forecasts of extreme weather events, the MIT team’s algorithm empowers policymakers and communities to make informed decisions about infrastructure development, resource allocation, and emergency response plans.

As climate change continues to intensify extreme weather events worldwide, the need for reliable and precise forecasting tools becomes increasingly crucial. The MIT team’s innovative approach offers a promising path forward, enabling us to better understand, predict, and prepare for the challenges of a changing climate.

The link to the original research story can be accessed here.

Written by


Dr. Ravindra Shinde is the editor-in-chief and the founder of The Science Dev. He is also a research scientist at the University of Twente, the Netherlands. His research interests include computational physics, computational materials, quantum chemistry, and exascale computing. His mission is to disseminate cutting-edge research to the world through succinct and engaging cover stories.

Responses (0 )

Related posts

We use cookies to improve your experience and for analytics. You can accept all or reject optional cookies. Learn more in our Privacy Policy.