Machine Learning And Extreme Weather Forecasting
AI and machine learning can do so much for the good of humanity. Scientists are using it to improve the taste of our beer and help people find their next best friend. And, more seriously, all of this cool stuff.
Scientists at MIT have developed an algorithm that can help predict weather events for specific locations over the next few decades. Traditionally, scientists would need to use a global climate model that can make predictions forward in time. For example, these models can predict the weather for the northeast US, but not specifically for New York.
Policymakers would then combine this global prediction model with a finer-resolution model to begin pinpointing the likelihood of extreme weather events in a specific location. However, this data is only as accurate as the future prediction from the global climate model.
MIT researchers have developed a method to “correct” the global climate model predictions by combining machine learning with dynamical systems theory. The machine learning model “nudges” the simulation into more realistic patterns over large scales and, when paired with smaller-scale models, can more accurately predict specific weather events for particular locations.
Themistoklis Sapsis, the William I. Koch Professor and director of the Center for Ocean Engineering in MIT’s Department of Mechanical Engineering, says scientists can use the new correction scheme with any global climate model.
“Climate change will have an effect on every aspect of human life, and every type of life on the planet, from biodiversity to food security to the economy,” Sapsis says. “If we have capabilities to know accurately how extreme weather will change, especially over specific locations, it can make a lot of difference in terms of preparation and doing the right engineering to come up with solutions. This is the method that can open the way to do that.”
The team published their results in the Journal of Advances in Modeling Earth Systems.
