Microsoft has a new machine learning tool to improve weather forecasts
Old-school physics-based weather forecasting and even new AI models typically become less reliable beyond two weeks, a problem Microsoft claims to have solved using machine learning. Researchers with the company say they've found a new, cost-effective way to make more accurate weather forecasts two to six weeks in advance - allowing farmers and emergency responders […]
New open-source software developed by Microsoft could make AI-enhanced weather forecasting more affordable and accurate. Traditional physics-based weather forecasting and AI models often lose reliability beyond two weeks, but Microsoft's machine learning tool aims to provide more precise predictions two to six weeks in advance. This breakthrough could help farmers and emergency responders plan ahead with greater certainty.
The software, known as Probabilistic Bias Correction (PBC), corrects bias in existing models by identifying and making corrections. PBC can run on CPUs, making it more accessible for resource-strapped weather agencies. In tests, PBC improved the accuracy of a leading physics-based weather model by 98 percent for precipitation and 92 percent for temperature.
The researchers, including Lester Mackey from Microsoft Research New England and Genevieve Flaspohler, believe PBC can be particularly beneficial for agencies in economically developing countries with limited resources. Flaspohler is now working on implementing AI for early warnings with the Kenya Meteorological Service Authority.
While physics-based and AI-based models struggle to predict weather beyond two weeks, PBC offers a more cost-effective solution for extended weather forecasting.
Written by urgent.news from The Verge's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
