Meteorologists have long struggled to accurately predict weather due to the challenge of capturing water vapor. However, a team from the Wrocław University of Environmental and Life Sciences has developed a deep learning technique using AI to create high-resolution 3D maps of humidity, reducing errors by 62% in Poland and 52% in California.

The technique, powered by NVIDIA GPUs, upscales low-resolution data from navigation satellites into detailed humidity maps. The AI, trained on global weather data, uses explainable AI tools like Grad-CAM and SHAP to show where decisions are made. This transparency builds trust in AI-driven weather forecasting, potentially providing crucial lead time for communities at risk of sudden severe weather.

Lead author Saeid Haji-Aghajany emphasizes the importance of high-resolution humidity data in improving weather forecasts. By incorporating this data into weather models, forecasters can better predict and warn of dangerous weather events like flash floods and downpours, ultimately saving lives and property. The focus on humidity as a key element in weather forecasting highlights its significance in accurate predictions.

Read more at NVIDIA: It’s the Humidity: How International Researchers in Poland, Deep Learning and NVIDIA GPUs Could Change the Forecast