Multistation AI Improves Regional Atmospheric Profiles
· newswise.com · popularity 21/100 · 0 similar stories today · AI & Data Science (Artificial Intelligence)
Read the original at newswise.com →
Summary
Accurate atmospheric temperature and humidity profiles are essential for weather prediction and climate simulation, yet sparse ground-based radiometer stations limit regional retrieval. This study introduces CTDformer (Convolutional Neural Network [CNN]-Transformer-DualBranchMLP), a geoscience-informed deep-learning framework that fuses multistation microwave brightness temperatures with spatial information. Across four Xi'an stations, it reduced temperature and humidity profile errors and improved convective-energy capture, supporting severe-weather monitoring and short-term forecasting.
Found in: Newswise. Rating basis: single outlet; outlet tier 2 (unlisted); 1 feeds; feed position 1.