Researchers have integrated Global Navigation Satellite System (GNSS) derived Zenith Wet Delay (ZWD) data into a weather foundation model called Aurora. This integration aims to improve precipitation forecasting, particularly for severe events, by addressing the underestimation of precipitation in current machine learning weather models (MLWM). The study found that the extended Aurora model learned ZWD effectively and showed systematic improvements in six-hour accumulated precipitation forecasts, with an 8.8% increase in Equitable Threat Score at the 99th percentile for severe precipitation. AI
IMPACT This research demonstrates how integrating novel data sources like GNSS-derived ZWD can enhance the accuracy of AI-driven weather prediction models, particularly for extreme events.
RANK_REASON The cluster contains an academic paper detailing a new method for improving weather forecasting models.
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
