S. 1378 — TAME Extreme Weather and Wildfires Act

This section defines terms related to artificial intelligence (AI) weather forecasting, including "artificial intelligence weather model" (i.e., a weather model based primarily on AI technology to project future Earth system conditions using machine learning on weather datasets), and states purposes to improve weather forecast accuracy, strengthen analytic capacity and public-private partnerships for hazard response, and accelerate AI adoption. This section directs the Under Secretary of Commerce (for NOAA), not later than 4 years after enactment and in consultation with specified agencies and experts, to develop and curate comprehensive weather forecasting training datasets using relevant Earth system data, quality information, and metadata, building to the greatest extent practicable on existing federal datasets. This section further authorizes the Under Secretary to (1) develop and test a global AI weather model using NOAA data and experiment with regional and local AI weather models; (2) explore AI use to enhance dissemination of weather and wildfire risk information; (3) evaluate cost functions in AI model training to address inequities in observing system coverage; and (4) develop uncertainty quantification research for probabilistic forecasts; requires continued NOAA support for observational data collection, Earth system research, numerical weather modeling, data post-processing, and data assimilation; and requires a report to Congress not later than 2 years after enactment.