Decoding the weather: How GRAF & New Computer Modeling can Change Everything
summary
GRAF is a global high‑resolution forecasting system that cuts grid size from 12 km to 3 km grid, runs 15‑hour forecasts and assimilates hourly instead of every six hours, ingesting satellites, radar, aircraft, phones and millions of observations. It uses roughly 300 GPUs on 90 IBM servers to run 5–6× faster than traditional models. That finer detail catches thunderstorm updrafts, tropical cyclone formation points and regional risks for places like Delhi, Bangladesh and Kenya, enabling better crop decisions, fuel routing and flight reroutes.
AI-written summary · about
chapters (7)
- 0:00Introduction to IBM weather
- 2:33The impact of weather
- 5:52Weather as big business
- 10:30Leveraging weather data
- 14:13IBM GRAF forecasting system
- 19:08Call for code initiative
- 26:26Conclusion and wrap-up
related talks
47mPowering global-scale predictive intelligence using HPC on
19mLightning Talk: Real-Time Weather Forecasting in AgS...
28mKubernetes Meets Climate Science: Building Large-scale Feature Detection From Climate Data Records
49mMeteorology, Climate and Python: desperately trying to forget technical details
21mGPU-accelerated Hydrology Algorithms for On-prem Computation: Flow accumulation,(…)
01h 02mAccelerate data science and analytics with GPUs
09mExponential Space Technology: Getting Weather Right
31mKeeping up with the renewable energy transition