Data-driven global weather predictions at high resolutions

Society has benefited enormously from the continuous advancement in numerical weather prediction that occurred over many decades driven by a combination of outstanding scientific, computational and technological breakthroughs. Here, we demonstrate data-driven methods are now positioned to contribute next wave major advances atmospheric science. We show models can predict important meteorological quantities interest society such as global high resolution precipitation fields (0.25°) deliver accurate forecasts future state atmosphere without prior knowledge laws physics chemistry. also how these be scaled run on supercomputers with up 1024 modern graphics processing units beyond resulting rapid training models, thus supporting cycle research innovation. Taken together, two results illustrate significant potential advance science operational forecasting.