{"id":"arxiv:1911.00630","kind":"preprint","title":"Predicting Weather Uncertainty with Deep Convnets","authors":["Peter Grönquist","Tal Ben-Nun","Nikoli Dryden","Peter Dueben","Luca Lavarini","Shigang Li","Torsten Hoefler"],"author_count":7,"year":2019,"published":"2019-11-02","venue":null,"cited_by":null,"doi":null,"arxiv_id":"1911.00630","nct_id":null,"source_url":"https://arxiv.org/abs/1911.00630","archive_url":"https://space-kitty.com/papers/arxiv:1911.00630","updated":"2019-12-04","pmid":null,"pmcid":null,"openalex_id":null,"pdf_url":"https://arxiv.org/pdf/1911.00630","abstract":"Modern weather forecast models perform uncertainty quantification using ensemble prediction systems, which collect nonparametric statistics based on multiple perturbed simulations. To provide accurate estimation, dozens of such computationally intensive simulations must be run. We show that deep neural networks can be used on a small set of numerical weather simulations to estimate the spread of a weather forecast, significantly reducing computational cost. To train the system, we both modify the 3D U-Net architecture and explore models that incorporate temporal data. Our models serve as a starting point to improve uncertainty quantification in current real-time weather forecasting systems, which is vital for predicting extreme events.","abstract_license":"CC0-1.0 (arXiv descriptive metadata)","license":null,"is_open_access":false,"retracted":false,"peer_reviewed":false,"categories":["cs.LG","physics.ao-ph","stat.ML"],"details":null,"provenance":"arxiv","attribution":"Abstract: arXiv descriptive metadata (CC0). Thank you to arXiv for use of its open access interoperability. This service was not reviewed or approved by, nor does it necessarily express or reflect the policies or opinions of, arXiv.","topics":[],"cited_in_dossiers":[]}