Reproduce “Observation Assimilation Transformer: Probabilistic Weather Analysis from Sparse Station Observations” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Christian Donner, Michele Cattaneo, Maxim Samarin, Francesco Zanetta, Icíar Lloréns Jover, Gabriela Aznar Siguan, Michele Volpi
OAT combines numerical weather prediction, terrain, and station observations to produce probabilistic local temperature and precipitation estimates. It uses geometry-aware sparse attention to pass information among observed stations and requested locations. On Europe, MAE/CRPS were 0.992/0.714 for temperature and 0.067/0.052 for precipitation, lower than the uninformed baseline; aggregate results do not guarantee improvement at every location or event.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.