New AI model PEAR uses equal-area grid for improved weather forecasting

Researchers have developed PEAR, a new transformer-based weather forecasting model that operates directly on the Hierarchical Equal Area iso-Latitude Pixelization (HEALPix) grid. This approach addresses limitations of previous models that used equiangular grids, which have finer resolution at the poles. PEAR demonstrates superior performance compared to models on equiangular grids and other baselines, without increased computational cost. The model also shows promise in climate model emulation. AI

IMPACT This research could lead to more accurate and efficient AI-driven weather forecasting models by utilizing a more suitable data discretization method.

RANK_REASON This is a research paper describing a new AI model for weather forecasting. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• arXiv
• Graphcast
• HEALPix
• Pangu Weather

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →