# Learning joint probabilistic weather forecasts from station observations alone

Preprint · Chaeyeon Yi, Yun Am Seo

- Archive id: arxiv:2610.09898
- Published: 2026-10-07
- arXiv: https://arxiv.org/abs/2610.09898
- Citations: 0
- Source URL: https://arxiv.org/abs/2610.09898
- PDF: https://arxiv.org/pdf/2610.09898
- Provenance: arxiv

## Abstract

Assessing compound weather risks requires forecasts representing dependence between variables. CLARA (Calibrated Advection-Routing Attention) learns joint Gaussian predictive distributions of five surface variables from station observations alone, without numerical weather prediction or reanalysis; the approximately 28,000-parameter model supports CPU training and prediction. Across six multi-year folds on 96 stations, its lead-mean energy score is 4.9% lower than that of a learned comparator with matched temporal inputs (4.7% with a similar parameter count) and 11-65% lower than those of statistical baselines. Holding marginal variances fixed, removing learned correlations worsens joint negative log-likelihood by 1.0-2.8 nats per station. A covariance-scale estimator, proved consistent under stated assumptions, improves short-lead calibration but over-corrects at long leads. Synthetic interventions show an attention-bias coefficient alone does not measure forecast influence. Retrained in ten regions on six continents, CLARA outperforms persistence in all 60 multi-year region-lead comparisons and a similarly sized learned model in 57 of 60.

_Abstract reproduced under CC0-1.0 (arXiv descriptive metadata)._

Topics: [AI Weather & Climate Modeling](https://space-kitty.com/topics/ai-weather-climate)

---
Space Kitty Archive · https://space-kitty.com · API: https://space-kitty.com/api/v1 · MCP: https://space-kitty.com/mcp · Premium: $0.20/request via x402 (Base Sepolia (testnet)). Metadata: arXiv (CC0), OpenAlex (CC0), Europe PMC (CC BY/CC0 works), ClinicalTrials.gov.