Research map / Time series / Long-term forecasting
ClimateAR: Multi-Scale Autoregressive Generative Modeling for Climate Forecasting
Yue Yu, Weiqi Chen, Binqing Wu, Dongliang Cui, Wanyi Jiang, Zongjiang Shang, Bo Wu, Liang Sun, Ling Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use ClimateAR instead of the strongest baseline | ClimateAR · global forecast, lead 1-10 months | ACC: 37.56% | Improved |
| use ClimateAR instead of baselines for probabilistic forecasting | ClimateAR · global forecast | CRPS: 23.33% | Improved |
| add intra-scale modeling (ablation) | ClimateAR · ERA5 z500 (6-month avg) | ACC: 0.152 -> 0.331 | Improved |
| add intra-scale modeling (ablation) | ClimateAR · ERA5 t2m (6-month avg) | ACC: 0.091 -> 0.285 | Improved |
| add tokenizer alignment (ablation) | ClimateAR · ERA5 z500 (6-month avg) | ACC: 0.300 -> 0.331 | Improved |
| add tokenizer part design (ablation) | ClimateAR · ERA5 z500 (6-month avg) | ACC: 0.282 -> 0.331 | Improved |
| add prompt condition as prefix (ablation) | ClimateAR · ERA5 z500 (6-month avg) | ACC: 0.304 -> 0.331 | Improved |
| add noise to teacher-forcing input (ablation) | ClimateAR · ERA5 z500 (6-month avg) | ACC: 0.296 -> 0.331 | Improved |
| add tokenizer alignment (ablation) | ClimateAR · ERA5 t2m (6-month avg) | ACC: 0.280 -> 0.285 | Improved |
| add prompt condition (ablation) | ClimateAR · ERA5 pr (6-month avg) | ACC: 0.107 -> 0.121 | Improved |
| add part design (ablation) | ClimateAR · ERA5 pr (6-month avg) | ACC: 0.120 -> 0.121 | No clear effect |
| add noise to teacher-forcing input (ablation) | ClimateAR · ERA5 psl (6-month avg) | ACC: 0.213 -> 0.258 | Improved |
Closest papers by meaning
- MSAR: Next-Scale Autoregressive Forecasting for Time Series via Modular Multi-Scale Decoupling (NeurIPS 2026 · similarity 0.89)
- XLinear: A Lightweight and Accurate MLP-Based Model for Long-Term Time Series Forecasting with Exogenous Inputs (AAAI 2026 · similarity 0.86)
- VITA: Variational Pretraining of Transformers for Climate-Robust Crop Yield Forecasting (AAAI 2026 · similarity 0.86)
- Deep Switching Auto-Regressive Factorization: Application to Time Series Forecasting (AAAI 2021 · similarity 0.86)
- OmniCast: A Masked Latent Diffusion Model for Weather Forecasting Across Time Scales (NeurIPS 2025 · similarity 0.86)
- EcoMapper: Generative Modeling for Climate-Aware Satellite Imagery (ICML 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
