Research map / Time series / Long-term forecasting
Benchmarking Physics-Informed Time-Series Models for Operational Global Station Weather Forecasting
Tao Han, Zhibin Wen, Zhenghao Chen, Dazhao Du, Song Guo, LEI BAI
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 |
|---|---|---|---|
| add dynamic core and physics-informed losses (PhysicsFormer vs Pyraformer, best TSF baseline) | PhysicsFormer · WEATHER-5K, Temperature | MAE: 2.49 -> 2.34 | Improved |
| PhysicsFormer vs ECMWF-HRES on temperature MAE | PhysicsFormer · WEATHER-5K, Temperature | MAE: 1.94 -> 2.34 | Hurt |
| PhysicsFormer vs ECMWF-HRES on sea level pressure MAE | PhysicsFormer · WEATHER-5K, Sea Level Pressure | MAE: 1.29 -> 3.53 | Hurt |
| PhysicsFormer vs ECMWF-HRES on wind speed MAE | PhysicsFormer · WEATHER-5K, Wind Speed | MAE: 1.56 -> 1.44 | Improved |
| add dynamic core and physics losses vs best TSF on 99.5th temperature extremes | PhysicsFormer · WEATHER-5K SEDI, Temperature 99.5th | SEDI (%): 26.2 -> 30.6 | Improved |
| PhysicsFormer vs ECMWF-HRES on extremes | PhysicsFormer · WEATHER-5K SEDI, Sea Level Pressure 99.5th | SEDI (%): 77.5 -> 62.6 | Hurt |
| increase parameter count (Timer, Corrformer vs Pyraformer) (ablation) | TSF models · WEATHER-5K 72h average | accuracy | No clear effect |
| use pretrained large time-series model Chronos for extreme events vs typical TSF models | Chronos · WEATHER-5K SEDI, Temperature 99.5th | SEDI (%): 14.1 -> 26.2 | Improved |
| longer forecast lead time beyond three days (ablation) | TSF models · WEATHER-5K | error | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- iTransformer — iTransformer: Inverted Transformers Are Effective for Time Series Forecasting (ICLR 2024) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- ChaosBench: A Multi-Channel, Physics-Based Benchmark for Subseasonal-to-Seasonal Climate Prediction (NeurIPS 2024 · similarity 0.87)
- Scaling transformer neural networks for skillful and reliable medium-range weather forecasting (NeurIPS 2024 · similarity 0.86)
- Generalizing Weather Forecast to Fine-grained Temporal Scales via Physics-AI Hybrid Modeling (NeurIPS 2024 · similarity 0.86)
- Tyche: One Step Flow for Efficient Probabilistic Weather Forecasting (NeurIPS 2026 · similarity 0.85)
- Fixing the Double Penalty in Data-Driven Weather Forecasting Through a Modified Spherical Harmonic Loss Function (ICML 2025 · similarity 0.85)
- WeatherSyn: An Instruction Tuning MLLM For Weather Forecasting Report Generation (ICML 2026 · similarity 0.84)
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.
