Research map / Large language models / AI and society
ClimateLearn: Benchmarking Machine Learning for Weather and Climate Modeling
Tung Nguyen, Jason Kyle Jewik, Hritik Bansal, Prakhar Sharma, Aditya Grover
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 deep learning models instead of nearest or bilinear interpolation | ERA5 to ERA5 and ERA5 to PRISM downscaling · ERA5 to ERA5 / ERA5 to PRISM | RMSE: 269.67 -> 43.84 (Z500); 2.91 -> 1.57 (daily max T2m) | Improved |
| Use UNet rather than CNN-LSTM | ClimateBench models · Diurnal temperature range | Total: 16.778 -> 11.030 | Improved |
| Train and test on the same dataset rather than cross-dataset | ResNet forecasting · ERA5 and CMIP6, 3-day and 5-day forecasts | RMSE | Improved |
| Evaluate deep models on the extreme split rather than the normal split | ResNet, U-Net, and ViT forecasting · Extreme-ERA5 T2m | RMSE: 1.50 -> 1.33 (ResNet, 3 days); 2.20 -> 1.86 (ResNet, 5 days) | Improved |
| Evaluate climatology on the extreme split rather than the normal split | Climatology forecasting · Extreme-ERA5 T2m | RMSE: 5.87 -> 6.51 (6 hours); 5.87 -> 6.58 (3 days) | Hurt |
| Use continuous forecasting instead of direct forecasting | ResNet forecasting · ERA5, 6-hour and 1-day versus 3-day and 5-day lead times | RMSE | Mixed |
| Evaluate continuous forecasting at an unseen 10-day lead time | ResNet continuous forecasting · ERA5, 10-day forecast | RMSE | Hurt |
| Use iterative forecasting instead of direct forecasting | ResNet forecasting · ERA5, longer forecast horizons | RMSE | Hurt |
| Train on CMIP6 including additional years 1850–1978 | Forecasting models · ERA5 evaluation | RMSE | Improved |
Closest papers by meaning
- OmniCast: A Masked Latent Diffusion Model for Weather Forecasting Across Time Scales (NeurIPS 2025 · similarity 0.88)
- DroughtSet: Understanding Drought Through Spatial-Temporal Learning (AAAI 2025 · similarity 0.88)
- Quantile-Regression-Ensemble: A Deep Learning Algorithm for Downscaling Extreme Precipitation (AAAI 2024 · similarity 0.88)
- ChaosBench: A Multi-Channel, Physics-Based Benchmark for Subseasonal-to-Seasonal Climate Prediction (NeurIPS 2024 · similarity 0.86)
- Probablistic Emulation of a Global Climate Model with Spherical DYffusion (NeurIPS 2024 · similarity 0.85)
- AtmosSci-Bench: Evaluating the Recent Advance of Large Language Model for Atmospheric Science (NeurIPS 2025 · similarity 0.85)
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.
