Research map / Image generation / Super-resolution
GeoFAR: Geography-Informed Frequency-Aware Super-Resolution for Climate Data
Chang Xu, Gencer Sumbul, Li Mi, Robin Zbinden, Devis Tuia
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 residual prediction (predict target minus input) (ablation) | ViT · CERRA T2m (x2) | RMSE: 0.380 -> 0.233 | Improved |
| add frequency-aware FCK (ablation) | ViT+Residual · CERRA T2m (x2) | RMSE: 0.233 -> 0.216 | Improved |
| add Geo-INR (location+elevation) over location-only 2D-INR (ablation) | ViT+Residual+FCK · CERRA T2m (x2) | RMSE: 0.198 -> 0.191 | Improved |
| replace FCK with DWT frequency decomposition (ablation) | ViT · CERRA T2m (x2) | RMSE: 0.380 -> 0.434 | Hurt |
| concatenate elevation map to input instead of Geo-INR (ablation) | ViT · CERRA T2m (x2) | RMSE: 0.380 -> 0.381 | No clear effect |
| add GeoFAR components (ablation) | ViT · ERA5 T2m (global x2) | RMSE: 1.125 -> 1.099 | Improved |
| add GeoFAR to U-Net | U-Net · CERRA T2m (x2) | RMSE: 0.272 -> 0.180 | Improved |
| add GeoFAR to ViT in joint multi-variable downscaling | ViT · CERRA surface pressure | RMSE: 277.719 -> 47.922 | Improved |
| add GeoFAR to ViT | ViT · ERA5-to-PRISM T2m | RMSE: 2.163 -> 1.745 | Improved |
| add GeoFAR to U-Net at x8 ratio | U-Net · CERRA T2m x8 | RMSE: 0.482 -> 0.393 | Improved |
| add GeoFAR to U-Net on Z500 | U-Net · ERA5 Z500 | RMSE: 49.060 -> 48.683 | Improved |
Closest papers by meaning
- No Location Left Behind: Measuring and Improving the Fairness of Implicit Representations for Earth Data (ICLR 2025 · similarity 0.88)
- Terra: A Multimodal Spatio-Temporal Dataset Spanning the Earth (NeurIPS 2024 · similarity 0.85)
- Earthfarsser: Versatile Spatio-Temporal Dynamical Systems Modeling in One Model (AAAI 2024 · similarity 0.84)
- GeoCore-9B: Towards Geo-Aware Generative Foundation Models in Earth Observation (NeurIPS 2026 · similarity 0.84)
- Scaling Storm-Resolving Atmospheric AI Simulation to the Entire Planet (NeurIPS 2026 · similarity 0.84)
- Precipitation Downscaling with Spatiotemporal Video Diffusion (NeurIPS 2024 · 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.
