Research map / Probabilistic and scientific modeling / Dynamical systems
CO2-Net: A Physics-Informed Spatio-Temporal Model for Global Surface CO2 Reconstruction
Hao Zheng, Yuting Zheng, Hanbo Huang, Chaofan Sun, Enhui Liao, Lin Liu, Yi Han, Hao Zhou, Shiyu Liang
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 spatial expert alongside temporal expert (simple fusion) (ablation) | CO2-Net (L) · CMIP6 | RMSE: 14.28 -> 13.01 | Improved |
| add layer-wise B-CAST connections between experts (ablation) | CO2-Net (L) · CMIP6 | RMSE: 13.01 -> 10.83 | Improved |
| use wind-based embedding instead of ClimODE embedding (ablation) | CO2-Net (L) · CMIP6 | RMSE: 27.06 -> 10.83 | Improved |
| add wind-based embedding versus none (ablation) | CO2-Net (L) · CMIP6 | RMSE: 23.40 -> 10.83 | Improved |
| add wind-based embedding to baseline (ablation) | ViT · CMIP6 | RMSE: 18.18 -> 21.51 | Hurt |
| add semi-supervised wind reconstruction loss (ablation) | CO2-Net (L) · CMIP6 | RMSE: 10.83 -> 5.31 | Improved |
| add semi-supervised wind loss to baselines (ablation) | ViT, SwinLSTM, CycleGAN · CMIP6 | RMSE: ViT 21.51 -> 19.78; SwinLSTM 27.80 -> 28.22; CycleGAN 28.43 -> 28.45 | Mixed |
| increase historic data points from 0 to 4 months (ablation) | CO2-Net (L) · CMIP6 | RMSE: 20.0 -> 5.31 | Improved |
| increase historic data points from 4 to 5 months (ablation) | CO2-Net (L) · CMIP6 | RMSE: 5.31 -> 8.03 | Hurt |
| use finer temporal resolution (3h vs 24h) (ablation) | CO2-Net (L) · CarbonTracker | RMSE: 5.57 -> 3.36 | Improved |
| keep wind variables UAS and VAS as inputs (ablation) | CO2-Net (L) · CMIP6 | RMSE: 44.7 -> 5.31 | Improved |
| keep static variables GPH and ORO as inputs (ablation) | CO2-Net (L) · CarbonTracker | RMSE: 3.43 -> 3.36 | No clear effect |
Closest papers by meaning
- CarbonSense: A Multimodal Dataset and Baseline for Carbon Flux Modelling (ICLR 2025 · similarity 0.86)
- ChaosBench: A Multi-Channel, Physics-Based Benchmark for Subseasonal-to-Seasonal Climate Prediction (NeurIPS 2024 · similarity 0.84)
- DiffDA: a Diffusion model for weather-scale Data Assimilation (ICML 2024 · similarity 0.84)
- Scaling Storm-Resolving Atmospheric AI Simulation to the Entire Planet (NeurIPS 2026 · similarity 0.83)
- Predict-Project-Renoise: Sampling Diffusion Models under Hard Constraints (NeurIPS 2026 · similarity 0.83)
- Causal Deciphering and Inpainting in Spatio-Temporal Dynamics via Diffusion Model (NeurIPS 2024 · similarity 0.83)
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.
