Chen et al. (2026) Structural Gating and Effect-aligned Lag-resolved Temporal Causal Discovery Framework with Application to Heat-Pollution Extremes

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: arXiv (Cornell University)
• Year: 2026
• Date: 2026-04-11
• Authors: Rui Chen, Jinsong Wu
• DOI: None

Research Groups

[Not specified in the provided text]

Short Summary

The study introduces SGED-TCD, a general framework for lag-resolved causal discovery in multivariate time series, and applies it to identify the drivers of compound heatwave-air-pollution extremes in China.

Objective

• To develop a robust and interpretable framework (SGED-TCD) for temporal causal discovery and evaluate its ability to recover physically consistent causal pathways in complex climate-environment systems.

Study Configuration

• Spatial Scale: Eastern and Northern China.
• Temporal Scale: Seasonal analysis (warm-season and cold-season).

Methodology and Data

• Models used: Structural Gating and Effect-aligned Discovery for Temporal Causal Discovery (SGED-TCD), incorporating structural gating, stability-oriented learning, and perturbation-effect alignment.
• Data sources: Large-scale climate indices, regional circulation variables, boundary-layer variables, and compound extreme indicators.

Main Results

• Regional/Seasonal Heterogeneity: Causal drivers of compound extremes differ significantly by region and season.
• Eastern China (Warm-season): Extremes are primarily linked to low-latitude oceanic variability, operating through circulation, radiation, and ventilation pathways.
• Northern China (Cold-season): Extremes are primarily governed by high-latitude circulation variability, associated with boundary-layer suppression and persistent stagnation.
• Framework Performance: SGED-TCD successfully reconstructed weighted causal networks with explicit dominant lags and relative causal importance.

Contributions

• Proposes a novel, general-purpose framework (SGED-TCD) for lag-resolved causal discovery that enhances the interpretability and robustness of inferred graphs in complex multivariate time series.
• Demonstrates the practical utility of the framework in resolving hierarchical causal pathways within the challenging domain of climate-environment interactions.

Funding

[Not specified in the provided text]

Citation

@article{Chen2026Structural,
  author = {Chen, Rui and Wu, Jinsong},
  title = {Structural Gating and Effect-aligned Lag-resolved Temporal Causal Discovery Framework with Application to Heat-Pollution Extremes},
  journal = {arXiv (Cornell University)},
  year = {2026},
  url = {https://openalex.org/W7154538853}
}