Liu et al. (2026) Benchmarking Model Complexity for Short-Term Surrogate Forecasting of High-Resolution Urban WRF Outputs During an Extreme Heatwave in Chongqing

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

Identification

• Journal: Land
• Year: 2026
• Date: 2026-09-16
• Authors: Yanan Liu, Maoyuan Chai, Xie Runjie, Hong Li, Ruiqing Du, Bao‐Jie He
• DOI: 10.3390/land15091726

Research Groups

• Department of Atmospheric Science, University of Colorado Boulder
• Laboratory for Climate and Weather Modeling, National Center for Atmospheric Research (NCAR)

Short Summary

This study evaluates eight pointwise temporal surrogates for recursive 24 h prediction of urban weather variables in Chongqing, China. The best-performing models varied depending on the target variable.

Objective

• Investigate the performance of different machine learning algorithms for predicting urban weather variables using high-resolution WRF simulations.

Study Configuration

• Spatial Scale: Urban area of Chongqing, China (approximate dimensions not provided)
• Temporal Scale: 120 h simulation with a focus on recursive 24 h prediction

Methodology and Data

• Models used:
  • Persistence
  • Previous-day same-hour
  • Ridge
  • Random Forest
  • XGBoost
  • Multilayer perceptron (MLP)
  • Gated recurrent unit (GRU)
  • Long short-term memory (LSTM)
• Data sources: 120 h WRF-BEP/BEM-LCZ heatwave simulation over Chongqing

Main Results

• Random Forest performed best for T2 (1.081 ± 0.002 °C), LSTM for Q2 (0.969 ± 0.063 g/kg), and MLP for AC (0.505 ± 0.187 W/m²)
• Deterministic Ridge was best for W10 (1.095 m/s)
• Recurrent models did not improve W10 accuracy over Ridge

Contributions

This study provides a comprehensive evaluation of different machine learning algorithms for predicting urban weather variables, highlighting the importance of model choice depending on the target variable and balance between accuracy, spatial fidelity, and computational cost.

Funding

• This research was supported by [project/program name] (reference code not provided)

Citation

@article{Liu2026Benchmarking,
  author = {Liu, Yanan and Chai, Maoyuan and Runjie, Xie and Li, Hong and Du, Ruiqing and He, Bao‐Jie},
  title = {Benchmarking Model Complexity for Short-Term Surrogate Forecasting of High-Resolution Urban WRF Outputs During an Extreme Heatwave in Chongqing},
  journal = {Land},
  year = {2026},
  doi = {10.3390/land15091726},
  url = {https://doi.org/10.3390/land15091726}
}