Luo et al. (2026) Gray‐Zone Simulation of the Dry Convective Boundary Layer Using Fourier Neural Operator

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

Identification

• Journal: Geophysical Research Letters
• Year: 2026
• Date: 2026-08-04
• Authors: Tengfei Luo, Zhijie Li, Jianchun Wang, Pak Wai Chan, Naigeng Wu, Xiaoming Shi
• DOI: 10.1029/2025gl120100

Research Groups

• Department of Atmospheric Science, University of California, Los Angeles (UCLA)
• Center for Climate Systems Research, Columbia University

Short Summary

This study develops three-dimensional Fourier neural operator (FNO) models to simulate the dry convective boundary layer at 800-m grid spacing, outperforming traditional gray-zone simulations in predicting flow statistics and instantaneous structures.

Objective

• Investigate the feasibility of using FNO models for simulating the dry convective boundary layer at high resolution

Study Configuration

• Spatial Scale: Local (dry convective boundary layer) with a 800-m grid spacing
• Temporal Scale: Not specified

Methodology and Data

• Models used: Fourier neural operator (FNO)
• Data sources: Filtered large-eddy simulation data of the CBL

Main Results

• FNO models outperform traditional gray-zone simulations in predicting flow statistics and instantaneous structures, including updrafts and downdrafts, vertical profiles, and energy spectra
• FNO models demonstrate generalization capability for surface heat fluxes within and outside the training range

Contributions

• The study presents a promising ML method (FNO) for fast and accurate weather prediction in the gray zone, filling the gap between global mesoscale forecasting and high-resolution local simulations

Funding

• Not specified

Citation

@article{Luo2026GrayZone,
  author = {Luo, Tengfei and Li, Zhijie and Wang, Jianchun and Chan, Pak Wai and Wu, Naigeng and Shi, Xiaoming},
  title = {Gray‐Zone Simulation of the Dry Convective Boundary Layer Using Fourier Neural Operator},
  journal = {Geophysical Research Letters},
  year = {2026},
  doi = {10.1029/2025gl120100},
  url = {https://doi.org/10.1029/2025gl120100}
}