Maynard et al. (2026) Turbulent coherent structures in the atmospheric surface layer: Detection on Doppler lidar observations by supervised machine learning

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

Identification

• Journal: Springer Link (Chiba Institute of Technology)
• Year: 2026
• Date: 2026-04-09
• Authors: Paul V. Maynard, Elsa Dieudonné, Anton Sokolov, Hervé Delbarre
• DOI: 10.1051/epjconf/202636216001/pdf

Research Groups

Not explicitly mentioned in the provided text. The study was conducted during a campaign in Dunkirk, France.

Short Summary

This study observed and classified turbulent coherent structures, specifically streaks, in the atmospheric surface layer using Doppler lidar scans in an industrial coastal city, developing an automated classification method that successfully discriminates between organized and disorganized streaks.

Objective

• To observe and classify turbulent coherent structures (streaks) in the atmospheric surface layer using Doppler lidar, and to develop an automated method for their classification from lidar images.

Study Configuration

• Spatial Scale: Atmospheric surface layer, covering horizontal areas scanned by a Doppler lidar in an industrial urban environment.
• Temporal Scale: 13-month campaign.

Methodology and Data

• Models used: Supervised learning classification algorithms, with Quadratic Discriminant Analysis (QDA) showing the best performance.
• Data sources: Approximately 40,000 quasi-horizontal scans from a Doppler lidar.

Main Results

• Two main types of coherent structures, organized and disorganized streaks, were identified from Doppler lidar scans.
• An automated classification method was developed using image features (Gray-level Co-occurrence Matrices, texture parameters, and curve parameters) and supervised learning algorithms.
• The Quadratic Discriminant Analysis classifier achieved the best performance, with a cross-validation error of 5.2 % on the training set.
• Approximately 61% of the lidar scans were classified as coherent structures, comprising about 31% organized streaks and 30% disorganized streaks.

Contributions

• Direct observation and characterization of turbulent coherent structures (streaks) in the atmospheric surface layer of an industrial coastal city using Doppler lidar.
• Development and validation of an automated, robust classification method for identifying organized and disorganized streaks from a large dataset of lidar images.
• Quantification of the prevalence of different types of coherent structures in the observed environment.

Funding

Not explicitly mentioned in the provided text.

Citation

@article{Maynard2026Turbulent,
  author = {Maynard, Paul V. and Dieudonné, Elsa and Sokolov, Anton and Delbarre, Hervé},
  title = {Turbulent coherent structures in the atmospheric surface layer: Detection on Doppler lidar observations by supervised machine learning},
  journal = {Springer Link (Chiba Institute of Technology)},
  year = {2026},
  doi = {10.1051/epjconf/202636216001/pdf},
  url = {https://doi.org/10.1051/epjconf/202636216001/pdf}
}