Luo et al. (2026) Enhancing the Usability of CALIPSO Low-Confidence Cloud Products Using a Multilayer Perceptron-Based Data Refinement Framework

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

Identification

• Journal: Atmosphere
• Year: 2026
• Date: 2026-04-18
• Authors: Xiaolu Luo, Wenkai Song, Shiqi Yan, Miao Zhang, Ge Han
• DOI: 10.3390/atmos17040413

Research Groups

Not specified in the provided text.

Short Summary

This study develops a multilayer perceptron (MLP) framework to refine low-confidence and "unknown" cloud-type labels in the CALIPSO V4.10 5 km product, enhancing the dataset's completeness for climatological and Earth system applications.

Objective

• To improve the practical usability and completeness of CALIPSO cloud-type observations by replacing "unknown" and low-confidence labels with probabilistic assignments of ice clouds, water clouds, and oriented ice crystals.

Study Configuration

• Spatial Scale: Global
• Temporal Scale: 2006–2021 (summer daytime)

Methodology and Data

• Models used: Multilayer Perceptron (MLP)
• Data sources: CALIPSO V4.10 5 km cloud-layer product (utilizing 11 variables including cloud optical depth, cloud thickness, depolarization ratio, and color ratio).

Main Results

• The MLP model demonstrated a high level of agreement (~94.99%) with operational CALIPSO classifications when tested on withheld high-confidence samples.
• The framework successfully generated probabilistic assignments for the low-confidence subset (~5% of the dataset), resulting in vertical structural characteristics that are physically coherent with established cloud thermodynamic regimes.

Contributions

• Provides a statistically robust and physically consistent approach to reduce the fraction of "unknown" cloud labels in CALIPSO products, thereby increasing the utility of the data for large-scale scientific modeling and climatological analyses.

Funding

Not specified in the provided text.

Citation

@article{Luo2026Enhancing,
  author = {Luo, Xiaolu and Song, Wenkai and Yan, Shiqi and Zhang, Miao and Han, Ge},
  title = {Enhancing the Usability of CALIPSO Low-Confidence Cloud Products Using a Multilayer Perceptron-Based Data Refinement Framework},
  journal = {Atmosphere},
  year = {2026},
  doi = {10.3390/atmos17040413},
  url = {https://doi.org/10.3390/atmos17040413}
}