Improved estimation of diurnal variations in near-global PBLH through a hybrid WCT and transfer learning approach

Li, Yarong; Liu, Zeyang; He, Jianjun

doi:10.5194/amt-19-1059-2026

Articles | Volume 19, issue 3

https://doi.org/10.5194/amt-19-1059-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/amt-19-1059-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 19, issue 3

Research article

|

16 Feb 2026

Research article |

| 16 Feb 2026

Improved estimation of diurnal variations in near-global PBLH through a hybrid WCT and transfer learning approach

Yarong Li, Zeyang Liu, and Jianjun He

Supplement

https://doi.org/10.5194/amt-19-1059-2026-supplement

Data sets

Refined satellite measurements of diurnal variation in near-global PBLH through a ResNet-transfer learning approach Yarong Li https://doi.org/10.5281/zenodo.16907935

ERA5 hourly data on single levels from 1940 to present Hans Hersbach et al. https://doi.org/10.24381/cds.adbb2d47

Short summary

An attention-augmented ResNet and a transfer training are implemented to derive diurnal variations in near-global planetary boundary layer height. The transfer-trained model shows superior performances compared to conventional algorithms and non-transfer trained mode. The model predicted more reliable diurnal behaviors, with daily amplitude and peak timing approaching radiosonde results.