A novel segmentation algorithm for the ARM user facility all-sky imagers using machine learning applications

Silber, Israel; Flynn, Donna M.; Comstock, Jennifer M.; Cromwell, Erol L.; Ermold, Brian D.

doi:10.5194/amt-19-5457-2026

Articles | Volume 19, issue 16

https://doi.org/10.5194/amt-19-5457-2026

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

https://doi.org/10.5194/amt-19-5457-2026

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

Articles | Volume 19, issue 16

Research article

|

25 Aug 2026

Research article |

| 25 Aug 2026

A novel segmentation algorithm for the ARM user facility all-sky imagers using machine learning applications

Israel Silber, Donna M. Flynn, Jennifer M. Comstock, Erol L. Cromwell, and Brian D. Ermold

Data sets

asiskycover (b1) I. Silber et al. https://doi.org/10.5439/1890629

All-sky imager cloud mask I. Silber et al. https://doi.org/10.5439/3005849

asiskycover (b0) Israel Silber et al. https://doi.org/10.5439/3005850

asiskyimage (a1) D. Flynn https://doi.org/10.5439/1890632

tsicldmask (a1) D. Flynn and V. Morris https://doi.org/10.5439/1992208

kazrmd (a1) Y.-C. Feng et al. https://doi.org/10.5439/1976091

ceil D. Zhang et al. https://doi.org/10.5439/1181954

Short summary

We describe ASISKYCOVER, a new machine learning algorithm for pixel segmentation of all-sky imager (ASI-16) data used by the Atmospheric Radiation Measurement (ARM) User Facility. ASISKYCOVER provides cloud cover and thickness estimates, detects artifacts, and reports uncertainties. Using one year of data from the ARM Southern Great Plains site and comparisons with other ARM datasets, we demonstrate its use and robustness, which will improve cloud cover analyses and data evaluation efforts.