Articles | Volume 14, issue 11
https://doi.org/10.5194/amt-14-7079-2021
https://doi.org/10.5194/amt-14-7079-2021
Research article
 | 
11 Nov 2021
Research article |  | 11 Nov 2021

The University of Washington Ice–Liquid Discriminator (UWILD) improves single-particle phase classifications of hydrometeors within Southern Ocean clouds using machine learning

Rachel Atlas, Johannes Mohrmann, Joseph Finlon, Jeremy Lu, Ian Hsiao, Robert Wood, and Minghui Diao

Data sets

University of Washington Ice-Liquid Discriminator single particle phase classifications and 1 Hz particle size distributions/heterogeneity estimate Johannes Mohrmann, Joseph A. Finlon, Rachel Atlas, Jeremy Lu, Ian Hsiao, and Robert Wood https://doi.org/10.26023/PA5W-4DRX-W50A

VCSEL 25 Hz Water Vapor Data, Version 2.0 M. Diao https://doi.org/10.26023/V925-2H41-SD0F

PHIPS-HALO Stereo Imaging Data, Version 1.0 M. Schnaiter https://doi.org/10.5065/D62B8WWF

PHIPS-HALO Single Particle Data, Version 1.0 M. Schnaiter https://doi.org/10.5065/D6639NKQ

Low Rate (LRT – 1sps) Navigation, State Parameter, and Microphysics Flight-Level Data, Version 1.3 EOL https://doi.org/10.5065/D6M32TM9

NSF/NCAR GV HIAPER Raw 2D-S Imagery EOL Data Support https://data.eol.ucar.edu/dataset/552.009

Model code and software

UWILD and Analysis Codes Johannes Mohrmann, Joseph A. Finlon, Jeremy Lu, Ian Hsiao, and Rachel Atlas https://doi.org/10.5281/zenodo.5197777

University of Illinois/Oklahoma Optical Array Probe (OAP) Processing Software G. M. McFarquhar, J. A. Finlon, D. M. Stechman, W. Wu, R. C. Jackson, and M. Freer https://doi.org/10.5281/zenodo.1285969

Download
Short summary
Many clouds with temperatures between 0 °C and −40 °C contain both liquid and ice particles, and the ratio of liquid to ice particles influences how the clouds interact with radiation and moderate Earth's climate. We use a machine learning method called random forest to classify images of individual cloud particles as either liquid or ice. We apply our algorithm to images captured by aircraft within clouds overlying the Southern Ocean, and we find that it outperforms two existing algorithms.