Articles | Volume 13, issue 11
https://doi.org/10.5194/tc-13-2915-2019
https://doi.org/10.5194/tc-13-2915-2019
Research article
 | 
08 Nov 2019
Research article |  | 08 Nov 2019

Estimating early-winter Antarctic sea ice thickness from deformed ice morphology

M. Jeffrey Mei, Ted Maksym, Blake Weissling, and Hanumant Singh

Data sets

Sea Ice Layer Cakes, PIPERS 2017 M. Jeffrey Mei et al. https://doi.org/10.15784/601207

NBP1704 CTD sensor data Stammerjohn, S. https://doi.org/10.15784/601192

Download
Short summary
Sea ice thickness is hard to measure directly, and current datasets are very limited to sporadically conducted drill lines. However, surface elevation is much easier to measure. Converting surface elevation to ice thickness requires making assumptions about snow depth and density, which leads to large errors (and may not generalize to new datasets). A deep learning method is presented that uses the surface morphology as a direct predictor of sea ice thickness, with testing errors of < 20 %.