ICESat-2 monthly gridded winter Arctic sea ice thickness

Monthly gridded (winter only) Arctic sea ice thickness estimates from ICESat-2 derived using ATL10 freeboards (https://nsidc.org/data/atl10) together with snow depth and density estimates from the NASA Eulerian Snow on Sea Ice Model (NESOSIM, https://github.com/akpetty/NESOSIM). Along-track data (from the three strong beams) are binned to the 25 km x 25 km NSIDC polar stereographic projection (EPSG:3411). The full processing chain is described in Petty et al., (2020) (code available at https://github.com/akpetty/ICESat-2-sea-ice-thickness) including several updates as detailed below.

Temporal range: November 2018 - April 2019, October 2019 to April 2020.

Data: A single netCDF file is included for each month. Variables include:

• Sea ice freeboard (from ATL10)
• Snow depth (redistributed NESOSIM)
• Snow density (redistributed NESOSIM)
• Bulk sea ice density
• Sea ice type (from OSI SAF)
• Sea ice thickness uncertainty
• Mean day of month in a given grid cell
• Number of freeboard segments in a given grid cell.

A summary of the differences between the version 1 and version 2 winter Arctic sea ice thickness estimates are being presented at AGU 2020 and prepared for publication.

Key changes from version 1 (Petty et al., 2020) to version 2 include:

• Use of release 003 ATL10 freeboards. A detailed assessment of the freeboard changes from release 002 to release 003 is provided in Kwok et al., (2020).
• Upgrade to NESOSIM v1.1: CloudSat scaling of ERA5 snowfall, a new atmospheric wind loss term, calibration against recent OIB snow depths, an extended Arctic Ocean domain and various bug fixes (https://github.com/akpetty/NESOSIM).
• Use of all three strong beams (instead of just strong beam #1).

The data have also been made available on a Google Cloud bucket to enable rapid data analysis from any cloud-based analytics platform: gs://sea-ice-thickness-data/v2/