It also features exciting sea ice draft data from mooring observations in the northwestern Barents Sea (2018–2021):

The Nansen Legacy data paper also includes ice and snow thickness measurements from both on-ice and helicopter-borne electromagnetic surveys:

Sea ice density data from Nansen Legacy show seasonality and temperature dependence very similar to the MOSAiC observations we recently analyzed: doi.org/10.5194/tc-19-1259-2025

Sea ice density data from Nansen Legacy show seasonality and temperature dependence very similar to the MOSAiC observations we recently analyzed: doi.org/10.5194/tc-1...

An example of how data from ROV-based multibeam sonar was used to quantify ridge-enhanced melt rates can be found here: doi.org/10.5194/tc-1...

The study includes an overview of >80 surveys including measurements of ice draft from sonar, solar irradiance and radiance (see below), hyperspectral images, physical, chemical (pH, nitrate, oxygen), and bio-optical (fluorometry, ultra-violet and visible absorbance spectroscopy) water properties.

And if you are interested in longer observations of snow thickness from drifting stations in 1937–1991, I can recommend a wonderful PhD thesis by @robbiemallett.bsky.social: discovery.ucl.ac.uk/id/eprint/10161766

Huge thanks to Don Perovich, Chris Polashenski, Cameron Planck, and others for leading and maintaining CRREL, Dartmouth @dartmouthears.bsky.social, and Cryosphere Innovation www.cryosphereinnovation.com ice mass balance buoy programs.

And if you are working with the data from the MOSAiC expedition, here are an additional 24 Digital Thermistor Chain (DTC) datasets led by Mario Hoppmann from @awi.de and processed by me: doi.org/10.1594/PANGAEA.964023

A similar but even larger dataset with 96 SIMBA ice mass balance buoys was published by Andreas Preußer @ando-puru.bsky.social for both Arctic and Antarctic ice during 2012–2023: doi.org/10.1594/PANGAEA.973193

The raw discrete data of sea ice physical properties, isotopes, and nutrients is available for first-year ice doi.org/10.1594/PANGAEA.959830, second-year ice doi.org/10.1594/PANGAEA.971385 for October 2019-July 2020, and doi.org/10.1594/PANGAEA.971266 for August-September 2020.

Finally, I wanted to thank my co-authors Odile Crabeck @universitedeliege.bsky.social, Niels Fuchs @cenunihh.bsky.social, Nils Hutter @awi.de
@geomarkiel.bsky.social, Philipp Anhaus @dlr-spaceagency.bsky.social, and Jack Christopher Landy (UiT), and Norwegian Polar Institute, where I made the study

What about historical observations? We measure sea ice density since 1927 by Malmgren. Seasonal evolution is complicated due to differences in melt onset timing in the Arctic Ocean. But the same data looks much tidier when plotted against ice temperature, with MOSAiC values fitting previous values:

The densities were similar from all used methods. This means that weighing is the most affordable and accurate way with minor errors from brine loss. Here we show that ice is typically not in hydrostatic balance on scales of 10 meters, which leads to artificially large spreads of density estimates:

But why does ice get lighter upon warming? We linked the increase in air volume to two factors: (1) internal melt, which creates voids, enlarges bubbles, and nucleates new bubbles, and (2) the replacement of liquid brine by air in drained inclusions. Air occupies much more than 10% of brine: