Dataset: MOSAIC Snow and Frost Flower Be7

Bulk Snow Sampling: For bulk snow samples, we collected the full snow column (snow surface to ice surface). First, the depth of snow was measured with a ruler. Next, a measured area of the snow surface was outlined with a stick or shovel. Then all of the snow (down to the ice surface) within the marked area was removed with a plastic ice scoop and placed in a plastic bucket.

Snow Profile Sampling: For snow profiles, we collected samples at different depth intervals within the snow column. After measuring the depth of snow and marking a measured area on the snow surface, the first layer of snow was carefully removed with a plastic flat bottomed ice scoop and placed in a bucket. The depth of the remaining snow was then measured with a ruler before collecting the next layer of snow. This process was repeated until the ice surface was reached. We sampled two to five layers per profile.

Frost Flower Sampling: On two occasions we collected frost flowers from a recently frozen lead before significant snow accumulation had occurred. After marking a measured area of the ice surface, all of the material was removed with a plastic ice scoop and placed in a plastic bucket.

Lead Snow samples are bulk snow and snow profiles collected from a frozen lead after frost flowers were covered with snow.

Falling Snow Sampling: On five occasions we were able to sample falling snow in buckets deployed on the Polarstern’s p-deck. Snow was collected in a plastic bucket (diameter = 26 cm, area = 531 cm2). The buckets were deployed during periods of active snowfall and light winds (minimal interference from blowing snow).

Shipboard Laboratory Procedure: After the snow samples in buckets to the Polarstern, we added the following solutions: 5 ml hydrochloric acid, 2 ml iron chloride solution, and 1 ml of 1000 ppm stable beryllium AA standard (chemical yield tracer). After melting at room temperature, the water volume was measured using a graduated cylinder. Next, the sample was precipitated using ammonium hydroxide. After allowing the precipitate to settle, excess water was removed by decanting and centrifugation. Finally, the precipitate was transferred to petri dishes and placed in an oven for drying.

Gamma Analysis: Dried samples in petri dishes were counted by gamma spectroscopy. Be-7 has a readily identifiable gamma peak at 478 keV. The counting system was calibrated for all samples by preparing a commercial standard in a geometry identical to the samples.

Chemical Yield Determination: After gamma counting, samples were brought up to 1 liter in 1% nitric acid solution. The beryllium concentrations were then determined using a Perkin Elmer Optima 7300 DV ICP-OES. Our chemical yields averaged over 80% (Kadko et al., 2019).