Seasonal changes in microbial community composition of Arctic sea-ice ridges

The yearlong MOSAiC ice drift (October 2019 to September 2020), with the research vessel Polarstern serving as the base, started in the eastern Eurasian Basin and crossed the Amundsen and Nansen basins towards the Fram Strait (Fig. 1) (e.g. Nicolaus et al., 2022; Fong et al., 2024). One dedicated research project (Ridges - Safe HAVens for ice-associated flora and fauna in a seasonally ice-covered Arctic Ocean (HAVOC) (Granskog and Müller, 2024) performed detailed and interdisciplinary observations of ridges during MOSAiC. During the drift, three different ridges were sampled at different times of the year. The changes between ridges were necessary, as logistical challenges and ice dynamics prevented the sampling of the same ridge throughout the entire period. The first ridge (R1) was investigated in winter, the second ridge (R2) was investigated in spring, and the third ridge (R3) was investigated in summer. Ice cores for temperature and salinity measurements as well as biogeochemical variables were extracted with a 9-cm (Mark II) internal diameter ice corer (Kovacs Enterprise, USA). Ice cores collected for biogeochemical variables were cut into 10 cm long sections in the field and collected in sterile plastic bags, with the focus on the three habitats: the ice of the roof and the floor of water-filled voids, the bottom of the ridge, and, when present, the frozen void and algae inclusions. Biogeochemical variables were, when possible, derived from pooled ice core sections of three replicate cores (R3), and during challenging weather periods (R1 and R2) from single ice cores. The core sections were kept dark and cool, transferred to the lab on board and melted in the dark after the addition of filtered seawater: 50 mL 0.22 µm filtered seawater was added per cm of sea ice thickness, and the sea ice samples melted within 24–36 hours in the dark at around 4°C. When possible, the water (20–30 L) inside the voids, below the ridge and below level ice, was sampled using a manual bilge pump with a silicon tube with a diameter of 20 mm into prewashed polyethylene containers. Both melted sea ice and water samples were filtered through a Sterivex 0.22 µm filter or, when volumes were < 500 mL (three sea ice samples), onto 0.22 µm Durapore filters. The filters were directly flash-frozen in liquid nitrogen, stored at -80°C, and at the end of the campaign, shipped to the University of Bergen on dry ice. DNA from Sterivex filters was extracted using the Qiagen PowerWater DNA kit, and the DNA from Durapore filters was extracted using the QIAGEN DNeasy Power Soil kit following the manufacturer's instructions. The extracted DNA was stored at -20°C until PCR amplification and prepared for paired-end MiSeq sequencing (Illumina, CA, USA) using a two-step PCR method to amplify the V4 region of the eukaryotic 18S rRNA genes (EukA7F, Euk570R) and the V4 region of the bacterial and archaeal 16SrRNA genes (519F, 806RB) before attaching MiSeq adapter and barcode sequences. Comparative 18S rRNA gene data from first-year and second-year level ice (FYI and SYI) and surface seawater used in this study are part of the MOSAiC main programme, obtained from the Alfred-Wegener institute and can be found at ENA under the project number PRJNA895866.