Helicopter-borne melt pond, sea ice, and ocean surface temperatures with surface type classifications during the Polarstern PS131 ATWAICE expedition in the Arctic during summer 2022

• Main Authors: Reil, Marlena, Huntemann, Marcus, Spreen, Gunnar
• Format: Dataset
• Language: English
• Published: PANGAEA 2024
• Subjects: AC3; airborne; Arctic; Arctic Amplification; ATWAICE; Comment of event; Date/Time of event; Event label; HELI; Helicopter; melt ponds; netCDF file; netCDF file (File Size); Polarstern; PS131; PS131_HELI220713T0753; PS131_HELI220713T1045; PS131_HELI220717T0759; PS131_HELI220717T1223; PS131_HELI220718T0812; PS131_HELI220718T1429; PS131_HELI220719T1049; PS131_HELI220719T1120; PS131_HELI220724T1313; PS131_HELI220730T0428; PS131_HELI220730T0852; PS131_HELI220730T1114; PS131_HELI220808T0849; Sea ice; Surface Temperature; Berichte zur Polar- und Meeresforschung; Fram Strait; Reports on Polar and Marine Research
• Online Access: https://doi.pangaea.de/10.1594/PANGAEA.971908; https://doi.org/10.1594/PANGAEA.971908

We provide 1 m resolution airborne thermal infrared imagery that captures sea ice, melt pond, and ocean surface temperatures during the Arctic summer in 2022. The data were acquired during the RV Polarstern PS131 ATWAICE expedition in the marginal sea ice zone of the Fram Strait region and north of it in July and August 2022. A total of 16 flights were conducted with an InfraTec Vario-CAM HD head 680 camera at a typical altitude of 300 m. Each image records the broadband infrared radiation (7.5 μm to 14 μm). The data is stored in NetCDF format, with one file per flight and 640x480 pixels per image at an acquisition frequency of 4 Hz (i.e., images have significant overlap). The data has been corrected for image gradients caused by the lens and sea ice drift during the flight. These data can be used to better understand sea ice surface temperatures and melt pond distribution. In addition, we provide a processed version where the surface temperatures are classified into sea ice, melt pond, and ocean surface types using a deep learning based model.