A multidimensional analysis of sea ice melt pond properties from aerial images

• Main Author: Fuchs, Niels
• Other Authors: Haas, Christian, Birnbaum, Gerit, Dierking, Wolfgang, Spreen, Gunnar
• Format: Doctoral or Postdoctoral Thesis
• Language: English
• Published: Universität Bremen 2023
• Subjects: polar regions; Arctic; Sea Ice; melt ponds; Airborne observations; High-resolution aerial imaging; Pond bathymetry; Albedo; image classification; Photogrammetry; 500; 500 Science; ddc:500; albedo; Sea ice
• Online Access: https://media.suub.uni-bremen.de/handle/elib/6928; https://doi.org/10.26092/elib/2249; https://nbn-resolving.org/urn:nbn:de:gbv:46-elib69280

Sea ice plays a fundamental role in Polar climate and ecosystems. Melt ponds, forming routinely on Arctic sea ice during summer, can cover and impact a considerable fraction of the ice area. However, data that allow a comprehensive understanding of pond evolution processes remain scarce. Consequently, we cannot yet predict how ponds will develop on the increasingly prevalent young ice in the future. Previous studies have drawn a very heterogeneous picture of pond coverage on young ice, which we can only improve with more detailed measurement data and analysis tools that allow the derivation of properties possibly driving pond evolution. The existence of over ten years of high-resolution aerial image data from AWI aircraft campaigns in the Arctic has motivated me to develop and refine evaluation methods for this dataset, the one-year drift campaign MOSAiC, and future measurement campaigns. I created a customized classification algorithm to classify images into sea ice surface classes with minimal manual intervention. By implementing cutting-edge photogrammetry tools and developing a spatially high-resolution albedo and pond depth retrieval method, I draw an unprecedented multidimensional picture of melt ponds. From this, I derived properties of the sea ice cover that favor and limit pond coverage. I found that within the observed areas, melt pond coverage was more constant than expected, ranging between 15% to 25%. The first-ever tracking of the evolution of the entire pond bathymetry shows that we have so far overlooked the deformability of the pond bottom ice. The multidimensional, high-resolution approach for long-range airborne measurements allowed me to make general recommendations for representative ground measurements. The tools presented, together with the refined insights into pond properties and evolution, will improve our understanding of summer sea ice and can help better assess the role and fate of ponds in the future Polar climate and ecosystems.