Fernerkundung des Meereises mit passiven und aktiven Mikrowellensensoren

• Main Author: Kaleschke, Lars
• Other Authors: Künzi, Klaus F., Lemke, Peter
• Format: Doctoral or Postdoctoral Thesis
• Language: German
• Published: Universität Bremen 2003
• Subjects: Sea ice; Microwave Remote Sensing; Ice Concentration; SSM/I; SAR; AMSR; ARTIST; ASI; NED; LVQ; Texture; Data Fusion; Frost Flowers; Pancake Ice; 80; ddc:80; Arctic; Asar; Pancake; Svalbard; Svalbard Archipelago; Storfjorden
• Online Access: https://media.suub.uni-bremen.de/handle/elib/1921; https://nbn-resolving.org/urn:nbn:de:gbv:46-diss000006281

The aim of this study is to reveal inconsistencies between the established sea ice retrieval algorithms as well as to develop and validate new algorithms. The main focus lies on two sensors, the Special Sensor Microwave/Imager (SSM/I) and the Synthetic Aperture Radar (SAR). The project ARTIST was conducted in the environment of the Svalbard archipelago in spring 1998. The core activity was an extensive field study with airborne measurements simultaneously to satellite overpasses. Thereby two new aircraft microwave radiometers operating at 19 and 37 GHz were used. The ARTIST Sea Ice (ASI) algorithm is a method to estimate the sea ice concentration from 85 GHz SSM/I data. Different versions of the algorithm (ASI3/ASI4/ASI5) which were adapted to different reference data are being used in this study. The ASI results are being compared to the ice concentrations obtained using established retrieval algorithms (NT/NT2), ship observations in the central Arctic during summer 2001 (ARK 17/2), and the results of a model of the Storfjorden polynya. The best agreement between ship observations and SSM/I data were obtained using the ASI3/ASI4 algorithm. The ASI5 algorithm achieved the best result compared to the polynya model. The sea ice concentration was underestimated by 10%-20% using the NT algorithm. The averaged NT2 ice concentration agreed well with observation and model data. However, the correlation was evanescent at high ice concentrations. The RABE algorithm presented here allows to estimate the sea ice type and concentration from SAR data. The ability to distinguish between pancake ice, young ice and old ice is being demonstrated on the basis of case studies. It was not possible to distinguish between first and multi-year ice. The potential of frost flower detection is being shown. The ASI and RABE ice concentration results have been found to be consistent. The methods developed in this study are suitable for application to new sensors (ENVISAT-ASAR and AMSR).