Mapping of seasonal sea ice in the west of the Antarctic Peninsula with Sentinel 1A images

Authors: Fernando Luis Hillebrand, Universidade Federal do Rio Grande do Sul, Cristiano Niederauer da Rosa, Universidade Federal do Rio Grande do Sul, Juliana Costi, Universidade Federal de Rio Grande, Ulisses Franz Bremer*, Universidade Federal do Rio Grande do Sul
Topics: Polar Regions, Remote Sensing, Oceanography
Keywords: Antarctica, Austral Ocean, Synthetic Aperture Radar, backscatter coefficients, sea ice classification
Session Type: Poster
Day: 4/4/2019
Start / End Time: 9:55 AM / 11:35 AM
Room: Lincoln 2, Marriott, Exhibition Level
Presentation File: No File Uploaded


Sea ice acts as a thermal insulation between the ocean and the atmosphere and without it, the ocean is free to emit heat into the atmosphere, causing changes in ocean circulation. Knowing the formation, thickness, and duration of sea ice in the Southern Ocean is a key process for understanding ocean-atmosphere interactions in Antarctica. Unlike the Arctic, the Southern Ocean does not have extensive coverage of permanent sea ice, as can be seen from summer satellite images of the southern hemisphere. Remote sensing of the sea ice cover comprises several techniques and operates over a wide range of frequencies in the electromagnetic spectrum. This work proposed the classification and identification of seasonal sea ice areas by means of SAR images in the C band, through Sentinel 1A satellite, under HH polarization and IW imaging mode. The supervised classification was performed using the Random Forest method, with sample training through texture analysis with Sentinel 2 optical images, K-Means unsupervised classification and analysis of HH / HV co-polarizations in Sentinel 1B images under EW imaging. Backscatter coefficients were identified from -10.2 to -13.5 dB for young ice and/or first-year ice, -17.5 to -19.6 dB for new ice, and -23.7 to -26.8 dB for open water in the western region of the Antarctic Peninsula during the austral winter.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login