Linear dune orientations in Antarctica over summers 2018--2021

EasyData

The dataset contains the most frequent linear dune orientations over November-December in the years 2018 to 2021 at 25km resolution (EPSG:3031 - WGS 84 / Antarctic Polar Stereographic) over Antarctica, excepted the observation gap at latitudes above 82°S and the disturbed areas (crevasses, mountains) . The dune orientation is retrieved by processing all available Sentinel-2 and Landsat-8 images over this period (around 33,000 images). Objects larger than a kilometer, especially megadunes, are excluded and objects smaller than 30m can not be detected due to the image resolution. Bed-instability and elongating mode orientation are calculated from wind data from ERA-5 Reanalysis data of the years 2018--2021 at 0.25°x0.25° resolution (EPSG:4326 - WGS 84) following Courrech et al. 2014, DOI:10.1130/G35657.1

Collection: Linear dune orientation in Antarctica

creationFeb 25, 2024releasedMay 31, 2024publicationMay 30, 2024

Temporal CoverageNov 1, 2018Dec 14, 2021

Overview

Lineage

The dune orientations are retrieved from the analysis of all Sentinel-2 (from Copernicus data center https://dataspace.copernicus.eu/ ) and Landsat-8 (from the online repository https://earthexplorer.usgs.gov/ ) satellite images over Antarctica. The dune orientation is retrieved by calculation of 2D spatial autocorrelation over sub-image (256x256 pixels) in each satellite image, following the methodology presented in Gao et al. 2015 https://doi.org/10.1038/srep14677 . The retrieved orientation for each sub-image at the different dates are then aggregated on a 25km resolution grid with a kernel density estimation giving the most frequent dune orientation over November-December and over the 4 years 2018--2021. Mountains, clouds and crevasses were masked out.
The bed-instability and elongation mode are retrieved from ERA-5 Reanalysis data https://cds.climate.copernicus.eu using the methodology described in Courrech et al. 2014 https://doi.org/10.1130/G35657.1 over the years 2018--2021.
This work is a part of the PhD Thesis of Marine Poizat (2024) done at Institut des Géosciences de l'Environnement, Grenoble, France.