Greenland Ice Sheet crevasse map from ArcticDEM

• Main Author: Chudley, Thomas R.
• Format: Dataset
• Language: English
• Published: 2022
• Subjects: crevasses; greenland; ice sheet; glacier; topography; remote sensing; Greenland; Ice Sheet; Sea ice
• Online Access: https://zenodo.org/record/6779088; https://doi.org/10.5281/zenodo.6779088

Data Description Parameter Title Greenland Ice Sheet crevasse map from ArcticDEM Version 1.00 Format GeoTiff Projection WGS84 / NSIDC Sea Ice Polar Stereographic North (EPSG: 3413) Resolution 2 m (binary) and 200 m (fraction) Size 3.5 GB (total binary) and 22 MB (total fraction) Citation Chudley et al. (2021) Contact Tom Chudley Email [e-mail] Products This dataset contains crevasse locations identified from the ArcticDEM v3 mosaic. There are two primary products: a 2 m binary crevasse map, and a 200 m crevasse fraction map. It is divided into the six IMBIE 'Rignot' drainage basins: central west (CW), southwest (SW), southeast (SE), northeast (NE), north (NO), northwest (NW). 2 m crevasse binary Byte GeoTiff product indicating derived crevasses at 2 m resolution. File naming convention is: crevasse_binary_XX_2m.tif where XX is the IMBIE basin code. Value Meaning 0 No data 1 No crevasses identified 2 Crevasses identified 200 m crevasse fraction Float32 GeoTiff product indicating fraction of 200 m grid cell identified as crevasses in the 200 m product. File naming convention is: crevasse_fraction_XX_2m.tif where XX is the IMBIE basin code. Value Meaning 0 -- 1 Fraction of grid cell identified as crevasse in 2 m dataset -9999 No data Method The full processing chain for data derivation is described in Chudley et al. (2021). A binary crevasse mask of the Greenland Ice Sheet is generated using ArcticDEM v3 mosaic data at 2 m resolution (Porter et al., 2018), with data processed in Google Earth Engine (Gorelick et al., 2017). The ArcticDEM is cropped to the GIMP ice mask (Howat et al., 2014), before a smoothed elevation model is generated by performing an image convolution with a circular kernel of 50 m radius. Residuals greater than 1 m between the smoothed and raw elevation values were identified as crevasses. To compare with public velocity datasets (and derived strain rates, stress, etc.), the 2 m dataset was aggregated (using GDAL) into grid cells to match the resolution (200 m) of the Making ...