Keyword

deep learning

2 record(s)

 

Provided by

Type of resources

Topics

Keywords

Contact for the resource

Update frequencies

From 1 - 2 / 2

  • The PolarLakes dataset provides bi-weekly observations of supraglacial lakes on Antarctic ice shelves, utilizing imagery from Sentinel-2 and Sentinel-1 to address time series gaps caused by frequent cloud cover. These observations detect the extents of supraglacial lakes with a U-Net model for every two weeks from November to March, with each sensor operating independently before the data is merged. The resulting bi-weekly product reflects the maximum lake extents for the first and second halves of each month. When combined for an entire season, the dataset consolidates all bi-weekly records over these five months, allowing for analysis of the maximum lake extent per season and the frequency of lake formation, which can occur up to ten times (5 months á two weeks). The year indicated in the dataset corresponds to January of the melt season, as this month typically experiences the highest melt rates (e.g., 2023 refers to the season from November 2022 to March 2023). The aggregation of all annual datasets creates a recurrence layer that illustrates the frequency of lake presence throughout the entire observation period, which spans from 2014 to 2024, depending on satellite data availability for each ice shelf. The PolarLakes dataset provides valuable insights into the dynamics of supraglacial lakes and serves as a crucial resource for hydrological and climate modeling.

  • IceCloudNet is a novel method based on machine learning able to obtain high quality vertically resolved predictions for ice water content and ice crystal number concentration of clouds containing ice. The predictions come at the spatio-temporal coverage and resolution of Meteosat SEVIRI and the vertical resolution of DARDAR. IceCloudNet consists of a ConvNeXt-based U-Net and a 3D PatchGAN discriminator model and is trained by predicting DARDAR profiles from co-located SEVIRI images. Despite the sparse availability of DARDAR data due to its narrow overpass, IceCloudNet is able to predict cloud occurrence, macrophysical shape, and microphysical properties with high precision. We release 5 years of vertically resolved ice water content (IWC) and ice crystal number concentration (Nice) of clouds containing ice with a 3 km×3 km×240 m×15 minute resolution on a spatial domain of 30°W to 30°E and 30°S to 30°N. The resulting data set increases the availability of vertical cloud profiles for the period when DARDAR is available by more than six orders of magnitude and moreover, is able to provide vertical cloud profiles beyond the lifetime of the recently ended satellite missions underlying DARDAR.