Abstract:
Mapping the distribution of subglacial water bodies (SWBs) in Antarctica using radio echo sounding (RES) data is of great significance for understanding subglacial hydrol...Show MoreMetadata
Abstract:
Mapping the distribution of subglacial water bodies (SWBs) in Antarctica using radio echo sounding (RES) data is of great significance for understanding subglacial hydrology, the mass balance of Antarctic ice sheet, biology, and so on. However, existing methods for identifying SWBs mostly focus on utilizing the characteristics of radar echoes at the individual trace level to pinpoint areas that consist of subglacial water, leading to discontinuities in the identification results. In this article, a novel multiscale segment-based SWB identification method according to basal dry-wet transition characteristics is proposed and applied to the RES data collected in the Antarctic’s Gamburtsev Province (AGAP) in East Antarctica. The identification results of SWBs with high, medium, and low confidence levels are provided and compared to the existing inventories of SWBs. It is found that 96.06% of the SWBs in the inventories can be identified, with 87.40% in quantity and 84.73% in length identified with high confidence, and 8.66% in quantity and 8.45% in length identified with medium confidence. In addition, some SWBs with high confidence that are not recorded in the inventories are identified by using the proposed method. The distribution of identified SWBs with high confidence is clustered and shows high consistency with inventorial SWBs. Three characteristics, including subglacial depth, hydraulic gradient, and abruptness index, of identified SWBs with different confidence levels are statistically analyzed, demonstrating rationality and consistency with the characteristics of the subglacial ice-water interface. The SWB identification results contribute valuable insights for a better understanding of subglacial environment in the AGAP region.
Published in: IEEE Transactions on Geoscience and Remote Sensing ( Volume: 62)