Assessment of Larsen C Ice Shelf stability: insights from surface and basal fracture

The ice shelf structure is influenced by surface and basal morphologies, making accurate detection of surface and basal fracture changes essential for assessing ice shelf stability. We used the U-KAN neural network model, which excels at capturing edge details, to detect surface fractures on the Larsen C Ice Shelf (LIS) over the past decade using optical remote sensing images. Potential basal fractures were validated using ice-penetrating radar data. Dynamic changes and development patterns of LIS fractures were analyzed from surface and basal perspectives. The study demonstrated that the U-KAN neural network model effectively enhances the continuity of individual fracture segmentation results while maintaining the independence of multiple fracture segmentation outcomes. The LIS surface fracture (surface crevasses and depressions) distribution was linked to strain rates and maximum buttressing. Basal melting predominantly influences basal fracture formation (basal crevasses), with the penetration rate increasing in the direction of ice flow. Surface and basal fractures influence ice shelf stability and its future dynamics to varying extents; the southern section of the LIS is at risk of further calving. Our results offer a more comprehensive understanding of the dynamic changes in the ice shelf and provides valuable insights for predicting its disintegration.