Glacier internal structure revealed by automatic image processing-powered classification of radar images
摘要
The evolution of glaciers’ internal structure, primarily driven by their thermal state, reflects ongoing climate change. Understanding the thermal structure is essential for studying glacier dynamics, mass balance, and hydrological system. Radio-echo sounding enabled the indirect identification of water-saturated temperate ice (W-STI) and water-free ice (W-FI). A novel automatic image processing method has been applied to eliminate the subjectivity and the time consumption of manual determining layer boundaries. The technique based on local binarization was used to assess the internal state and changes of the Arctic glacier Hansbreen from 2007 to 2021, revealing its evolution from a two-layer to a nearly homogeneous structure containing only temperate ice. The average difference between manual and automatic structure recognition results is 1.8 percentage points. The primary mechanisms of the disappearance of the W-FI layer are surface ablation, reducing the W-FI thickness, and the expansion of drainage zones, diminishing the range of W-FI.