https://hdl.handle.net/10012/21382

Lake ice is a significant aspect of the physical landscape at northern latitudes and plays a crucial role in regulating weather, climate, and supporting various socio-economic aspects. The sensitivity of lake ice development to air temperature makes it an effective indicator of climate change. This phenomenon not only impacts regional weather patterns but also contributes to global environmental shifts. It is important to closely observe variations in lake ice to gain insight into the potential impact that lakes might have on regional weather patterns and environmental conditions. This study focuses on monitoring lake ice dynamics using RADARSAT Constellation Mission (RCM) compact polarimetric data, employing a combination of the first difference method and compact polarimetric (CP) parameter analysis. The research centers on three Canadian sub-Arctic lakes, with a primary focus on Kluane Lake, and examines phenological transitions through changes in backscatter intensity and CP parameters such as Degree of Polarization (DoP), Circular Polarization Ratio (CPR), Relative Phase, and Alpha Angle. The results reveal that freeze-onset and melt-onset events, mapped using the first difference method, align well with validation data from Landsat thermal imagery and meteorological records, highlighting the reliability of RCM data. Additionally, CP parameter analysis provides deeper insights into ice surface and subsurface conditions. DoP and CPR captured variations in surface roughness, while Relative Phase and Alpha Angle distinguished subsurface scattering and ice-water transitions. Mid-winter stability, reflected by minimal variability in CP values across parameters, emerged as the most consistent pattern, signifying coherent solid or snow-covered ice surface. The results highlight the capacity of RCM data to capture key phenological events, demonstrating strong consistency with optical and thermal data sources. However, the research also identifies limitations related to the temporal resolution of RCM data.

https://hdl.handle.net/10012/21382