12:00 PM - 12:15 PM
[ACG52-12] Permafrost thaw subsidence around Beaver Creek, Yukon Territory: Comparative validation of InSAR time-sesires analysis and GNSS observations
Keywords:Permaftost, Thermokarst, Wildfire, InSAR, synthetic aperture radar, GNSS
The recent rise in temperature in the Arctic region has been progressing at a rate higher than the global average. Climate simulations predict continued temperature rises and an increase in wildfire frequency. However, Earth System Models incorporating permafrost thaw processes remain limited, particularly considering abrupt thaw events associated with subsidence and interactions between vegetation changes, wildfires, and permafrost degradation. To accurately assess these impacts and understand environmental changes, it is essential to conduct large-scale spatial observations of permafrost thaw processes, including abrupt thaw events.
Previous studies have identified massive ground ice in the southern area of Beaver Creek (BC) in Yukon Territory, indicating a potential for abrupt permafrost thawing and ground subsidence. This study focused on a post-wildfire area (~28 km²) burned in 2019, establishing an in-situ validation site. Interferometric Synthetic Aperture Radar (InSAR) analysis was conducted to derive a ground displacement signal using data from the JAXA's L-band SAR satellite ALOS-2/PALSAR-2. Initial analysis using the StripMap 10 m (SM3) mode revealed limitations in detecting spatially heterogeneous subsidence patterns around BC as local decorrelation issues (Yanagiya et al., JpGU2023). Therefore, high-resolution observations using the StripMap 3 m (SM1) mode were initiated in 2023 summer, allowing for detailed seasonal subsidence analyses within the validation site (Yanagiya et al., JpGU2024).
This study performed an InSAR time-series analysis using two years (2023–2024) of high-resolution observation data to detect both seasonal and annual ground deformation. Additionally, GNSS observations conducted in 2023–2024 directly measured ground displacement at the validation site, enabling a comparative validation with InSAR-derived displacement. The in-situ observations were conducted as part of the international research project PRISMARCTYC, which aims to investigate the impacts of permafrost thaw on soil-water interaction and carbon cycling.
The InSAR time-series analysis detected seasonal and annual ground subsidence on the north-facing slope within the post-fire area. Subsidence processes varied depending on topographic conditions: no significant deformation was observed in the upper slope (where bedrock is present), whereas, in the middle slope, the largest subsidence was detected—up to 14 cm in 2023 and 5 cm in 2024. In contrast, subsidence was smaller in the lower slope, where maximum subsidence signals of 7 cm in 2023 and 3 cm in 2024 were detected. The spatial distribution of thaw subsidence was consistent with the thaw depth distribution. Notably, the middle slope exhibited the deepest thaw depth (>100 cm), with a deepening trend from 2022 to 2024. In the lower slope, the thaw depth was approximately 60 cm, which was 4–20 cm shallower in 2024 compared to the past two years.
Furthermore, a time series of the Normalized Difference Vegetation Index (NDVI) derived from Sentinel-2 data was observed to assess post-fire vegetation recovery. NDVI values dropped immediately after the 2019 fire but showed a subsequent increasing trend. The recovery rate was higher in the lower slope, where subsidence was smaller, and the thaw depth was shallower than the middle slope. Minimal annual ground deformation from late summer 2023 to 2024 suggests only the active layer thawed in 2024 without subsequent permafrost thaw. GNSS observations during the same period showed no significant ground displacement at most observation points, consistent with InSAR results. The permafrost thaw processes within the post-fire area in BC exhibited spatial and temporal heterogeneity, even within the same slope. This study suggests a relationship between topography, post-fire vegetation recovery, and rapid permafrost degradation.
Previous studies have identified massive ground ice in the southern area of Beaver Creek (BC) in Yukon Territory, indicating a potential for abrupt permafrost thawing and ground subsidence. This study focused on a post-wildfire area (~28 km²) burned in 2019, establishing an in-situ validation site. Interferometric Synthetic Aperture Radar (InSAR) analysis was conducted to derive a ground displacement signal using data from the JAXA's L-band SAR satellite ALOS-2/PALSAR-2. Initial analysis using the StripMap 10 m (SM3) mode revealed limitations in detecting spatially heterogeneous subsidence patterns around BC as local decorrelation issues (Yanagiya et al., JpGU2023). Therefore, high-resolution observations using the StripMap 3 m (SM1) mode were initiated in 2023 summer, allowing for detailed seasonal subsidence analyses within the validation site (Yanagiya et al., JpGU2024).
This study performed an InSAR time-series analysis using two years (2023–2024) of high-resolution observation data to detect both seasonal and annual ground deformation. Additionally, GNSS observations conducted in 2023–2024 directly measured ground displacement at the validation site, enabling a comparative validation with InSAR-derived displacement. The in-situ observations were conducted as part of the international research project PRISMARCTYC, which aims to investigate the impacts of permafrost thaw on soil-water interaction and carbon cycling.
The InSAR time-series analysis detected seasonal and annual ground subsidence on the north-facing slope within the post-fire area. Subsidence processes varied depending on topographic conditions: no significant deformation was observed in the upper slope (where bedrock is present), whereas, in the middle slope, the largest subsidence was detected—up to 14 cm in 2023 and 5 cm in 2024. In contrast, subsidence was smaller in the lower slope, where maximum subsidence signals of 7 cm in 2023 and 3 cm in 2024 were detected. The spatial distribution of thaw subsidence was consistent with the thaw depth distribution. Notably, the middle slope exhibited the deepest thaw depth (>100 cm), with a deepening trend from 2022 to 2024. In the lower slope, the thaw depth was approximately 60 cm, which was 4–20 cm shallower in 2024 compared to the past two years.
Furthermore, a time series of the Normalized Difference Vegetation Index (NDVI) derived from Sentinel-2 data was observed to assess post-fire vegetation recovery. NDVI values dropped immediately after the 2019 fire but showed a subsequent increasing trend. The recovery rate was higher in the lower slope, where subsidence was smaller, and the thaw depth was shallower than the middle slope. Minimal annual ground deformation from late summer 2023 to 2024 suggests only the active layer thawed in 2024 without subsequent permafrost thaw. GNSS observations during the same period showed no significant ground displacement at most observation points, consistent with InSAR results. The permafrost thaw processes within the post-fire area in BC exhibited spatial and temporal heterogeneity, even within the same slope. This study suggests a relationship between topography, post-fire vegetation recovery, and rapid permafrost degradation.