1 Introduction
Unlithified Arctic coastlines are situated on the boundary of three rapidly changing and intertwined systems – terrestrial, oceanic, and atmospheric. Near surface terrestrial permafrost temperatures are increasing (Biskaborn et al., 2019), and active layer depths (ALDs) are growing (Letterly, 2018). Sea ice cover is in a state of rapid decline, and ocean temperatures are increasing (Markus et al., 2009; Steele et al., 2008; Stroeve et al., 2014). Surface air temperatures are warming at an accelerated rate relative to the rest of the planet (Johannessen et al., 2016; Serreze and Francis, 2006). All these changes are expected to continue through the 21st century in response to anthropogenic climate change (AMAP, 2017; Collins et al., 2013; Comiso, 2006). These profound transformations of the Arctic environment are already resulting in significant widespread degradation of coastal permafrost (Fritz et al., 2017; Günther et al., 2013, Günther et al., 2015; Jones et al., 2018; Lewkowicz and Way, 2019; Mars and Houseknecht, 2007; Novikova et al., 2018; Pizhankova et al., 2016; Ramage et al., 2018). One of the most active forms of thermokarst are Retrogressive Thaw Slumps (RTSs), a form of slope failure in which thawed soils and ice melt water flow along a massive ice (MI) body or layer of ice rich permafrost. Active thaw slumps are traditionally characterised by a distinctive “C” shaped scar zone up to 1,000 m wide (Lantuit et al., 2012), containing three main elements (Figure 1);
  1. A near vertical headwall consisting of ice poor permafrost and the active layer
  2. A steep angled headscarp with exposed ice – the ablation of which drives the back wasting of the RTS
  3. A low angled slump floor, where thawed permafrost material from the headwall combines with meltwater to form a muddy mixture which flows downslope
The areal extent of RTS affected terrain has undergone a dramatic increase in the last two decades across the western Canadian Arctic, where it is now believed to be the dominant driver of geomorphic change in the region (Kokelj et al., 2015; Lantuit et al., 2012; Lewkowicz and Way, 2019; Ramage et al., 2018; Segal et al., 2016). However, attempts to link RTS activity (using metrics such as headwall retreat rates [HWR]) to temperature indices, such as thawing degree days, have proved inconsistent and typically suited to a narrow range of meteorological and geomorphic conditions (Heginbottom, 1984; Jones et al., 2019; Lewkowicz, 1987a; Robinson, 2000, Zwieback et al., 2018). Two key limitations have constrained our understanding of RTS development and evolution, namely: (1) a lack of topographic data with sufficient spatial and temporal resolution and (2) knowledge of subsurface variability in overburden and MI thicknesses.
The recent increase in LiDAR use (Obu et al., 2016; Ramage et al., 2017) and Structure from Motion-Multiview Stereo (SfM-MVS) derived digital elevation models (Cunliffe et al., 2018, Westoby et al., 2012) have provided new opportunities to better constrain topographic controls on RTSs activity. However, detection and mapping of subsurface MI and overburden variability remains primarily limited to visual observations of their exposures along cliffs and headwalls, sporadic borehole measurements and identification of surface features that may act as proxy indicators for the presence of MI. Interpolation and extrapolation of these values to produce regional MI models (Couture and Pollard, 2017; Couture et al., 2018) may result in significant inaccuracies, especially on local scale, placing a limitation of estimates of carbon loss and management of vulnerable infrastructure. Only recently have new methods emerged for mapping massive ice non-invasively using passive seismic recordings (Lim et al., 2020). Here we present a combination of inter-annual, high spatial resolution SfM-MVS data in combination with passive seismic monitoring of subsurface variability to address these critical problems.