Pre-collapse motion of the February 2021 Chamoli rock-ice avalanche, Indian Himalaya

. Landslides are a major geohazard that cause thousands of fatalities every year. Despite their importance, identifying unstable slopes and forecasting collapses remains a major challenge. In this study, we use the 7th of February 2021 Chamoli rock-ice avalanche as a data-rich example to investigate the potential of remotely-sensed datasets for the assessment of slope stability. We investigate imagery over the three decades preceding collapse and assess the precursory signs exhibited by this slope prior to the catastrophic collapse. We evaluate monthly slope motion from 2015 to 2021 through feature tracking of 5 high-resolution optical satellite imagery. We then combine these data with a time series of pre-and post-event DEMs, which we use to evaluate elevation change over the same area. Both datasets show that the 26.9 Mm 3 collapse block moved over 10 m horizontally and vertically in the ﬁve years preceding collapse, with particularly rapid motion occurring in the summers of 2017 and 2018. We propose that the collapse results from a combination of snow-loading in a deep headwall crack and permafrost degradation in the heavily jointed bedrock. Despite observing a clear precursory signal, we ﬁnd ﬁnd that the timing of the

Several factors have contributed to raising landslide risk across the region in recent decades: first, climatic warming has driven rapid thinning and retreat of Himalayan glaciers -which are currently losing over 10 Gt of mass per year (e.g. Kääb et al., 2012;Brun et al., 2017;Shean et al., 2020;Jakob et al., 2021;Hugonnet et al., 2021). Glacier retreat may contribute to a range of factors conducive to landslides, including a reduction in slope buttressing and an increase in meltwater availability (Holm et al., 2004;Fischer et al., 2006;Huggel et al., 2012;Kos et al., 2016;Coe et al., 2018;Dai et al., 2020a;Glueer et al., 2020). In addition to glacier retreat, permafrost degradation has also been documented to reduce slope stability (Gruber 55 and Haeberli, 2007;Allen et al., 2011;Fischer et al., 2012;Krautblatter et al., 2013;Haeberli et al., 2017;Magnin et al., 2019;Pörtner et al., 2019;Patton et al., 2019;Deline et al., 2021). Second, increasing populations, economic growth, and infrastructure development in high-mountain valleys have greatly expanded the potential consequences of landslides. This second point is apparent for the Chamoli disaster, in which the majority of deaths occurred at hydropower plants that were recently built or were under construction (Shugar et al., 2021). Other factors, including changes in precipitation pattern (e.g. Li 60 et al., 2018; Kirschbaum et al., 2020) and land use (Cummins, 2019) may also contribute to evolving landslide hazard potential.

The 2021 Chamoli hazard cascade
During the morning of 7 February 2021, a 26.9 [95% confidence interval 26.5-27.3] Mm 3 wedge of rock and ice detached from the north face of Ronti, a 5500 m peak in the Uttarakhand Himalaya (Fig 1.). This wedge then dropped around 1800 m to the Ronti Gad valley floor, where it continued down-valley towards the Rishiganga and Dhauliganga rivers and transformed into a debris flow (Shugar et al., 2021;Cook et al., 2021). The collapse block was composed of approximatively 80% bedrock and 20% glacier ice. Frictional heat generation calculations suggest that most or almost all of the glacier ice melted during the 3400 m drop from the collapse source to the hydropower stations (Shugar et al., 2021). This melting of the ice faction, combined with major sediment deposition at the confluence of the Ronti Gad and Rishiganga, increased the initial rock-ice avalanche's water content and converted it into a debris flow. The resulting debris flow caused further downstream damage, leaving 204 70 missing or killed and destroying two hydropower stations.

Feature tracking
Optical feature tracking is a versatile technique, which can be used to track surface motion by evaluating the relative position of features or patterns in repeat satellite images or aerial photos. Feature tracking has been applied to a variety of problems, 75 including tracking post-seismic ground deformation (e.g. Leprince et al., 2007), measuring glacier flow velocities (e.g. Bindschadler and Scambos, 1991;Heid and Kääb, 2012;Millan et al., 2019;Van Wyk de Vries and Wickert, 2021), and measuring landslide displacements (e.g. Behling et al., 2014;Lucieer et al., 2014;Manconi et al., 2018;Dai et al., 2020a;Dille et al., 2021).

Stereo-DEM generation 80
Stereo-DEM generation uses two or more overlapping optical images to reconstruct surface topography. These images are acquired at the same time but from different viewing angles. Software implementing photogrammetric principles can then be used to derive elevation products (such as DEMs) from these images. With the recent availability of very high resolution satellite stereo imagery, these approaches can now be used to generate detailed DEM products over large spatial areas (e.g. Korona et al., 2009;Morin et al., 2016;Shean et al., 2016;Porter et al., 2018;Howat et al., 2019). Repeat DEMs obtained 85 at different time periods can provide precise estimates of surface elevation change associated with many processes, including glacier change (e.g. Brun et al., 2017;Willis et al., 2018;Zheng et al., 2019;Shean et al., 2020), snow accumulation/melt (e.g. Deschamps-Berger et al., 2020;McGrath et al., 2019;Bhushan et al., 2021), volcanic deformation (e.g. Bisson et al., 2021Schaefer et al., 2012), and landslide or debris flow events (e.g. van Westen and Lulie Getahun, 2003).

90
Satellite-based InSAR is a powerful tool for detecting small changes at the Earth's surface from space. It has been widely used to quantify ground displacements caused by processes such as earthquakes (e.g. Massonnet et al., 1993;Barba-Sevilla et al., features, InSAR can provide measurements of ground deformation at millimeter and centimeter scales. Active radar sensors can image the Earth's surface through clouds and darkness, a major advantage over passive optical sensors (e.g. Massonnet and Feigl, 1998). Leveraging InSAR data for the detection and assessment of mass movements, however, is not without challenges.
The oblique viewing geometry of radar satellites means that radar data can be rendered useless in areas of steep topography due to the effects of shadowing, foreshortening, and layover (Massonnet and Feigl, 1998;Wasowski and Bovenga, 2014). Finally, 100 in case of rapid displacements that surpass the phase-aliasing thresholds or dramatic changes in the surface cover or geometry, a loss of interferometric coherence can prohibit the quantification of (the full) ground deformation (Manconi, 2021). Despite these drawbacks, many studies have shown that InSAR can be successfully applied to assess stability of slopes even in high relief terrain (e.g. Manconi et al., 2018;Handwerger et al., 2019;Bekaert et al., 2020;Jacquemart and Tiampo, 2021).

105
The objective of this study is to evaluate the pre-collapse conditions of the 7 February 2021 Chamoli rock-ice avalanche, in particular: 1. What was the scale and geometry of pre-collapse surface change, and what insight do these changes provide into the collapse mechanisms?

Qualitative observations of slope change
We investigated three decades of pre-collapse optical satellite imagery to gain a preliminary understanding of pre-landslide changes. We documented changes in the north-facing slope of Ronti peak, which sourced the February 2021 rock-ice avalanche, 125 using all available data from Landsat 5 (TM), Landsat 7 (ETM+), Landsat 8 (OLI), and Sentinel-2 with a cloud cover of less than 60%. We focused our observations on surface changes, including deformation and fracturing, and rock or ice avalanches originating from the collapsed block or surrounding area.
Our ability to detect change is limited by the spatial resolution of the imagery used (15-30 m for Landsat and 10 m for Sentinel-2). We examined a 31-year (1990-2021) time series of satellite imagery (Fig 2), including 122 Landsat 5 images, 43 130 Landsat 7 images, 34 Landsat 8 images, and 155 Sentinel-2 images. A full list of images is provided in the supplementary material, along with a brief description of any anomalous features.

Optical feature tracking
We used feature tracking with a range of medium (10 m) to high (2.5 m) resolution satellite imagery to evaluate the pre-collapse motion of the Ronti peak north slope. We used two different feature-tracking toolboxes: GIV (Van Wyk de Vries and Wickert, 135 2021) and AutoRIFT (Lei et al., 2021). Both GIV and AutoRIFT are based on three core components: a pre-processing module which applies one or more filters to images to enhance distinct surface features for tracking, a multipass 2D image correlator, and a post-processing module to identify and filter erroneous displacement values (Van Wyk de Vries and Wickert, 2021;Lei et al., 2021). The GIV toolbox is written in MATLAB and performs image cross correlation in the frequency domain, while AutoRIFT is written in python/C++ and performs the cross correlation in the spatial domain. Using GIV, we pre-processed the 140 imagery using an orientation filter and ran the cross-correlation with a reducing window size from 20 to 5 pixels and a window overlap of 50%. In AutoRIFT we pre-processed the imagery with a Laplacian filter and used adaptive window sizes between 32 and 64 pixels with a skip rate of 8 pixels for the cross-correlation.
We calculated velocities using all available Sentinel-2 images through February 2021, excluding any images with a local cloud cover greater than 60% (based on the L1-C QA band cloud mask). A total of 155 images were available, for a total changing illumination, as the images were captured by different satellites during different times of the day/year. To compensate for this higher background noise, we chose a higher minimum temporal separation between Planet image pairs when calculating time-averaged velocity maps. We also calculated displacements (using both GIV and AutoRIFT) on one pair of high-resolution Cartosat-1 images (Oct 2017 to Nov 2018).
We used this velocity data to evaluate whether the collapsed block moved prior to collapse -with a null hypothesis that 160 the block moved no more than the surrounding 'stable' (non-glacierized) bedrock. A medial bedrock ridge near the center of the collapsed block provides motion of the underlying rock, rather than simply flow of the overlying glaciers. We divided the collapse block into three different regions alongside a zone of stable ground, and create a time series of average displacement for each zone. 165 We produced multiple pre-event and post-event DEM products from very high-resolution (Maxar/DigitalGlobe WorldView-1/2/3, GeoEye-1 and Airbus/CNES Pleiades, 0.3 to 0.5 m GSD) and high-resolution (Airbus SPOT-7 and ISRO CartoSat-1, 1.5 m to 2.5 m GSD) satellite imagery captured between 2015 and February 2021. The DEM products were used to calculate the vertical motion of the collapse block from 2015 to February 10, 2021.

DEM generation
We used the NASA Ames Stereo Pipeline (Shean et al., 2016;Beyer et al., 2018) to process all of the images. For this DoD provides insight into vertical changes in the hillslope prior to failure, while the second DoD provides the volume and geometry of the collapsed block. We calculated an empirical uncertainty estimate for each DoD using the tiling method (Berthier et al., 2016;Miles et al., 2018;Jacquemart et al., 2020).

InSAR maps
We analyzed Sentinel-1 data from the ascending and descending orbit tracks 56 and 63, respectively, to investigate whether the 180 precursory motion of the collapse block could have been detected from radar interferometry. All radar data was downloaded from the Alaska Satellite Facility Distributed Active Archive Center (ASF DAAC). Because the descending track is heavily affected by layover artefacts, we only performed the full processing with data from the ascending orbit. We processed the data with the InSAR Scientific Computing Environment (ISCE; Rosen et al., 2012), removed the topographic phase using the 2015 pre-event WorldView DEM (Bhushan and Shean, 2021) composite (resampled to 8m), and masked out all pixels 185 with an interferometric coherence of less than 0.3. Single look complex (SLC) images were multi-looked to 1 and 3 looks in azimuth and range, respectively. We generated 108 interferograms covering the period of January 2017 to November 2020, each spanning 12 days. We manually selected the best interferograms and performed unwrapping with the Statistical-Cost, Network-Flow Algorithm for Phase Unwrapping (SNAPHU; Chen and Zebker, 2002).

190
Where observation was possible, the different methods are in agreement: the slope fractured and was displaced on the order of tens of metres prior to the eventual collapse. The 2021 rock-ice avalanche was also preceded by several other large avalanches -although these were primarily sourced from an adjacent hanging glacier.

Qualitative observations of slope change
We identified four main types of processes in our 31 year optical satellite image time series: 2. Minor snow or ice avalanches (2005, 2006, 2007, 2008, 2012, and 2015): Smaller volume avalanches, which may either have originated from the adjacent hanging glacier or the seasonal snowpack. These did not appear to infill the underlying valley with any significant quantity of material (with the exception of one ∼500 m long snow/ice deposit in May 2006). 3. Minor landslides avalanches (2007, 2009, 2011, 2012, 2013, and 2015): Minor rockfalls or rock avalanches originating from Ronti peak, or the weak sediment on the flanks of Ronti Gad. These also do not appear to have deposited major volumes of sediment.
4. Opening and widening of cracks at the headwall of the collapse block (2016-2021): Gradual opening of a wide crack in the north face of Ronti peak. 205 We only interpret the 4th process type (crack opening) is a real sign of pre-collapse conditions. Minor rockfalls and snow/ice avalanches are a common feature of high relief, high slope active landscapes. The major ice avalanches represent a serious geohazard in the upper Ronti Gad, but appear to relate to internal dynamics of the western hanging glacier rather than instability in the underlying bedrock. The area of these 2000 and 2016 major ice avalanches was estimated at 0.16 km 2 and 0.2 km 2 , with melting and/or redistribution of the the resulting valley floor deposits within three years of the event (Shugar et al. (2021); but less rapidly than the opening in 2016-2018. We confirmed these observations with several very-high ( 0.5 m) resolution images (Fig 3).

Optical feature tracking
Feature tracking provides the most complete spatio-temporal assessment of displacement of the methods used in this study -  in an apparent increase of ∼ 36 %, or ∼ 25-40 m. Overall, the feature tracking results demonstrate that the collapse block was mobile several years prior to its collapse in 2021.

InSAR maps
Even with knowledge of the location of the failed block, the processed interferograms do not allow for a pre-collapse identifica-250 tion of the instability on Ronti Peak. Of the 108 available interferograms, roughly half exhibited a complete loss of coherence, largely due to snow cover (November through May). Good quality interferograms are limited to summer months, and on the collapse block, coherence is only retained on the ice free part at the bottom of the wedge. The upper, glacier-covered part of the collapse block remains decorrelated, likely due to shadowing and glacier/snow cover. Figure 6a  A high quality interferogram from July 2020 (Fig. 6b) does not indicate any motion on the collapse block in the summer prior to the failure, but this cannot be assessed in other interferograms due to high noise levels. In less steep terrain north-west of the collapse block, the motion of a rock glacier (on the order of cm yr −1 ) can consistently be detected in the interferograms 260 (Fig 6b). Despite its sensitivity, InSAR is not able to provide any conclusive information about the pre-failure conditions of the collapse block in this challenging terrain.  Figure 6. Sentinel-1 radar backscatter amplitude from the ascending orbit (a), wrapped phase (0 to 2 π) ascending orbit interferogram from July 2020 (b) and corresponding false color image (c). Large areas of low coherence (masked as white) and patchy coverage illustrate the complexities of InSAR monitoring in high alpine terrain. The avalanche block is outlined in black and red.

Discussion
The pre-collapse motion of the avalanche block raises important questions about the causes and timing of the slope failure. In this section, we explore the answers to these questions using our multi-dataset observations, and then discuss the potential and 265 limitations of satellite data for remote hazard monitoring.

Three-dimensional block motion
We examined the three-dimensional motion of the collapse block as a first step towards understanding the Chamoli rock-ice avalanche collapse mechanism(s). Rotation and translation are the two primary modes of landslide motion (e.g. Záruba and Mencl, 2014), with each having a distinct surface displacement pattern. We used a combination of horizontal displacement

A possible avalanche triggering mechanism
A viable triggering mechanism for the Chamoli landslide must explain both the lag between the initial instability and collapse, and the timing of the collapse -in the middle of the winter. Syn-collapse seismic signals show that there was no seismic trigger 285 for the collapse (Pandey et al., 2021;Shugar et al., 2021;Cook et al., 2021). Nearby meteorological stations and reanalysis data reveal heavy snowfall and a 5 K positive temperature anomaly in the week preceding collapse, as well as a temperature inversion in the valley (e.g. Pandey et al., 2021;Dandabathula et al., 2021;Zhou et al., 2021;Shugar et al., 2021). On the longer term, this region has warmed ∼0.14 K per decade (Qi et al., 2021;Shrestha et al., 2021). Zhou et al. (2021) and Dandabathula et al. (2021) propose that this sudden temperature increase may have triggered the The stability of a slope can be described by the balance between two terms: driving forces (F D ) and resistive forces (F R ).
Driving forces are primarily gravitational, while resistive forces are primarily related to slope cohesion and friction. For a detached wedge such as the Chamoli collapse block, dominant resistive forces are likely friction along the margins and base of 300 the collapsed block. The balance between these two forces is known as the factor of safety F S: A slope is considered unstable when its factor of safety falls below 1 (e.g. Záruba and Mencl, 2014;Das and Sivakugan, 2016).
The Chamoli collapse area is composed of heavily jointed bedrock (e.g. Shugar et al., 2021). A pure translational pre-305 collapse motion is consistent with a collapse block basal shear plane along a single bedding plane. High-resolution postcollapse satellite imagery also suggests that the detachment occurred along a bedding plane. This failure plane may have been superficially weakened by freeze-thaw fracturing (Qi et al., 2021;Kropáček et al., 2021;Shrestha et al., 2021), or at greater depth by changes in permafrost conditions (e.g. Gruber and Haeberli, 2007;Krautblatter et al., 2013). The surface velocity peaks in summers 2017 and 2018 suggest that surface meltwater may have reached into the later failure surface. Meltwater 310 infiltration may directly impact friction (F D ), and in a delayed way also alter ground temperatures through advection of heat and release of latent heat upon refreezing. Gruber and Haeberli (2007) note that advection-driven melt of permafrost thaw corridors may drive destabilization of large volumes of rock. Deep permafrost thaw may occur over long timescales (e.g. Gruber and Haeberli, 2007;Krautblatter et al., 2013), and provides one potential explanation for the 5-year lag between initial instability and collapse.

315
The deep headwall crack provides accommodation space for cumulative snow accumulation and loading, and also limits the melting of accumulated snow by reducing its surface exposure. Observations of elevation change over 2015-2018 show the opening of a crack at least 25 m deep at the collapse block headwall (Fig 7a), although DEMs may underestimate the true depth of the crack due to viewing angle, slope geometry, and stereo DEM processing parameters. The purely translation model of block motion (Fig 7b) suggests that the true crack depth would have been closer to 150 m. Snow, ice, or rock debris loading 320 within a headwall crack would exert a horizontal force on the collapse block. This horizontal force ('push') acts to reduce the factor of safety both by directly increasing the driving force of the collapse block, and reducing the angle between the driving force vector and slope direction (equivalent to an increase in slope, see Appendix B).
Accumulation of snow or ice in the crack is visible in optical satellite imagery, with additional input from snow/ice avalanches from the overlying slope (e.g . Fig 3b-d). A storm in the days preceding the 7th February collapse brought substan-325 tial snowfall to the Chamoli region, with local snowfall estimates ranging from 8.5 to 48 mm water equivalent of precipitation (Shugar et al. (2021); estimates from local weather stations and Weather Research and Forecasting Model). We use these data to calculate the potential range of snow loading on the collapsed block, which is equivalent to a slope-parallel force of 7000-40,000 kN (Appendix A3). Considering the total precipitation between crack initiation (March 2016) and collapse (February 2021) this rises to 6.3 × 10 9 N to 9.9 × 10 9 N, or 2-3% of the total driving force of the collapse block.

330
In the absence of in-situ instrumentation and observations, it may not be possible to determine the exact cause of the failure at Chamoli. Nevertheless, we propose a mechanism which is compatible with both the lag between initial instability and collapse, and the timing of the eventual collapse. Snow and ice loading in the headwall crack would progressively increase the driving force of the collapse block, while meltwater infiltration and permafrost degradation in a bedrock fracture would steadily reduce its resistive forces (basal friction). The combination of these two processes would reduce the factor of safety and pre-condition 335 the block for failure, with the early February positive temperature anomaly and loading from snowfall providing a final driver for mid-winter collapse.

Future perspectives : remote-sensing based hazard monitoring
Our work on the Chamoli avalanche took place after the collapse, with the full knowledge of the position of the avalanche source. This work is useful for better understanding the conditions of the slope collapse. However, to be directly useful for 340 hazard monitoring and prevention, these techniques must identify avalanche locations and sizes before -rather than afterthey occur. The key questions therefore remain: would it have been possible to identify the Chamoli landslide prior to its collapse using the methods used in our study, and can these methods be applied elsewhere to identify future failures?
Several factors suggest that the available pre-collapse data may have been useful for identifying the Chamoli rock-ice instability. Careful qualitative analyses of optical satellite images, feature tracking, and DEM analysis show clear precursory signs 345 of slope failure around the Chamoli collapsed block. Satellite images show a crack growing over the 5 years prior to failure (Fig 3), feature tracking reveals tens of metres of horizontal displacement of the collapsed block, and DEM differences show tens of metres of vertical elevation change over the collapsed block. Combining this information with background knowledge about this region, such as the extreme relief, steep slopes, and historic avalanches, it would in principle have been possible to identify this as an unstable slope with high collapse potential.

350
While the data are sufficient to identify precursory signs of this rock-ice avalanche, there are important limitations to their use. The first key limitation is the very low signal to noise ratio of these data in the steep terrain most susceptible to slope failure. For feature tracking, the noise level of the composite 2016-2021 mean velocity maps is low (<1 m per year). However, the background noise level (as evaluated over stable bedrock) of individual velocity maps is much higher -and in some cases comparable to the magnitude of the signal (∼5-20 m per year). For the DEMs, artifacts range from metres to tens of metres in 355 scale, and additional "noise" is introduced by real elevation changes from glacier and snowpack change (Fig 8b). While these issues with false positives can be mitigated, this is challenging without knowing the signal of interest.
InSAR, while also being susceptible to false positives, is additionally prone to false negatives. The north-facing aspect of Ronti peak provides a twofold challenge: the illumination of the slope is limited (low backscatter , Fig 6a), and any motionassuming it is largely in the direction of the steepest slope -is oriented in the direction in which the radar instrument is least 360 sensitive. Additionally, the non-glacierized area of the collapse wedge is small, making it challenging to identify fringe patterns amongst the noise. Furthermore, with the largest velocities reaching tens of meters per year, the InSAR measurements are prone to phase aliasing and underestimation of the true displacement. Sentinel-1 InSAR would not have provided an adequate tool for monitoring in this case, even with knowledge of the location of the instability. The second key limitation is that none of the datasets produced in this work could predict the timing of collapse. While most 365 methods pick up precursory signs of slope failure, these begin almost five years prior to eventual collapse. In addition, the largest magnitude changes did not occur immediately prior to failure, but rather preceded failure by around three years. Even with the knowledge that the collapse occurred on 7 February 2021, there are no obviously anomalous signs that a failure was imminent in late 2020 or early 2021.
One final limitation is related to the immense size of hazardous areas relative to the scale of hazards themselves. The Chamoli 370 collapsed block had an area of around 0.25 km 2 , while the Himalaya cover over half a million km 2 . Any methods aimed at automatically detecting hazards prior to their occurrence must have a low 'false positive' (identified as a hazard in the database, but not of real concern) rate, or any resulting database will be populated primarily with these incorrectly flagged regions. This becomes a major challenge when considering the high incidence of noise and artifacts in feature-tracking derived displacement or DoD maps (e.g. Figure 8).

375
Overall, forecasting the 7 February 2021 Chamoli rock-ice avalanche prior to its occurrence from remotely sensed datasets would have been very challenging, and certainly not routine work using well-established methods. Current image resolution, characteristics, and processing algorithms result in noise levels on a similar order to the signal itself -although joint interpretation of feature tracking results, DEM differences, and satellite images does reveal clear precursory signs of slope instability.
In addition, none of the data in this study are able to adequately forecast the timing of collapse. As such, current archives 380 of satellite images do not currently appear to be practical for forecasting individual events. At the same time, this should not prevent remote monitoring of hazardous zones, particularly when adjacent to vulnerable areas. Every slope failure will exhibit a different range of pre-collapse signals, and new instabilities might be recognized in some cases. Even though the forecasting of individual events remains a challenge, these data have value for identifying zones of highest risk for in-situ monitoring or the installation of early-warning systems (Cook et al., 2021).

385
Feature-tracking, DEM difference, and InSAR datasets can be processed and analyzed on a regional or even global scale -and in many cases pre-processed datasets are already available online (e.g. Morin et al., 2016;Gardner et al., 2018). While these pre-processed datasets are not generally produced for slope stability monitoring, they can be used to improve hazard maps and reduce landslide related damage. Future advances in Earth observation satellite capabilities and processing algorithms will improve the quality of such products.

Conclusions
The deadly 7 February 2021 Chamoli rock-ice avalanche was initiated by failure of >25 Mm 3 of rock and ice high in the Uttarakhand Himalaya. We investigated the conditions of the avalanche source zone over the decades preceding collapse through a combination of optical and radar satellite images. We used feature tracking to calculate horizontal slope displacements, and differenced photogrammetrically generated DEMs to investigate vertical displacements. We showed that the collapsed block 395 moved 20-30 m prior to its collapse, with most rapid motion occurring around 3 years prior to failure. Comparison between our datasets and synthetic displacement maps shows that the motion occurred primarily via down-slope translation, opening up a deep crack at the headwall. A combination of permafrost degradation and snow and ice debris loading within this headwall crack may explain both the lag between initial instability and collapse, and the mid-winter timing of the collapse. Finally, we assessed the potential of these datasets and approaches for monitoring other unstable slopes. While they were effective at 400 identifying precursory signals at a known collapse site, it remains very challenging to predict such collapses with sufficient levels of confidence in high-mountain areas.
The factor of safety F S is calculated from the balance driving and resistive forces (e.g. Záruba and Mencl, 2014;Das and Sivakugan, 2016).: In which A is slip surface area, C is cohesion, M is the mass of the unstable region, g is gravity, α is slope, and φ is the friction angle. A system may be considered unstable when the factor of safety falls below 1.
Introducing an additional horizontal force F H modifies this balance in two ways: firstly by increasing the driving force, and secondly by altering the angle between the driving force vector and resistive forces vector: The change in angle of the driving force vector α is then given by α = arctan( F H M gsin(α) ). In our situation, for a given mass accumulated in the headwall crack M C we have F H = M C sin(α).
The pre-event storm brought 8.5 to 48 mm water equivalent of precipitation (Shugar et al. (2021); estimates from local weather stations and Weather Research and Forecasting Model). We may use this data to calculate possible loading of this snow on the collapsed block -considering a 500 m long, 70 m wide crack with a 500 m long and fed by a 180 m wide 445 avalanche zone. Assuming that all of the snowfall was channeled into the crack, total loading M C would be equal to: With A A being the accumulation area feeding the crack, P being precipitation (in metres), ρ P being the density of the precipitation. Total snow loading in the headwall crack associated with this single precipitation event would therefore be 10000-60000 kN, equivalent to a slope-parallel horizontal force of 7000-40000 kN.

450
GPM IMGERG precipitation data suggests that around 9 ± 2 m of precipitation fell in the collapse area between crack initiation in 2016 and collapse in 2021. Using the same calculation, maximum snow load in the headwall crack is equal to 8.6-13.5 × 10 9 N, equivalent to a slope-parallel horizontal force of 6.3-9.9 × 10 9 N. For reference, the estimated total driving force of the collapse bloc, composed of 21 Mm 3 of rock and 6 Mm 3 of ice, is 4.0 × 10 1 1 N.
Author contributions. All authors designed the study and conducted the research. MVWDV wrote the paper, with input from all co-authors.

455
The final version has been approved by all co-authors.
Competing interests. The authors declare no competing interests.