Article Figures & Data
Figures
- Fig. 1.
Surface of a debris-covered glacier: Ngozumpa Glacier in the Nepal Himalaya. Photo taken in 2008; credit: A. Racoviteanu.
- Fig. 2.
Components of a landsystem model for debris-covered valley glaciers. Relative positions of different surface features are indicative, as supraglacial features can exist in numerous configurations (credit: Gareth Evans and Naomi Lefroy).
- Fig. 3.
Decadal evolution of lakes in the Zemu basin of Sikkim Himalaya based on remote sensing: (a) 1962 panchromatic Corona KH4 imagery (7.5 m); (b) 2001 ASTER image; (c) 2006 Quickbird image (2.4 m) (d) 2020 PlanetScope image (3 m). All multispectral images are shown as colour composites (bands 3, 2 and 1) (revised and expanded from Racoviteanu et al. 2015).
- Fig. 4.
Stages in the evolution of the surface and equilibrium line altitude of a Himalayan debris-covered glacier. Adapted from Benn et al. (2012). Credit: Gareth Evans.
- Fig. 5.
Example plot of the output of a hazard ranking scheme shown from the analysis of 41 glacial lakes in the Pumqu catchment in Tibet.
- Fig. 6.
Elements of a hazardous moraine-dammed glacial lake showing the key stages of a glacier lake outburst flood: (1) propagation of displacement or seiche waves in the lake, and/or piping through the dam; (2) breach initiation and breach formation; (3) propagation of resultant flood wave(s) down-valley. Key triggers are labelled A to F: (A) glacier calving; (B) icefall from hanging glaciers; (C) rock/ice/snow avalanches; (D) dam settlement and/or piping; (E) ice-cored moraine degradation; (F) rapid input of water from supra-, en-, or subglacial (including subaqueous) sources. Conditioning factors are labelled a to d: increased lake volume, low dam width to height ratio, ice degradation, minimal freeboard (modified from Richardson and Reynolds 2000; Westoby et al. 2014).
Tables
- Table 1.
Debris properties of interest for various applications, and remote sensing techniques used on previous studies to estimate them
Property Data source/technique Existing studies Debris lithology Medium to high optical remote sensing and hyperspectral data combined with in situ field spectrometry Casey et al. 2012; Casey and Kääb 2012 Debris grain size High-resolution imagery, SfM Miles et al. 2017b; Detert and Weitbrecht 2020 Debris layer porosity or bulk density No remote sensing method known yet; may be possible with polarimetric SAR Water content No remote sensing method known yet; may be possible with passive microwave Collier et al. 2014; Giese et al. 2020 Supraglacial vegetation Normalized Difference Vegetation Index, NDVI; spectral unmixing Fickert et al. 2007; Racoviteanu et al. 2021 Thermal conductivity No remote sensing method known yet — Broadband albedo Formulae for anisotropy correction and narrow to broadband conversion, though not developed for rock debris specifically Knap et al. 1999; Liang 2001; Greuell and Oerlemans 2004; Naegeli et al. 2017; Xu et al. 2020 Technique Data source Notes Existing studies Manual delineation of ice cliff crest or area Satellite imagery (high resolution) including UAV Time consuming and subjective; limited application at large scales Sakai et al. 2002; Han et al. 2010; Brun et al. 2016; Steiner et al. 2019; Stefaniak et al. 2021 Use of thermal imagery ASTER, Landsat Spatial resolution is challenging Herreid and Pellicciotti 2018 Feature detection (OBIA) High-resolution imagery including DEM Somewhat time consuming to set up the rules and may need some post-processing Kraaijenbrink et al. 2016; Watson et al. 2017; Mölg et al. 2019 DEM slope thresholding/topography relief metrics High-resolution DEM Can be used as proxy for ice cliff presence; high resolution DEMs limited in some areas Reid and Brock 2014; Herreid and Pellicciotti 2018; King et al. 2020a SAR intensity tracking High-resolution SAR, e.g. TerraSAR-X, IceEye Potentially useful in areas where cloud-free optical data not available; other confounding surfaces no studies yet Optical broadband adaptive thresholding High-resolution optical imagery Fast, requires optimization of threshold Anderson et al. 2021 Multispectral thresholding; adapted linear spectral unmixing High-resolution multispectral optical imagery Accurate, requires optimization Kneib et al. 2020 - Table 3.
Existing remote sensing techniques to map supraglacial ponds on debris-covered glaciers
Technique Data source Notes Existing studies Manual delineation Any high-resolution data (<10 m) Time consuming and subjective but probably most accurate Iwata et al. 2000; Salerno et al. 2012; Thompson et al. 2012; Watson et al. 2016; Miles et al. 2017b Band ratios or normalized-difference such as NDWI Landsat, Sentinel-2, ASTER, etc. (>10 m) Manual/semi-automatic thresholding; can be improved using Planet microsatellite constellation which enables rapid-repeat monitoring at 3 m resolution Wessels et al. 2002; Bolch et al. 2008a; Gardelle et al. 2011; Liu et al. 2015; Narama et al. 2017; Miles et al. 2017a; Watson et al. 2018a; Kneib et al. 2020 SAR backscatter intensity Sentinel-1, TerraSAR-X, ENVISAT ASAR, ALOS PALSAR, ERS-1, 2 etc. The oblique geometry of radar imaging creates challenges for pond identification (and especially for monitoring) due to the highly variable topography of debris-covered glaciers Strozzi et al. 2012; Wangchuk and Bolch 2020; Zhang et al. 2021 Feature extraction via decision-trees and/or OBIA Landsat, Sentinel-2, Pleiades etc. Rules needed are subjective; it often commercial software Panday et al. 2011; Liu et al. 2015; Kraaijenbrink et al. 2016 Sub-pixel spectral analysis Landsat, Sentinel, ASTER Upon careful end-member collection, ideally from field spectrometry; can be applied automatically at large scales Panday et al. 2011; Scherler et al. 2018; Kneib et al. 2020; Racoviteanu et al. 2021 Thermal imaging Landsat, ASTER, UAV • Supraglacial ponds are usually considerably cooler than surrounding debris (during the day, debris can reach 25°C or more) • • Application of thermal imagery is limited by relatively coarse resolution (90–100 m), potentially most suitable at present for UAV thermal imagery when existent Suzuki et al. 2007 - Table 4.
Trigger potential and threshold parameters for glacial hazard assessment (modified from RGSL 2003; Reynolds 2014)
Parameter affecting hazard Score 0 2 10 30 50 Threshold factors 1 Effective volume of lake water available for flood N/A Low Moderate Large Very large 2 Height of freeboard relative to lake level No dam Very high High Moderate Low 3 Width/height ratio of terminal moraine dam w ≫ h w > h w ≈ h w < h w ≪ h 4 Gradient of distal moraine dam Flat-5° 5–10° 10–25° 25–40° >40° Triggering factors 5 Height of glacier ice cliff and calving potential No cliff Low Moderate High Very high 6 Ice/rock avalanche into lake Open basin Low Moderate High Very high 7 Thermokarst degradation of ice within terminal moraine No ice Low Moderate High Very high 8 Buoyancy of submerged stagnant ice based on possible ice volume No ice Low Moderate High Very high
Debris-covered glacier systems and associated glacial lake outburst flood hazards: challenges and prospects
- Article
- Abstract
- The debris-covered glacier landsystem: concept and components
- Tools for observing and monitoring the debris-covered glacier landsystem and its components
- Response of the debris-covered glacier landsystem to climate change
- Strategies for assessing the hazard potential of a glacial lake
- Remaining challenges and limitations
- Conclusions and outlook
- Acknowledgements
- Author contributions
- Funding
- Competing interests
- Data availability
- References
- Figures & Data
- Info & Metrics