1. Introduction
Mass loss from glaciers worldwide has accelerated over recent decades, compounding concerns relating to water security as mountain glaciers diminish (Immerzeel and others, Reference Immerzeel2020). Glacier behaviour at high elevation is modified by spatially variable factors, such as the dominant climatic regime, local relief, and the presence or absence of a debris cover. Debris-covered glaciers – those with a terminus mantled by a largely continuous layer of supraglacial debris (Miles and others, Reference Miles2020) – occur across all glacierised mountain ranges. While supraglacial debris thicker than a few centimetres suppresses surface ablation, debris-covered glaciers are experiencing rates of net mass loss similar to their clean-ice counterparts (Brun and others, Reference Brun2019). Spatial patterns of mass loss from debris-covered glaciers are more complex than those from climatically equivalent clean-ice glaciers, due to the presence of surface features (e.g. debris, ponds, and ice cliffs) that promote differential ablation and influence englacial conditions and glacier dynamics (Miles, ES and others, Reference Miles, Steiner, Buri, Immerzeel and Pellicciotti2022). While supraglacial properties and processes at high-elevation debris-covered glaciers have been relatively well studied, englacial measurements are limited. This lack of empirical data raises challenges for the accurate representation of ice flow in models of glacier response to anticipated climate change.
Here, we first outline field-based studies that have contributed to current understanding of the interior of high-elevation debris-covered glaciers, focusing largely on Himalayan glaciers. Combining this with our own experience during the EverDrill project, we then suggest directions for future field-based research that will provide data to improve model projections of the behaviour of high-elevation debris-covered glaciers.
2. Summary of previous research on the interior of high-elevation debris-covered glaciers
2.1 Ice thickness
The thickness of debris-covered glaciers is likely affected by their distinctive mass balance and dynamic characteristics, and, in the Himalaya, their unique climatic and geomorphic context. Despite the importance of these data for calibrating glacier models, few direct measurements exist for high-elevation debris-covered glaciers, with only 11 records over High Mountain Asia in the recent GlaThiDa v3 database (Welty and others, Reference Welty2020). Ground-penetrating radar (GPR) or radio-echo sounding has been used at Khumbu and Lirung Glaciers, Nepal Himalaya (Gades and others, Reference Gades, Conway, Nereson, Naito and Kadota2000), and Koxkar and Southern Inylchek Glaciers, Tien Shan (Macheret and others, Reference Macheret1993; Zhen and others, Reference Zhen, Shiyin, Shiqiang and Honglang2013) to identify the glacier bed in the ablation area and thus estimate ice thickness. Gravimetric methods have given ice thickness profiles on Khumbu and Rongbuk Glaciers, located on the slopes of Mt. Sagarmatha (Chomolungma/Everest) (Sinica, Reference Sinica1975; Moribayashi, Reference Moribayashi1978). A limited number of boreholes (e.g. Miles and others, Reference Miles2019b) or ice cores (e.g. Liu and others, Reference Liu, Hou, Wang and Song2009) have reached the bed of high-elevation debris-covered glaciers, yielding point measurements of ice thickness and allowing the installation of sensors to measure englacial properties through the full ice column.
2.2 Ice temperature
High-elevation debris-covered glaciers are likely to be thermally distinctive because of some combination of having: a high-elevation accumulation area; considerable mass accumulation via avalanching; an icefall; a quasi-stagnant tongue hosting numerous, often large, supraglacial ponds; a cover of surface debris of variable thickness; and, in some regions, a dominance of summer accumulation. The englacial temperature field is important for determining ice viscosity/softness and also the presence and location of temperate ice, which affects both glacier hydrology and motion, and is thus critically important in understanding how glaciers are responding to climate change. Direct measurements of ice temperature have been made at depths of <20 m in boreholes on Khumbu Glacier (Mae and others, Reference Mae, Wushiki, Ageta and Higuchi1975; Miles and others, Reference Miles2021b) and Rongbuk Glacier (Sinica, Reference Sinica1975; TSECAS, 1975; Zhang and others, Reference Zhang2013), recording seasonally influenced ice temperatures. However, only a few measurements have been made in deeper boreholes, thereby recording longer-term ice temperature, such as in the accumulation areas of East Rongbuk Glacier (coldest $-8.9^\circ$
C at 108.8 m depth) (Hou and others, Reference Hou2007), Gyabrag Glacier ($-9.0^\circ$
C at 68.5 m depth) (Liu and others, Reference Liu, Hou, Wang and Song2009) and Dasuopo Glacier, Chinese Himalaya ($-13.8^\circ$
C at the glacier bed, ~150-168 m depth) (Thompson and others, Reference Thompson2000). Deep englacial ice temperature measurements in the ablation area of Khumbu Glacier demonstrated a polythermal regime, with warm ice towards the bed and the coldest ice (only $-3.3^\circ$
C) at ~20 m depth in the upper debris-covered area (Miles, KE and others, Reference Miles2018).
2.3 Ice deformation
Ice deformation is one of the primary glacier motion components and can be used to isolate basal sliding from net surface velocity. However, englacial ice deformation is difficult to measure directly; to our knowledge, no direct measurements of englacial ice deformation have been reported for high-elevation debris-covered glaciers. Instead, the ice deformation field of such glaciers has almost exclusively been reconstructed from theory or inferred from field-based structural mapping (which is hampered by the supraglacial debris covers of these glaciers). For example, interpretations of wave ogives on Khumbu Glacier have informed on ice motion through the Khumbu Icefall (Iwata and others, Reference Iwata, Watanabe and Fushimi1980; Hambrey and others, Reference Hambrey2008), while planes interpreted as thrust faults on the same glacier have been used to infer ice pathways (Fushimi, Reference Fushimi1977; Hambrey and others, Reference Hambrey2008). Subsurface structural features were observed in boreholes on Khumbu Glacier, including rotated layers of primary stratification (layers of ice and debris) and relict layers of basally derived sediment, indicating a formerly more dynamic glacier regime (Miles, KE and others, Reference Miles, Hubbard, Miles, Quincey and Rowan2022).
2.4 Englacial hydrology
The drainage and retention of meltwater within high-elevation debris-covered glaciers has important implications for the provision of water resources and the moderation of seasonal flow in rivers downstream. Khumbu and Ngozumpa Glaciers in the Nepal Himalaya are well studied in terms of their subsurface hydrology. Glacio-speleological studies have revealed features such as englacial cut-and-closure channels and an englacial water base-level controlled by the depth of an ice-marginal proglacial lake at the terminus of Ngozumpa Glacier (Gulley and Benn, Reference Gulley and Benn2007; Benn and others, Reference Benn2017). Field-based investigations have inferred englacial meltwater storage and transit through Khumbu and Lhotse Glaciers, Nepal Himalaya (Rounce and others, Reference Rounce, Byers, Byers and McKinney2017; Miles, ES and others, Reference Miles2018) and englacial channels controlling the base level of closely located supraglacial ponds on a number of debris-covered glaciers in the Tien Shan (Narama and others, Reference Narama2017). GPR surveys have located englacial and subglacial streams on Koxkar Glacier (Zhen and others, Reference Zhen, Shiyin, Shiqiang and Honglang2013), and dye-tracing studies have demonstrated subglacial to supraglacial meltwater transit and temporary meltwater storage in shallow englacial reservoirs within the debris-covered tongue of Khumbu Glacier (Miles and others, Reference Miles2019a).
2.5 Sediment properties
Knowledge of englacial debris concentrations is necessary to quantify high-elevation debris-covered glacier erosion rates and to decipher the contribution of englacial debris melt-out to the supraglacial debris layer, as well as to test simplified sediment input approaches to numerical models (Rowan and others, Reference Rowan, Egholm, Quincey and Glasser2015). Most such measurements are based on surface ice layers, such as an englacial debris concentration of 0.12% by weight over an area of 0.028 km2 on Djankuat Glacier, Caucasus (Bozhinskiy and others, Reference Bozhinskiy, Krass and Popovnin1986). Deeper englacial debris concentrations, reconstructed from optical televiewer logs of boreholes up to 192 m deep on Khumbu Glacier, varied with depth, with a mean of 0.7% by volume across four boreholes and a maximum of 6.4% by volume in a borehole located nearest (~1 km) to the terminus (Miles and others, Reference Miles2021a). Characterisation of debris facies (e.g. Hambrey and others, Reference Hambrey2008) has informed on debris sources from surrounding hillslopes and relocation within the glacier system.
3. Future research directions
3.1 Englacial data for glacier evolution models
The rationale for recording field-based data from high-elevation debris-covered glaciers is to improve model projections of the future behaviour of these glaciers, the results of which are paramount for addressing regional issues such as water security (Immerzeel and others, Reference Immerzeel2020). Dynamic glacier models both calculate glacier mass balance over time and simulate ice flow and consequent geometric change (Zekollari and others, Reference Zekollari, Huss, Farinotti and Lhermitte2022). Although such models have been widely used in glaciology, their application to high-elevation debris-covered glaciers requires a bespoke approach (e.g. Rowan and others, Reference Rowan, Egholm, Quincey and Glasser2015; Anderson and Anderson, Reference Anderson and Anderson2016) due to: (i) the interaction of mountainous topography with regional weather systems; (ii) differential ablation regimes that depend on feedbacks between mass balance, ice flow, and debris transport, commonly involving a low- or even reverse-angle lower ablation area with abundant surface meltwater ponds; and (iii) complex stress regimes reflecting ice flow through high-relief topography, including extensive lateral and terminal moraines.
Empirical englacial data are most useful if available at sufficiently high spatial resolution to be able to guide, calibrate, and evaluate higher-order models, which can represent the geometrical and associated complexities of high-elevation debris-covered glaciers. Such a resolution would ideally involve detailed studies of individual glaciers, with data collected systematically over the glacier surface and with depth through the ice column, allowing the efficient gridding of data into three dimensions. In particular, observations on a single glacier should be made across the full glacier width at multiple elevations, as well as at locations where measurements might be expected to vary, such as either side of confluences with tributary glaciers and around supraglacial ponds. While more systematic measurements are needed from more glaciers to understand the expected spatial variation in all englacial features, comprehensive studies on a single glacier will be more valuable for modelling than sparse measurements across many glaciers. Large-scale, multi-component, integrated projects could allow detailed measurements on individual glaciers to be upscaled for model input.
The value of empirical datasets is increased further if they can be incorporated into models at physically meaningful temporal scales and resolutions. First, long-term decadal to centennial data are needed because glacier response time to adjust to external forcing is typically tens to hundreds of years, though their reaction time is typically multiple years. Second, seasonal weather, and hence mass-balance patterns, are so distinctive at many high-elevation debris-covered glaciers (e.g. monsoon-influenced areas) that models should aim to resolve at the seasonal as well as the annual scale. Sub-seasonal data collection could be facilitated by the use of methods that transmit data remotely, allowing data access and equipment monitoring through seasons that are more challenging for conducting fieldwork (e.g. winter or monsoon months).
3.2 Future research directions
Based on the above, we make the following suggestions for future research directions that would allow significant advances in modelling capabilities.
• Ice thickness measurements, which could be achieved through wider application of GPR systems mounted on helicopters (Pritchard and others, Reference Pritchard2020) or uncrewed aerial vehicles (UAVs). Full-depth boreholes could evaluate such spatially extensive geophysical surveys by giving point-based ice thicknesses, as well as allowing further englacial measurements from borehole wall logs and ice cores. Enhanced community efforts to gather and compile unreported observations could further increase the coverage of such datasets (e.g. Welty and others, Reference Welty2020).
• Ice temperature measurements of the basal thermal field at high-elevation debris-covered glaciers, which would inform on meltwater production and percolation, latent heat release, and the proportion of glacier motion by basal sliding. This could be recorded at high spatial resolution by borehole-based distributed temperature sensing using fibre-optic cables. The long-term installation of such cables in full ice depth boreholes in both the accumulation and ablation areas of high-elevation debris-covered glaciers would provide an extremely valuable dataset.
• Basal motion, resolved into basal sliding and ice deformation components by measuring: (i) surface velocity using a network of GNSS stations; and (ii) subsurface deformation through distributed acoustic fibre-optic sensing in full ice depth boreholes, calibrated with discrete borehole probes measuring ice tilt.
• Mass balance from seasonal and long-term accumulation records and trends, measured by combining meteorological stations with analysis of snow and firn cores from accumulation areas and/or borehole logging using an optical televiewer or a borehole camera.
• Englacial meltwater storage, both over the short term in shallow englacial reservoirs and over longer timescales from refreezing in crevasses, firn, or the supraglacial debris layer, to better predict timing of future meltwater delivery downstream. This could be achieved through glacier-scale water balance studies, supported by dye tracing to elucidate timescales of water storage in the system.
• Runoff quality and quantity, including pollution and sediment levels, measured through water sample analyses and long-term hydrological stations installed on proglacial streams.
• Englacial sediment concentrations, which could be measured directly by obtaining deep ice cores or indirectly by logging a borehole with an optical televiewer. Debris emergence could be further investigated by measuring the sediment concentration of shallow cores along a flowline.
• Subglacial debris properties, measured by sampling ice and sediment in the proglacial area and at the base of full ice depth boreholes.
Acknowledgements
This research was supported by the ‘EverDrill’ Natural Environment Research Council Grant awarded to Aberystwyth University (NE/P002021) and the Universities of Leeds and Sheffield (NE/P00265X). A. V. R. was supported by a Royal Society Dorothy Hodgkin Research Fellowship (DHF/R1/201113). We thank Himalayan Research Expeditions, Mahesh Magar, and Sagarmatha National Park for supporting fieldwork in Nepal.
Open access
