Contextualizing lobate debris aprons and glacier-like forms on Mars with debris-covered glaciers on Earth

2.1 Glacier mass balance and ice flow

Terrestrial glaciology was founded in observing and understanding how ice deforms, how ice responds to climate, and how these processes control glacier, ice-sheet, and glacial landform evolution. This body of work across glaciology is represented in numerous textbooks (e.g. Benn and Evans, 2010; Cuffey and Paterson, 2010; van der Veen, 2013), as well as in recent review papers (e.g. Adhikari and Marshall, 2012; Mackintosh et al., 2017; Marshall, 2005). In addition, aspects of glaciology relevant to this review intersect with glacial geology (e.g. Bennett and Glasser, 2009; Hudleston, 2015), glacial geomorphology (e.g. Anderson and Anderson, 2010; Bingham et al., 2010; Hubbard and Glasser, 2005) and studies of the periglacial environment (e.g. French, 2018). Here, the concepts of glacier mass balance and ice flow are summarized in order to set the stage for the following discussion on active debris-covered glaciers; see the preceding textbooks and literature for a more comprehensive review. It is important to note that past and present mass balance are not known for ice masses on Mars, and assumptions about Martian ice-flow history should be considered in context with what we know about ice flow on Earth.

The mass balance of a glacier accounts for gains and losses of mass at the glacier surface (surface mass balance), within the glacier (internal mass balance), and at the glacier bed (basal mass balance). For land-terminating glaciers, key processes control the mass exchanged between the atmosphere, glacier, and land. These processes operate as a function of climate conditions, altitude, and environmental setting, which differ between mountain glaciers, large ice caps, the Greenland Ice Sheet, and the Antarctic Ice Sheet. Cogley et al. (2011: 61) defined mass balance as “the change of the mass of the glacier, or part of the glacier, over a stated span of time,” where the dimension is typically taken to be a mass and where the mass–balance rate has dimension mass per time.

In general, surface mass exchange may include snowfall accumulation, hoar frost, diamond dust, freezing rain, avalanche accumulation, snow/ice melt, refreezing of meltwater, sublimation, wind scouring or redeposition, and calving of ice blocks at the glacier terminus. Temperature is a dominant control on spatial gradients in surface mass exchange. In general, the net of these processes is tallied as surface accumulation (mass gains) and surface ablation (mass losses), the sum of which is the surface mass balance. The zone where net surface accumulation occurs in a given year is referred to as the accumulation zone, and the zone where net ablation occurs is referred to as the ablation zone, and the point that divides these zones is often referred to as the equilibrium line. Internal mass exchange occurs primarily from refreezing of meltwater and ice melt driven by either water flow or strain heating of ice during ice deformation. Basal mass exchange for land-terminating glaciers may include freeze-on at a land-ice boundary (accreted ice), as well as melting of ice at the bed from water flow, basal sliding, and/or elevated heat flux. Surface, internal, and basal mass balance all affect the glacier state.

Glacier mass balance is often calculated annually from point stake measurements on the glacier surface, where the height of the surface in winter and in summer is calculated to estimate the net gains and net losses at specific locations, and then extrapolated over the glacier area. Geodetic estimates are also made when digital elevation models are available and can be differenced to get the net surface-elevation change over a time period, where the snow/firn/ice density must be assumed in order to convert elevation change to mass change. Satellite gravimetry has also enabled systematic, regional-to-global-scale measurements of changes in surface snow and ice since 2003 (e.g. Gardner et al., 2013; Jacob et al., 2012; Tapley et al., 2019, and references therein). Observed mass-balance fluctuations are archived from glaciers around the world (e.g. Zemp et al., 2009; hosted by the World Glacier Monitoring Service) and analyzed with respect to climate fluctuations (e.g. Zemp et al., 2015, 2019). The general factors controlling the relationship between regional/local climate and glacier mass balance include surface mass and energy exchange, calving, accumulation, and ablation – which include feedbacks with ice flow that are influenced by glacier slope, thickness, and surface altitude (Cuffey and Paterson, 2010: 93). Owing to non-linear ice dynamics, glacier response lags the precipitation and temperature forcing (e.g. Jóhannesson et al., 1989; Roe and Baker, 2014). Although discussed in terms of a balance state, glaciers are unlikely to be exactly in balance in any given year but may be close to balance averaged over multiple years; since interannual variations are typical, it is the multi-year to multi-decade state that is most relevant when considering if a glacier is in or out of balance with regional and local climate. Important to consider with respect to debris-covered glaciers is that the mass-balance gradient as a function of altitude can be different compared with debris-free glaciers (e.g. Bisset et al., 2020), indicating that debris cover insulates ice from ablation.

The flux of ice delivered by glacier flow can be considered using glacier mechanics (dynamic flux) and glacier kinematics (mass flux). Ice flows due to the force of gravity, and the flow rate is affected by conditions at the basal and sidewall boundaries, as well as ice properties and the glacier geometry. Ice deformation rate can be altered due to variations in the physical properties of ice, such as grain size, crystal orientation, and impurity content that can enhance or retard the strain rate compared to the value calculated by the isotropic creep relation. When these variations are concentrated in the basal layer of a glacier or ice sheet, their effect on ice-flow rate can be especially significant (e.g. Hubbard and Sharp, 1989; Knight, 1997). Glacial ice also flows in response to changes in climate, where thickness and length change when snowfall adds mass and when melting removes mass; in general, a glacier tries to achieve a state near equilibrium with its boundary conditions. Active glaciers undergo active flow, and active glaciers are typically not far out of equilibrium with the climate and environmental conditions that drive their evolution. However, since it takes decades for most mountain glaciers to fully respond to changes in climate, sustained warming over the past century means that many glaciers are far from being in equilibrium with the present-day climate (e.g. Christian et al., 2018). When glaciers are in a significantly negative balance state they will retreat, thin, and possibly stagnate. Depending upon a glacier’s history and upon environmental conditions, an active and ice-rich glacier may evolve into a multitude of glacier states and/or glacial landforms, some of which are discussed in subsequent sections of this review.

Internal deformation of ice occurs by creep and also by large-scale folding and faulting when creep cannot support the stresses incurred in the ice (e.g. Hudleston, 2015). Sliding can occur at a glacier bed, and typically depends on the basal thermal regime and basal material. The stress–strain relationship for ice (i.e. the flow law) and conservation of mass, momentum, and energy govern how a glacier moves, and how the flow rate may change along the glacier length, as a function of depth within the glacier over time. Driving and resistive forces can be summed to analyze the force balance, components of which can be calculated separately to help determine the dominant flow processes and may be simplified depending on the glacier state and on the problem being addressed. Our understanding of ice dynamics has been applied to interpret the evolution of glacial and icy masses on not just Earth and Mars, but throughout the Solar System (e.g. Sori et al., 2017b; Umurhan et al., 2017).

2.2 Debris-covered glaciers

All mountain glaciers and ice-sheet outlet glaciers typically have some fraction of debris, including dust and/or volcanic ash, but select glaciers have a larger fraction of material that can influence the rate of ice flow and the amount of ice melt in the ablation zone. The role of debris is significant: a thin cover on the surface can enhance melt, a thick cover may insulate underlying ice, and entrained material may increase or decrease the flow rate (e.g. Østrem, 1959). Debris-covered glaciers are defined as having part of the ablation zone covered with dust, debris, and/or ash across its width (see, e.g., Kirkbride, 2011). The material covering ice-rich debris-covered glaciers can range from large rocky debris to finer volcanic ash, and may have accumulated on the surface from the top down (e.g. rockfall, rock avalanches, landslides, volcanic eruptions) or from the bottom up (e.g. entrained englacial/subglacial material that emerges due to ice flow and melt). Debris-covered glaciers on Earth do not have to be entirely covered by debris, and it is common that debris predominantly covers the lower glacier and/or terminus region. Some authors refer to “partially debris-covered glaciers,” which can mean that the boundary between debris cover and clean ice is either distinct or gradual between the upper and the lower glacier, or that only part of the glacier along its length has debris cover. This may arise naturally based on debris source and ice flow, due to moraine formation or due to avalanche and/or ashfall events. While many active debris-covered glaciers could be referred to as partially debris-covered, in the literature some glaciers have been specifically designated in this way (e.g. Kellerer-Pirklbauer et al., 2008; Lundstrom et al., 1993).

Debris-covered glaciers are found around the world (Figure 1; Herreid and Pellicciotti, 2020; Scherler et al., 2018), typically in mountain ranges with high relief and with significant supplies of debris, such as in the Himalaya (e.g. Kääb et al., 2012; Scherler et al., 2011a), Karakoram (e.g. Senese et al., 2018), Caucasus (e.g. Stokes et al., 2007), and parts of the Andes (e.g. Janke et al., 2015), as well as in Alaska (e.g. Berthier et al., 2010), New Zealand (e.g. Anderson and Mackintosh, 2012; Kirkbride, 1995), Antarctica (e.g. Bockheim, 2014; Shean and Marchant, 2010), Iceland (e.g. Björnsson and Pálsson, 2008), Greenland (e.g. Humlum, 2000), Svalbard (e.g. Etzelmüller et al., 2000), Canada (e.g. Bolch et al., 2010), and the west coast of the continental United States (e.g. Fountain et al., 2017; Moore et al., 2019; Pelto, 2000). Figure 2 showcases a variety of forms from glacial environments around the world, where differences in climate conditions, debris sources, and landscape geometry can all contribute to the evolution of different modern debris-covered glacier structures and associated landforms.

OPEN IN VIEWER

Figure 1.

Debris coverage on glaciers around the world given as % by area and estimated from Landsat-8 satellite imagery between 2013 and 2015 with data binned by area with a tile size of 1°×1°.

OPEN IN VIEWER

Figure 2.

Examples of debris-covered glacier termini from around the world. (a) Aerial photo of Mullins Glacier in Beacon Valley, Antarctica (Source: USGS aerial photo TMA 3080/275 available from Polar Geospatial Center). (b) Miage Glacier, Italy in 2012 (Source: Smiraglia et al., 2015; Figure 8). (c) Emmons Glacier, Mount Rainier, WA, USA in July 2020; note person for scale (Photo: C. Todd). (d) Ngozumpa Glacier, Nepal in 2010 (Source: D. Breashears, GlacierWorks), which is the longest glacier in the Himalaya at ∼36 km.

2.3 Rock glaciers and ice-cored rock glaciers

Rock glaciers are differentiated from debris-covered glaciers based on being entirely covered by rock (where rock can be taken as synonymous with debris, and characterized as having a component that is coarse and angular), with an ice-rich or inter-layered ice/debris core, and moving downslope by creep deformation at a rate typically slower than debris-covered or clean-ice glaciers. When active, rock glaciers continue to lengthen by creep (e.g. Kääb and Reichmuth, 2005) and are prevalent in periglacial domains where debris-covered and/or clean-ice glaciers may also be found; for example, in the Western USA (e.g. McCabe and Fountain, 2013; Millar et al., 2013), northern Iceland (e.g. Fernández-Fernández et al., 2020; Lilleøren et al., 2013), western Greenland (e.g. Humlum, 2000), the Nepalese Himalaya (e.g. Jones et al., 2018a), the European Alps (e.g. Springman et al., 2012), the Andes (e.g. Janke et al., 2015; Monnier and Kinnard, 2015), and Antarctica (e.g. Bockheim, 2014; Guglielmin et al., 2018; Swanger et al., 2010). See also the recent review by Jones et al. (2019a).

The origin of the ice-rich cores of the tens of thousands of different rock glaciers around the world (see Jones et al., 2018b for an inventory) is not necessarily a simple and settled matter. Reviews on the topic address this decades-long discussion in the scientific community between the role of periglacial and glacial processes in rock-glacier formation, and whether the definition should follow morphology or origin (e.g. Anderson et al., 2018; Berthling, 2011; Clark et al., 1998; Haeberli, 2000; Haeberli et al., 2006; Jones et al., 2019b; Knight, 2019; Millar and Westfall, 2008). As described in these reviews and recent work, a periglacial origin for a rock glacier means that the action of permafrost is required and rock glacier core ice is derived from the freezing of water in the pore space. A glacial origin means that the ice-rich core is derived from a glacier (that may have been debris covered), and is typically considered as an end member of glacier retreat; these are separately termed ice-cored rock glaciers, where Galena Creek Glacier in Wyoming is a well-studied example (e.g. Ackert, 1998; Petersen et al., 2020; Potter, 1972). A debris and/or rock avalanche event may also be involved in rock-glacier formation. Knight (2019) addressed how rock-glacier types can be distinguished and that they can be polygenic in origin – where the dynamic controls on rock-glacier evolution may be due to periglacial and/or glacial processes, as well as paraglacial processes. Rock glacier classification types are interconnected with the dynamics of their formation and evolution, and are primarily a function of water, temperature, and sediment (Knight, 2019; Knight et al., 2019). As global glacier recession continues on Earth due to climate warming, landform transitions occur and rock glaciers develop (e.g. Anderson et al., 2018; Jones et al., 2019b), but with varying sensitivity to climate change (Knight et al., 2019).

Figure 3 diagrams significant glacial landforms that evolve in mountainous environments containing ice and debris, and the schematic illustrates some of the terminology mentioned here, and also indicates where multiple terms may refer to the same (or similar) landform. Forms related to rock glaciers include multi-lobate rock glaciers (Degenhardt, 2009), piedmont or spatulate rock glaciers based on their unbounded terminus form, as well as pronival ramparts and protalus ramparts (e.g. Hedding, 2016). A pronival rampart is a ridge (or series of ridges) at the base of a cliff that has been sourced by upslope debris and snow, and is a depositional periglacial landform. In older literature, pronival ramparts have been used synonymously with protalus ramparts, but more recently the term “protalus rampart” has been used to describe an “embryonic rock glacier” of periglacial origin (e.g. Matthews et al., 2017; Scapozza, 2015). See also Whalley (2014) for a discussion of “protalus features,” where protalus lobes are defined as “snowbanks buried by debris from cliffs above” and protalus ramparts are defined as “talus and finer debris accumulating from rockfalls and avalanches at the foot of perennial snow patches and firn banks.” Key to understanding the development of these forms is terrain geometry, headwall erosion rate, any additional debris source, temperature, and precipitation, as well as the state of a given landform within its regional environment (e.g. as informed by the presence of permafrost, debris-covered glaciers, and clean-ice glaciers). Such site-specific studies can elucidate the dynamic and thermal history of individual rock glaciers (e.g. Frauenfelder and Kääb, 2000), and numerical modeling can elucidate the dominant dynamic and thermal processes (e.g. Müller et al., 2016).

OPEN IN VIEWER

Figure 3.

Illustration shows significant glacial landforms containing ice and debris that are found in mountainous environments.

Another terminology consideration is that regions with rock glaciers on Earth exhibit both active and inactive forms, as well as relict forms, often in close proximity to each other (e.g. Bockheim, 2014; Humlum, 1988; Jones et al., 2019a; Lilleøren et al., 2013), and which are not always distinguishable from morphology alone. Morphological characterization is an important step, but as on Earth, robust understanding of the controls on landform evolution requires a suite of data from specific study sites. With enough site-specific understanding, it is becoming possible to generalize across the observed spectrum of landforms (e.g. Knight, 2019).

In addition to surface morphology, the extensive application of geophysical methods to image the interior structure of permafrost and glaciers using different methods is particularly relevant to the understanding of landform evolution. For rock glaciers, Maurer and Hauck (2007) evaluated the application of diffusive electromagnetics, geoelectrics, seismics (also see Draebing, 2016), and ground-penetrating radar to resolve: near-surface structure in the active layer of the debris/rock cover; internal structure and boundaries between ice, water, and debris through the glacier depth; and bedrock topography. They concluded that multiple data sets complemented each other and that the joint views aided interpretation (Maurer and Hauck, 2007). Geophysical investigations have been conducted at sites around the world, and an important finding at multiple locations has been the correspondence of surface structure with internal structure (e.g. Emmert and Kneisel, 2017; Florentine et al., 2014; Fukui et al., 2007, 2008; Leopold et al., 2011; Mackay and Marchant, 2017; Monnier et al., 2009; Petersen et al., 2020; Shean and Marchant, 2010). In addition to constraining the past from internal structure, the fact that modern glaciers are retreating and that the retreat has been observed over the past decades means that we may be directly witnessing landform transitions from a glacier to a debris-covered glacier to a rock glacier (e.g. Jones et al., 2019b; Monnier and Kinnard, 2015; Shroder et al., 2000).

The term “continuum” has been applied in addressing landform diversity together with landform evolution. While not the focus of this review, the occurrence of landform transitions on Mars, and understanding how they may relate to terrestrially established continuum process models, is an important research goal. For Earth, Anderson et al. (2018) presented a model that captures the co-evolution of ice and debris under climate forcing that leads to glaciers evolving across clean ice (relatively debris-free), debris-covered, and rock-covered forms, as well as developing ice-cored moraines in alpine glacier environments. Rock glaciers as part of a continuum have also been proposed for depositional periglacial forms (e.g. Kirkbride, 1989; Matthews et al., 2017; Shakesby et al., 1987), where talus supply, climate forcing, and deformation drive transitions between pronival ramparts, rock glaciers, and moraine ridges. In recent work, Knight (2019) presented a transitional model between periglacial, glacial, and paraglacial rock glaciers. As another example, Matthews et al. (2017) presented a periglacial-glacial landform continuum model that links push/dump moraines, ice-cored moraines, and rock glaciers. These conceptual models bring together how the local landscape and the local-to-regional environmental history are controls on landform evolution. Continuum process models are also motivated in part by regional investigations where a diversity of landforms have been identified and sometimes referred to as “active,” “inactive,” and “relict” (e.g. Barsch, 1996; Bockheim, 2014). For example, active rock glaciers and/or debris-covered glaciers actively deform, exhibit a steeper ice front and may also exhibit surface ridges and furrows due to variable debris input and/or ice dynamics (e.g. Kääb and Weber, 2004; Whalley and Martin, 1992). Whereas deformation is significant in shaping active rock glacier landforms, other surface-modification processes dominate over deformation for inactive forms, and an unknown amount of core ice remains. Relict forms have a further deflated surface profile, are immobile, and are presumed to have a limited amount of core ice. In general, rock glaciers move primarily by internal deformation concentrated at a shear zone near the bed, and debris-covered glaciers move by internal deformation, basal sliding, and if underlain by soft sediments, can undergo soft-bed deformation (see, e.g., Jones et al., 2019a).

2.4 Lobate debris aprons and glacier-like forms on Mars

A large body of work on Mars is available based on initial views from Mariner 9 (1971–1972) imagery, subsequent planetary mapping during the Viking era (1975–1980), and primarily since the late 1990s using a suite of satellite imagery, topography, and radar data largely from the following missions: Mars Global Surveyor (MGS: 1997–2006), Mars Odyssey (2001–present), Mars Express (2003–present), and Mars Reconnaissance Orbiter (MRO: 2005–present). Satellite imagery with a resolution from hundreds of meters per pixel to tens of centimeters per pixel provides the major source of data for surface landform studies, complemented by topographic measurements and available radar data.

The focus of this review on debris-covered glaciers connects to the general category of viscous flow features on Mars. Viscous flow has long been proposed as a process controlling “fretted terrain” landforms observed in Mariner 9 (Sharp, 1973) and Viking (Squyres, 1978) images. Squyres and Carr (1986), following upon the work of Squyres (1979), mapped the global distribution of viscously “softened” terrains, and found that lobate debris aprons (LDA), lineated valley fill (LVF), and concentric crater fill (CCF) are all widespread at mid-latitudes between 30°and 60° in both hemispheres. While these landforms are typically referred to using established acronyms, in an attempt to retain clarity across research communities, we will use the full names here. Figure 4 shows the currently mapped distributions of lobate debris aprons (Levy et al., 2014), glacier-like forms (Brough et al., 2019), and recessional glacier-like forms (Brough et al., 2016a) across the mid-latitudes of Mars; these are the Martian flow features most related to debris-covered glaciers. The latitudinal clustering stands out, as well as how features are often regionally co-located.

OPEN IN VIEWER

Figure 4.

Distribution of select viscous flow features on Mars. Background is the High Resolution Stereo Camera (HRSC) Mars Orbiter Laser Altimeter (MOLA) blended global Digital Elevation Model (DEM) (Fergason et al., 2018). Red dots show the locations of mapped glacier-like forms (Brough et al., 2019), black dots show recessional glacier-like forms (Brough et al., 2016a) and yellow dots show the locations of mapped lobate debris aprons (Levy et al., 2014). These features are primarily found within 30–60° north and south latitudes. For interpretation of the references to colours in this figure legend, refer to the online version of this article.

Lobate debris aprons are broad, thick accumulations of ice-rich material that are commonly found at the base of prominent topographic features such as massifs and valley/crater walls (e.g. Carr and Schaber, 1977; Head et al., 2005). Based on MRO Shallow Radar (SHARAD) observations across Mars (Holt et al., 2008; Petersen et al., 2018; Plaut et al., 2009), lobate debris aprons most likely correspond to hundreds-meter thick deposits of nearly pure (>80%) water ice (Petersen et al., 2018) with a relatively thin dust and debris cover (<10 m; Holt et al., 2008). Numerous studies have been conducted to analyze the shape, morphology, distribution, history, and inter-relationships between lobate debris aprons and other surface features (e.g. Baker et al., 2010; Baker and Head, 2015; Berman et al., 2015; Chuang and Crown, 2005; Dickson et al., 2012; Fastook et al., 2014; Li et al., 2005; Mangold, 2003; Parsons and Holt, 2016; Pierce and Crown, 2003; Sinha et al., 2017). Separately categorized ghost lobate debris aprons have been identified based on a lack of remaining ice, but with a remnant debris apron surrounding a massif (Hauber et al., 2008). Concentric crater fill is a crater-filling deposit that is concentrically lineated (Squyres, 1979) and, similar to lobate debris aprons, likely comprise hundreds of meters of ice below a surficial debris layer (Levy et al., 2010). Lineated valley fill is a deposit on valley floors that is often located in the vicinity of other viscous flow features (Squyres, 1978) with which it may exhibit complex fold patterns consistent with flow (e.g. Head et al., 2006b). Lineations parallel to the flow direction (e.g. Dickson et al., 2008; Morgan et. 2009) and lineations orthogonal to the flow direction (e.g. Levy et al., 2007), as well as surface crevasses (e.g. Hubbard et al., 2014) are also visible on viscous flow features.

Milliken et al. (2003: 2) using Mars Orbiter Camera (MOC) images and Mars Orbiter Laser Altimeter (MOLA) topography, identified a thinner class of landforms that they called viscous flow features, which “have characteristics including surface lineations, compressional ridges, and flow fronts similar to characteristics associated with larger scale viscous flow features, such as lobate debris aprons or lineated valley fill.” Milliken et al. (2003) observed such small-scale features predominantly within the 30°–50° N and S latitude bands. Following Hubbard et al. (2011), we will refer to these small-scale viscous flow features as glacier-like forms (also previously called “glacier-like flows” by Arfstrom and Hartmann, 2005). Souness et al. (2012) inventoried smaller “alcove-fed” glacier-like forms using MRO Context Camera (CTX) imagery to map features centered on 40° latitude in the northern and southern hemispheres, and Brough et al. (2019) validated 1243 of these features. They found that these features averaged 4.7 km in length and 1.3 km in width and are located within similar topographic settings of relatively high relief (Brough et al., 2019; Souness et al., 2012). Glacial-geology and glacial-geomorphology style analyses have been applied to individual landforms, making the case that these features are physically (as well as visually) like alpine glaciers on Earth (e.g. Brough et al., 2016a, 2016b; Dickson et al., 2008; Hubbard et al., 2011, 2014; Souness and Hubbard, 2013). Figure 5 shows an example of a lobate debris apron and a glacier-like form. Recessional glacier-like forms exhibit landform evidence of recession in length and/or in extent, and this population makes up about one-third of the total inventory of glacier-like forms (Brough et al., 2016a). Superposed glacier-like forms are found to overlay lineated valley fill or lobate debris aprons, and are indicative about the timing and extent of the most recent stages of alpine glaciation on Mars (e.g. Hepburn et al., 2020).

OPEN IN VIEWER

Figure 5.

(a) Lobate debris apron at ∼41 S, 103 E shown in perspective false color from HRSC image 0451 (ESA/DLR/FU Berlin). The diameter of the apron is ~50 km. (b) Glacier–like form at ∼41 N, 54 E shown in perspective from HiRISE image PSP_009455_2215 draped over topography.

The identification and understanding of debris-covered glaciers on Mars involves drawing qualitative comparisons with analogous features on Earth. Martian “debris aprons” were first identified by Carr and Schaber (1977), who also noted the similarity of these features to terrestrial flows resulting from frost creep and gelifluction. Squyres (1978) showed that surface lineations indicative of flow upon Martian lobate debris aprons resembled the parallel ridges and furrows observed on terrestrial rock glaciers by Wahrhaftig and Cox (1959). Squyres (1979) mapped the distribution of viscous flow features in Viking images and found that they clustered in the mid-latitudes, indicating a climatic control on their formation and that it was likely that the buried ice was deposited by airfall. Lucchitta (1981) compared cold-climate features on Mars and Earth, and concluded that many Martian features could be attributed to periglacial activity involving buried ice, including talus aprons, debris avalanches, rock glaciers, and moraine-like ridges. Whalley and Azizi (2003) systematically considered three different models (permafrost, glacial, and landslide) of terrestrial rock glacier and protalus lobe/rampart formation, and showed that these features exhibit “equifinality” (i.e. any of the three models could produce the observed topographic form), thereby complicating attempts to interpret the origin of analogous Martian landforms. In general, equifinality must be considered when diagnosing landform-evolution processes from qualitative associations alone.

Arfstrom and Hartmann (2005) noted the similarity of ridges associated with glacier-like forms to terrestrial moraines produced by active glaciers, and argued that these Martian moraine-like ridges must have formed during recent periods of higher obliquity. This is consistent with the analysis of Marchant and Head (2007), who showed that the Antarctic Dry Valleys are a good terrestrial analogue for hyperarid cold-desert climates on Mars, because both exhibit similar microclimate zonation in the expression of landforms at the macroscale, mesoscale, and microscale (such as gullies, debris-covered glaciers, and surface pitting, respectively), which they attribute to a changing climate on Mars caused by recent ice ages driven by Martian obliquity variations (Head et al., 2003). Similarly, Jawin et al. (2018) argued that Mars is currently undergoing a paraglacial period corresponding to a gradual decay in ice-related activity, based on identification of a suite of landforms (including polygons, spatulate depressions, and washboard terrain) that on Earth are indicative of transitional post-glacial processes. In addition, quantitative investigations have yielded new understanding. For example, terrestrially derived glacier-ice dynamics have been applied to: interpret the former extent of a glacier-like form (e.g. Brough et al., 2016a), estimate the volume of the lobate debris aprons (Karlsson et al., 2015), investigate the evolution of concentric crater fill in Utopia Planitia (Weitz et al., 2018), estimate the volume of glacier-like forms (Brough et al., 2019), and evaluate the evolutionary history and deformational properties of lobate debris aprons (e.g. Fastook et al., 2014; Parsons et al., 2011; Schmidt et al., 2019).

While lineated valley fill and concentric crater fill are relatively thick, ice-rich landforms that are likely of glacial origin (e.g. Head et al., 2010), throughout the rest of this review we focus on lobate debris aprons and glacier-like forms because their environmental settings and present-day characteristics indicate a history that may most relate to active debris-covered glaciers on Earth. Part of the motivation for this review is that the application of ice dynamics to Mars introduces assumptions about physical processes that influence how landforms are interpreted, and those processes need to be understood. Here we expand on how supraglacial debris is emplaced and evolves, and the effects of debris on topography and ice flow in order to provide context from Earth to studies on Mars.

2.5 Complexity with and without similarity

Baker et al. (2010) addressed the challenges in applying terrestrial periglacial and glacial terminology to Mars. As previously discussed, since the terms “rock glacier” and “debris-covered glacier” on Earth do not necessarily imply a single origin and subsequent evolution, using these terms to describe features on Mars does not uniquely refer to the processes that are of interest in the study of both planets. Terminology in the planetary literature has shifted toward more often evaluating Martian features as debris-covered glaciers (instead of rock glaciers), especially for lobate debris aprons where the majority of subsurface radar profiles available over select features indicate a composition of at least 80% water ice (Petersen et al., 2018), and for glacier-like forms that can exhibit similar morphologies to alpine glaciers on Earth. While this is not a simple consideration, we note that using “ice-cored rock glacier” to describe features on Mars may be more widely representative, but until systematic evaluation of debris and ice content is available for Martian features, we cannot know what term(s) are actually the most appropriate and how to classify all features – this is an important future research direction. Regional mapping of the extensive mid-latitude glaciation (and subsequent deglaciation) that occurred in the last 100 million to 1 billion years suggests that many of the viscous flow features previously discussed may have a common regional system-scale origin (e.g. Baker et al., 2010 and references therein; Fassett et al., 2014). Separately, Hepburn et al. (2020) found that flow of superposed glacier-like forms was active more recently in at least two distinct episodes between 2 and 65 million years ago. The number of glacial cycles and the extent and modes of ice activity during these cycles is still being reconstructed. Given the variety of landforms on Mars spanning a range of environments, and therefore past conditions and subsequent evolution, the application of Earth-based terminology – and, by extension, understanding – must be carefully considered. In addition, it is important to note that this can be a challenge on Earth. Especially in mountainous environments and when interpreting debris assemblages, similar terrestrial landforms can develop by different processes, including landform transitions; as noted by Hedding et al. (2018), landform identification and the relationship of origin processes to paleoclimatic reconstruction need to be considered carefully. Regardless of this challenge, the diversity of glacial landforms identified on Mars with morphological similarity to glacial landforms on Earth (including debris-covered glaciers) is a striking starting point for deeper understanding.

On Earth, some important open research questions focus on how to place individual debris-covered glaciers in the proper regional context: how to robustly interpret past climate and landform evolution from glaciological and morphological features? Furthermore, how will debris-covered glaciers continue to evolve as a function of future climate warming and contribute to regional runoff and landscape evolution? An appreciation of landform complexity and diversity is necessary to diagnose key processes. Given the inevitable complexity on Earth, our initial aim in this review is to build understanding from simpler scenarios consistent with terrestrial data and with relevance to Mars. However, while questions similar to Earth may apply to Mars, the inferred climate histories are different. For example, extant (remnant) glacier-like forms on Mars are mostly associated with destabilization, retreat, or ice loss in a past epoch, whereas debris-covered glaciers on Earth span a continuum from relict to inactive to active and are still undergoing significant changes today and into the future (e.g. Herreid and Pellicciotti, 2020). Other open questions regarding Martian dust and debris-covered ice masses include: ascertaining the atmospheric and orbital controls on the cycling of water between the poles and the mid-latitudes; determining how the debris cover was emplaced (by localized erosion such as rockfall or by larger-scale deposition of a dust-rich mantle); and quantifying the number of episodes recorded in images and radargrams of the present-day structure of viscous flow features. During warmer climate periods, how did the combination of accumulation, ablation, dust/debris supply, and ice flow shape ice masses on Mars? These questions lead to local investigations that include mapping and modeling surface structure, but also include placing individual features in a regional and global context. Details of the processes controlling debris emplacement, debris-cover evolution and the effects of debris on topography and ice flow are not fully established for Mars, which is not surprising since these are key research areas for debris-covered glaciers on Earth. Process-based understanding of landforms requires an understanding of the main processes that drive landform evolution, and the following main section of this review aims to put knowledge of debris-covered glaciers on Earth in context to interpret lobate debris aprons and glacier-like forms on Mars.

3.1 Supraglacial debris emplacement

Throughout this review we use “debris” as a general term, but when discussing origins of debris we differentiate between locally sourced rock/soil, volcanic tephra, and windblown dust. Deposition of debris on glacier surfaces primarily occurs by hillslope and mountain-slope erosion, often from a headwall or sidewall source, including lateral moraines. This can include sustained deposition on the surface, as well as subsequent burial in the accumulation zone and eventual redeposition on the surface as ice is lost in the ablation zone. There are snow avalanches entrained with debris, as well as rock/debris avalanches that are episodic events that can alter a significant fraction of the glacier area. Debris may also be sourced by bedrock erosion, and different sources of debris can typically be distinguished based on clast angularity, polish, and striations (e.g. Reheis, 1975). We review work on volcanic ashfall alteration of glaciers, especially in Iceland where this is a common occurrence. Windblown dust alteration of glacier surfaces is also discussed. Note that somewhat separate attention is paid to debris redistribution (e.g. Moore, 2018). Sorting by melt, local surface slope, and ice flow typically results in finer debris near the ice interface and coarser debris on top; processes driving debris redistribution are important but mentioned only briefly here.

The influences of debris cover on ice ablation, ice topography, and ice flow depends on debris character. Independent of the origin and character of the debris cover, we refer to all glaciers as “debris-covered glaciers”. In relation to understanding features on Mars, it is necessary to separately consider erosion-sourced debris, avalanche-sourced debris, and atmospherically derived debris on Earth. We first review these three primary ways that debris is emplaced on terrestrial glaciers, and then provide a perspective on supraglacial debris emplacement in relation to Mars.

3.1.1 Terrain erosion and headwall source

Debris-covered glaciers are typically found in regions where the flux of debris is comparable to the flux of glacier ice, which is common in tectonically active mountain ranges such as the Himalaya-Karakoram, Andes, and Southern New Zealand Alps. In general, hillslope and mountain-slope erosion as sources of debris to glaciers is a function of surface slope, rock type, glacier action, and climate forcing, including freeze–thaw processes. Headwall sources of debris that are deposited in the accumulation zone of a glacier will first be buried by subsequent accumulation, entrained and transported within the ice, and then emerge in the ablation zone after an amount of time that can be estimated by the characteristic response time of the glacier as given by the characteristic thickness divided by the characteristic ablation rate at the terminus (Jóhannesson et al., 1989). Erosion at the glacier bed can also be a source of debris to the surface in the ablation zone. A consistent source of debris that is near the head of the glacier and is limited in spatial extent with respect to the glacier width may form a medial moraine that extends along the glacier length. An eroding sidewall slope may form a lateral moraine along the glacier length, and the lateral moraines of two merged tributary glaciers can form a medial moraine in the downstream portion of a glacier system. Moraines can lead to heterogeneity in glacier surface elevation, and moraine-derived debris can have a significant influence on the glacier state if such debris extends over most of the glacier width as it widens over time (e.g. Anderson, 2000). Lateral moraine degradation may be significant, especially when the glacier has detached from the sidewall as it thins during retreat (van Woerkom et al., 2019).

Each glacier environment is distinct, even within high-mountain Asia (i.e. the mountain ranges bounding the Tibetan Plateau) where many studies have been conducted. For example, Scherler et al. (2011a) analyzed 287 high-mountain Asia glaciers using satellite data to illustrate the effects of topography upon glacier state, where low-relief Tibetan Plateau glaciers have minimal debris cover and glaciers in this region with steep headwalls have an extensive debris cover due to avalanche activity. Scherler et al. (2011b) showed that Himalayan debris-covered glaciers typically have steep accumulation areas with an ice-free mountain area that has a mean slope >25°, which is the minimum critical slope for snow avalanching (taken as 25–50° from Luckman, 1977; Scherler et al., 2011a). Figure 6 shows the main features of this distinctly Himalayan glacial environment. The regional analyses from Scherler et al. (2011b) suggested that, as steep headwall avalanches feed debris and ice to these glaciers, debris cover affects the surface shape, surface velocity, and landscape erosion potential. If the glacier exists primarily below the snow line (the elevation where seasonal snow is still found at the end of summer) then the areal extent of debris can increase as debris input exceeds avalanched snow input and “the debris-covered glaciers may only act as debris conveyors” (Scherler et al., 2011a); these are among the types of headwall-sourced glaciers that may exist along a continuum of glacier forms similar to those explored by Anderson et al. (2018).

OPEN IN VIEWER

Figure 6.

Debris-covered glaciers and surrounding topography in the Khumbu Himal, Nepal. NASA Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) image from November 10, 2004 that has been draped over a 90-meter resolution Shuttle Radar Topography Mission (SRTM) Digital Elevation Model with ∼1.5 times vertical exaggeration. Snow accumulates primarily on the steep headwalls and avalanches down to low-gradient valley glaciers that have high debris content and cover.

Mackay et al. (2014) showed that headwall slope erosion rates can vary not just spatially but also temporally. Using radar data along two adjoining cold-based debris-covered glaciers in the McMurdo Dry Valleys of Antarctica, they found that the upper 1–2 km of each glacier is debris free (corresponding to >99% pure ice), but the rest of both glaciers contain englacial debris bands that are observed as arcuate ridges intersecting the ice surface (Mackay et al., 2014; see Figure 10 and section 3.3.3). Given that basal entrainment is negligible since cold-based glaciers do not significantly erode their beds, Mackay et al. (2014) argued that these englacial debris bands most likely formed as supraglacial lags during periods of lower net ice accumulation driven by episodic climate change. An interesting analogue extension is that this may also be occurring within Martian cold-based ice masses that are morphologically similar to those in the McMurdo Dry Valleys (e.g. Marchant and Head, 2007).

3.1.2 Rock/debris avalanches

We focus primarily on the literature addressing rock/debris avalanches from higher-elevation mountain (bedrock) topography that is surrounding glaciers. However, many glaciers are fed by snow avalanches, which may include some entrained rock/debris if the snow avalanche is erosive. This is especially true for glaciers fed by icy debris fans, where snowfall and rockfall are delivered together to the glacier surface (e.g. Kochel et al., 2018). In addition, in high-mountain Asia where debris-covered valley glaciers with low surface slopes abut steep headwall topography there is a consistent source of snow and debris, which is important for the glacier mass balance (e.g. Benn and Lehmkuhl, 2000; Hambrey et al., 2008; Laha et al., 2017; Nagai et al., 2013; Scherler et al., 2011a). There can also be a debris flow, which is a separate category of event commonly triggered near the terminus of a glacier that is usually due to water-fed instability (glacial outburst, heavy rainfall). While debris flows are not typically sources of supraglacial debris, they are important in debris/sediment delivery downstream of the glacier and to the glacier forefield (e.g. Chiarle et al., 2007; Lane et al., 2017). In addition, the role of liquid water in rock/debris avalanches may become important if a slide causes substrate liquefaction, as can occur in a hyperconcentrated flow (see, e.g., Calhoun and Clague, 2018; Hewitt et al., 2008; Pierson, 2005).

Rock avalanches on glaciers have been studied more recently as their major influence on glacier evolution has become better understood (e.g. Dufresne et al., 2016; Menounos et al., 2013; Reznichenko et al., 2011; Shugar and Clague, 2011; Shugar et al., 2012; Vacco et al., 2010a). Deline et al. (2015) reviewed rock avalanches on glaciers, focusing on large events that have been documented, but acknowledging that many additional events have likely occurred compared with those documented due to remoteness of many glaciers, high snowfall rates in many areas, and possible sequestration of debris in the glacier ice (Dunning et al., 2015). The occurrence of rock avalanches has been noted for the past few thousand years in some areas (e.g. Deline, 2009). If the rock avalanche covers much of the glacier below the equilibrium line and the debris cover is an efficient insulator, then the glacier will thicken due to reduced melting (e.g. Kirkbride, 1995), which can alter the ice velocity (e.g. Vacco et al., 2010a) and can advance (e.g. Deline and Kirkbride, 2009; Hewitt, 2009). Figure 7 shows an example of a major rock avalanche on the Morsárjökull Glacier in Iceland and how debris can spread across the glacier (Sæmundsson et al., 2011). Then, after an adjustment period, the extended glacier will likely retreat and produce a terminal moraine (e.g. Reznichenko et al., 2011; Shulmeister et al., 2009; Vacco et al., 2010a): these responses are typically out-of-phase with climate conditions.

OPEN IN VIEWER

Figure 7.

Rock avalanche on Morsárjökull Glacier, Iceland that occurred in 2007 and has significantly altered the glacier mass balance (see Sæmundsson et al., 2011).

Therefore, rock avalanche-derived moraines may be difficult to distinguish from climate-driven moraine formation (e.g. Hewitt, 1999; Reznichenko et al., 2011, 2012), unless the avalanche event is large and the moraine thickness and clast-size distribution are distinct (e.g. Shulmeister et al., 2009). In addition, rock-avalanche scars can also aid interpretation when they are evident (e.g. Carter, 2015; Deline, 2009). Cook et al. (2013) found that if a glacier advances over a preexisting rock avalanche deposit then the reworked avalanche material incorporated in a moraine could be distinguished from purely glacier-derived material, especially for rock avalanches that distribute large, angular boulders that may be more resistant to erosion. Rock avalanches from glacier/ice-covered slopes also leave distinct zones from source to deposit (e.g. Lipovsky et al., 2008), but features may not be retained on the landscape due to subsequent modification and burial by snowfall. However, extensive rock avalanches that alter glacier and moraine sequences are challenging to interpret from remnant morphology alone (e.g. Deline and Kirkbride, 2009; Kirkbride and Winkler, 2012).

Rock avalanches on glaciers not previously debris covered are more rare. Sherman Glacier, Alaska, is a case where the debris cover on the lower glacier formed suddenly due to a huge rock avalanche after the Great Alaska Earthquake in 1964 (e.g. Reznichenko et al., 2011). Such events also occur off massifs due to mountain permafrost degradation and slope failure, and in some cases the trigger is unidentifiable, but all events deposit debris on the glaciers below (e.g. Nagai et al., 2013). Landslides and rock avalanche events will likely continue to become more frequent in a warming climate as glacier ice loss and permafrost degradation continue to destabilize slopes (e.g. Allen et al., 2011; Coe et al., 2018; Gariano and Guzzetti, 2016). Reznichenko et al. (2011) made a distinction between the properties of englacial debris that has “melted out” and been left behind as lag deposit on the surface and those of rock avalanches, within which debris is typically thicker and has a higher thermal inertia because of the grain-size distribution, which includes very fine grains. This spread in grain size – resulting from the intense fragmentation during the avalanche – also decreases porosity and permeability, thereby making rock avalanches far more effective inhibitors of ablation from underlying ice than typical debris cover (Reznichenko et al., 2011). Isolated large events likely contribute more debris by volume than more frequent smaller-to-medium sized events (e.g. Korup and Clague, 2009), and especially for large events there is a longer run-out distance when the rock avalanche occurs on a glacier surface (e.g. Deline et al., 2015; Sosio et al., 2012). Since the rock avalanche material can be spread more effectively on a glacier surface, the resulting deposit is typically thinner compared with an event in a non-glacial environment, with large events on Earth spreading material non-uniformly, based on available measurements that are typically a meter to a few meters in thickness (Deline et al., 2015).

Recently, field-based geomorphological and sedimentological analyses have been conducted in combination with high-resolution satellite imagery (e.g. Shugar et al., 2013). Longitudinal surface banding due to shearing within the debris flow (and clast-size sorting) can result from both the initial rock avalanche event and subsequent reworking (e.g. Shugar and Clague, 2011). Finer debris can be blown by wind or washed away by water, and an originally poorly sorted distribution can homogenize and organize over time due to weathering and glacier flow (e.g. Black Rapids Glacier, Alaska; Shugar and Clague, 2011). In addition, freeze–thaw cycles can also act to break larger clasts into fragments and affect rockfall activity (e.g. Matsuoka, 2008). Multiple rock avalanches can occur upon a single glacier, and in some cases coalesce to form a spatially continuous debris cover (e.g. Miage Glacier, Mont Blanc massif; Deline et al., 2015); proximal topography such as moraines can alter the path and distribution of avalanche debris.

3.2 Supraglacial debris evolution

Following emplacement of debris and the development of an extensive debris cover across a glacier surface, the co-evolution of debris and ice must be addressed in order to understand the dynamics of a debris-covered glacier. In this section we focus on the evolution of the debris cover, including subsections 3.2.1 “Surface-atmosphere exchange,” 3.2.2 “Enhancement or inhibition of ablation,” and 3.2.3 “Thickness of the debris cover,” which are discussed in relation to section 3.2.4 “Debris-cover evolution on Mars.”

3.2.1 Surface-atmosphere exchange

The general exchange of the ice surface with the atmosphere has already been discussed with regards to the glacier mass balance, where the state of snow accumulation and ice ablation constrain the annual balance state of the glacier. In relation to debris-covered glaciers, how debris alters the energy balance – and therefore the ablation rate and the mass balance – is what is critically important. Here we review key concepts of surface energy balance on a glacier and introduce energy-balance models applied to terrestrial glaciers. Although the same atmosphere-glacier feedbacks on Earth may not apply directly to Mars, we highlight the significant differences in the energy balance on a clean-ice surface compared with a dirty-ice surface. Figure 8 diagrams the significant terms mentioned here. The energy balance is the sum of all energy fluxes at the ice surface (e.g. Cuffey and Paterson, 2010: 140; Hock, 2005):

Q melt = Q N + Q H + Q E + Q R + Q G

1

OPEN IN VIEWER

Figure 8.

Glacier in cross-section with ice-flow paths and ice velocity structure indicated, and showing the significant glacier surface energy-balance components. The accumulation zone (C) and the ablation zone (A) are separated by the equilibrium-line altitude (ELA). Ice-surface velocity (u_S) and basal sliding velocity (u_b) are a function of glacier-ice thickness (H) and slope (θ); basal shear stress (τ_b) is the product of ice density (ρ), gravity (g), thickness, and sin(θ). Surface energy balance components include incoming shortwave (S_in) and incoming longwave (L_in) radiation, outgoing shortwave (S_out) and outgoing longwave (L_out) radiation, precipitation heat flux (Q_R), latent heat flux (Q_E) and sensible heat flux (Q_H).

Net solar radiation Q _N is usually dominant and includes the four components of incoming/outgoing shortwave (S_in and S_out) and incoming/outgoing longwave radiation (L_in and L_out), where the amount of shortwave radiation absorbed is a function of the surface albedo. Sensible heat flux Q _H and latent heat flux Q _E (together also known as turbulent heat fluxes) are determined by air mixing near the ice surface. Here, Q _R is due to precipitation, especially rainwater that refreezes, but this flux is typically small and can be ignored on Earth and we do not expect it is relevant for Mars. The ground flux Q _G through the glacier surface corresponds to the energy flux through the bottom of this surface layer due to conduction and to shortwave radiation; on Earth, this flux is typically set to zero if the ice is near temperate but may be non-trivial for the cold, dry regions of Antarctica (Lewis et al., 1998) and especially those that exhibit “blue ice” exposed at the surface (Bintanja et al., 1997), as well as being relevant for Mars. Thus when considering the energy balance for debris-covered ice on Mars, the total energy flux depends primarily on the net solar radiation, sensible heat flux, latent heat flux, and ground flux.

Typical values of albedo range from 0.85 for fresh dry snow to 0.2 for debris-rich ice, with the classification between “old clean dry snow” and “old debris-rich dry snow” affecting albedo by at least 30% (Cuffey and Paterson, 2010: 146, table 5.2). Surface albedo varies throughout the melt season as well as year-to-year due to the presence of impurities, melt water, and grain-scale processes in the snowpack and the firn (the intermediate layer between fresh snow and glacier ice). Note that debris tends to accumulate over time, thus lowering the albedo. Since albedo decreases through the melt season the timing of maximum surface melting is delayed compared with the timing of maximum insolation, especially on mid-latitude glaciers where winter snow is typically present during summer solstice and an albedo decrease through the summer leads to an increase in the amount of absorbed shortwave radiation by up to a factor of four (Cuffey and Paterson, 2010: 147). Microtopography and daily meteorology are also important for the net shortwave flux, whereas the temperature and humidity of the atmosphere are particularly important for the net longwave flux. It is the net flux that must be considered; for example, cloud cover will reduce the amount of incoming shortwave radiation but will increase the amount of incoming longwave radiation.

Individual components of the energy balance are represented in numerical energy-balance models (e.g. see Oerlemans, 2001), and can be constrained by detailed energy-balance measurements (e.g. Hogg et al., 1982; Klok and Oerlemans, 2002; Shea et al. 2015) of various parameters including net radiation (or radiation components), air temperature, air humidity, and wind speed. However, in many cases these values are poorly known in space and/or time for a given study site, so parameterizations of ice melting are used. A simple approach is to apply a regression equation between widely available air temperature measurements and locally observed melt rate (e.g. Braithwaite, 1981, 1995; see also Hock, 2003), and though the real relationship cannot necessarily be so simply reduced, mean summer temperature has been shown to be a reliable index if the melt rate is controlled by summer ablation (Braithwaite and Olesen, 1989; Ohmura, 2001).

“Positive-degree day models” have been developed that relate the amount of time an ice surface is above the melting temperature (0°C) as an index summed over the year instead of represented as a mean value in the initial temperature-index models (e.g. Braithwaite and Olesen, 1989). The total surface melt is related to the number of positive-degree days using an empirically derived degree-day melt factor (e.g. Hock, 2005). While these relationships vary regionally they have been applied successfully to a variety of glaciers, and this general class of “temperature index models” have had continuous application and adaptation for decades (e.g. Carenzo et al., 2016; Hock, 1999; Wake and Marshall, 2015).

Although separate in approach, numerical models and empirical models have developed in tandem. For example, distributed temperature-index models take into account how variations in surface topography may affect ice melt rates (e.g. Hock, 1999, 2003), introducing spatial and/or temporal variations as realistically as possible. Locations that require separate consideration are those where melt rarely occurs, for example in the cold, dry interior of Antarctica where ablation occurs by sublimation and the influence of katabatic winds is significant (e.g. Bliss et al., 2011). In addition, tropical glaciers are complicated by diurnal and seasonal variations in the controls on ablation (or sublimation), which are not driven by temperature, and so positive-degree day models are not typically applied.

As with melt models for clean-ice glaciers, both numerical energy-balance models and empirical temperature-index models have been applied to debris-covered glaciers. Point-based energy-balance models have been used in numerous cases (e.g. Brock et al., 2010; Nakawo and Young, 1981; Nicholson and Benn, 2006; Reid and Brock, 2010) to calculate the energy flux at the debris surface and the conduction of heat into the underlying ice. Including debris cover means that additional variables such as debris albedo, surface roughness, debris thermal conductivity, and water content in the debris must be prescribed, and in distributed energy-balance models these parameters are prescribed over an entire glacier or glaciated region in order to capture spatial variations in surface melt. This is an active area of research, where new data are being collected (e.g. Miles et al., 2017; Steiner and Pellicciotti, 2016; Yang et al., 2017), models continue to be refined (e.g. Aubry-Wake et al., 2018; Evatt et al., 2015; Fyffe et al., 2014; Groos et al., 2017; Lejeune et al., 2013; Rounce et al., 2015; Shaw et al., 2016), and data-model evaluation progresses (e.g. Collier et al., 2015). However, spatial and temporal estimates of debris physical properties – including debris thickness – still remain as significant unknowns.

Distributed temperature-index models also continue to improve and are applied in regions where energy-balance data remain limited, and have been validated in regions where output can be compared to an energy-balance calculation (e.g. Carenzo et al., 2016). Poorly constrained spatial and temporal variations in debris thickness are also significant unknowns in these modeling efforts, but remote sensing and ground-based analyses of debris thickness are advancing in step with the melt models (e.g. Fujita and Sakai, 2014; McCarthy et al., 2017; Nicholson and Mertes, 2017; Rounce and McKinney, 2014; Schauwecker et al., 2015), and inverse modeling can be employed to improve estimates of debris thickness (e.g. Rounce et al., 2018). In addition to uncertainties and variability in debris-cover thickness (e.g. Nicholson et al., 2018), limited availability of meteorological data and limited understanding about the role of meltwater are research targets needed to improve melt-rate calculations for debris-covered glaciers.

3.3 Effects of debris on glacier topography and ice flow

The first order influence of supraglacial debris on a glacier surface is due to the difference in albedo between debris and ice, and many studies have investigated how the response of debris-covered glaciers to climate change is different compared to clean-ice glaciers. Included in this response is the rate of ice melting, surface shape and length evolution, and debris distribution, all of which contribute to glacial and proglacial landform development, especially for a retreating glacier; meltwater-driven supraglacial lakes and rivers are not covered here. The following discussions are covered in this section: 3.3.1 “Debris-covered glacier shape,” 3.3.2 “Deformation of ice containing debris,” and 3.3.3 “Debris transport by ice flow.” While modern changes in debris-covered ice on Mars are minimal (if there are any changes at all), spatial variations in modern topography of glacier-like forms and lobate debris aprons indicate that debris cover may have contributed to differential ablation and differences in deformation in the past; in the last subsection 3.3.4 “Effects of debris on topography and ice flow on Mars” are discussed.

3.3.1 Debris-covered glacier shape

Measurements of a glacier surface can be readily obtained from ground-based, airborne, and satellite platforms, so it is important to understand what can be learned about glacier states from surface data. In comparison, planetary studies are primarily based upon remotely sensed surface measurements. Since consolidated debris cover thicker than a few centimeters typically insulates the underlying ice, the dynamics of debris-covered glaciers are different from clean-ice glaciers. For example, debris cover often leads to lower surface slopes and surface velocities across the lower glacier (e.g. Scherler et al., 2011a, 2011b; Thompson et al., 2016), and the reduced surface gradient can affect surface thinning, mass balance, meltwater ponding, and drainage (e.g. Salerno et al., 2017). Figure 9 shows two examples from the compilation of Scherler et al. (2011b), comparing velocity and topography profiles for the Duofeng Glacier (relatively low debris coverage) and for the Rongbuk Glacier (relatively high debris coverage). The important point is that there is a dynamic and a kinematic response to the distribution of debris, which both affect the glacier shape.

OPEN IN VIEWER

Figure 9.

Influence of debris coverage on glacier surface velocity, surface elevation, and surface slope for two different glaciers in High-Mountain Asia. (a) and (b) show values for Duofeng Glacier in the West Kunlun Shan region that has relatively low debris coverage, and (c) and (d) show values for Rongbuk Glacier in the Khumbu Himal region that has relatively high debris coverage. (a) and (c) show surface velocity along the glacier length over the given time interval, where black crosses are velocity measurements, the red line is the mean velocity, and shading corresponds to the number of measurements (see original paper for further details). (b) and (d) show the surface elevation (values on the left axis) and surface slope (values on the right axis); the shaded region is the estimated location of the equilibrium line altitude (also referred to as the snow line).

As debris-covered glaciers retreat, surface lowering may be more evident than pull back of the terminus. Surface lowering as the predominant mass-loss mechanism is significant in the Himalaya, and influences how depositional landforms are interpreted (e.g. Benn and Owen, 2002; Hambrey et al., 2008). Different stages of glacier response to climate can be evaluated as a function of hillslope relief and glacier topography. After analyzing the form and flow of 287 glaciers across High Mountain Asia, Scherler et al. (2011b) developed a conceptual model of “hillslope-glacier coupling” where terrains evolve from less-relief to more-relief over time as snow/debris avalanching, erosion, ice-flow dynamics, and climate processes are coupled in the evolution of these landscapes. Following their model, as debris cover thickens in the ablation zone and the debris extent expands upglacier over time, the maximum glacier velocity is also shifted upglacier and headwall zones retreat to form steeper slopes, thereby delivering even more debris to the lower glacier surface. It is worth noting that, while most High Mountain Asia glaciers are retreating, some glaciers – especially those in the Karakoram range – evolve differently due to dynamic-driven terminus stability and glacier surge activity, in addition to climate drivers of glacier change; see Farinotti et al. (2020) and references therein for a discussion of the so-called “Karakoram glacier Anomaly.”

Although glacial evolution is generally driven by climate change-induced variations in snowfall and melting, how an individual glacier changes is also a strong function of size, elevation, aspect, and debris cover. Within a given region, glaciers may all exhibit retreat, but the magnitude of response often varies due to differences in local environmental settings. With that known, higher spatial and temporal resolution of satellite imagery has allowed for regional analyses of glacier motion as we continue to learn about the complexity in topographic controls on glacier change. For example, Sam et al. (2018) highlighted how elevation and seasons control ice velocity across the Western Himalaya. Surface heterogeneity and instability driven by the development of depressions, supraglacial melt ponds, and ice cliffs (e.g. Steiner et al., 2019) can also lead to local changes in the glacier shape.

3.3.2 Deformation of ice containing debris

Debris-covered glaciers that contain englacial debris can flow at a rate different from clean-ice glaciers. Debris content, especially basal debris, influences vertical profiles of ice deformation and subglacial sediment transport. Moore (2014) discussed that the key rheological variables governing the flow of debris-ice mixtures are debris concentration, particle size, temperature, salinity, and stress, as well as any presence of unfrozen water. The role of these variables is primarily determined from laboratory experiments that are used to develop an ice-flow law that is used to extrapolate these relationships to conditions found in nature.

For most crystalline solids, the relationship between axisymmetric stress σ and strain rate ∊ ˙ can be expressed by a constitutive relationship, referred to as the “flow law,” of the form (Durham and Stern, 2001)

∊ ˙ = A d i − P σ n exp − Q R T exp − n b d ϕ d

2

where d_i = grain size; R = the gas constant; T = temperature; ϕ_d = the volumetric particle fraction; and A, p, n, Q, and b_d = laboratory-measured material constants that depend upon the mechanism of deformation. Glen (1955) conducted the first systematic laboratory measurements of the rheology of pure water ice under compressive stress, obtaining a value of n ∼ 3 that is still the default stress exponent widely used in terrestrial glaciology (e.g. Cuffey and Paterson, 2010: 58).

However, more detailed laboratory experiments indicate that the apparent applicability of Glen’s (n = 3) flow law may be attributable to an averaging effect of four different water ice deformation mechanisms (Goldsby and Kohlstedt, 2001; Peltier et al., 2000): dislocation creep (disl, n = 4), basal slip (bs, n = 2.4), grain boundary sliding (GBS, n = 1.8), and diffusional flow (diff, n = 1). Since grain boundary sliding is accommodated by basal slip and vice-versa (i.e. both processes must occur to drive deformation, the rate of which is governed by the slower mechanism), the total flow of ice can be better characterized by a composite flow law (Goldsby and Kohlstedt, 2001)

∊ ˙ t o t a l = ∊ ˙ d i f f + ∊ ˙ d i s l + 1 ∊ ˙ b s + 1 ∊ ˙ G B S − 1

3

where each strain rate on the right-hand side of equation (3) can be expressed as a flow law in the form of equation (2). In order to characterize the flow of debris-covered glaciers, the rheological effects of debris particles within ice must be explicitly considered.

Laboratory results (e.g. Durham et al., 1992, 2009; Qi et al., 2018) and theoretical work (e.g. Goldsby and Kohlstedt, 2001) to describe the rheological effects of debris particles in ice are broadly consistent with the conceptual framework developed by Moore (2014) in an attempt to explain the, at times, paradoxical field observations and laboratory measurements of the constitutive properties of debris-laden ice. Field measurements often indicate that the presence of debris can weaken ice (enhancing the flow rate), in contrast to the strengthening implied by most laboratory experiments (as reviewed in detail by Moore, 2014; see also Hopkins et al., 2019). Moore (2014) identified three main components of this framework: (1) in the “dirty ice” regime (i.e. debris < 40% by volume), stress- and temperature-dependent ice creep predominate (as in Durham et al., 1992); (2) in the “ice-rich debris” (i.e. fully pore ice saturated) to “ice-poor debris” regimes, ranging from the onset of inter-particle friction up to the maximum packing density of debris, deformation at small strains is primarily governed by shear opposed by friction (which is consistent with Durham et al., 2009); and (3) at all debris concentrations, the debris-ice interaction will be dramatically affected by the presence of unfrozen water (the volume of which is dependent on solute concentration and debris particle size).

Modeling debris-covered and debris-laden ice on Earth and Mars requires applying an ice-flow law; for example, as given in equation 2 or equation 3. Different assumptions may be more-or-less valid for different problems, but it is often the case that reliable estimates of debris concentration, particle size, temperature, salinity, and stress are not available, and that temporal variations in climate and ice rheology are also unknown, so simplified models are applied to interpret specific landforms. This includes simplifying the flow law itself where only the flow-law exponent is an unknown parameter, as well as limiting the dimension of the deformation along flowlines and assuming the ice mass is in steady state (e.g. Schmidt et al., 2019).

3.3.3 Debris transport by ice flow

Debris delivered to a glacier surface by ice flow can be transported supraglacially upon the surface, as well as englacially within the glacier (e.g. Bennett and Glasser, 2009: ch. 7). Surface debris directly affects surface mass balance and englacial debris can directly affect ice deformation. Debris transport rate depends on the glacier state, where generally a glacier moves faster and experiences less ablation when the mass balance is positive, and a glacier moves slower and experiences more ablation when the mass balance is negative. Debris that is emplaced on a glacier surface in the accumulation zone can be buried by subsequent accumulation, transported englacially by ice flow, and emerge in the ablation zone as isolated surface deposits or contribute to an extensive debris cover. Debris may also be incorporated at the glacier bed (Alley et al., 1997; Knight, 1997). Glaciers may simply and “passively” transport debris along flow trajectories within the ice; deliver debris to and detach debris from the glacier bed; orient debris-banded stratigraphy near the glacier bed by thrust-type faulting from compressive flow; develop debris-rich basal layers; concentrate debris as transverse debris bands on the surface; and, depending on debris sources and flow, develop debris mounds, stripes, and medial moraines on the surface. Debris structures and distribution in the ablation zone of a glacier also depend on the state of the glacier terminus, and particularly whether there are any obstructions to flow (Kirkbride and Deline, 2013). The activity of these processes can be determined through structural glaciology field studies (e.g. Goodsell et al., 2005; Hambrey and Lawson, 2000; Jennings et al., 2014), laboratory analyses of field samples (e.g. Moore et al., 2013), historical photos (e.g. Kellerer-Pirklbauer et al., 2008), and geophysical measurements of the glacier interior (e.g. Florentine et al., 2014; Mackay et al., 2014; Nicholson et al., 2018).

The structure of englacial debris may be possible to relate to debris source as a function of climate and ice-flow history (e.g. Mackay and Marchant, 2017; Mackay et al., 2014; Petersen et al., 2020). Figure 10 shows how radar reflections from glaciers in the Dry Valleys of Antarctica can be interpreted as a function of debris/ice inputs over time as the glaciers evolve (MacKay et al., 2014). Identifying englacial structure is informative about processes of headwall debris delivery, subsequent debris burial, ice flow, and how debris emerges on the surface. It is important to consider how climate controls, environmental conditions, and ice flow led to the observed glacier state, and the englacial and supraglacial structures provide key constraints on those processes over time. Further analysis of the origin and evolution of these structures comes from ice-flow modeling, where processes and the timescales leading to their development can be explored. The structures observed at this cold-based glacier system may have direct relevance to interpreting the ice and climate history of landforms on Mars, especially if englacial observations on Mars become available in the future.

OPEN IN VIEWER

Figure 10.

An approximately 500-meter longitudinal transect of Friedman Glacier, Dry Valleys, Antarctica. (a) Unmigrated 80 MHz radar data, (b) migrated 80 MHz data, and (c) interpretation of reflections and major englacial features. Left axis gives both depth in meters and two-way travel time (TWTT) in microseconds. (d) Schematic of glacier structure based on radar-data interpretation and field study. Acronyms given for interpretation include Friedman internal reflection (FIR), arcuate surface discontinuity (ASD), and inclined debris layer (IDL).

It is clear that modeling debris transport is key to understanding supraglacial and englacial debris distribution, debris-cover development and overall glacier evolution, but how has this effort progressed? Modeling debris transport by ice flow, and therefore modeling debris-covered glacier evolution, can be done using models of varying complexity that may invoke different assumptions depending on the application. For example, Anderson and Anderson (2016) modeled glacier evolution under the influence of steady debris deposition as a start to understand the coupled debris-ice system. In addition to glacier ice dynamics and climate forcing, understanding glacier response to debris deposition requires considering numerous parameters, such as the location and amount of debris deposition, englacial advection of debris, emergence of debris in the ablation zone, supraglacial advection of debris, sub-debris melting of ice, and debris contribution to terminus landform development (Anderson and Anderson, 2016). Wirbel et al. (2018) took another approach using a 3D model with time-evolving englacial debris transport, which advanced upon prior work that included simplified (typically 2D) ice-flow models subject to spatially restricted and prescribed debris deposition (e.g. Konrad and Humphrey, 2000; Menounos et al., 2013; Vacco et al., 2010a), with a parameterized deposition rate (Jouvet et al., 2011) or a simplified deposition or ice-flow scheme (Anderson and Anderson, 2016; Rowan et al., 2015). Advances in ice-flow modeling are one component of studying the debris-ice system and need to be coupled to surface mass-balance schemes and supraglacial debris-transport schemes that adequately represent conditions for target glaciers. On Earth, modeling the co-evolution of ice and debris is done to diagnose glacial landform development, understand controls on modern glacier state, and project glacier behavior. As an example, Figure 11 illustrates processes that are important to consider in a model of debris-covered glacier change due to climate warming (net mass loss). A model that includes these processes can be applied to constrain debris delivery, debris accumulation on the surface, surface elevation change and terminus change over time, as well as proglacial landform development where the co-evolution of ice and debris is critical (e.g. Rowan et al., 2015).

OPEN IN VIEWER

Figure 11.

Schematic of key processes and landforms for a generic debris-covered glacier in the Himalaya. (a) Illustrates the case when the glacier is in balance with climate forcing: debris is delivered to the glacier from hillslopes, gets incorporated into the ice in the accumulation area and emerges in the ablation area due to ice flow and ablation. Debris can also be transported on the glacier surface (supraglacially) and be deposited at the terminus to form a terminal moraine. (b) Illustrates the case when the glacier experiences net mass loss: the equilibrium line altitude (ELA) moves up glacier, debris accumulates on the surface, the glacier surface lowers even while the terminus position remains stable, and the debris-covered tongue of lower glacier ice can dynamically detach from the upper glacier.

4.1 Toward process-based understanding of glacial landform evolution on Earth and Mars

Process-based understanding of landforms requires an understanding of the main processes that drive landform evolution, and we have focused our discussion in previous sections around processes specific to active debris-covered glaciers on Earth and related this to what is known on Mars. But, of course, a broader view is also needed. Debris-covered glaciers are found in mountain environments, where slope, aspect, ice thickness, ice-flow speed, bedrock type, and erosional history of the landscape are all important toward understanding sediment sources and sediment transport within an individual (local) glacial system. Glaciers can abrade the landscape (where entrained rock scrapes the bedrock as ice flows), fracture the landscape along weaknesses in the bedrock, and erode the landscape as ice flows (e.g. Benn and Evans, 2010: 263). This understanding is relevant to annual glacier change, decadal glacier evolution, and landform development during long-term (multi-decadal to centennial to millennial) glacier retreat that leads to ice loss and/or ice preservation. For example, the distribution of buried (remnant) ice in the proglacial area affects how paraglacial processes lead to sediment redistribution, ice degradation or preservation, and landform development (e.g. Cook et al., 2013; Irvine-Fynn et al., 2011). Paraglacial adjustment involves how landscape evolution has been primed by the transition from glacial to non-glacial conditions, particularly with regard to sediment availability (Ballantyne, 2002). For example, modern rock-slope failures are one way that slopes continue to relax in response to glacial sediment loading and deglacial climate forcing – both of which can mobilize sediment and affect glacier and landscape evolution.

Glacier type and glacier state are a function of the glacier’s regional setting and not only the local characteristics of its nearby environment; for example, equatorial, mid-latitude, and high-latitude debris-covered glaciers on Earth evolve quite differently from one another. While glacial landform generation (e.g. moraines, cirques) can be described based on general physical processes, if and how these processes manifest differs both regionally and temporally. As an example, Figure 11 highlights the regional imprint and impact of glacier retreat in the Himalaya, where debris-covered glaciers typically lose most of their mass by surface lowering because the terminus region is covered with a thick debris layer (Rowan et al., 2015). When the glacier is in balance with the climate forcing, debris is primarily sourced in the accumulation area from headwall and hillslope erosion and then exhumed in the ablation area, where resurfaced debris can be transported by glacier flow and form a terminal moraine. During a period of net mass loss, the lower-elevation portions of the glacier experience significant surface lowering and the extent of debris coverage increases. A large portion of the lower glacier may become inactive and can dynamically detach from the upper region of the glacier; eventually, it may be possible for the thick debris-covered lower glacier to physically detach from any remaining active portion of the glacier (e.g. as demonstrated in modeling by Rowan et al., 2015 and Vacco et al., 2010b). In general, as debris cover increases, ice velocity decreases, and surface lowering occurs with or without terminus retreat that can lead to extended tongues of ice isolated from the atmosphere by thick debris cover (e.g. Rowan et al., 2015). Hence the presence of a detached debris-covered ice tongue may constrain the climate conditions under which it formed (e.g. Vacco et al., 2010b), and this structure may evolve into other landforms as ice continues to be lost (e.g. Krüger et al., 2010).

How to interpret glacier type on Mars, where the modern state of landforms primarily reflects processes that occurred in the past, is a challenge. We acknowledged the difficulty on Earth of classifying glacial landforms, and then determining the processes and past climate conditions that led to the evolution of that landform. This is even more difficult on Mars, where interpretations are based primarily from remote sensing data and where a long temporal history of surface processes can be preserved. However, what has already been discovered about the ice and climate histories on Mars has led to relatively sophisticated questions about how ice responds to past climate changes.

A major advance in planetary science was the recognition that glaciation has been widespread on Mars, and the leading hypothesis is that the global accumulation and recession of ice are orbitally controlled (e.g. Head et al., 2003, 2005). The significant mid-latitude ice reservoirs that are buried beneath the surface and covered with debris provide evidence for these geologically recent Martian glaciations. Related to the formation and evolution of mid-latitude lobate debris aprons and glacier-like forms, the processes we have to better understand for the suite of landforms observed include: external controls on glacial flow (i.e. accumulation, ablation, air temperature, basal conditions), internal controls on glacial flow (i.e. ice deformation, debris loading), glacial erosion, glacial deposition, the extent of glacial activity, and the timing of glacial cycles (see also Hubbard et al., 2014). In addition, we need to better understand if and how landform transitions may have occurred on Mars (including activation, stagnation, and any reactivation of the same landform type) and their relationship to continuum process models developed on Earth.

As an example of how individual features, as well as regional landform relationships, may be possible to interpret in terms of key processes, Whalley (2009) provided a general framework to evaluate different periglacial and glacial landforms as a function of ice activity and debris input (Figure 12). This framework highlights how each landform state can be informative about the controls on landform evolution. While much progress has been made on Mars, we still need to determine the number of landform distinctions that are possible to identify (also advocated by Whalley and Azizi, 2003), especially with respect to small-scale remnant forms like moraines, lobes, and ramparts. Moreover, additional transitions and landform instability can occur during a paraglacial period (the unstable interglacial period), so features can reflect multiple phases of evolution, in addition to multiple phases of glaciation. Evidence for polyphase glaciation has been identified on Mars in select regions (e.g. Brough et al., 2016b; Dickson et al., 2008) and in relation to select landform types such as superposed glacier-like forms (e.g. Hepburn et al., 2020; Levy et al., 2007). Constraints on the timing and phasing of glaciations that can be inferred from landform morphologies are critical toward our ability to address the fundamental question on both Earth and Mars of “What paleoclimate conditions permit significant ice deposition and flow?”

OPEN IN VIEWER

Figure 12.

Periglacial and glacial landforms can be evaluated as a function of ice activity and debris input, and additional transitions and landform instability can occur during a paraglacial period. Most definitions are given in Section 1, but see also Whalley (2009) for complete discussion. Dots indicate relative debris content and shading broadly groups glacier types. Ice activity and debris input do not have to be constant in time, and these types are not necessarily distinct (but can be distinguished in many regions). Considering any given deposit as a function of debris and ice sources can help to identify formation processes as a function of climate controls.

Understanding how climate, ice dynamics, and environmental conditions control the evolution of ice-rich and ice-remnant landforms on Mars is an interdisciplinary research goal. While robust ice-rich and ice-remnant feature categorizations (including lobate debris aprons and glacier-like forms) are critical to that understanding, more work can be done to map and classify all ice-remnant features, including better constraining the volume of buried ice and considering the spatial distribution of all features (including those not yet categorized) to understand how they jointly evolved over time. Detailed local mapping will continue to be important (e.g. Brough et al., 2016b; Tsibulskaya et al., 2020), especially as new data sets become available and new interpretations become possible. As part of this effort, a glacial landsystem approach might be usefully applied to Mars that can comprehensively analyze the landscape including the landforms and sediments involved in all process relevant to landscape generation (e.g. Benn and Evans, 2010: 707; Evans, 2003). Such system-scale studies will likely be an important part of understanding supraglacial debris emplacement, supraglacial debris evolution, and the effects of debris on glacier topography and ice flow on Mars.

4.2 Outlook and recommendations for future research

Landforms formerly or presently containing ice, are used as indicators of climate conditions, and landform morphology can be used to constrain key processes controlling landform evolution. Studying active glacial, periglacial, and paraglacial environments that exhibit characteristic landforms due to the interaction of ice, climate, soil, and rock leads to a better understanding of the internal and external controls on the evolution of these landforms in response to the processes that shaped them. Advances in our understanding of how surface processes develop ice-associated surface landforms will continue as new field and remote-sensing data are collected, laboratory studies are conducted, and numerical models progress. Remote-sensing data are critical to our understanding of landform distribution and diversity on Earth and Mars, but on Earth, field-based results or new field studies at specific sites are often targeted in order to further develop and/or evaluate regional and system-scale relationships. Ground-based observations on Mars are much more limited, so extending our terrestrial knowledge to analogous extraterrestrial environments is a necessary part of advancing our understanding of processes that control landform evolution on terrestrial planets.

In this review we have focused on how active debris-covered glaciers on Earth may inform our understanding of debris-covered ice on Mars, particularly glacier-like forms and lobate debris aprons. The co-evolution of debris and ice is especially important on Mars because all Martian mid-latitude ice-rich and ice-remnant features have a debris cover that protects the remaining ice from rapid insolation-driven loss. We have reviewed how supraglacial debris can be emplaced on terrestrial glaciers by slope erosion, headwall sources, rock/debris avalanches, volcanic tephra, and windblown dust. The thickness and distribution of debris alters the glacier surface-atmosphere exchange, and debris affects surface topography primarily by enhancing or inhibiting ablation, as well as altering the ice-flow field within a debris-covered glacier. Three key takeaways in relation to possible avenues for future research include work toward how landforms are classified, advances in comparative planetology, and new understanding that may be gained from future missions.

1) How landforms are classified. Glaciers on Earth are currently retreating, and in many cases climate warming, as well as terrain degradation, are leading to debris-covered glacier generation or debris-cover expansion, and possibly eventual glacier stagnation, exhibiting a continuum of forms, from active to inactive to relict, that are still undergoing significant changes today. Many studies of individual glaciers have been conducted in order to address site-specific questions – we have learned the ways in which landscape heterogeneity around the world contributes to how different glacial forms evolve as a function of different environmental and climatic controls. For example, in parts of the Himalaya steep headwalls with consistent rock and ice avalanches distinctly source and control glaciers in that region. In Iceland, ashfall can alter glacier ablation. In the Aconcagua and Maipo basins in the central zone of the Andes, debris-covered glaciers and rock glaciers outnumber “uncovered glaciers” by a factor of 2.3 to 1 (see Janke et al., 2015). Despite regional differences, satellite observations and advances in numerical modeling have led to improvements in generalized system-scale understanding. Transitions between clean-ice glaciers, debris-covered glaciers and rock glaciers that have been observed over the past decades can be modeled as a function of debris input, climate forcing, and ice flow. Furthermore, site-specific field studies have informed our understanding of system-scale feedbacks between ice and debris that control glacier and glacial landform evolution. Classification of landforms on Mars in relation to landforms on Earth, for example as “debris-covered glaciers,” matters because of how the processes controlling landform evolution are implied by the type of landform that is classified. While glacier-like forms and lobate debris aprons on Mars are the debris-covered ice-rich or ice-remnant landforms that may be best informed by our understanding of debris-covered glaciers on Earth, future work could include new mapping and new work to subdivide existing classifications by landform attributes (see section 2.4; e.g. Brough et al., 2019; Levy et al., 2014; Milliken et al., 2003; Souness et al., 2012). This is a recognized research priority (e.g. Whalley and Azizi, 2003) that would involve diagnosing whether landforms are ice-rich and active, or if the surface form is a remnant from past ice activity, as well as whether observable attributes uniquely relate to terrestrial landforms.

2) Advances in comparative planetology. Debris cover alters the flow rate, shape and length of glaciers. From repeat imagery and topography alone we can determine that the lower portions of debris-covered glaciers often have lower surface slopes and lower ice-flow velocities. In addition, an insulating debris cover may enable glaciers to extend to (or remain at) lower elevations because ablation is reduced. The shape of an ice mass can be diagnostic of ice-flow history, and if the accumulation and temperature histories can be reasonably estimated then models can be used to evaluate the sequence of glacier retreat and/or advance over time. The problem is often posed the other way, where moraine sequences and/or decadal-to-centennial records of glacier-length change are used to constrain past climate (e.g. Oerlemans, 2001; Mackintosh et al., 2017). However, glaciers can respond to any climate variations – including those that occur in a constant (but noisy) climate and on interannual timescales – and on Earth they respond with a time lag that is often multiple decades long, which can complicate the attribution of an observed change in glacier length to the magnitude and timing of climate change (e.g. Roe and Baker, 2014). But what we have learned from Earth can be applied to Mars when attempting to interpret moraine-like sequences and evidence for recession of glacier-like forms (e.g. Brough et al., 2016a). High-resolution topography across debris-covered glaciers makes quantitative analyses increasingly possible. Future work could include more detailed evaluation of moraine-like assemblages on Mars, and to the extent possible, scrutinizing if mapped or unmapped moraine-like ridges are most likely glacial deposits or could be pronival ramparts from moraines, protalus rock glaciers, or slope-failure deposits (e.g. Hedding, 2016). These analyses will support an understanding of the extent of glacial and periglacial activity, and also possibly paraglacial activity, as well as constrain whether there is evidence of landform deposits or landform transitions on Mars that may have developed similarly to landforms associated with debris-covered glaciers on Earth. While equifinality is possible on Earth and Mars, comparative planetary geomorphology may boost confidence in landform interpretation on both planets.

3) New understanding from future missions. Debris cover typically thickens toward the terminus over time, and in general the distribution of debris can be diagnostic of debris-loading and ice-flow histories. Current radar imaging with SHARAD (15–25 MHz; ∼10 m vertical resolution) has detected the basal contact of select lobate debris aprons and constrains the ice content (e.g. Petersen et al., 2018). Future radar sounding using frequencies of hundreds of MHz, and with a resolution of 1 meter or less, would greatly elucidate the upper tens of meters across buried ice masses on Mars, perhaps revealing spatial heterogeneity or homogeneity in debris thickness – measurements that could be interpreted as a function of governing processes at the glacier-scale as well as regionally. In addition, debris distribution in combination with surface landform characterization is important toward constraining past activity from the observed modern state. On Mars, the relative timing of glacial and paraglacial activity may be possible to constrain using the spatial distribution of surface features, their properties (including possible debris sources; debris-cover thickness; surface slope; length; area; and distributions of crevasses/cracks, moraines, mounds, and surface lineations), and any regional relationships between features and their host terrain (e.g. alcoves, massifs, craters). Some of this mapping has been done (e.g. Brough et al., 2019; Levy et al., 2014; Souness et al., 2012), but higher-resolution imagery, topography and radar observations would continue to further our understanding and support new work.

Future missions to Mars have the chance to collect orbital data, ground-based measurements and sample sequestered subsurface glacial ice (a number of priorities related to Mars’ ice and climate history have been identified; for example, MEPAG ICE-SAG, 2019; Smith et al., 2020). Some important new datasets to be considered include orbital radar sounding with a vertical resolution of centimeters to meters that can detect: debris-ice interfaces of buried ground-ice deposits; ice and any debris bands that may remain in glacier-like forms; and structures in the debris layer, including the contact with the buried ice within lobate debris aprons. Present-day atmospheric and surface sources and sinks of water ice and dust/debris are important to measure, especially at lower altitudes and over as much time as possible. Sample return is a key NASA objective in the coming decades, and different forms of in-situ sampling have also been considered (e.g. MEPAG, 2018). Linking present-day processes and landforms to past processes on Mars will likely take an approach analogous to terrestrial studies, where orbiter data at a global-to-regional scale are used to develop specific hypotheses about landform evolution that can be evaluated by ground-based data at a local scale. Indeed, if local-scale observations on Mars are regionally (or globally) representative, then some questions may be easier to answer on Mars than on Earth. The questions discussed here that relate primarily to the evolution of Martian glacier-like forms and lobate debris aprons should be considered along with key objectives that have been defined more generally in relation to all ice reservoirs on Mars (e.g. Smith et al., 2020). In addition to providing insight to physical processes, understanding debris-covered ice on Mars has important resource ramifications for future landed investigations and any future crewed missions to Mars planning to extract ice that is currently covered by dust/debris (e.g. Abbud-Madrid et al., 2016).

Landform complexity is observed on Earth and Mars. While the governing processes may or may not be similar on both planets, interplanetary associations that can be made are qualitatively strong, and existing quantitative applications of terrestrial knowledge to Mars are also compelling. Continuing to apply our state-of-knowledge, as well as ongoing advances in understanding, is critically important for both terrestrial and planetary science in the ways that they intersect – there are exciting opportunities for future surface-process studies that are interdisciplinary and interplanetary. On both Earth and Mars, new data that advance how we connect site-specific observations to physical processes that act locally, regionally and globally is a robust research framework from which to generate new knowledge.