Basal sediment: The debris layer that exists at the ice–bedrock interface of a glacier or ice sheet and that is produced by glacial erosion of the underlying bedrock.
Basal sediment evacuation: The removal of basal sediment from the ice–bed interface by glacial and/or glaciofluvial sediment transport; most frequently used to describe the flushing of basal sediment from the ice–bedrock interface by subglacial water.
Subglacial drainage system: The network of drainage pathways at the ice–bed interface that convey water to the glacier margin or terminus.
Contemporary observations and theoretical analyses demonstrate that, for Temperate Glaciers (qv), the entrainment and transport of basal sediment by fluvial processes dominates all other glacial sediment transport ( Sediment Entrainment, Transport, and Deposition , qv) processes (see Hallet et al., 1996; Alley et al., 1997). Basal sediment evacuation by the Subglacial Drainage System (qv) therefore dominates in glacial Sediment Budgets (qv) (Alley et al., 1997), and the high Suspended Sediment Load (qv) of most glacier-fed rivers presents a major difficulty for the management and use of meltwater (Østrem, 1975; Bezinge et al., 1989; Bogen, 1989).
Sediment entrainment by subglacial water is difficult to observe directly. However, the relatively rapid transit of water through the subglacial drainage system means that the suspended load of glacier-fed streams should reflect closely the processes of sediment entrainment at the glacier bed. Analysis of the sediment load of glacier-fed streams has therefore been applied widely as a means of elucidating the processes of basal sediment evacuation (e.g., Gurnell et al., 1992; Clifford et al., 1995; Hodgkins, 1996; Willis et al., 1996; Hodson and Ferguson, 1999; Swift et al., 2005a).
Controls on basal sediment entrainment and evacuation
Investigation of basal sediment evacuation has been confined largely to the analysis of suspended sediment time series from glacier-fed rivers in temperate alpine environments (see, e.g., reviews by Gurnell [1987a, 1995] and Gomez ). These studies have highlighted considerable temporal variation in basal sediment evacuation at a wide range of scales (e.g., Bogen, 1988, 1996; Collins, 1990; Clifford et al., 1995) ( Suspended Sediment Dynamics , qv) that has been linked to a complex array of subglacial processes. Nevertheless, hydrologically-driven seasonal evolution of the subglacial drainage system ( Glacier Hydrology , qv) has been shown to exert a key control on rates and patterns of basal sediment evacuation because it greatly influences how and where subglacial water accesses, entrains, and transports basal sediment (e.g., Gurnell et al., 1992; Swift et al., 2005a).
The processes of subglacial erosion and subsequent comminution of erosion products at the ice–bedrock interface ( Glacial Erosion , qv) generate significant quantities of debris characterized by an extremely wide size distribution. The extent to which such debris accumulates at the ice–bedrock interface in the form of a basal sediment layer ( Till , qv) will depend largely on the efficacy of sediment evacuation by glacial and fluvial processes. Fluvial sediment transport is widely appreciated to dominate basal sediment evacuation (e.g., Alley et al., 1997), but the efficiency of basal sediment evacuation by fluvial processes is dependent on subglacial water having sufficient competence and capacity to entrain and transport available sediment, as well as being able to access basal sediment sources and/or stores.
Fluvial sediment entrainment and transport processes in the subglacial environment are largely similar to those that operate in subaerial streams (e.g., Alley et al., 1997). Notably, the efficiency of sediment transport or evacuation reflects, on the one hand, the competence and capacity of flow, and on the other, the size and accessibility/availability of sediment. Flow strength is therefore critical because it determines both flow competence (the largest particle a given flow is able to carry) and flow capacity (the total quantity of sediment of a given size that the flow can carry). Although flow strength is best defined in terms of flow energy or boundary shear stress, many sediment transport studies use flow velocity or discharge ( Discharge/Streamflow , qv), which in general provide convenient and acceptable approximations.
Both discharge and suspended sediment transport can easily be measured in glacier-fed streams ( Suspended Sediment Concentration , qv). As a result, the predictably strong relationship between these two factors has been demonstrated by a very large number of studies. However, many such studies have also reported changes in basal sediment transport that have occurred irrespective of discharge, and these range from short-term flushes of sediment lasting minutes or hours, to longer-term changes over a single melt season or a series of melt seasons. Such changes have been attributed to both the subglacial processes that determine basal sediment availability and the ease with which available sediment sources and/or stores can be accessed by elements of the subglacial drainage system.
Importance of subglacial drainage system morphology
Two morphologically distinct types of subglacial drainage exist beneath most temperate glaciers and, in the case of polythermal glaciers, beneath warm-based ice. The most widespread system, both spatially and temporally, will be some form of distributed system. Such systems convey small volumes of water derived largely from basal melting via a physically tortuous flowpath that may include flow within linked cavities, thin water films, and/or porous sediments. Changes in discharge within this system are accommodated largely by changes in the number or size of individual flowpaths and not by changes in flow velocity. Hence, despite being spatially extensive, such systems entrain and transport only the very finest fractions of available sediment (e.g., Humphrey and Raymond, 1994; Willis et al., 1996; Alley et al., 1997).
The efficiency of basal sediment evacuation increases significantly if a channelized drainage system is present. Channel systems typically form in summer in response to the very large quantities of surface melt that are conveyed to the glacier bed via Moulins (qv) and Crevasses (qv). Although physically and temporally discrete, channels convey large discharges at very high velocities, and are also sufficiently large to transport all but the very largest fractions of basal sediment. The resulting competence and capacity of the channel system is such that the annual glacial sediment budget may be dictated almost entirely by the annual spatial and temporal extent of channelized drainage (e.g., Alley et al., 1997; Swift et al., 2005a). Furthermore, a number of mechanisms related to the inception and growth of the channel network are considered to influence temporal patterns of basal sediment evacuation.
One important reason why the channelized system is so efficient at transporting sediment is because the increase in flow velocity that occurs as a function of discharge is extremely nonlinear when flow is confined within a subglacial channel. Alley et al. (1997) demonstrated theoretically that the transport capacity of unconfined channels typically varies as Q 2 because changes in discharge can be accommodated equally by changes in flow depth, width, and velocity. However, for subglacial channels, which can only be enlarged by melting of the channel walls, rapid changes in discharge will be accommodated almost exclusively by changes in flow velocity, such that transport capacity will vary up to Q 4·5. Swift et al. (2005a) demonstrated that such relationships were a major factor behind seasonal changes in the efficiency of basal sediment evacuation at Haut Glacier d’Arolla, Switzerland, where sediment transport during periods of distributed drainage was shown to vary as Q ∼2.3, and during periods of channelized drainage was shown to vary as Q ∼3.2.
Observations at surging glaciers ( Glacier Surging , qv), such as Variegated Glacier, Alaska, also demonstrate the importance of drainage system morphology for basal sediment evacuation (e.g., Humphrey et al., 1986; Kamb and Engelhardt, 1987; Humphrey and Raymond, 1994). Surging occurs when the channelized drainage system fails to develop as expected, promoting high basal water pressures and hence rapid basal sliding that further discourages channel formation. These studies show that basal sediment evacuation remains low during each surge, despite distributed drainage having wide access to the glacier bed, and despite high rates of basal sliding that should enhance rates of both basal sediment disturbance and subglacial erosion. In contrast, flushing of basal sediment occurs as the surge ceases and channelized drainage is reestablished.
Importance of basal sediment availability
Changes in sediment availability (and also accessibility) are believed to be significant because the sediment loads of glacier-fed streams are often only weakly correlated with discharge at annual scales. For example, Bogen (1989, 1996) linked poor correlation between annual sediment load and discharge from various Norwegian glaciers to changes in subglacial erosion rate, but also changes in the location of arterial subglacial channels across successive melt seasons. Similarly, Gurnell (1995) linked annual sediment load from Tsidjiore Nouve, Switzerland to the rate of glacier advance, citing changes in sediment availability as a result of the rate of disturbance of subglacial drainage.
Studies of sediment transport in glacier-fed streams at shorter timescales have identified changes in sediment availability during specific periods of the melt season. For example, enhanced availability of basal sediment has commonly been observed in spring and has been explained in terms of: (1) a build-up of basal sediment over winter as a result of the absence of surface melt reaching the glacier bed (e.g., Liestøl, 1967; Hooke et al., 1985; Collins, 1989, 1990); (2) unprecedented access of water to the glacier bed as surface melt swells the distributed system (see previous references); (3) enhanced disturbance of existing subglacial drainage pathways and/or sediments as a result of enhanced rates of Glacier Sliding (qv) (e.g., Collins, 1989; Hooke et al., 1985); and (4) drainage system reorganization associated with the rapid incision and growth of a channelized drainage network (e.g., Collins, 1989, 1990; Anderson et al., 1999).
Another common observation is an overall decline in sediment availability during the melt season. This has been attributed both to the reduced ability of water to access basal sediment as flow becomes increasingly confined to discrete channels (e.g., Collins, 1989, 1990; Gurnell, 1987b; Gurnell et al., 1992) and to the exhaustion of sediments that are accessible to such channels (e.g., Østrem, 1975; Hooke et al., 1985). Nevertheless, a number of studies have demonstrated increasingly efficient evacuation of basal sediment during the melt season (e.g., Clifford et al., 1995; Vatne et al., 1995). Swift et al. (2005a) demonstrated that this may be a result of the increasing peakedness of surface melt production during the melt season, as this should promote: (1) an associated increase in the sediment transport capacity of subglacial channel flow (Clifford et al., 1995; Swift et al., 2005b); and (2) an increasingly strong diurnally reversing subglacial hydraulic gradient that should encourage basal sediment to flow or deform toward channels (Vatne et al., 1995; Swift et al., 2005b).
Short-term controls on basal sediment evacuation
On hourly to weekly timescales, temporal patterns of sediment transport are controlled largely by discharge, particularly the diurnal changes in discharge that reflect the temporal pattern of surface melt production (e.g., Gurnell, 1987b; Clifford et al., 1995; Willis et al., 1996). However, diurnal patterns of sediment transport are complicated both by Hysteresis (qv), where sediment concentrations at equivalent discharges on the discharge hydrograph are observed to be higher on the rising limb than on the falling limb, and by flushes of sediment on a variety of temporal scales that may or may not be related to changes in discharge.
Hysteresis is commonly evident both during individual diurnal discharge cycles (e.g., Leistøl, 1967; Østrem, 1975; Collins, 1979; Bogen, 1980; Hammer and Smith, 1983; Gurnell, 1982, 1987b; Willis et al., 1996) and over multiple discharge cycles or events, notably where the magnitude of a discharge peak fails to equal or exceed that of a previous peak (e.g., Leistøl, 1967; Østrem et al., 1967; Church, 1972; Collins, 1979; Gurnell, 1982, 1987b; Richards, 1984; Schneider and Bronge, 1996). Leistøl (1967) suggested that such hysteresis occurred because changes in discharge in a subglacial channel are accommodated at least partly by changes in flow width, such that flow on the rising limb accesses successively larger areas of the glacier bed, whereas flow on the falling limb is confined to areas that have been stripped of basal sediment. Transport hysteresis might also reflect velocity hysteresis, which describes the characteristically higher flow velocities that occur on the rising limb (e.g., Richards, 1984).
Flushes of sediment lasting hours or days have been linked to rapid changes in flow routing. For example, flushes commonly occur during spring as surface melt promotes the reorganization of distributed drainage pathways into discrete channels (Collins, 1989, 1990) and as incipient channels release stored water and sediment from within a swollen distributed system (Warburton and Fenn, 1994; Anderson et al., 1999). Flushes have also been linked to short-term injections of water into the distributed system resulting from intense rainfall (e.g., Raymond et al., 1995; Denner et al., 1999), the drainage of ice-marginal or intra-glacial water stores (e.g., Collins, 1979; Beecroft, 1983), or temporary blockages in subglacial channels (e.g., Collins, 1990). The injection of water into the distributed system may further facilitate flushing by enhancing ice-bed separation and basal sliding (e.g., Anderson et al., 1999; Hooke et al., 1985; Raymond et al., 1995; Swift et al., 2005a).
Shorter pulses of sediment, often lasting only seconds or minutes, can also contribute significantly to the total annual sediment load (Gurnell and Warburton, 1990). These have been linked largely to instabilities in elements of the subglacial drainage system, including: (1) the establishment of new drainage pathways or the migration of existing ones (e.g., Collins, 1979, 1989, 1990; Gurnell, 1982, 1995; Gurnell and Warburton, 1990; Gurnell et al., 1992; Willis et al., 1996; Hodson et al., 1997; Vatne et al., 1995; Hodson and Ferguson, 1999); (2) disturbance of basal sediment or subglacial flowpaths by sudden changes in the rate of basal sliding (e.g., Willis et al., 1996; Swift et al., 2005a); (3) blockage of channels by channel-roof failure (e.g., Lawler et al., 1992); and (4) the deformation and/or collapse of basal sediment toward or into adjacent channels (Collins, 1979; Vatne et al., 1995; Clifford et al., 1995; Bogen, 1996).
Feedbacks and significance for glacial erosion rates and sediment yields
Efficient evacuation of basal sediment is likely to be very important in glacial systems ( Glacial Ecosystems , qv) because erosion of subglacial bedrock will be limited where debris at the ice–bedrock interface accumulates to form a basal sediment layer (Alley et al., 1997). Efficient flushing of debris is likely to maintain glacial erosion by enhancing ice–bedrock interaction that is also likely to enhance the comminution of larger debris ( Crush , qv) that is produced at the ice–bedrock interface by subglacial quarrying (Swift et al., 2002). The efficiency of basal sediment evacuation by subglacial drainage is therefore likely to be a dominant control on glacial Erosion Rate (qv) and Sediment Yield (qv).
The high sediment competence and transport capacity of subglacial channels indicates that the extent to which a spatially extensive channelized drainage network evolves during the melt season is also likely to be extremely important in dictating erosion rates and annual sediment yields (Swift et al., 2002; Swift et al., 2005b). Glaciers that evolve limited channelized systems and have predominantly distributed drainage are likely to have thicker and more extensive basal sediment layers that discourage ice–bedrock interaction and promote the entrainment and transport of sediment by glacial processes. The efficiency of basal sediment evacuation by subglacial drainage is therefore likely also to influence rates and styles of proglacial sediment deposition (Swift et al., 2002).
Areas of controversy and future research potential
Present understanding of basal sediment evacuation by subglacial drainage is limited largely to temperate ice masses. Few studies have been conducted at polar or polythermal glaciers (e.g., Hodgkins, 1996; Hodson and Ferguson, 1999), especially those glaciers that emanate from larger ice masses such as the Greenland ice sheet, even though the potential for surface melt to reach the ice–bed interface of such glaciers has now been demonstrated widely. As a result, the processes of basal sediment evacuation in such environments, including those that are associated with the seasonal evolution of subglacial drainage system morphology, may play an important but as yet poorly understood role in a wide array of subglacial processes at ice sheet scales.
The significance of fluvial sediment evacuation in glacial geomorphic systems is also poorly understood, most notably in terms of the importance of basal sediment evacuation in controlling rates of glacial erosion and resulting patterns of landscape evolution ( Glacial Geomorphology and Landforms Evolution , qv). Widely variable sediment yields (e.g., Hallet et al., 1996) indicate that glacier-to-glacier variation in the efficiency of basal sediment evacuation may be a critical control on rates of glacial erosion. Sediment transport by subglacial drainage may also play a critical role in the evolution of glacier-bed overdeepenings ( Glacial Overdeepening , qv) (Alley et al., 2003) and their geomorphic impacts ( Glaciohydraulic Supercooling , qv) (e.g., Swift et al., 2002; Cook et al., 2006). However, understanding is limited by an absence of sufficiently long-term records of sediment transport in glacier-fed streams, particularly with respect to bed-load (e.g., Alley et al., 2003).
Basal sediment evacuation by subglacial meltwater plays a key role in glacial systems. Notably, the efficiency of basal sediment evacuation is likely to determine the extent to which the products of glacial erosion are retained at the ice–bed interface (Alley et al., 1997), which is likely to dictate not only rates and patterns of subglacial erosion and glacial sediment yield, but also the proportion of debris in glacial versus fluvial sediment transport pathways, and as a result will greatly influence rates and styles of sediment deposition in ice-marginal environments ( Moraine , qv) (see Swift et al., 2002). Nevertheless, the importance of basal sediment evacuation by subglacial water for glacial geomorphic processes has largely until recently been neglected by the glaciohydrological and glaciogeomorphological literature.