Australia: The Land Where Time Began

A biography of the Australian continent 

Regulation of Ice Stream Flow Through Subglacial Formation of Gas Hydrates

A primary control of the stability of ice sheets are variations of the flow of ice streams and outlet glaciers, but a comprehensive understanding of the key processes that operate at the ice-bed interface has remained elusive. For the maintenance of force balance in a subglacial environment that is well-lubricated/high slip basal resistance is critical, especially at sticky spots, local zones of high basal traction (Alley, 1993). In this paper Winsborrow et al. considered the influence of subglacial gas-hydrate formation on the dynamics of ice streams, and the potential of gas hydrates to initiate and maintain sticky spots. It is documented by geophysical data that about 20,000 years ago a major palaeoice-stream drained the Barents Sea-Fennoscandian ice sheet. In this paper Winsborrow present the results of this study that reveal a sticky spot covering an area of about 250 km² that coincided with subsurface shallow gas accumulations, fluid expulsion on the seafloor and a fault complex that is associated with deep hydrocarbon reservoirs. It is proposed by Winsborrow et al. that gas that migrated from these reservoirs formed gas hydrates under subglacial conditions of high pressure and low temperature. The subglacial sediments were desiccated and stiffened by the gas hydrate which thereby strengthened the subglacial sediments which promoted high traction – a sticky spot – which regulated the flow of the ice stream. It is common to find deep hydrocarbon reservoirs beneath both present and past glaciated areas, which implies that regulation of subglacial dynamics by gas hydrates could be a widespread phenomenon.

The degree to which gravitational driving stress is resisted by lateral shear at its margins, and friction at its base, largely determines the flow if ice streams. The spatial distribution of subglacial traction is critical to the overall stability of ice streams as lateral drag from adjacent ice that is moving more slowly is the primary resistive force, but this is not sufficient to balance the driving stress. As a consequence, variations in the velocity of the ice stream will be caused by variations in the basal friction, with the result that fast flow is promoted by a well-lubricated weak bed. It is indicated by observations from the Siple Coast ice stream, West Antarctica, that the basal shear stress is focused disproportionately on localised areas of high basal traction, so-called stick spots, which are surrounded by a deformable bed that is weaker and well-lubricated (Alley, 1993). A critical role is therefore played by the distribution and extent of sticky spots in regulating the flow of ice streams, with the shutdown of ice stream flow observed to follow on the widespread development of sticky spots (Tulaczyk et al., 2000; Anandakrishnan & Alley, 1997).

If subglacial sediments are desiccated and stiff, or if freezing conditions prevail at the ice-bed interface, sticky spots can form (Alley, 1993; Stokes et al., 2007). The shear strength of subglacial sediment is related directly to normal effective stress (ice load minus water pressure, though pore-water pressure is referred to explicitly in the case of basal sediments). As a result of this a significant increase of sediment strength is caused by a small reduction of basal pore-water pressure. There are 2 processes that can trigger this initial water pressure change. Reorganisation of, or changes in, the availability of subglacial meltwater ‘water piracy’ (Anandakrishnan & Alley, 1997; Alley, Anandakrishnan, Bentley & Lord, 1994), is the first, or the second, basal freeze-on that is caused by ice thickness changes, or the increase by advection of cold surface to the bed (Tulaczyk, Kamb & Engelhardt, 2000; Bougamont, Tulaczyk & Joughin, 2003; Christoffersen & Tulaczyk, 2003). Winsborrow et al. introduce a 3^rd mechanism for the formation of sticky spots based on observation from a palaeoice-stream bed on the continental shelf of Norway: pore-water piracy and stiffening of sediment as a result of the accumulation of subglacial gas hydrate.

Håkjerringdjupet is a cross-shelf trough in the southwest Barents Sea that has been glacially overdeepened. Håkjerringdjupet drains ice westwards from the mainland to an ice margin that was located at the edge of the shelf (Ottesen, Dowdeswell, & Rise, 2005) at the time of the Last Glacial Maximum (LGM) approximately 24,000 years ago, when an ice sheet covered the entire Barents Sea and the Norwegian continental shelf. The Troms-Finnmark Fault Complex dissects the trough, and a complex glacial landform assemblage characterises the seafloor. Highly elongated streamlined sedimentary deposits, more than 50 km in length and with a width in the order of 500 m, which are aligned to the axis of the trough. These are observed across the entire trough, except for the northern part, west of the fault complex and mega-scale glacial lineations (MSGLs), which formed beneath an ice stream that was flowing rapidly (King, Hindmarsh & Stokes, 2009; Stokes & Clark, 1999). There are 2 broad deposits with a wedge shape that have asymmetrical profiles, steeper on the down-ice slopes, and on their surface are MSFLs. These have been interpreted as being the grounding zone wedges (GZWs), which had formed during standstills or readvances in the course of the retreat of the ice margin (Ottesen et al., 2008; Winsborrow et al., 2012). Landform assemblages of MGSLs and GZWs have been identified across wide areas of continental shelves that were previously glaciated, their formation being attributed to the ice stream retreat that was rapid and episodic (Dowdswell et al., 2008).

The hummocky morphology of the northern part of the Håkjerringdjupet, which is comprised of a broad depression bounded by the lateral margins of the Troms-Finnmark Fault Complex and by an area of irregular hills that are located beyond the western margin of the fault complex, which is unusual for a palaeoice-stream. The depression, which has an area of approximately 250 km², represents an overdeepening of 90 m from the seafloor to the east of the fault complex. Beneath the seafloor depression the boundary between the bedrock and the glacigenic sediment is also overdeepened and rough. The irregular hills have a maximum elevation of approximately 80 m relative to the surrounding seafloor and cover an area similar to that of the depression. Across much of Håkjerringdjupet there is a uniform thickness and stratified acoustic signature of older glacigenic sediments that were deposited prior to the deglaciation (Rokoengen et al., 1979), though they are largely absent from the depression, and within the hummocky hills they have a thickness that is unusually great and varying. Also, within the hills the older glacigenic sediments have a seismic reflection pattern that is chaotic that is comprised of blocks of sediment that are oriented randomly and are acoustically stratified. Winsborrow et al. interpreted this to be a glacitectonic hill-hole pair that formed from rafted subglacial sediment and bedrock which have been displaced, creating a source depression, which was transported with the overlying ice, being deposited to form irregular hills, which is consistent with earlier work (Sættem, 1994). According to Winsborrow et al. glacitectonic rafting of sediment/bedrock occurs along a decollement of low strength which may represent the boundary between consolidated sediment and soft sediment, or sediment that is frozen or non-frozen. Slow flow of ice is implied by these processes in a subglacial environment that is high traction/low-slip, which contrasts with the remainder of the bed of the Håkjerringdjupet palaeoice-stream, which records rapid ice flow on a subglacial environment of low-traction/high-slip. Sticky spots are suggested in the geomorphic record to appear as a disruption to a classic pattern of parallel MSGLs that have been formed by streaming ice flow (Stokes et al., 2007). Winsborrow et al. therefore conclude that the glacitectonic source depression and sediments that have been rafted are the geologic imprint of a sticky spot on the bed of the Håkjerringdjupet ice stream.

Across the seafloor of Håkjerringdjupet there are many small circular depressions, typically of less than 50 m diameter, and less than 7 m deep that have been interpreted as pockmarks, which were formed by focused flow of fluid through the seafloor. Winsborrow et al. suggest the presence of pockmarks is consistent with migration in the past of gas from hydrocarbons in the reservoir below (Hovland & Judd, 1988). The highest density and largest examples are found in the source depression, within the fault complex, their distribution and size not being uniform. Pockmarks can also be observed to the east of the fault complex, where there is a tendency for them to be less dense and smaller. A sharp boundary is present between seafloor areas with and without pockmarks, and pockmarks are to a large extent not present in grounding wedges, rafted sediments and the area of MSGLs in the southern part of the trough.

It has been shown by seismic data that there are multiple faults in the central part Håkjerringdjupet and high-amplitude, phase-reversed reflections along the dipping layers that are classic indicators of accumulations of gas and the migration along faults from deeper reservoirs. The shallow accumulations of gas have been shown to be restricted to the substrate beneath the pockmarked surface depression, which is also the zone of maximum fault displacement. The fault escarpment is less abrupt in the southern parts of the trough, with no deep seafloor being observed and there is a lack, to a large extent, of pockmarks and shallow gas accumulations. Winsborrow et al. therefore infer there is a link between gas migration from deeper reservoirs that is structurally controlled and the formation of the Håkjerringdjupet sticky spot. Gas was found by a discovery well drilled 15 km north of Håkjerringdjupet (Wellbore 7019/1-1) in 2 reservoirs, Middle Jurassic and Lower Cretaceous, and the Troms Fault Complex has been found to be associated with major hydrocarbon reservoirs in the southwestern Barents Sea.

Thermogenic gas from deep reservoirs migrates upwards along faults and permeable layers of sediments as it is not as dense as either water or sediments. In the Arctic, gas migrating from leaking petroleum systems may be trapped in the shallow subsurface as gas hydrates, ice-like solids, that form under conditions of high pressure and low temperature, the conditions present in marine sediments and permafrost (Sloan & Koh, 2008). At the present temperature and pressure conditions in Håkjerringdjupet are just outside the boundary for the formation of methane hydrate. Under glacial conditions, however, ice that was grounded in Håkjerringdjupet, as is evidenced by subglacial bedforms such as glacial lineations on a mega scale, produced the conditions of high pressure and low temperature necessary for a thick zone, of approximately 400 m, of methane hydrate stability beneath the ice stream. As gas migrating into the zone of methane-hydrate stability encountered pore water within the subglacial sediments and shallow sedimentary bedrock that was under high pressure, the methane combined with the pore water to form methane hydrate. Winsborrow et al. say gas hydrate is expected to have formed in high saturation patches that were localised around faults and fractures and areas of porous sediment, given the bedrock that was highly fractured and the heterogeneous nature of the subglacial sediments within Håkjerringdjupet (Rokoengen et al., 1979), instead of forming a continuous layer. The host sediments are desiccated and stiffened by gas hydrate as a result of pore water piracy into the gas-hydrate cage structure, connection of grains, and of the greater strength and lower volume of the gas hydrate compared with its constituent gas and ice (Walte et al., 2009; Winters et al., 2004; Hyodo et al., 2013; Durham et al., 2003). In the McKenzie Delta, northern Canada, sediments bearing gas hydrates display an increase in shear strength of more than 8-fold when gas hydrate develops, from initial shear strength of 0.8 to 6.7 MPa with 60 % hydrate saturation (Winters et al., 2004). The strength of sediments that host gas hydrate and shallow bedrock in Håkjerringdjupet would have been far in excess of the “weak” till that was recovered from beneath the Siple Coast ice stream, West Antarctica (Kamb, 1991) and driving stresses that are typical of active streams, which have a tendency to not exceed 20 kPa, assuming a similar relative shear strength increase.

Winsborrow et al. suggested there would be basal friction that would have been locally enhanced by stiff sediments and shallow bedrock beneath the Håkjerringdjupet ice stream which would provide resistance to the flow of the overlying ice. A strong friction bond would probably have been formed with the overlying ice by the sediments of high shear strength which would favour glacitectonic decollement deeper within the substrate. The glacitectonic hill-hole pair that is visible at present on the seafloor was formed by thrusting and subsequent deposition of sediments and bedrock containing hydrate by the ice, which was slow moving.

There are major hydrocarbon reservoirs across the shelf areas that were previously glaciated by the Eurasian ice sheet complex, and several palaeoice streams flowed over substrates that were associated with confirmed hydrocarbon systems. Widespread gas hydrate formation, particularly at reservoirs that were leaking, that had focused migration of gas, would have been promoted by the subglacial conditions of high pressure and low temperature prevailing beneath these ice streams, and consequent regulation of ice stream basal thermomechanics and flow. Across the footprint of the palaeoice sheet multiple observations may now require to be reinterpreted on the basis of this hypothesis. The large glacitectonic landforms and till that was heavily overconsolidated associated with the Sklinnadjupet ice stream (Ottesen et al., 2005; Sættem et al., 1996) flowing over the extensive reservoirs of hydrocarbon, and the widespread presence of overconsolidated glacial sediments across the southwestern part of the Barents Sea (Sættem et al., 1992) are consistent with the regulation of the flow of ice streams by gas hydrates.

Based on the presence of extensive sedimentary basins and modelling studies (Wadham et al., 2012; Wallmann et al., 2012) it is proposed that abundant gas hydrate accumulations are present beneath the ice sheets of Greenland and Antarctica. Also, gas hydrates have been identified in ice core samples obtained from above the subglacial Lake Vostok in East Antarctica (Uchida et al., 1994). The role of potentially widespread gas hydrate reservoirs in the modification of the thermomechanical regime at the base of contemporary ice sheets, which makes them critically sensitive, as well as their impact on ice steam force balance and dynamics has, so far, not been recognised. This control that was previously unforeseen, given the current lack of knowledge with regard to the distribution of gas hydrate, represents a significant unknown in attempts to model the current and future discharge and evolution of contemporary ice sheets, as well as their contribution to rising global sea levels.

Sources & Further reading

1. Winsborrow, M., K. Andreassen, A. Hubbard, A. Plaza-Faverola, E. Gudlaugsson and H. Patton (2016). "Regulation of ice stream flow through subglacial formation of gas hydrates." Nature Geosci 9(5): 370-374.