Heart of Avalonia |
Reverse Ice Segregation
Reverse ice segregation occurs when stress acting on a macroscopic volume of ice causes water molecules from the ice to exit the crystalline state and migrate as pore water into an adjacent rigid permeable medium. The process closely parallels conventional ice segregation running in reverse, but includes an added requirement that a permeable medium be sufficiently rigid to resist mechanical deformation under the stress imparted by ice loading. Thus reverse ice segregation can occur when ice bears against well-indurated permeable rock, but the process would not be expected to occur when ice bears against a deformable granular medium like silt or soil.
Water confined to a capillary space can be considered a separate phase from bulk liquid water. Ice segregation and reverse ice segregation describe a pressure-mediated phase transition between crystalline water and pore water. These two complimentary processes somewhat resemble the process of regelation that operates in temperate glacial ice. Ice segregation and reverse ice segregation differ fundamentally from regelation in that regelation describes a reversible pressure-mediated phase transition between between crystalline water and bulk liquid water. Regelation depends on the pressure-induced reduction in the thermodynamic stability of crystalline water, while reverse ice segregation does not depend on this unusual water characteristic. Unlike regelation, which can occur in nature only over a narrow temperature range extending downward a fraction of one degree from zero degrees C, reverse ice segregation in a subglacial environment can theoretically operate at temperatures extending several degrees below zero degrees C.
Water confined to a capillary space can be considered a separate phase from bulk liquid water. Ice segregation and reverse ice segregation describe a pressure-mediated phase transition between crystalline water and pore water. These two complimentary processes somewhat resemble the process of regelation that operates in temperate glacial ice. Ice segregation and reverse ice segregation differ fundamentally from regelation in that regelation describes a reversible pressure-mediated phase transition between between crystalline water and bulk liquid water. Regelation depends on the pressure-induced reduction in the thermodynamic stability of crystalline water, while reverse ice segregation does not depend on this unusual water characteristic. Unlike regelation, which can occur in nature only over a narrow temperature range extending downward a fraction of one degree from zero degrees C, reverse ice segregation in a subglacial environment can theoretically operate at temperatures extending several degrees below zero degrees C.
The above diagram summarizes the main elements of reverse ice segregation. The diagram closely parallels diagrams shown in the section Crystal Growth Pressure, but with the direction of migration of pore-water molecules reversed. Accordingly, the mass (hence, volume) of the segregated ice participating in the process diminishes as the process advances. The process of reverse ice segregation was postulated in Technical Note 12 when discussing plucking by a cold glacier. However, the discussion presented in Note 12 did not include arguments that reverse ice segregation was anything more than a hypothetical process. It will be argued here that evidence for the occurrence of reverse ice segregation can be obtained from field observations.
Qualitative thermodynamic considerations
Reverse ice segregation is an unavoidable consequence of the equilibrium thermodynamics describing ice crystal growth pressure. Clearly, ice crystal growth pressure must have a limit. A totally confined region of ice being augmented by incoming molecules of pore water will accumulate pressure until crystallization is no longer thermodynamically favored. At this point, incoming molecules that augment the crystallized mass are statistically matched in number by outgoing molecules that decrement the crystallized mass. The outgoing molecules constitute reverse ice segregation. The equilibrium process is somewhat analogous to a process of ice sublimation into a water-saturated atmosphere. In a sublimation equilibrium, molecules bound to an ice surface escape and depart into the atmosphere, while a statistically equivalent number of air-borne water molecules condense onto the ice surface so as to leave the mass of ice unchanged.
A key question becomes: At what limiting pressure does the water-molecule migration underlying ice crystal growth pressure reach equilibrium? This question is critical in determining whether or not reverse ice segregation can occur under natural conditions.
Crystal growth pressure results, in part, from the enthalpy difference between a unit mass of water bound in an ice crystal and a unit mass of water confined in an adjacent pore space. Generally, the process of water transitioning from capillary confinement to crystallization in a zone of segregated ice is assumed to be an exothermic process. This assumption is supported by observations that frost heave powered by ice segregation is a spontaneous process over a wide range of conditions. The assumption that ice segregation is an exothermic process depends on the heat of interaction of water with capillary walls being less than the heat of crystallization of water. While the heat of crystallization of water changes little over a substantial range of natural conditions, the heat of interaction of water with pore walls depends on details of the placement and polarization of pore-wall molecules (hence rock composition) and on the size of the pores. Pore-wall interaction energy per unit mass of water will rise as pore size diminishes and surface contact area increases. Thus, limits to ice crystal growth pressure will vary both with variations in rock composition and with variations in rock texture.
A key question becomes: At what limiting pressure does the water-molecule migration underlying ice crystal growth pressure reach equilibrium? This question is critical in determining whether or not reverse ice segregation can occur under natural conditions.
Crystal growth pressure results, in part, from the enthalpy difference between a unit mass of water bound in an ice crystal and a unit mass of water confined in an adjacent pore space. Generally, the process of water transitioning from capillary confinement to crystallization in a zone of segregated ice is assumed to be an exothermic process. This assumption is supported by observations that frost heave powered by ice segregation is a spontaneous process over a wide range of conditions. The assumption that ice segregation is an exothermic process depends on the heat of interaction of water with capillary walls being less than the heat of crystallization of water. While the heat of crystallization of water changes little over a substantial range of natural conditions, the heat of interaction of water with pore walls depends on details of the placement and polarization of pore-wall molecules (hence rock composition) and on the size of the pores. Pore-wall interaction energy per unit mass of water will rise as pore size diminishes and surface contact area increases. Thus, limits to ice crystal growth pressure will vary both with variations in rock composition and with variations in rock texture.
The above diagram follows from the preceding discussion and illustrates the change in potential energy of a single molecule transitioning from an internal crystal environment to a pore environment. The indicated change in enthalpy, deltaH, reflects a difference in the interaction energy of the molecule with adjacent molecules. On the left side of the diagram, the interaction is with appropriately aligned water molecules in an ice crystal matrix. On the right side of the diagram, the interaction is with rock molecules lining the pore wall and with other water molecules that are not statistically optimally aligned. Because of the near-optimal hydrogen bonding arrangement in ice, it is reasonable to assume that the potential energy of the illustrated water molecule would be lower in the environment represented on the left side of the diagram.
Assuming approximate constancy of the heat of crystallization of ice, the value of deltaH becomes a function of ice pressure and of rock composition and texture. An upper limiting value for deltaH is the value corresponding to a transition from ice to bulk liquid-phase water since interactions of water molecules with polarized capillary walls would tend to lower the potential energy of water molecules and thus reduce deltaH. The heat of crystallization of water, calculated on a per-unit-volume basis (reasonable since the ice-water phase transition does not involve a large change in density), can be expressed directly as a pressure. The heat of crystallization, ~300 MPa, corresponds to the pressure at which ice segregation and reverse ice segregation would reach equilibrium (ignoring entropy) if water molecules experienced negligible pore-wall interaction energy.
Entropy changes must be considered alongside enthalpy changes when determining whether water molecules will statistically tend to accumulate in pores or in the crystalline state, given specified conditions of temperature and stress. It can be assumed that the entropy of crystalline water is lower than that of pore water because water molecules are relatively confined in ice and are relatively mobile in pores. Thus entropy considerations alone would favor reverse ice segregation over ice segregation. Ice segregation dominates at low pressure because ice segregation is an exothermic process while reverse ice segregation is an endothermic process. At sufficiently high ice pressure, the thermodynamic advantage afforded to the exothermic process diminishes because the added mechanical energy of the pressurized ice offsets its lower internal potential energy. At an ice pressure less than that needed to make to make the reverse ice segregation process exothermic, the process might nevertheless run spontaneously because of the increase in entropy. Cooling of the ice could be compensated by heat absorbed from the environment and by the conversion of mechanical energy (ice pressure X volume reduction) to heat.
In light of the qualitative thermodynamic considerations discussed above, it is reasonable to conclude that under natural conditions, the limit to ice crystal growth pressure might be reached since this limit is less than 300 MPa, and possibly, much less. Equivalently, reverse ice segregation could then reasonably be expected to be a naturally occurring process in a cold, deep, high-stress subglacial environment. Analysis of field observations is a necessary step in determining whether reverse ice segregation was or was not a significant factor affecting bedrock disruption by ice.
Assuming approximate constancy of the heat of crystallization of ice, the value of deltaH becomes a function of ice pressure and of rock composition and texture. An upper limiting value for deltaH is the value corresponding to a transition from ice to bulk liquid-phase water since interactions of water molecules with polarized capillary walls would tend to lower the potential energy of water molecules and thus reduce deltaH. The heat of crystallization of water, calculated on a per-unit-volume basis (reasonable since the ice-water phase transition does not involve a large change in density), can be expressed directly as a pressure. The heat of crystallization, ~300 MPa, corresponds to the pressure at which ice segregation and reverse ice segregation would reach equilibrium (ignoring entropy) if water molecules experienced negligible pore-wall interaction energy.
Entropy changes must be considered alongside enthalpy changes when determining whether water molecules will statistically tend to accumulate in pores or in the crystalline state, given specified conditions of temperature and stress. It can be assumed that the entropy of crystalline water is lower than that of pore water because water molecules are relatively confined in ice and are relatively mobile in pores. Thus entropy considerations alone would favor reverse ice segregation over ice segregation. Ice segregation dominates at low pressure because ice segregation is an exothermic process while reverse ice segregation is an endothermic process. At sufficiently high ice pressure, the thermodynamic advantage afforded to the exothermic process diminishes because the added mechanical energy of the pressurized ice offsets its lower internal potential energy. At an ice pressure less than that needed to make to make the reverse ice segregation process exothermic, the process might nevertheless run spontaneously because of the increase in entropy. Cooling of the ice could be compensated by heat absorbed from the environment and by the conversion of mechanical energy (ice pressure X volume reduction) to heat.
In light of the qualitative thermodynamic considerations discussed above, it is reasonable to conclude that under natural conditions, the limit to ice crystal growth pressure might be reached since this limit is less than 300 MPa, and possibly, much less. Equivalently, reverse ice segregation could then reasonably be expected to be a naturally occurring process in a cold, deep, high-stress subglacial environment. Analysis of field observations is a necessary step in determining whether reverse ice segregation was or was not a significant factor affecting bedrock disruption by ice.
Field observations
The direct observation of reverse ice segregation occurring in a natural environment or in a laboratory environment is outside the scope and capability of the Heart of Avalonia Frost-heaved Bedrock project. Alternatively, observations of relict ice-disrupted bedrock features can be used to infer the past operation of a reverse ice segregation process. The most abundant evidence is potentially found in numerous instances of apparent plucking by a cold-based glacier. However, this evidence is ambiguous because a combination of groundwater-driven ice extrusion and glaciotectonic action can, in principle, generate all observed features that are interpreted as cold-based glacial plucking. The question that necessarily arises is: Was the groundwater supply for ice extrusion sourced from cold glacial ice via reverse ice segregation, or was the groundwater present deep in the rocks and then drawn toward the surface by declining hydrostatic pressure as deglaciation progressed? The large volumes of displaced rock seen in many relict cold-based plucking occurrences suggests that the requisite groundwater did not travel a long distance. This observation implies a nearby glacial origin for the groundwater, but such a determination remains necessarily ambiguous.
Observations of grooves in bedrock formed by subglacial ice extrusion can be used to infer the past occurrence of reverse ice segregation with a lesser ambiguity. The approach involves looking for instances of grooved erosion by ice extrusion on the vertical sides of a frost-heaved monolith. Presumably the vertical sides of a rectangular frost-heaved monolith would have remained obstructed by bedrock and thereby immune to erosion by ice extrusion until the monolith accumulated enough ice beneath its base to raise its side surfaces into an exposed position. By the time the side surfaces became exposed, a frost-heaved monolith would be disconnected from pore water emanating from bedrock. Any pore water needed to cause erosion by ice extrusion on side surfaces would necessarily be sourced from pressurized ice reentering the monolith, that is, by reverse ice segregation.
Observations of grooves in bedrock formed by subglacial ice extrusion can be used to infer the past occurrence of reverse ice segregation with a lesser ambiguity. The approach involves looking for instances of grooved erosion by ice extrusion on the vertical sides of a frost-heaved monolith. Presumably the vertical sides of a rectangular frost-heaved monolith would have remained obstructed by bedrock and thereby immune to erosion by ice extrusion until the monolith accumulated enough ice beneath its base to raise its side surfaces into an exposed position. By the time the side surfaces became exposed, a frost-heaved monolith would be disconnected from pore water emanating from bedrock. Any pore water needed to cause erosion by ice extrusion on side surfaces would necessarily be sourced from pressurized ice reentering the monolith, that is, by reverse ice segregation.
Example 1
The above photo shows a low frost-heaved bedrock monolith occurring in an area strongly affected by erosion by subglacial ice extrusion. The side surface of the monolith visible facing the camera is a failure surface along a cross joint. A closeup of the side surface is shown below.
The top surface of the block shown above was eroded by ice extruded from bedrock, the grooved erosion pattern presumably forming before frost heave commenced and before an ice lens developed beneath the block. There is no evidence of erosion by ice extrusion on the vertical sides of the block. Apparently the developing ice lens prevented pore water from reaching the vertical sides after they became exposed. This example provides a baseline illustrating a complete lack of occurrence of reverse ice segregation according to the criteria discussed above in this section.
Example 2
The frost-heaved block seen in the above photo shows evidence of delamination of well-indurated ancient sedimentary bedrock by subglacial groundwater penetration and ice segregation. Two large central grooves (see photo below) in the block were apparently formed by erosion by subglacial ice extrusion. Close examination of the block indicated that the block was completely intact and that the prominent large grooves seen in the photo below were not appreciably levered apart by frost wedging during the Holocene.
A closeup of the region outlined by the cyan rectangle in the above photo is shown below. As with the previous example, the block seen above demonstrates a flat vertical side surface facing the camera that apparently resulted from bedrock failure along a cross joint.
Although the block shown above demonstrates wide, ice-eroded grooves on a vertical side surface, the observation falls short of demonstrating evidence for reverse ice segregation. The grooves appear to indicate preferred channels for groundwater transport and ice segregation. The channels were likely formed before the block was displaced from bedrock substrate. The grooves were potentially etched wider by Holocene erosion (small-scale frost wedging, rainwater and vegetation) that exploited weakness previously induced in the rock by ice crystal growth pressure and upward-directed subglacial ice extrusion. Note the numerous narrow white lines running vertically on the flat face of the rock. These lines follow joints that have been widened enough to provide a preferential foothold for lichen. The joints parallel tectonic foliation and reflect the overall delamination of the rock. An extensive partial delamination process is commonly observed to be closely associated with large scale bedrock frost heave and with erosion by ice extrusion.
Example 3
Several frost-heaved bedrock monoliths are shown above. The grooved surfaces and layered appearances of the monoliths indicate that erosion by ice extrusion was occurring prior to, or in conjunction with, a frost heave process. Note the monolith just to the left of the hammer. This block is grooved on the top, as might be expected, but is also grooved on the end facing the camera. These vertical-side grooves, if the result of erosion by ice extrusion, suggest that ice was extruded from the block after the block became underlain by an ice lens. This interpretation would imply that reverse ice segregation took place beneath the block, during or after frost heave, or that reverse ice segregation took place at the opposite end of the block as a result of glacial ice loading during or after frost heave.
A small monolith is highlighted by the cyan-colored circle shown in the photograph. A fissure adjacent to this monolith made it possible to compare the degree of erosion by ice extrusion that occurred on the above-ground portion versus the below-ground portion of this monolith.
A small monolith is highlighted by the cyan-colored circle shown in the photograph. A fissure adjacent to this monolith made it possible to compare the degree of erosion by ice extrusion that occurred on the above-ground portion versus the below-ground portion of this monolith.
The two photos above link roughly together to form a composite image of the vertical side surface of the monolith that was circled in cyan on the preceding photo. The illustrated surface demonstrates an abrupt transition from a grooved surface to a flat surface, with the transition occurring approximately at ground level. This distinct transition militates against an interpretation that Holocene erosion might be responsible for the well-defined grooves on the upper portion of the block. Given that the grooves were formed by subglacial ice extrusion, then the absence of the grooves below ground level implies that confinement by pressurized ice in the fissure prevented ice extrusion on the below-ground side surface. Furthermore, the grooves on the above-ground side surface imply that reverse ice segregation led to pore water continuing to enter the block after frost heave had taken place.
Example 4
Two prominent frost-heaved monoliths are shown in the above photo. Both blocks show evidence of erosion by ice extrusion on their top surfaces, but only the block just to the right of the hammer shows a grooved erosion pattern on a vertical side surface. Both blocks are composed of homogeneous fine-grained Ediacaran sedimentary rock that had been regionally metamorphosed. There was no discernible difference in the composition of the two blocks. The sharply different erosion patterns on the vertical surfaces facing the camera imply that erosion by ice extrusion occurred on the side surface of the background block, but not on the side surface of the foreground block. The background block was examined in detail as a potential example of the occurrence of reverse ice segregation.
A view of the top corner and adjacent vertical side of the block seen beside the hammer in the preceding photo is shown above, viewed from the opposite direction. The general layout for this block resembles the layout seen in Example 3, with a fissure alongside the end of the block and a vertical surface adjacent to the fissure that transitions from a grooved upper portion to a relatively flat lower portion. An overhead view of the fissure is shown below.
The measured depth of the fissure was about 1 m, with the frost-heaved block extending about 70 cm above ground level at its tallest point. The fissure appears to have formed after failure along a cross joint, in response to ice pressure driving rock rightward (direction as seen in the photo). The fissure was about 10 cm wide at the top (left edge to right edge in photo) and did not widen or narrow significantly with depth.
The above two closeup photos, taken at opposite ends of the block and at a similar scale (about 15 cm wide), show differences in the near-ground erosion patterns seen at the two opposing ends of the block. Grooves are deeper, wider and generally more pronounced on the end of the block opposite to the fissure. The grooves seen in the first (left) photo above become less pronounced with increasing depth below ground level, although the entire block from top to bottom appears to have undergone a discontinuous partial delamination parallel to the plane of foliation.
The tape shown in the above photo marks approximate ground level at the end of the block adjacent to the fissure. The block is about 25 cm thick at ground level. Note how the surface facing the camera changes from a predominantly flat surface to a grooved surface several centimeters above ground level.
The top corner of the block on the end opposite to the fissure is shown above. A grooved erosion pattern extending along the top of the block, and then 60 cm down the side of the block, is visible in the photo. Unlike the surface shown on the end of the block adjacent to the fissure, the vertical face shown above does not demonstrate a transition to a flat surface. It is quite probable that a transition to a flat surface occurs at ground level and that a flat surface extends below ground level, but observations to support this conjecture were not available.
The surface shown above can reasonably be interpreted as a grooved erosion surface formed by subglacial ice extrusion. The portion of the eroded surface now lying at ground level was at least 70 cm deep in bedrock before development of an ice lens under the block initiated the frost heave process that raised the block to its present height. The erosion by ice extrusion that occurred on the vertical surface of the block at ground level provides evidence suggesting reverse ice segregation. Ice pressure needed to drive the reverse ice segregation is further evidenced by the formation of the deep fissure seen on the opposing side of the block. The indicated ice pressure could have originated in groundwater travelling as pore water through bedrock beneath the block, or it could have been the result of the deformational flow of cold glacial ice acting against the opposing side of the block, or it could have been the result of a combination of both processes. In all cases, reverse ice segregation becomes a necessary step before erosion by ice extrusion could arise on the surface seen above.
Example 5
The following example differs from the above four examples in that no attempt is made to either support the hypothesis that reverse ice segregation was occurring nor to exclude that possibility. Rather, the example illustrates why reverse ice segregation is a process that could be important for understanding the geomorphology of cold-ice affected areas of the Avalon Peninsula.
The above photo shows a glacially eroded hillside that, on cursory inspection, might be mistaken for an instance of plucking by a warm-based glacier. However, the top of the hill is marked by numerous instances of frost-heaved bedrock. Some frost-heaved bedrock monoliths (not visible in photo) have been tilted toward the camera, apparently by glacial ice loading. Since plucking by a warm-based glacier and vertically directed bedrock frost heave are generally not mutually compatible, an interval of cold-based glacial activity is indicated at this site.
The horizontally-layered sections of rock seen toward the left side of the photo do not reflect sedimentary bedding. Rather the layered appearance is presumably the result of bedrock erosion by subglacial ice extrusion. The apparent layering follows the distinct foliation in the phyllosilicate-rich rock induced by ancient regional metamorphism. The bedrock is a fine-grained well-indurated Ediacaran sedimentary rock lacking discernible bedding planes. The foliation in the bedrock comprising the hill dips uniformly at a vertical angle across the entire site. The plane of foliation strikes roughly left-right with respect to the camera's perspective.
The sections of rock showing horizontal layering paralleling foliation are sections of rock that have been rotated 90 degrees by glaciotectonic action. These sections of bedrock were rotated about a horizontal axis with the original top surfaces of the blocks moving toward the camera. The grooves visible on the rotated blocks, attributed to erosion by subglacial ice extrusion, were presumably formed before the blocks were separated from bedrock substrate and rotated.
The magenta-colored circle marked "1" on the above photo highlights an occurrence of frost-heaved bedrock. A closeup view of this occurrence is shown below.
The horizontally-layered sections of rock seen toward the left side of the photo do not reflect sedimentary bedding. Rather the layered appearance is presumably the result of bedrock erosion by subglacial ice extrusion. The apparent layering follows the distinct foliation in the phyllosilicate-rich rock induced by ancient regional metamorphism. The bedrock is a fine-grained well-indurated Ediacaran sedimentary rock lacking discernible bedding planes. The foliation in the bedrock comprising the hill dips uniformly at a vertical angle across the entire site. The plane of foliation strikes roughly left-right with respect to the camera's perspective.
The sections of rock showing horizontal layering paralleling foliation are sections of rock that have been rotated 90 degrees by glaciotectonic action. These sections of bedrock were rotated about a horizontal axis with the original top surfaces of the blocks moving toward the camera. The grooves visible on the rotated blocks, attributed to erosion by subglacial ice extrusion, were presumably formed before the blocks were separated from bedrock substrate and rotated.
The magenta-colored circle marked "1" on the above photo highlights an occurrence of frost-heaved bedrock. A closeup view of this occurrence is shown below.
Large frost-heaved bedrock monoliths, strongly delaminated and grooved by subglacial ice extrusion are present on the top of the hill and provide evidence of upward-directed ice stress and ice motion. The block to the right of the group shown above appears to demonstrate two stages of frost heave, with thin central slabs of rock apparently shifted upward from an already-raised base. If the central slabs comprise a second episode of frost heave, following the initial uplift of the main block, then reverse ice segregation presumably occurred at the base of the main block. This interpretation is supported by considering the reduction in confinement of the main block that would follow from the main block being shifted upward. It is also reasonable to interpret the raised central slabs as the result of an initial stage of frost heave, preceding the uplift of the main block. This interpretation does not require the occurrence of reverse ice segregation.
The feature shown above is marked with the magenta-colored circle numbered "2" on the photo at the beginning of Example 5. The feature comprises ice-disrupted bedrock that has undergone substantial rotation, with a small element (estimated 0 to 2 m) of horizontal displacement. It is possible that the feature is mainly or entirely the result of cold glaciotectonic action, but this explanation fails to account for the high degree of cavitation seen in the arrangement of the blocks. The illustrated site is located about 4 km inland and any ice path to the coast is obstructed by topographic features with vertical relief of the order of 100 m or more. Several similar ice-disrupted hillsides in the local area imply ice movement in a similar direction toward the same coast. It must therefore be assumed that hundreds of meters of cold glacial ice overlay the site when the above-pictured feature was formed. Cavitation in such a deep subglacial environment would entail several MPa of negative ice pressure generated relative to ambient hydrostatic pressure.
It is reasonable to assume that cavitation did not occur when the feature shown above was formed. Rather, ice entered all available space between shifted pieces of detached bedrock. Deformational flow of cold glacial ice could account for some ice entry, but narrow, irregular channels extending 2 to 3 m in depth militate against ice creep being solely responsible. It appears more likely that the feature shown above is, at least in part, a subglacial bedrock frost-heave feature. This description would link the occurrence of ice-disrupted bedrock extending outward from the base of the hill with frost-heave monoliths seen at the top of the hill.
If the feature shown above is a bedrock frost-heave feature, promoted in its development by several MPa of negative relative ice pressure, then infilling ice originated either from groundwater emanating from within the hill or from reverse ice segregation creating pore water from ambient glacial ice. The large volume of ice needed to fill voids in a feature such as is seen above suggests that reverse ice segregation played a significant role in providing the ice. This interpretation would make the above-pictured feature an example of plucking by a cold-based glacier. There are numerous features similar to that shown above scattered over the frost-heave affected areas of the Avalon Peninsula.
It is reasonable to assume that cavitation did not occur when the feature shown above was formed. Rather, ice entered all available space between shifted pieces of detached bedrock. Deformational flow of cold glacial ice could account for some ice entry, but narrow, irregular channels extending 2 to 3 m in depth militate against ice creep being solely responsible. It appears more likely that the feature shown above is, at least in part, a subglacial bedrock frost-heave feature. This description would link the occurrence of ice-disrupted bedrock extending outward from the base of the hill with frost-heave monoliths seen at the top of the hill.
If the feature shown above is a bedrock frost-heave feature, promoted in its development by several MPa of negative relative ice pressure, then infilling ice originated either from groundwater emanating from within the hill or from reverse ice segregation creating pore water from ambient glacial ice. The large volume of ice needed to fill voids in a feature such as is seen above suggests that reverse ice segregation played a significant role in providing the ice. This interpretation would make the above-pictured feature an example of plucking by a cold-based glacier. There are numerous features similar to that shown above scattered over the frost-heave affected areas of the Avalon Peninsula.
Related topic: Crystal growth pressure
heartofavalonia.org Exploring Geologic History