Friday, 12 August 2016

Pingos, Palsas, Frost Mounds and Hummocks as thermal ratchets

The ratchet process believed to be involved in the development of both open and closed pingos has been described in rather greater detail elsewhere (Croll 2004a, 2004b; Croll et alia 2004). For this reason the following will be brief and attempt to show how the mechanics described for the development of pingos could also help to explain how the related features of hummocks, frost mounds and palsas could derive from closely related processes.

Pingos
In earlier drafts of reports outlining a thermal ratchet model for the development of pingos the first author was clearly incorrect in his dismissal of the important role played by excess pore water pressure created by the aggrading lower permafrost boundary. It is now recognised that the process leading to the build-up of pore water pressure in the unfrozen talik underlying the permafrost has two crucial effects. First, and especially where as is common the pingos develop in recently forming and relative to its surroundings thin permafrost, the aggrading lower permafrost boundary might induce differential frost heave. This differential frost heave will be likely to contain significant components of the characteristic geometric shape of the deformation that will eventually emerge in the form of the pingo. For the development of the upheaval thermal buckle, believed to be partially responsible for the initiation and growth of the pingo, this geometric imperfection plays a crucial role. Second, the excess pressure while not necessarily being sufficient to produce the distorted pingo geometric shape, will have the effect of reducing the effective weight of the permafrost sheet. With the temperatures required to induce an incremental growth of the upward distortion of a thermally upheaving permafrost sheet being directly related to this effective weight, the excess pore water pressure plays a second crucial part in the process of thermal uplift buckling of pingos. However, it remains the authors’ view that pressure alone would not be capable of providing the energy required to cause the gradual growth of pingos. This view conflicts with the widely accepted models outlined in the literature dealing with the growth of pingos (see for example Porsild 1938; Mackay 1973a, 1998; Washburn 1979; French 1996). So what is the thermal uplift buckling process and in what sense might it be involved in the thermal ratchet uplift of pingos?

Figure 8(b) depicts a typical situation from which pingos are likely to emerge. A recently drained lake might have exposed unfrozen underlying soil. Initially with considerable rapidity the depth of perennially frozen ground will increase over the first few years. This will be associated with general heave relative to the more stable surrounding and possibly deeper permafrost. During this initial period of thickening it is clear that the seasonal temperature cycles will penetrate through the entire, relatively thin, permafrost sheet; clear, because if it did not then the lower boundary would not aggrade downward. Hence, on average the permafrost layer at the site of the potential pingo will be undergoing substantial seasonal temperature variations. Clearly these thermal cycles will be greatest at the surface. However, at least over the first few decades the average temperatures through the permafrost thickness are likely to be a significant fraction of those experienced at the surface, with the fraction decreasing as aggrading of the lower permafrost boundary stalls. During the seasonal warming period, up to point (3) in Figure 8(a), the permafrost layer will want to expand horizontally outwards. Being prevented from doing so by the restraint from the surrounding, possibly older and deeper, permafrost it will develop significant compressive stresses. It has been shown elsewhere (Croll 2004a,b; Croll et alia 2004) that these stresses could produce an incremental growth of the upward deformation, similar to that depicted for time (3) in Figure 8(b), through a process of upheaval sheet buckling (Hobbs 1989, 1990, 2004; Croll 2005a). The underlying void created by the upwardly buckled permafrost sheet will be likely to have pressurised pore water extruded into it, with a proportion becoming attached to the underside of the permafrost in the form of an ice lens. The associated lowering of the pore water pressure and increase of thickness of the permafrost layer will tend to inhibit the extent of the uplift experienced in any one year. That in the following year a further incremental growth of upward deformation can occur will be the result of the again increased pore water pressure and, crucially, the now increased magnitude of the uplift, which in structural mechanics language would be an increased initial geometric imperfection (Croll 1997, 1998). Annual growth will only become stalled when the effects of incremental increase of imperfection become cancelled out by the increased thickness of the permafrost and ice lens. While the former has the effect of reducing the temperatures required for an increment of upheaval buckling to occur, the latter will increase the temperatures so required. Eventually, the two effects will compensate each other leading to a period of reduced annual growth.
                        
 Why the upward deformation of the pingo occurring during the warming period, up to point (3) Figure 8(a), is not recovered during the following cooling period is again believed to be the result of two important and interacting actions. First, the growth of the underlying ice lens will prevent the permafrost sheet from fully settling back into its previous position. This means that a proportion of the most recent upward deformation will have become lock-in, presenting an increased level of imperfection for the following seasonal, compression, growing cycle. Second, the differential properties of ice in compression and tension will ensure that during the following cooling season much of the thermal energy will almost immediately go into producing tensile strain in the permafrost sheet. During the compressive buckling a major part of the internal strain energy will have been expended in the lifting of the permafrost sheet. This means that when the temperatures start to tumble in early autumn the permafrost will almost immediately want to develop membrane tensile stress associated with the now restrained contraction of the permafrost ice. Being weak in tension the ice will develop fissures and cracks whose aggregate widths will be in proportion to the drop in temperature. As in the ratchet described for ice-wedges, and that for side-shift of stones in creating stone circles, polygons, etc, these cracks will attract moisture and other precipitation which will be converted to ice. This has the effect of re-establishing the continuity of the permafrost sheet so that at the commencement of the next warming cycle the restraint to the outward expansion will once again start to develop the high compressive stresses needed to again propagate the upward growth of the pingo.
Crucial evidence which would help to establish whether this thermo-mechanical ratchet process might be providing a major contribution to the growth of pingos, as opposed to the build-up of underlying pore water pressure currently considered to provide the energy source for growth, is the timing during the annual seasonal cycle at which upward growth occurs. Allowing for the time lag for the propagation of the thermal wave into the permafrost, but taking account of the consideration that the greatest intensity of thermal energy will be in the top few metres, the thermo-mechanical ratchet process would predict the greatest growth to occur between spring and late summer. Little if any growth would be expected during the cooling autumn to late winter period; indeed, the thermal contractions over this period could result in some recovery of the previous incremental growth. In contrast, the process requiring the build-up of pore water pressure to initiate further growth would be anticipated to produce the greatest growth rate when in late winter the cold front starts to aggrade the lower permafrost boundary. Unfortunately, the writer has been unable to find any published data that would resolve this crucial aspect. Unpublished data collected over the period 1997 to 2004 (private correspondence with Burn, 2005), does however, strongly suggest that maximum growths occur over the spring to late summer period, consistent with the predictions of the present thermo-mechanical ratchet model. 

Palsas
Palsas appear to be similar to pingos with the main distinction being in the scale, the role of the active layer, and the extent of organic material contained in the soil - typically they form out of peat bogs. While pingos are not always of the symmetric dome form, this shape does tend to dominate especially for closed pingos. Palsas on the other hand are more commonly found to have mounds in the form of long, often straight ridges or humps. In the context of thermal buckling of heavy sheets the symmetric dome form requires essentially isotropic stress to develop whereas the ridge type of deformation would be more typical of situations where the thermal stress in one direction exceeds that in the orthogonal direction. The one dimensional, ridge-type, of thermal buckle requires thermal stresses to be roughly 10% higher than those required when an isotropic stress produces the two-dimensional, dome-type, of buckle. The direction of the ridge where this occurs is likely to be at right angle to the direction of the principal compressive stress. In ice sheets, and therefore possibly in bogs where palsas commonly form, the principal compression during warming will tend to be greatest normal to any restraining edge. In these circumstances the buckling ridges might be anticipated to be roughly parallel with any such relatively rigid restraint such as a rising hill or rock face. In similar situations pingos appear to also change their preferred geometric forms. A further distinction between pingos and palsas seems to be the extent of permafrost. Palsas are reported to form in areas of at best discontinuous permafrost and are often formed within seasonally frozen ground overlying a deeply buried upper permafrost boundary. But there appears to be a continuum between seasonal frost mounds and perennial pingos. Palsas seem to be sandwiched between these two extremes of what appear to be similar processes involved in their formation. There seems to be good reasons for supposing that the thermo-mechanical ratchet described above for the formation of pingos could also be playing an active role in the growth of both palsas and frost mounds.

Frost Mounds
As described by Mackay (Mackay 1998) in relation to the various growths that occurred at the sides of a number of pingos, frost mounds can come and go during anything from a single season through to almost perennial growths. However, they appear to be associated with rather shorter cyclic thermal timescales than pingos and palsas, and seem to be confined much more to the active frost layers. In some of the cases reported (Mackay 1998) they formed within an active layer that overlies permafrost. Most of the observed features would seem to fit the following thermo-mechanical ratchet process. 

     Clearly they cannot form until a frost layer has been developed in late autumn. Current hypotheses for their development are similar to those for pingo growth, namely that a build-up of pore water pressure beneath the frozen and impermeable frost layer reaches a level where it forces the upward growth. Why this does not lead to a general heave of the downwardly aggrading frozen layer is not made clear but some models rely upon a form of differential heave to account for the localised shapes that are characteristic of frost mounds. An alternative model would envisage a short term warming of the frost layer, caused by a short term increase in surface temperatures. Being relatively thin this short term, perhaps a few weeks, thermal wave could easily result in average temperature increases in the frost layer that would be sufficient to induce an upheaval buckle. The high pore water pressures would play the vital role of injecting water into the cavity created by the upwardly deformed frost sheet, which might then be turned to ice when temperatures again drop. When the temperature drops the frost layer could be expected to develop cracks which will eventually be filled with ice. The frost mound is then locked into its deformed shape and of course becomes vulnerable to any further fluctuations in temperature during the period in which the frost layer remains unthawed. Whether the frost mound will disappear during the thaw would presumably depend upon the depth to which the lower frost boundary had reached when the frost mound was initiated. Upward freezing from an adjacent, underlying, permafrost could conceivably lead to circumstances whereby the frost mound is substantially preserved during the summer thaw. Alternatively, the shape could be largely preserved by liquified soil being intruded into the cavities beneath the frost mounds, similar to that experienced by hummocks. Either way the previous shape would provide a natural location for additional growth during the following seasonal cycle. It seems likely that a related thermo-mechanical ratchet could in some cases be driving the formation of hummocks.

Hummocks
Are essentially features of the active layer with the frost affected soil having high organic content. Excellent field data reproduced at Fig 5.31, (Washburn 1979), suggests that at least some forms of hummock may have strong causal relationships with a form of thermo-mechanical ratchet similar to that outlined above for pingos, palsas and frost mounds. When in autumn the saturated ground starts to freeze contraction cracks will likely be preferentially located in the troughs where the soil is somewhat more saturated. Water filling these cracks as well as the detritus or intruded, perhaps liquified, loamy materials filling the cavities beneath the humps shown in for example Figs 5.30 and 5.31, (Washburn 1979) will likewise lock-in the shape against contraction subsidence during the cooling period. Warming prior to thaw will develop compression in the frozen soil that at the early stages could be sufficient to induce incremental uplift. Just as for the pingo an annual growth increment could be expected until the total effective thickness, including any growing ice-lens in the cavity underlying the hummock, reaches the point where the compressions are no longer sufficient to cause growth. The gradual increase in total length of the upwardly deformed hummock would seem to be accommodated by a gradual infilling of the trough area by eroded surface material. In some cases this is strongly supported by the lengths of the stratified sections in adjacent hummocks, (see for example the apparently delaminated beds of volcanic ash and dust in Fig 5.31, (Washburn 1979) that closely approximate to what they would have been when horizontally deposited. On any section through the hummocks the growth in total length seems to be accommodated through the build-up of new trough material, probably from surface erosion of organic material from adjacent hummocks. This mechanism is illustrated in Figure 9.

While it is not suggested that all hummocks have their origin in the form of thermo-mechanical ratchet described, it would appear more than possible that at least some are likely to be the result of this process.                   

 

No comments:

Post a Comment