Geology Reference
In-Depth Information
motion of loose soil towards the crack relative to ice-cemented soil at depth. It is also clear
that recurrent cracking and continued addition of wind-transported sand into the crack
causes long-term deformation of the surrounding permafrost. This must take the form of
a net aggradation, or “infl ation,” of the ground surface within the polygon. The inferred
displacement fi elds are indicated by arrows in Figure 6.5. According to Sletten et al. (2003,
p. 15-4), net ground surface aggradation could average 0.05 to 0.1 mm/year.
Additional studies are needed in order to substantiate this inferred model of Antarctic
landscape recycling. This may be diffi cult because few areas have experienced the same
long period of uninterrupted cold-climate conditions as the Dry Valleys.
6.2.5. Controls over Cracking
Undoubtedly the most comprehensive investigations into the nature of thermal-
contraction cracking have been undertaken at a number of sites in the Western Canadian
Arctic by J. R. Mackay (1974a, 1975a, 1978a, 1984c, 1986d, 1992a, 1993a, 1993b, 2000;
Mackay and Burn, 2002). Breaking cables and electronic crack detectors have established
the timing, frequency, and direction of cracking. In addition, the speed and sound of
cracking has been investigated, as have the relationships between cracking, snow cover,
air and ground temperatures, and creep of frozen ground.
It appears that cracking is not related simply to a rapid drop in air temperatures in
early winter, as fi rst suggested by Lachenbruch (1962) and others. The best correlation
between air temperature and cracking occurs in localities of thin snow cover. The favored
duration and rate of temperature drop that results in cracking is about 4 days, at a rate of
about 1.8 °C/day (Mackay, 1993b). It follows that tensions which cause cracking originate
not at depth but either at the top of permafrost or in the frozen active layer. Proof of this
can be seen by the repeated cracking that can be observed beneath any shallow tundra
pond (Figure 6.6). Because water does not possess a “memory” which would permit crack-
ing to occur in exactly the same location the following year, cracking must commence in
frozen ground and then propagate both upwards and downwards.
Observations also indicate that less than half the fi ssures in any given area crack annu-
ally (Mackay, 1975a, 1989b; Harry et al., 1985). The frequency of cracking is also site-
specifi c; for example, observations indicate that while a fi ssure may crack nearly every
year at one locality, it may crack only once in every ten years or so at a site just several
meters away. Moreover, years of exceptionally heavy snowfall inhibit cracking; along the
western Arctic coast, an average snow depth of 60
cm is thought to be suffi cient.
In refl ecting upon his data, Mackay (1992a, p. 244) points out that theoretical analyses
of thermal-contraction cracking (e.g. Grechishchev, 1970; Lachenbruch, 1962) assume
uniform conditions. However, the reality is that crack formation is accompanied by changes
in micro-relief associated with polygon development, and by associated vegetation and
snow-cover changes. Cracking may be regarded as a random process, and Mackay (1992a)
refers to “chaos” theory, and the complexity which follows growth, to explain the discrep-
ancies between theoretical considerations and the fi eld situation.
Few comparable data sets exist. Most refl ect either one or two years of semi-quantitative
observations inferences. For example, crack frequencies at Barrow, Alaska, ranged from
37% to 64% (Black, 1974) and only 38% of ice wedges examined in coastal exposures along
the northern Yukon coast showed signs of cracking during the previous winter (Harry
et al., 1985).
Several recent studies (Allard and Kasper, 1999; Christiansen, 2005; Fortier and Allard,
2005) permit a slightly more precise determination of the temperature requirements for
+
