You are on page 1of 15

PERMAFROST AND PERIGLACIAL PROCESSES Permafrost Periglac. Process. 12: 299313 (2001) DOI: 10.1002/ppp.

393

Microgelivation versus Macrogelivation: Towards Bridging the Gap between Laboratory and Field Frost Weathering
Norikazu Matsuoka*
Institute of Geoscience, University of Tsukuba, Tsukuba, Ibaraki 305-8571, Japan

ABSTRACT The application of laboratory criteria for frost weathering to eld problems needs caution, because a number of discrepancies lie between the laboratory and eld conditions. This paper reviews thresholds for microgelivation of soft, intact rocks and macrogelivation of hard, jointed rocks, aiming at proposing better criteria in accordance with eld conditions. The temperature at which ice segregation induces microgelivation varies signicantly with lithology, ranging from about 1 C in high porosity rocks to below 4 C in low porosity rocks. Microgelivation can occur in initially unsaturated rocks when slow (seasonal) freezing drives prolonged water migration from surrounding rock or an external moisture source, while the occurrence requires a high degree of saturation (>80%) or a nearby moisture source when a rock undergoes rapid (diurnal) freezing. Rocks with a high internal surface area and low tensile strength favour microgelivation. These criteria are invalid for macrogelivation that tends to take place just below 0 C in water-lled joints. In addition, because the depth reached by cracking varies with the type of freeze-thaw action, the analysis of thermal regimes should be based on data at the depth of actual cracking. Future targets for macrogelivation studies include the formation of new cracks in hard, intact rocks, as indicated by in situ shattering of clasts or bedrock typically observed in optimal moisture environments. Copyright 2001 John Wiley & Sons, Ltd. RESUM E Lapplication de crit` eres de laboratoire dans des probl` emes de terrain exige des pr ecautions car les conditions de part et dautre sont diff erentes. Le pr esent article passe en revue les seuils intervenant dans la microg elivation de roches tendres, intactes et dans la macrog elivation de roches dures ` proposer de meilleurs crit` pr esentant des joints, de fa con a eres en accord avec les conditions de ` laquelle la glace de s terrain. La temp erature a egr egation provoque la microg elivation, varie de mani` ere signicative avec la lithologie et est comprise entre 1 C dans des roches tr` es poreuses et 4 C dans des roches faiblement poreuses. La microg elivation peut se produire dans des roches au d epart non satur ees quand un gel lent (saisonnier) induit une migration prolong ee de leau depuis la roche environnante ou une source dhumidit e ext erieure; tandis que la microg elivation se produit aussi quand une roche qui subit un gel rapide (diurne), pr esente un haut degr e de saturation (>80%) ou une lev source dhumidit e voisine. Une surface interne e ee et une coh esion faible des roches favorisent la microg elivation. Ces crit` eres ne sont pas valides pour la macrog elivation qui prend place juste
Correspondence to: Professor N. Matsuoka, Institute of Geoscience, University of Tsukuba, Tsukuba, Ibaraki 305-8571, Japan. E-mail: matsuoka@atm.geo.tsukuba.ac.jp

Copyright 2001 John Wiley & Sons, Ltd.

Received 6 February 2001 Revised 4 May 2001 Accepted 28 May 2001

300 N. Matsuoka

en dessous de 0 C dans des joints remplis deau. En outre, parce que la profondeur atteinte par la tre bas ssuration varie avec le type de gel et de d egel, lanalyse des r egimes thermiques devrait e ee ` la profondeur o` tudes sur des donn ees correspondant a u se produit la fracturation. Les objectifs des e futures de macrog elivation comprennent la formation de nouvelles ssures dan des roches dures intactes comme cela est obverv e dans la ssuration in situ de d ebris ou du bedrock se produisant dans des environnements ayant une humidit e optimale. Copyright 2001 John Wiley & Sons, Ltd.
KEY WORDS:

frost weathering; microgelivation; macrogelivation; ice segregation; volumetric expansion; laboratory criteria

INTRODUCTION Frost weathering plays an important role in rock breakdown and regolith formation in cold-humid regions (e.g. Lautridou, 1988; Ballantyne and Harris, 1994). Operating very slowly in the eld, frost weathering has often been approached by laboratory simulations (e.g. Tricart, 1956; Potts, 1970; Murton et al., 2000). The laboratory studies have documented some empirical relationships between the frost sensitivity of rocks and such variables as temperature, moisture and rock properties (e.g. Lautridou and Ozouf, 1982; Matsuoka, 1990a). These laboratory criteria have been used to interpret eld data (e.g. Thorn, 1979; Fahey and Lefebure, 1988; Matsuoka, 1990b). However, there seems to remain a signicant gap between laboratory and eld weathering. This is not only due to the difference in the size of rock, but also to an essential difference in the nature of rock breakdown (cf. Whalley, 1984). Figure 1 illustrates how laboratory conditions differ from eld conditions. Whereas some laboratory studies address the inuence of pre-existing cracks or aws on rock fragmentation (e.g. Davidson and Nye, 1985; Douglas et al., 1987; Nicholson and Nicholson, 2000), most studies use intact rock samples with medium (520%) to high (>20%) porosity. This is because, rstly, intact rocks allow

us to correlate rock properties with frost sensitivity and, secondly, high porosity rocks are generally vulnerable to weathering (e.g. Matsuoka, 1990a). In contrast, many eld studies concern weathering of low porosity (<5%) but jointed rocks, which compose most of the high mountains in the world. Accordingly, the nature of fragmentation differs between the two conditions. Most of the laboratory studies simulate microgelivation, which involves granular disintegration and small aking, the process leading to production of ne debris (m-to-cm scale). This contrasts with eld studies which mainly deal with macrogelivation, or opening (wedging) of preexisting macrofractures (joints) that tend to produce pebble-size or coarser materials (cm-to-m scale). The governing mechanism may also be different. Two processes, the 9% volumetric expansion upon phase change and ice segregation, possibly contribute to frost weathering (e.g. Hallet et al., 1991). A number of recent theories and laboratory experiments have addressed the prime role of ice segregation in intact rocks (e.g. Walder and Hallet, 1985; Akagawa and Fukuda, 1991; Murton et al., 2000). Field observations have also attributed the origin of layered fractures in bedrock or stones to ice segregation (e.g. Murton, 1996; Mackay, 1999). Nevertheless, volumetric expansion cannot be discounted in bursting or macrogelivation of low porosity rocks in the eld (e.g.

Figure 1

Contrasts between laboratory and eld frost weathering. Permafrost and Periglac. Process., 12: 299313 (2001)

Copyright 2001 John Wiley & Sons, Ltd.

Laboratory and Field Frost Weathering 301

Michaud et al., 1989; Matsuoka, 2001). Although such discrepancies suggest the limited applicability of laboratory criteria to eld problems, there is still confusion in the analysis of eld data. Aiming at bridging the gap between laboratory and eld studies, this paper reviews dynamic approaches to frost weathering by comparing the two contrasting c onditions: (1) microgelivation of soft, intact rocks and (2) macrogelivation of hard, jointed rocks. Particular attention is focused on the application of proper laboratory criteria to the analysis of eld data. The nal part of the paper explores a future target for frost weathering studies.

freezing of pore water, which is ensured by the zerocurtain period seen on the rock temperature records (e.g. Mellor, 1970; Matsuoka, 1990a; Pissart et al., 1993; see also Figure 2). Experimental data fullling this requirement can propose physical conditions for microgelivation. Parameters controlling microgelivation include, in broad terms, temperature, moisture and rock properties (e.g. Matsuoka, 1991). Thermal Controls Rock temperature is the most popular parameter measured in the eld with regard to cold climate weathering (e.g. Thorn, 1979; Coutard and Francou, 1989; Gardner, 1992; Matsuoka, 1994; Hall, 1997; Anderson, 1998). Four variables, freezing intensity, rate and duration, and number of freeze-thaw cycles, are evaluated from the eld data. Freezing intensity, represented by the minimum subzero temperature that the rock surface experiences during a freeze-thaw cycle, gives a basis for counting effective freezethaw cycles occurring in the bedrock, while freezing duration inuences the depth reached by freeze-thaw action. Freezing rate mainly controls the magnitude of expansion of rocks and its effect is variable in combination with moisture regimes (see below). The critical freezing intensity (the upper threshold temperature) for microgelivation has not yet unequivocally dened. A basic problem is that there is a confusion in the recognition of the freezing temperature. Supercooling to a few degrees Celsius below

MICROGELIVATION OF SOFT, INTACT ROCKS A large number of laboratory freeze-thaw tests have highlighted microgelivation of soft rocks. Visual breakage of rocks, however, does not necessarily manifest rock disintegration by freeze-thaw action, because other processes are also operative within the temperature range used in the freeze-thaw tests. These include hydration shattering of argillaceous rocks in which water is possibly unfreezable as low as 20 C (Dunn and Hudec, 1972) and thermal shock generated when a rock undergoes a rapid temperature change (Hall and Hall, 1991). Freeze-thaw action is indicated when expansion or crack generation in a rock follows

Figure 2 Temporal variations in rock temperature and strain when a saturated porous shale specimen (shale-g in Matsuoka, 1990a; porosity 34%) is cooled at a rate of 6 C h 1 under (A) sealed (closed system) and (B) submerged (open system) conditions. Supercooling is dened by the difference between the stable equilibrium freezing temperature TF and the spontaneous freezing temperature TS , while the freezing point depression is given by that between the datum freezing temperature (0 C) and TF . Note that supercooling is absent in the submerged specimen. Copyright 2001 John Wiley & Sons, Ltd. Permafrost and Periglac. Process., 12: 299313 (2001)

302 N. Matsuoka

0 C usually precedes ice nucleation in stationary pure water. Pore water in rocks also experiences supercooling when cooled to subzero temperatures under a closed system condition (Figure 2A). When the spontaneous freezing temperature TS is reached, supercooling terminates and pore water temperature suddenly rises to the stable equilibrium freezing temperature TF due to latent heat release by ice nucleation (e.g. Lock, 1990, pp. 4648). In frost weathering studies, the critical freezing intensity is sometimes dened by TS (e.g. Thorn, 1979; Douglas et al., 1983). However, when a rock specimen is cooled under an open system condition, freezing of pore water usually begins at TF without supercooling (Figure 2B), triggered by an external ice nucleus. Similarly, in most eld conditions, supercooling is likely to be disturbed by a variety of triggers including wind, snow crystals, impurities in water and ice already formed elsewhere on the rock surface. Even where supercooling does occur at the rock surface, it most likely diminishes within the top few centimetres of rock; below this depth, the ice front reaches TF without supercooling. These conditions suggest that TS tends to underestimate the freezing temperature of eld rocks. Accordingly, the following discussion does not consider supercooling. Expansion of a rock usually begins at a temperature slightly lower than TF (Figure 2), in response to increasing ice volume in pores. Thus, the minimum requirement for microgelivation is that the rock is cooled below TF . Table 1 summarizes temperature ranges for effective frost weathering, which are delimited by the upper and lower thresholds (TH , TL ). Since methodology is different between researchers, some care should be given to compare the listed values. The TH

values for acoustic emission (AE) generation (Hallet et al., 1991) and ice lens formation (Akagawa and Fukuda, 1991) delimit the highest temperature for actual cracking in a rock sample (Figure 3A). In contrast, length change of a sample (Mellor, 1970; Matsuoka, 1990a; Weiss, 1992; Pissart et al., 1993; Murton et al., 2000) includes pre-failure expansion unaccompanied by cracking (Figure 3B); hence, actual cracking must begin at a slightly lower temperature than the listed TH values. Moreover, laboratory data are not available for rocks with low porosity, which, unless containing pre-existing aws, rarely fail even after hundreds of freeze-thaw cycles under an optimum moisture condition in the laboratory (e.g. Lautridou and Ozouf, 1982; Matsuoka, 1990a). With these limitations in mind, Table 1 suggests that rocks having high porosity crack at a high subzero temperature. Cracking begins at 0 to 1 C and terminates at about 5 C in these rocks (Figure 3A). The cracking temperature falls between 3 C and 6 C in medium porosity rocks (Hallet et al., 1991) and probably below 4 C in low porosity rocks. Most of the experiments show that dilatation and/or cracking progresses most rapidly just below TH and decelerates toward TL . It should be noted that the temperature range for effective crack growth ( 4 C to 15 C) theoretically suggested by Walder and Hallet (1985) is applicable only to intact rocks with low porosity (<3%) and has to be validated by experiments. A couple of conditions seem to contribute to this apparent relationship between porosity and cracking temperature. First, the freezing point of pore water generally lowers with decreasing porosity, depending on the amount of smaller pores (e.g. Williams and Smith, 1989, p. 192). Second, ice segregation

Table 1 Temperature range for effective microgelivation, dened by the upper limit TH and lower limit TL . Lithology Low porosity rocks Granite and marble Medium porosity rocks Berea sandstone Berea sandstone Indiana limestone German sandstones High porosity rocks Tuff, shale and andesite Ohya tuff Br ez e chalk Br ez e chalk Porosity (%) 13 20 20 14 419 2346 38 47 47 TH C 4 3 0.2 0.5 0 0 1.4 0 0.2 TL C 15 6 8 8 5 5 5 11 2 Indicator Moisture supply Open Open Closed Closed Closed Open Open Closed Open Reference

Theoretical AE generation Length change Length change Length change Length change Ice lens formation Length change Length change

Walder and Hallet (1985) Hallet et al. (1991) Mellor (1970) Mellor (1970) Weiss (1992) Matsuoka (1990a) Akagawa and Fukuda (1991) Pissart et al. (1993) Murton et al. (2000)

Copyright 2001 John Wiley & Sons, Ltd.

Permafrost and Periglac. Process., 12: 299313 (2001)

Laboratory and Field Frost Weathering 303

Figure 3 Experimental temperature range for effective microgelivation of Ohya tuff. (A) Ice lens distribution after 1000 h of freezing with the top and bottom temperatures maintained at 14 C and 4.5 C, respectively. From Akagawa and Fukuda (1991). John Wiley & Sons Limited. Reproduced with permission. (B) Dilatation during a 12-h freeze-thaw cycle. From Matsuoka (1990a). Reproduced with permission from Elsevier Science.

is optimized by the balance between the suction force, which increases with lowering temperature, and the water permeability, which decreases with lowering temperature (e.g. Williams and Smith, 1989, pp. 224225). Furthermore, the strength of rock generally increases with decreasing porosity, so that larger suction force (hence, lower temperature) is required for low porosity rocks to crack. As regards the second condition, part of the pore water, supported by capillarity or adsorption, remains unfrozen after the initial freezing. Change in the

unfrozen water content with temperature is known for Ohya tuff and Berea sandstone (Figure 4), both often used for the frost weathering experiments. The comparison between Table 1 and Figure 4 allows us to estimate the pore ice/water content ratio at which these rocks crack. In saturated Ohya tuff, pre-failure expansion begins just after the initial freezing at 0.2 C (Matsuoka, 1990a). In this rock, ice segregation is most active at 1.4 C when 65% of the pore water is frozen and becomes rare below 5 C when ice occupies 80% of the total pore space (Akagawa and Fukuda, 1991). Perhaps the former temperature represents the optimum balance between suction force and permeability, while the latter relates to the minimum permeability for ice segregation. In saturated Berea sandstone, expansion begins with the initial freezing just below 0 C as well (Mellor, 1970), while cracking begins below 3 C (Hallet et al., 1991) when more than 90% of the pore water is frozen. The unfrozen water content for intensive cracking differs between the two rocks. The reason for this is unclear but may partly reect the difference in the tensile strength. In summary, the critical freezing intensity at which intact rocks crack varies signicantly with rock properties, hence an appropriate value should be used for each lithology when eld data are analysed. Moisture Controls

Figure 4 Unfrozen water content of rocks as a function of temperature. Compiled from Mellor (1970) and Akagawa and Fukuda (1991). Copyright 2001 John Wiley & Sons, Ltd.

Moisture regimes play a decisive role in promoting ice action in rocks, while rocks lacking moisture are sensitive to salt action or thermal stress rather
Permafrost and Periglac. Process., 12: 299313 (2001)

304 N. Matsuoka

than ice action even in cold regions (e.g. Matsuoka et al., 1996; Hall, 1997). Two moisture parameters, the degree of saturation before freezing and the amount of water migration during freezing, govern frost weathering (e.g. Matsuoka, 1990a). The initial degree of saturation constrains expansion of rock if water migration is minimum during freezing, the condition favoured by short-term, rapid freezing or the lack of an external water source (Matsuoka, 1990a). Such a situation occurs on rocks lying distantly from open water (e.g. stream, lake or sea) or far above the subsurface water table, as well as on those undergoing frequent diurnal freeze-thaw cycles. A number of experiments have evaluated the critical degree of saturation, SCR , above which frost weathering is active (Mellor, 1970; Fagerlund, 1979; Houiou, 1979; Matsuoka, 1990a; Weiss, 1992; Prick, 1997). In these experiments, rock samples with different water contents are sealed and submitted to repetitive freeze-thaw cycles; then SCR is determined from the nick point on the relationship between the degree of saturation and frost sensitivity of the samples. The SCR value varies from 58% to 93%, but mainly falls around 80% (Table 2). This critical value is slightly lower than that expected from the 9% volumetric expansion, suggesting that some moisture redistribution occurs within a sealed rock sample during freezing, or simply reecting uneven moisture distribution before freezing. An external moisture source close to a freezing rock contributes to ice segregation, signicantly expanding the rock. A long-term, slow freezing test under an open system condition, which simulates seasonal freezing, highlighted continuous water migration and resulting ice segregation in porous Ohya tuff (Akagawa and Fukuda, 1991). The open system condition would accelerate disintegration of porous rocks also in short-term freeze-thaw regimes; in fact, partially submerged samples were broken by alternative freezing and thawing much more rapidly than sealed samples despite the same initial water

content (e.g. Matsuoka, 1990a). Water migration from an external moisture source intensies ice segregation even in slowly thawing porous rocks (Akagawa, 1993; Murton et al., 2000). The latter situation implies that frost heave possibly takes place during early summer in the seasonally frozen rock and during late summer in the top of perennially frozen rock, where downward percolation of rain or snowmelt water is available (cf. Mackay, 1983). Figure 5 summarizes the inuence of the moisture availability on frost weathering of a rock undergoing downward freezing. If the water table is too deep, water migration from the underlying unfrozen part would be minimum so that an initial water content in excess of SCR is required for signicant expansion (model A). In contrast, where the water table is shallow, water migration from the saturated zone eventually raises the water content of the frozen part above SCR and subsequent prolonged ice segregation would induce expansion much larger than expected

Figure 5 The inuence of (A) deep and (B) shallow water levels on frost weathering of partially frozen rock with the initial degree of saturation, SO .

Table 2 Critical degree of saturation, SCR , for microgelivation of porous rocks and other materials. Lithology Berea sandstone Indiana limestone Concretes and brick Concretes Japanese tuffs and sandstone German sanstones French limestones ND, no data.
Copyright 2001 John Wiley & Sons, Ltd. Permafrost and Periglac. Process., 12: 299313 (2001)

Porosity (%) 20 14 ND 1521 2839 1319 947

SCR (%) 60 70 80 8188 75 6387 5893

Indicator Length change Length change Reduction in E-modulus Length change Reduction in P-wave velocity Reduction in E-modulus Reduction in E-modulus

Reference Mellor (1970) Mellor (1970) Fagerlund (1979) Houiou (1979) Matsuoka (1990a) Weiss (1992) Prick (1997)

Laboratory and Field Frost Weathering 305

from full saturation (model B). Volumetric expansion associated with rapid (diurnal) freezing favours model A, while ice segregation associated with slow (seasonal or even inter-annual) freezing plays a major role in model B. The transition from the model A to B depends on how high the capillarity or adsorption can draw water from the saturated zone to the frozen rock. Laboratory simulations show that the frozen part draws water from the water table at least shallower than 10 cm (e.g. Akagawa and Fukuda, 1991; Hallet et al., 1991); the critical height is unlikely to be greater than a few decimetres, reecting the low permeability of rocks. Rock Controls A number of rock properties would contribute to microgelivation of intact rocks. These properties are classied into (1) those affecting ice formation and the expansive force and (2) those constraining the resistance (Matsuoka, 1991). As already pointed out, pore geometry inuences the critical temperature while permeability controls water migration; both belong to the former properties. Other properties affecting the expansive force include porosity which, by delimiting the capacity of ice volume, governs volumetric expansion (Mellor, 1970) and specic surface area which contributes to suction (Matsuoka, 1990a). The resistance is simply given by the tensile strength of a bulk rock, on the assumption of homogeneity of rock (Matsuoka, 1990a). When the growth of microfractures in rocks is considered, the fracture toughness controls the resistance (Walder and Hallet, 1985). A laboratory freeze-thaw test conducted using 47 lithologies under the same thermal and moisture regimes showed that frost sensitivity is roughly proportional to the ratio of specic surface area to tensile strength, demonstrating the major role of ice segregation in microgelivation of intact rocks (Figure 6). Other parameters associated with the resistance include the mineral composition, which mainly controls the granulometry of the debris produced by microgelivation (e.g. Etlicher and Lautridou, 1999). Properties of a bulk sample do not always indicate the frost sensitivity of heterogeneous rocks containing microscopic fractures, aws, bedding and schistosity (e.g. Hall, 1987; Douglas et al., 1991). For instance, the specic surface area of a bulk sample is unlikely to be indicative of suction generated in the fracture system. Nicholson and Nicholson (2000) illustrated the growth of visible fractures. Quantifying the geometry of such pre-existing fractures and their rates of growth with freeze-thaw cycles will allow us to evaluate more realistic frost sensitivity. The rate
Copyright 2001 John Wiley & Sons, Ltd. Figure 6 Experimental rate of frost weathering, represented by the decay of P-wave velocity per freeze-thaw cycle, as a function of the ratio of the specic surface area per volume (SV ) to the tensile strength (ST ) of rocks. Data from Matsuoka (1990a).

of crack growth can be represented by decreasing P-wave velocity or increasing crack porosity (e.g. Matsuoka, 1990a; Remy et al., 1994). Application of Laboratory Criteria to Field Problems There have been numerous visual observations on granular disintegration or small aking of rocks exposed to cold climates (e.g. Tricart, 1956; Washburn, 1969; Robinson and Jerwood, 1987; see also Figure 7). Since other physical weathering processes may also produce similar features (Hall, 1995), any of these features cannot be attributed to frost weathering

Figure 7 Granular disintegration on schist boulders near an alpine lake, Wallis, Swiss Alps. Permafrost and Periglac. Process., 12: 299313 (2001)

306 N. Matsuoka

without direct evidence documenting disintegration by ice action. Direct evidence can be obtained in the eld by simultaneous monitoring of rock disintegration and associated parameters. Unfortunately, such an attempt is so far rare for soft, intact rocks, because of technological immaturity and the need for longterm operations. In this context, laboratory criteria will continue to be useful for interpretation of eld observations. Field problems which can be approached by the extant laboratory criteria or simulated by laboratory models involve small-scale fragmentation of soft rocks on a rock outcrop (e.g. Fahey and Lefebure, 1988; Pancza and Ozouf, 1988) or that in the active layer and upper permafrost (Williams, 1987; Murton, 1996), as well as granulometry of periglacial sediments (e.g. Lautridou, 1988). Field researches on microgelivation should examine at least three factors: water accessibility (e.g. distance from open water or depth of the water table), dominant freezethaw regime (diurnal, annual or inter-annual) and rock properties (e.g. porosity; strength). In addition, the use of laboratory criteria should incorporate the following limitations. (1) Most of the criteria are applicable only to porous, soft rocks (porosity more than 5%). (2) Thresholds (e.g. critical temperature) may vary with rock properties. (3) Laboratory criteria account mainly for fragmentation up to the centimetre scale. (4) Even though a proper criterion is chosen for the eld rock, it is applicable only to the surcial fragmentation; in fact, ice lens formation at a certain depth requires cooling below the critical temperature experienced not at the rock surface but at this depth. As regards the nal limitation, the analysis of microgelivation at depth should be based on heat conduction which governs the depth and time reached by frost penetration (cf. Matsuoka, 1994; Anderson, 1998; Wegmann et al., 1998) and on moisture proles in bedrock (cf. Trenhaile and Mercan, 1984). MACROGELIVATION OF HARD, JOINTED ROCKS Laboratory Simulation The laboratory criteria derived from intact rocks are generally invalid for weathering of jointed bedrock in the eld. For instance, whereas hard, intact rocks tend to crack below 4 C (Table 1), ice action may widen joints in the same lithology just below 0 C (see below). Similarly, the degree of saturation of a rock sample taken from the inter-joint area may underestimate the humidity in joints, since
Copyright 2001 John Wiley & Sons, Ltd.

moisture would concentrate in joints with the intact rock containing a relatively low amount (McGreevy and Whalley, 1985). Experimental evaluation of physical conditions for macrogelivation thus requires rock samples with pre-existing fractures. However, the complexity of naturally-induced cracks has prevented dynamic approaches to this process. Instead, simplied, articial crack models have been used to simulate frost wedging under the closed system. Davidson and Nye (1985) measured pressure produced by volumetric expansion when ice propagates into a 1-mm-wide crack made in a lucite block. During one-sided freezing, pressure perpendicular to the crack increased near and below the ice front, reaching about 1 MPa (Figure 8A); this ice-water contact pressure is equivalent to the freezing point at 0.07 C in terms of the Clausius-Clapeyron equation (e.g. Williams and Smith, 1989, p. 189). Davidson and Nye also predicted that the maximum pressure rises further with decreasing aspect ratio (crack widthto-depth ratio: Tharp, 1987) to a value sufcient to extend cracks in natural rocks. In fact, although the tensile strength of hard, intact rocks is of the order of 10 MPa, the strength decreases signicantly in jointed bedrock of the same lithology (e.g. Matsuoka, 1990b; cf. Selby, 1989). Figure 8B illustrates a laboratory test of frost wedging in a 2-mm-wide crack made in an impermeable granite block (see Matsuoka, 1995, for methodology). Crack widening occurred most rapidly when the ice temperature fell from 0 C to 1 C. The maximum expansion was about 0.1%, much lower than predicted from the phase change (9%), suggesting that the rigid crack wall forced ice to extrude both upwards and sidewards. These simple experiments suggest that, although pressure is relieved considerably from the value expected from volumetric expansion, frost wedging may occur near the ice front in water-lled cracks at 0 C or a slightly lower temperature. Experimental data have not yet been acquired on the critical moisture condition for macrogelivation, but a joint must be lled at least partially with water to minimize ice extrusion to open spaces and to feed expansive force against the joint wall (cf. Davidson and Nye, 1985; Tharp, 1987). There is also a lack of experimental data on the rock properties controlling macrogelivation. Crack geometry such as the aspect ratio and crack-tip morphology (Tharp, 1987) rather than physical properties of the inter-joint rock would be a more relevant control on the frost sensitivity of joints.
Permafrost and Periglac. Process., 12: 299313 (2001)

Laboratory and Field Frost Weathering 307

Figure 8 Laboratory models for macrogelivation. (A) Pressure distribution after 2 h of downward freezing of water in a crack. From Davidson and Nye (1985). Reproduced with permission from Elsevier Science. (B) Dilatation of a crack due to volumetric expansion, as a function of ice/water temperature at two depths.

Future experiments should also target ice segregation through the joint system in bedrock. The operation of this process is equivocal, because while the high permeability in open joints favours water migration, such wide joints are unlikely to produce large suction. A relevant aspect of this process is that widening of joints can occur in rocks with low water content before freezing (cf. Figure 5B), where slow seasonal freezing promotes water migration and eventually lls the joints with ice (e.g. Hallet et al., 1991). The laboratory simulation will require a large climatic cabinet in which a jointed rock mass is
Copyright 2001 John Wiley & Sons, Ltd.

submitted to slow freezing under the open system. Alternatively, centrifuge techniques would permit us to reduce the size of a rock sample (Davis et al., 2001). Field Monitoring In the light of the scale of process, macrogelivation merits eld monitoring rather than laboratory simulations which are restrained by the capacity of climatic cabinets. Field monitoring is essential also to predict future rates of rock weathering. However,
Permafrost and Periglac. Process., 12: 299313 (2001)

308 N. Matsuoka

Figure 9 Rock joint widening on a rockwall (2950 m ASL), Hosozawa cirque, Japanese Alps, showing a number of short-term and seasonal widening events which also resulted in inter-annual net widening. Apparent expansion due to thermal change was removed from the original extensometer data. Temperature data were lacking from 7 April to 21 May, 1998, and from 16 June to 20 August, 1999. Data for the former period are substituted by those at a nearby monitoring site (the shaded line in B). Expansion for the latter period (the shaded line in A) was uncorrected with reference to temperature.

it was only in recent years that eld technology has enabled direct observations of macrogelivation. The measured parameters involve widening of joints on a rock face and internal stress/strain in bedrock. Surface and internal rock temperatures are simultaneously recorded. Widening of surcial rock joints was monitored with crack extensometers in order to evaluate the role of diurnal and/or annual freeze-thaw action in alpine rockwall instability (Matsuoka et al., 1997; Matsuoka, 2001). Signicant widening occurred in joints subject to abundant rain or snowmelt water. For instance, four years of monitoring on the shale bedrock (inter-joint porosity 3%) in the Japanese Alps indicated frequent widening events in autumn, synchronized with short-term (diurnal or multi-day) temperature oscillations across 0 C (Figure 9). The widening amount was nearly proportional to the frost penetration depth (Matsuoka, 2001). These shortterm events were followed by a larger widening event associated with seasonal freezing. A large event also occurred upon seasonal thawing, when meltwater from the thawing snowcover is likely to have entered the joint and refrozen. Although
Copyright 2001 John Wiley & Sons, Ltd.

a large part of individual widening was recovered upon subsequent thawing, the repetition of the widening events resulted in permanent widening at a mean rate of 0.07 mm yr 1 (Figure 9). These results demonstrate that macrogelivation is active in the jointed bedrock situated in an optimal moisture regime and submitted at least to shallow freezing, thus highlighting volumetric expansion as a contributing process. In contrast, the widening activities are rarely associated with the temperature range suggested for ice segregation in hard, intact rocks (< 4 C). Wegmann and his co-workers have monitored expansion in frozen bedrock at Jungfraujoch, the Swiss Alps, in an assessment of the inuence of construction activities on the thermal regime of permafrost (Wegmann, 1998; Wegmann and Gudmundsson, 1999). A number of thermal probes and extensometers were installed in the gneiss bedrock to a depth of about 10 m below the rock surface. The permafrost temperature rose to the melting point (c. 0.2 C) during summer and fell again to a few degrees below the melting point in winter (Figure 10A). Correspondingly, annual cycles of summer contraction and winter expansion
Permafrost and Periglac. Process., 12: 299313 (2001)

Laboratory and Field Frost Weathering 309


(A)

(B)

Figure 10 (A) Rock temperatures and (B) deformation at different depths in bedrock permafrost, monitored on a southeast-facing rockwall (3600 m ASL) near Jungfraujoch, Swiss Alps. From Wegmann and Gudmundsson (1999). Reproduced with permission from Springer-Verlag. Annual cycles of winter expansion and summer contraction and resulting inter-annual net expansion are indicated.

occurred presumably in response to partial melting of frozen rock and refreezing (Figure 10B). The summer expansion was probably fed by migration of meltwater from the active layer to permafrost. Wegmann attributed the expansion events to opening of several joints across the borehole. Permanent expansion accumulated in permafrost with these cycles (Figure 10B). These results indicate that temperature change of only a few degrees can activate macrogelivation in the bedrock at a few metres depth. In summary, eld monitoring indicates that rock joints lled with water can expand when the rock is cooled just below the freezing point. Expansion occurs near the rock surface (shallower than 50 cm) mainly by diurnal and multi-day freeze-thaw action and at the top of permafrost by annual freeze-thaw action. Field monitoring has also to be designed
Copyright 2001 John Wiley & Sons, Ltd.

to explore macrogelivation in the lower part of the seasonally frozen bedrock (at a few metres deep), which has rarely been investigated.

A FUTURE TARGET: FROST WEATHERING OF HARD, INTACT ROCKS The above review highlights the constraints on frost weathering for the two contrasting conditions. Between the two extremes a number of problems remain unsolved or virtually untouched. Of these, among the most important but least understood is the way ice action produces a new crack in hard, intact rocks. In addition to internal stress originating from tectonic activity and unloading, subaerial weathering would also generate pre-existing cracks used in
Permafrost and Periglac. Process., 12: 299313 (2001)

310 N. Matsuoka

macrogelivation (e.g. Whalley et al., 1982). The crack generation by ice action, if any, may occur at a condition distinct from that governs macrogelivation in pre-existing cracks in the same rock. The previous freeze-thaw experiments suggest that hard, low porosity rocks break largely by the growth of pre-existing fractures, while the intact samples of the same lithology rarely show disintegration as indicated by the reduction in the P-wave velocity or elastic modulus even under an optimum moisture (saturated and partially submerged) condition (e.g. Matsuoka, 1990a). This result, however, disagrees with visual eld evidence for intensive, in situ shattering of hard rocks like most of the metamorphic, crystalline and Palaeozoic sedimentary rocks. Such a feature is mainly reported from bedrock or stones along lakes, streams, late-lying snowpatches and rocky coasts in cold regions (e.g. Taber, 1950; Matthews et al., 1986; Dionne and Brodeur, 1988; Berrisford, 1991; McEwen and Matthews, 1998; see also Figure 11). Intensively shattered pebbles are also found in delta or other near-shore deposits along Norwegian coasts and used as a palaeoenvironmental indicator (Blikra and Longva, 1995). The mutual factors for these environments are (1) the proximity to the water table and (2) the occurrence of deep seasonal freezing (>1 m) whether or not permafrost is present. As regards the former factor, intensive shattering features are conned to an altitudinal zone in association with the water level (Matthews et al., 1986; Blikra and Longva, 1995). One of the possible explanations for the intensive shattering is that long-term ice segregation during seasonal freezing may induce supersaturation, eventually producing expansion in excess of the tensile failure strain for the intact rocks. The low thermal conductivity of open water would help provide long-lasting supply of unfrozen water towards the frozen rock (Shakesby

and Matthews, 1987). In coastal areas, the presence of salts may enhance ice segregation (e.g. Williams and Robinson, 1991). Intensively shattered rocks may also originate from sudden rock bursting due to the 9% expansion or thermal shock (e.g. Michaud et al., 1989; Hall, 1997; Mackay, 1999), although this process accounts only for surcial shattering. The discrepancy between the laboratory and eld evidence originates partly from the experimental methodology simulating short-term freeze-thaw action. The verication of ice segregation and resulting crack growth in hard, intact rocks would require a laboratory simulation of annual freeze-thaw action. For this purpose, a rock specimen should be submitted to slow freezing (or thawing) under the open system, such that Akagawa and Fukuda (1991) and Akagawa (1993) conducted with porous tuff. Measurements of expansion or AE generation in the specimen may allow us to evaluate the cracking temperature for these rocks and to compare the result with the theoretical model by Walder and Hallet (1985). Field monitoring for this process is worth conducting but technically more difcult. The simplest methodology is to investigate fragmentation of marked bedrock or boulders near the water level, concurrent with monitoring of rock temperature (and also moisture, if possible). However, measurable fragmentation of hard, intact rocks probably needs long-term monitoring over decades or more, although Mackay (1999) reported rapid shattering of boulders within a few years in the Canadian Arctic. The most recommended methodology such as Wegmann (1998) was conducted at Jungfraujoch, to install extensometers and thermal probes in newly exposed bedrock lacking visible fractures and lying close to open water. This methodology permits detection of subsurface cracking in the bedrock. A technical problem is that such a measurement requires a large electric auger allowing boring 1 m or deeper in the bedrock, which is expensive and difcult to transport to most of the periglacial mountains.

CONCLUSIONS A review of recent experimental studies on frost weathering suggests that the application of the laboratory results is mostly limited to eld problems on microgelivation of soft, intact rocks. The following criteria are recommended for the analysis of microgelivation. With respect to thermal parameters, the critical freezing intensity for effective cracking varies from about 1 C in high porosity (>20%) rocks to below 4 C in low porosity (<5%) rocks.
Permafrost and Periglac. Process., 12: 299313 (2001)

Figure 11 Intensive shattering of gneiss boulders along an alpine stream, Upper Engadin, Swiss Alps. A rucksack (60 cm long) at the centre for scale. Copyright 2001 John Wiley & Sons, Ltd.

Laboratory and Field Frost Weathering 311

The duration of freezing affects the depth to which microgelivation occurs. Whereas the repetition of freeze-thaw cycles is responsible for shallow but intensive fragmentation that produces ne debris, prolonged slow freezing near the permafrost table leads to continuous ice segregation that results in layered structures. The presence of a nearby moisture source plays a fundamental role in ice segregation. Where the moisture source is distant, water migration is limited so that effective microgelivation requires an initial degree of saturation in excess of 80% and rapid freezing. In contrast, rocks proximal to a moisture source (e.g. along a lake shore or river bank) can soak up water during slow freezing and expand signicantly by ice segregation; thus, a high initial water content is unnecessary. If environmental factors are similar, microgelivation prevails in rocks with a large internal surface area and/or low tensile strength. The evaluation of frost sensitivity in terms of rock properties, however, needs caution because hydration rather than ice action may govern shattering of some argillaceous rocks involving very small pores in which water is unfreezable even at 20 C. The laboratory criteria derived from intact rocks are rarely applicable to macrogelivation of hard, jointed rocks. Both laboratory and eld studies indicate that joint widening is most active just below 0 C, highlighting at least the effect of the 9% volumetric expansion. The critical moisture condition for macrogelivation is yet to be dened, but a joint must be lled at least partially with water to minimize ice extrusion to open spaces and to feed expansive force against the joint wall. The water content of the inter-joint rock may not indicate the humidity in joints, because permeable joints are more vulnerable to change in humidity than the impermeable interjoint area. Physical properties in the inter-joint rock may also fail to indicate the frost sensitivity of the jointed bedrock. Crack geometry would be one of the more relevant parameters. Attention is also given to the analysis of rock temperature data. Rock surface temperatures are only indicative of surcial fragmentation. Because of the timelag originating from heat conduction, cracking activity in the interior of bedrock may occur when the rock surface is cooled well below the threshold temperature. This warrants the evaluation of the depth reached by cracking based on temperature data at different depths in bedrock and/or on heat conduction models, when the size and volume of frost shattered debris are discussed. Future targets for frost weathering studies include the physical condition in which ice action generates a new crack in hard, intact rocks. The condition may
Copyright 2001 John Wiley & Sons, Ltd.

differ from that governing widening of pre-existing macrofractures. A long-term, slow freezing test in the laboratory may help understanding of this process.

ACKNOWLEDGEMENTS This paper was initially presented at a Periglacial Workshop held on 67 September, 2000 at St Andrews, Scotland. I have beneted from discussion with the participants of the Workshop and would particularly like to thank Julian Murton and Colin Ballantyne for useful comments.

REFERENCES
Akagawa S. 1993. Initiation of segregation freezing observed in porous soft rock during melting process. In Proceedings, 6th International Conference on Permafrost, Beijing, China, Vol. 2. South China University of Technology Press: Wushan; 10501053. Akagawa S, Fukuda M. 1991. Frost heave mechanism in welded tuff. Permafrost and Periglacial Processes 2(4): 301309. Anderson RS. 1998. Near-surface thermal proles in alpine bedrock: implications for the frost weathering of rock. Arctic and Alpine Research 30(4): 362372. Ballantyne CK, Harris C. 1994. The Periglaciation of Great Britain. Cambridge University Press: Cambridge. Berrisford MS. 1991. Evidence for enhanced mechanical weathering associated with seasonally late-lying and perennial snow patches, Jotunheimen, Norway. Permafrost and Periglacial Processes 2(4): 331340. Blikra LH, Longva O. 1995. Frost-shattered debris facies of Younger Dryas age in the coastal sedimentary successions in western Norway: palaeoenvironmental implications. Paleogeography, Paleoclimatology, Paleoecology 118(12): 89110. Coutard JP, Francou B. 1989. Rock temperature measurements in two alpine environments: implications for frost shattering. Arctic and Alpine Research 21(4):399416. Davidson GP, Nye JF. 1985. A photoelastic study of ice pressure in rock cracks. Cold Regions Science and Technology 11(2): 143153. Davis MCR, Hamza O, Harris C. 2001. The effect of rise in mean annual temperature on the stability of rock slopes containing ice-lled discontinuities. Permafrost and Periglacial Processes 12(1): 137144. Dionne JC, Brodeur D. 1988. Frost weathering and ice action in shore platform development with particular reference to Qu ebec, Canada. Zeitschrift f ur Geomorphologie, NF, Supplementband 71: 117130. Douglas GR, McGreevy JP, Whalley WB. 1983. Rock weathering by frost shattering processes. In Proceedings, 4th International Conference on Permafrost. National Academy Press: Washington; 244248.
Permafrost and Periglac. Process., 12: 299313 (2001)

312 N. Matsuoka Douglas GR, McGreevy JP, Whalley WB. 1987. The use of strain gauges for experimental investigations on frost weathering. In International Geomorphology 1986 Part II , Gardiner V (ed). Wiley: Chichester; 605621. Douglas GR, Whalley WB, McGreevy JP. 1991. Rock properties as controls on free-face debris fall activity. Permafrost and Periglacial Processes 2(4): 311319. Dunn JR, Hudec PP. 1972. Frost and sorption effects in argillaceous rocks. Highway Research Record 393: 6578. Etlicher B, Lautridou JP. 1999. G elifraction exp erimentale dar` enes de roches cristallines: bilan dessais de longue dur ee. Permafrost and Periglacial Processes 10(1): 116. Fagerlund G. 1979. Studies of the destruction mechanism at freezing of porous materials. In Problems Raised by Frost Action. Fundamental and Applied Researches (Rocks and Articial Building Materials), AguirrePuente J (ed). Fondation Fran caise dEtudes Nordiques: Paris; 167196. Fahey BD, Lefebure TH. 1988. The freeze-thaw weathering regime at a section of the Niagara Escarpment on the Bruce Peninsula, Southern Ontario, Canada. Earth Surface Processes and Landforms 13(4): 293304. Gardner JS. 1992. The zonation of freeze-thaw temperatures at a glacier headwall, Dome Glacier, Canadian Rockies. In Periglacial Geomorphology, Dixon JC, Abrahams AD (eds). Wiley: Chichester; 89102. Hall KJ. 1987. The physical properties of quartz-micaschist and their application to freeze-thaw weathering studies in the maritime Antarctic. Earth Surface Processes and Landforms 12(2):137149. Hall KJ. 1995. Freeze-thaw weathering: the cold region Panecea? Polar Geography and Geology 19: 7987. Hall KJ. 1997. Rock temperature and implications for cold region weathering, I: new data from Viking Valley, Alexander Island, Antarctica. Permafrost and Periglacial Processes 8(1): 6990. Hall KJ, Hall A. 1991. Thermal gradients and rock weathering at low temperatures: some simulation data. Permafrost and Periglacial Processes 2(2): 103112. Hallet B, Walder JS, Stubbs CW. 1991. Weathering by segregation ice growth in microcracks at sustained sub-zero temperatures: verication from an experimental study using acoustic emissions. Permafrost and Periglacial Processes 2(4): 283300. Houiou A. 1979. M ethodes dessai pour la g elivit e des b etons. In Problems Raised by Frost Action. Fundamental and Applied Researches (Rocks and Articial Building Materials), Aguirre-Puente J (ed). Fondation Fran caise dEtudes Nordiques: Paris; 213224. Lautridou JP. 1988. Recent advances in cryogenic weathering. In Advances in Periglacial Geomorphology, Clark MJ (ed). Wiley: Chichester; 3347. Lautridou JP, Ozouf JC. 1982. Experimental frost shattering: 15 years of research at the Centre de G eomorphologie du CNRS. Progress in Physical Geography 6(2): 215232. Loch GSH. 1990. The Growth and Decay of Ice. Cambridge University Press: Cambridge.
Copyright 2001 John Wiley & Sons, Ltd.

Mackay JR. 1983. Downward water movement into frozen ground, western Arctic coast, Canada. Canadian Journal of Earth Science 20(1): 120134. Mackay JR. 1999. Cold-climate shattering (1974 to 1993) of 200 glacial erratics on the exposed bottom of a recently drained Arctic Lake, Western Arctic Coast, Canada. Permafrost and Periglacial Processes 10(2): 125136. Matsuoka N. 1990a. Mechanisms of rock breakdown by frost action: an experimental approach. Cold Regions Science and Technology 17(3): 253270. Matsuoka N. 1990b. The rate of bedrock weathering by frost action: eld measurements and a predictive model. Earth Surface Processes and Landforms 15(1): 7390. Matsuoka N. 1991. A model of the rate of frost shattering: application to eld data from Japan, Svalbard and Antarctica. Permafrost and Periglacial Processes 2(4): 271281. Matsuoka N. 1994. Diurnal freeze-thaw depth in rockwalls: eld measurements and theoretical considerations. Earth Surface Processes and Landforms 19(5): 423435. Matsuoka N. 1995. A laboratory simulation on freezing expansion of a fractured rock: preliminary data. Annual Report of the Institute of Geoscience, the University of Tsukuba 21: 58. Matsuoka N. 2001. Direct observation of frost wedging in alpine bedrock. Earth Surface Processes and Landforms 26(6): 601614. Matsuoka N, Moriwaki K, Hirakawa K. 1996. Field experiments on physical weathering and wind erosion in an Antarctic cold desert. Earth Surface Processes and Landforms 21(8): 687699. Matsuoka N, Hirakawa K, Watanabe T, Moriwaki K. 1997. Monitoring of periglacial slope processes in the Swiss Alps: the rst two years of frost shattering, heave and creep. Permafrost and Periglacial Processes 8(2): 155177. Matthews JA, Dawson AG, Shakesby RA. 1986. Lake shoreline development, frost weathering and rock platform erosion in an alpine periglacial environment, Jotunheimen, southern Norway. Boreas 15(1): 3350. McEwen LJ, Matthews JA. 1998. Channel form, bed material and sediment sources of the Sprongdla, southern Norway: evidence for a distinct periglacio-uvial system. Geograska Annaler 80A(1): 1736. McGreevy JP, Whalley WB. 1985. Rock moisture content and frost weathering under natural and experimental conditions: a comparative discussion. Arctic and Alpine Research 17(3): 337346. Mellor M. 1970. Phase Composition of Pore Water in Cold Rocks. US Army Cold Regions Research and Engineering Laboratory, Research Report 292. Michaud Y, Dionne JC, Dyke LD. 1989. Frost bursting: a violent expression of frost action in rock. Canadian Journal of Earth Sciences 26(10): 20752080. Murton JB. 1996. Near-surface brecciation of chalk, Isle of Thanet, south-east England: a comparison with icerich brecciated bedrocks in Canada and Spitsbergen. Permafrost and Periglacial Processes 7(2): 153164.
Permafrost and Periglac. Process., 12: 299313 (2001)

Laboratory and Field Frost Weathering 313 Murton JB, Coutard JP, Lautridou JP, Ozouf JC, Robinson DA, Williams RBG, Guillemet G, Simmons P. 2000. Experimental design for a pilot study on bedrock weathering near the permafrost table. Earth Surface Processes and Landforms 25(12): 12811294. Nicholson DT, Nicholson FH. 2000. Physical deterioration of sedimentary rocks subjected to experimental freeze-thaw weathering. Earth Surface Processes and Landforms 25(12): 12951307. Pancza A, Ozouf JC. 1988. Contemporary frost action on different oriented rock walls: an example from the Swiss Jura Mountains. In Proceedings, 5th International Conference on Permafrost, Trondheim, Norway, Vol. 2, Senneset K (ed). Tapir: Trondheim; 830833. Pissart A, Prick A, Ozouf JC. 1993. Dilatometry of porous limestone undergoing freezing and thawing. In Proceedings, 6th International Conference on Permafrost, Beijing, China, Vol. 1, South China University of Technology Press: Wushan; 523528. Potts AS. 1970. Frost action in rocks: some experimental data. Transactions of the Institute of British Geographers 49: 109124. Prick A. 1997. Critical degree of saturation as a threshold moisture level in frost weathering of limestones. Permafrost and Periglacial Processes 8(1): 9199. Remy JM, Bellanger M, Homand-Etienne F. 1994. Laboratory velocities and attenuation of P-waves in limestone during freeze-thaw cycles. Geophysics, 59(2): 245251. Robinson DA, Jerwood LC. 1987. Frost and salt weathering of chalk shore platforms near Brighton, Sussex, U.K. Transactions of the Institute of British Geographers NS 12(2): 217226. Selby MJ. 1989. Rock slopes. In Slope stability, Anderson MG, Richards KS (eds). Wiley: Chichester; 475504. Shakesby RA, Matthews JA. 1987. Frost weathering and rock platform erosion on periglacial lake shorelines: a test of hypothesis. Norsk Geologisk Tidsskrift 67(3): 197203. Taber S. 1950. Intensive frost action along lake shores. American Journal of Science 248(11): 784793. Tharp TM. 1987. Conditions for crack propagation by frost wedging. Geological Society of America Bulletin 99(1): 91102. Thorn CE. 1979. Bedrock freeze-thaw weathering regime in an alpine environment, Colorado Front Range. Earth Surface Processes 4(3): 211228. Trenhaile AS, Mercan DW. 1984. Frost weathering and the saturation of coastal rocks. Earth Surface Processes and Landforms 9(4): 321331. xperimentale du probl` Tricart J. 1956. Etude e eme de la g elivation. Biuletyn Peryglacjalny 4(4): 285318. Walder J, Hallet B. 1985. A theoretical model of the fracture of rock during freezing. Geological Society of America Bulletin 96(3): 336346. Washburn AL. 1969. Weathering, frost action, and patterned ground in the Mesters Vig district, Northeast Greenland. Meddelelser om Grnland 176(4): 1303. Wegmann M. 1998. Frostdynamik in Hochalpinen Felsw anden am Beispiel der Region JungfraujochAletsch. Mitteilungen der Versuchsanstalt f ur Wasserbau, Hydrologie und Glaziologie, Eidgen ossische Technische Hochshule Z urich 161. Wegmann M, Gudmundsson GH. 1999. Thermally induced temporal strain variations in rock walls observed at subzero temperatures. In Proceedings, 6th International Symposium on Thermal Engineering and Sciences for Cold Regions, Lecture Notes in Physics 533, Hutter K, Wang Y, Beer H (eds). Springer-Verlag: Heidelberg; 511518. Wegmann M, Gudmundsson GH, Haeberli W. 1998. Permafrost changes in rock walls and the retreat of Alpine glaciers: a thermal modelling approach. Permafrost and Periglacial Processes 9(1): 2333. Weiss G. 1992. Die Eis- und Salzkristallisation im Porenraum von Sandsteinen und ihre Auswirkungen auf das Gef uge unter besonderer Ber ucksichtigung gesteinsspezischer Parameter. M unchner Geowissenshaftiche Abhandlungen 9B: 1118. Whalley WB. 1984. Rockfalls. In Slope Instability, Brunsden D, Prior DB (eds). Wiley: Chichester; 217256. Whalley WB, Douglas GR, McGreevy JP. 1982. Crack propagation and associated weathering in igneous rocks. Zeitschrift f ur Geomorphologie NF 26(1): 3354. Williams PJ, Smith MW. 1989. The Frozen Earth: Fundamentals of Geocryology. Cambridge University Press: Cambridge. Williams RBG. 1987. Frost weathered mantles on the Chalk. In Periglacial Processes and Landforms in Britain and Ireland , Boardman J (ed). Cambridge University Press: Cambridge; 127133. Williams RBG, Robinson DA. 1991. Frost weathering of rocks in the presence of salts: a review. Permafrost and Periglacial Processes 2(4): 347353.

Copyright 2001 John Wiley & Sons, Ltd.

Permafrost and Periglac. Process., 12: 299313 (2001)

You might also like