Réf. Matsuoka & al. 1998 - P

Référence bibliographique complète

MATSUOKA, N., HIRAKAWA, K., WATANABE, T., HAEBERLI, W., KELLER, F. 1998. The role of diurnal, annual and millennial freeze-thaw cycles in controlling alpine slope instability. Seventh International Conference on Permafrost, Yellowknife (Canada), Collection Nordicana No 55, pp. 711–717. [Etude en ligne]

Abstract: The instability of rock and debris slopes in the Swiss Alps was evaluated in light of the temporal and spatial scales of freeze-thaw processes. Diurnal freezing and thawing penetrate to centimeter-to-decimeter scale depths, producing rock debris mainly of pebble size or smaller on rock slopes and miniature patterned forms on debris slopes. Annual freeze-thaw cycles result in weathering and soil movement up to meter scale, supplying boulders to rock glaciers and developing solifluction lobes with risers of 30 cm or higher. The growth and decay of permafrost, originating from long-term climatic change, lead to freeze-thaw activity reaching meter-to-decameter scale depths. Permafrost melting can trigger cliff falls and debris flows in the thawing phase of millennial freeze-thaw cycles.




Organismes / Contact

• Institute of Geoscience, University of Tsukuba, Ibaraki 305-8571, Japan, e-mail: matsuoka@atm.geo.tsukuba.ac.jp
• Graduate School of Environmental Earth Science, Hokkaido University, Sapporo 060, Japan
• Department of Geography, University of Zurich, Zurich 8057, Switzerland
• Academia Engiadina, Samedan 7503, Switzerland


(1) - Paramètre(s) atmosphérique(s) modifié(s)

(2) - Elément(s) du milieu impacté(s)

(3) - Type(s) d'aléa impacté(s)

(3) - Sous-type(s) d'aléa


Permafrost Rockfalls, Debris flows



Pays / Zone

Massif / Secteur

Site(s) d'étude



Période(s) d'observation

Eastern Switzerland

Upper Engadin






(1) - Modifications des paramètres atmosphériques










Informations complémentaires (données utilisées, méthode, scénarios, etc.)



(2) - Effets du changement climatique sur le milieu naturel










Sensibilité du milieu à des paramètres climatiques

Informations complémentaires (données utilisées, méthode, scénarios, etc.)




(3) - Effets du changement climatique sur l'aléa





Freeze-thaw action induces both rock weathering and mass wasting, destabilizing rock and debris slopes in high mountain regions. Two types of freeze-thaw cycles, diurnal and annual, are normally recognized according to the period for the completion of one cycle. In addition, recent global warming has highlighted a third type, which has a much longer period. Corresponding to the growth and decay of permafrost, this type of freeze-thaw is completed typically in many centuries or millennia (e.g., Haeberli, 1996) and here is termed the millennial freeze-thaw cycle. The relationship between the freeze-thaw types and the magnitude and nature of resulting geomorphic processes, however, has been poorly understood because of the lack of long-term, continuous monitoring of processes and variables.

The periglacial belt in a mountain area is usually subdivided into permafrost and seasonal frost zones, mainly in relation to elevation and aspect. Between the two zones, a transient permafrost zone can be defined in which permafrost has grown and decayed repeatedly in response to climatic change during the Holocene. The transient permafrost zone is, therefore, characterized by the occurrence of millennial freeze-thaw cycles, as well as of diurnal and annual freeze-thaw cycles. Millennial freeze-thaw cycles can also operate in the permafrost zone as a result of melting and refreezing of the top and bottom of the permafrost body, although their effects would be less dramatic than in the transient permafrost zone. During the Little Ice Age, a large part of the transient permafrost zone was probably characterised by a freezing phase of a millennial cycle. The 20th Century warming will have switched this zone into a thawing phase.

The prediction of future geomorphic changes due to global warming requires the distinction of effects due to millennial cycles from those due to shorter cycles. The distinction is necessary, in particular, in the permafrost and transient permafrost zones where permafrost melting is in progress and the three freeze-thaw types are superimposed, causing slope instability. (…)

Diurnal freeze-thaw cycles


The magnitude and frequency of diurnal freeze-thaw cycles depend partly on the aspect of slopes. This tendency is enhanced on steep rockwalls. Figure 2 displays the contrast of rock surface temperatures between north and south-facing rockwalls (TFN and TFS sites). Both are located at 2850 m ASL. Covered with thick snow from winter to spring, the north-facing rockwall experiences continuous subzero temperatures. Even during summer months, the minimal insolation leads to small ranges of diurnal fluctuation in the rock surface temperature. As a result, diurnal freeze-thaw cycles take place only in early autumn. In contrast, because of the lack of snowcover, the south-facing rockwall undergoes large diurnal fluctuation in the surface temperature throughout the year. This thermal condition favours the high frequency of diurnal freeze-thaw cycles during all seasons except midsummer.

Temperature fluctuations across 0°C, however, do not always indicate freeze-thaw alternations effective in rock breakage. An abundant moisture supply is necessary for frost damage (e.g., Matsuoka, 1991; Prick, 1997). Subzero temperatures, following the infiltration of water into the joints and pores in the bedrock, may cause effective freezing expansion. Consequently, the effective diurnal freeze-thaw cycles must be considerably fewer than those counted from the fluctuation in rock temperature.

Frost (or thaw) penetration in the bedrock is usually 30 cm or shallower during a diurnal freeze-thaw cycle (Figure 2). In response to the amount of the latent heat exchange, however, wet rocks favourable for freezing expansion are subjected to much shallower freeze-thaw. Furthermore, frost damage can occur at depths cooled to a few degrees below 0°C (e.g., Matsuoka, 1994). Thus diurnal frost weathering is considered to be active with in 10 to 20 cm of the rock surface and able to produce rock debris up to cobble size. Controlled by joint spacing, the size of the released rock debris can be smaller. In fact, pebbles are the major components of screes below the south-facing rockwalls. Observations of scaling from the painted bedrock also showed that a number of rock fragments smaller than 5 cm were produced every year.


Large parts of the debris slopes in the study area are covered with snow for half of the year. Diurnal freeze-thaw cycles are most frequent in early autumn and are prevented by the late-lying snowcover in spring (Figure 3). Windy crest slopes lack snowcover and experience frequent freeze-thaw cycles in both autumn and spring (Matsuoka et al., 1997).

Debris slopes experience shallower freeze-thaw depths than rock slopes, because of the lower thermal conductivity and larger latent heat. Diurnal frost depth is typically about 5 cm and rarely in excess of 15 cm (Figure 3). Frost heaving usually accompanies diurnal freeze-thaw cycles. The heave amount depends upon the soil granulometry, but rarely exceeds 2 cm. Despite such small individual heaves, the cumulative amount is considerable. Thin debris mantles and insignificant snowcover combine to make diurnal frost heaving prevail on crest slopes where small sorted stripes and lobes dominate. These landforms are considered to originate mainly from diurnal freeze-thaw cycles. In fact, the sorting depth of the stripes is about 5 cm and the risers of the lobes are about 10 cm high, values similar to the depth of soil movement induced by diurnal frost heave activity.

Annual freeze-thaw cycles


Regardless of the aspect and the presence of permafrost, rock slopes in the periglacial belt are subjected to deep seasonal freezing and thawing. Direct determination of annual frost or thaw penetration is difficult. Equations derived from the thermal conduction theory, however, permit us to estimate the depth using the freezing or thawing index at the rock surface (Matsuoka, 1994). The modified Berggren equation (Aldrich, 1956), one of the Stefan-type equations, was used for the calculation of the frost (or thaw) penetration depth in the rockwalls. The thermal conductivity, a parameter involved in this equation, was determined from temperature curves at different depths. The calculation includes assumptions of the vertical gradient of the mean annual rock temperature being negligible and the freezing point at 0.0°C. Such a simplification does not seem to lower the accuracy of calculation significantly (Matsuoka, 1994). This model was applied to TFN and TFS sites (Figure 2). The mean annual surface temperature was negative at both rockwalls, indicating the presence of permafrost. The maximum thaw depth in 1995 was computed to be 4.3 m at TFN site and 6.8 m at TFS site. These values are comparable with the computed frost depth in a seasonal frost environment of the Japanese Alps (Matsuoka, 1994). Thus, the annual freeze-thaw depth in the alpine periglacial belt is typically 5±2 m, rarely reaching the decameter scale.

The annual freeze-thaw depth defines the boundary to which frost weathering can operate annually. The above calculation predicts that a rock mass up to about 5 m thick is detachable from the rockwalls. The locations at which frost damage happens, however, depend on several factors including the joint patterns, moisture distribution, and temperature range at which ice segregation occurs. The concurrent monitoring of rock temperatures and joint opening on the rockwall behind the Murtèl rock glacier, the Upper Engadin, showed that the largest opening at the rock surface occurred during an early thawing period when meltwater infiltrated the joint and refroze (Matsuoka et al., 1997). Since only minor opening was recorded during seasonal freezing in winter, moisture supply is considered to play a major role in near-surface frost weathering. In permafrost areas, segregational freezing may lead to intensive frost damage at the base of active layer where moisture availability is high (Hallet et al., 1991), although this idea has yet to be verified by field evidence.

Frost damage associated with seasonal freezing is often followed by rockfalls on thawing. For instance, in early June 1997, a block of rock about 100 m3 was detached from the rockwall behind the Murtèl rock glacier. The slip plane lay at about 2 m deep. This block fall happened during a high temperature period after snowmelt (Matsuoka, 1997). These conditions indicate that rapid seasonal thawing triggered the block detachment. The block was broken into numerous boulders and smaller debris, which were deposited on a scree and a rock glacier. Despite low frequency, such a block fall would be an important source in terms of the debris supply on screes and rock glaciers.


Seasonal freezing penetrates to about 2 m deep in debris slopes located just below the lower limit of permafrost (Figure 3). Thaw penetration over the rock glacier permafrost is slightly deeper, reaching 3 to 5 m (e.g., Barsch, 1996), probably because of the small latent heat exchange and large cold air drainage through the open-work clasts. Where a large part of the freeze-thaw layer consists of fine debris, seasonal freezing is associated with a large frost heave (5 cm or more). Formation of ice lenses tends to be concentrated in the upper part of the annual freeze-thaw layer, because the progressive downward freezing may cause desiccation of the lower part.

Thawing of the heaved ground, often aided by snowmelt, raises the moisture content and mobility of the thawed soil, resulting in solifluction or small debris flows. In response to the locations of ice lenses, soil movements due to annual freeze-thaw cycles would occur mainly in the upper part of seasonal frost. In fact, many solifluction lobes in the study area have a riser 30 to 50 cm high. Solifluction lobes with similar riser heights seem to reflect the movement of soil mass some decimeters thick (e.g., Smith, 1987). Since the thickness of the mobile layer estimated from the riser height far exceeds the diurnal freeze-thaw depths, these lobes are considered to have developed as a result of repeated annual freeze-thaw cycles. In the permafrost zone, upward freezing from the top of permafrost can produce ice-rich layers in the lower part of the active layer (e.g., Mackay, 1981). This process could lead to deeper soil movements near the active layer-permafrost inter face, although no data have yet been obtained in the Alps.

Millennial freeze-thaw cycles

Annual freezing and thawing are unlikely to reach depths in excess of 5 m on debris slopes and 10 m on rock slopes, depths to which only permafrost can penetrate. Segregational freezing tends to produce ice rich-layers in the uppermost part of permafrost (e.g., Cheng, 1983) and possibly near the base of permafrost. Large-scale mass movements sometimes take place in the transient permafrost zone. Some of these movements following abnormally warm summers are possibly associated with thawing of the top of permafrost. Thawing may also occur within or at the base of the permafrost body, causing much deeper changes.


There are a number of recent records of large-scale cliff falls in the Alps. The starting zones of these falls were mainly located near the lower limit of permafrost. For instance, in October 1988, a rock mass fell from the north-facing rockwall of Piz Morteratsch, the Upper Engadin, the fragments being deposited on a glacier (Haeberli et al., 1992). The detached rock mass had a volume of about 3•10 5 m3 and thickness in excess of the depth reached by annual freeze-thaw cycles. Coinciding with the period of the maximum seasonal thawing, the cliff fall might have been triggered by the partial melting of permafrost and/or the penetration of meltwater.

Other recent cliff falls (or rockslides) possibly associated with permafrost melting in the Alps include the Val Pola landslide in 1987 (Dramis et al., 1995), Randa rock slide in 1991 (Schindler et al., 1993) and Zuetribistock rockslide in 1996. The volume of the Val Pola landslide is two orders of magnitude larger than the cliff fall at Piz Morteratsch. Intense precipitation is considered to have triggered this slide. However, the presence of ice-cemented rock blocks among the landslide debris indicates that permafrost melting possibly enhanced the mobility of the rock mass prior to the heavy rain (Dramis et al., 1995).


[In the Trais Fluors region] Figure 4 displays complicated landforms developed near the lower limit of permafrost. The BTS values indicate that permafrost underlies the rock glacier (probably inactive), while permafrost is rare in the east-facing debris slope. Lying near the borderline, however, the latter slope can be subject to permafrost growth with minimum cooling. The climatic fluctuation during the Holocene would have allowed this debris slope to experience millennial freeze-thaw cycles. The major processes modifying the debris slope are debris flows and solifluction which resulted respectively in alluvial cones and numerous lobes. Upslope of a large alluvial cone, lies a landslide scar, 150 m long, 100 m wide and 5 m deep. The total volume of the debris flow deposit in the alluvial cone is on the order of 104 m3. The size of the landslide is likely to exceed that originating from the annual freeze-thaw action. Millennial freeze-thaw cycles may have intensified weathering of the porous calcareous rocks and destabilized the debris layer.

Zimmerman and Haeberli (1992) reported that many large-scale debris flows originated recently near the lower limit of permafrost. Permafrost melting appears to have triggered directly or affected indirectly part of these debris flows. Debris flows, possibly related to recent permafrost melting, are also observed on the frontal and side slopes of a number of rock glaciers (e.g., Haeberli et al., 1993).






Paramètre de l'aléa

Sensibilité des paramètres de l'aléa à des paramètres climatiques

Informations complémentaires (données utilisées, méthode, scénarios, etc.)



This report aims at evaluating the effects of the three kinds of freeze-thaw cycles on alpine slope instability, based on studies of contemporary periglacial processes in the Swiss Alps. Attention is focused on the scales of geomorphic change caused by each freeze-thaw type. The study area is located in the Upper Engadin, eastern Switzerland. The lower limit of permafrost lies at about 2400 m ASL on northern exposures, rising to about 3000 m ASL on southern exposures. The periglacial belt, lying above the timberline at 2000 to 2200 m ASL, includes both present-day permafrost and non-permafrost areas. The most extensive periglacial landscape is the talus-to-rock glacier sequence, which develops on slopes covered by coarse debris. Patterned ground and solifluction features are also common, and are characteristic of the slopes underlain by fine debris (Matsuoka et al., 1997).


(4) - Remarques générales



(5) - Syntèses et préconisations

Summary and conclusions

Alpine slopes are subjected to three kinds of freeze-thaw cycles which may be completed in a day, a year or a century. These freeze-thaw cycles influence the slope stability over different temporal and spatial scales (Table 1). Diurnal frost weathering is significant where rainfall or snowmelt supplies abundant moisture to rock surface, producing rock debris mainly of pebble size or smaller. Regardless of the aspect and elevation, most of the debris slopes experience high frequencies of diurnal freeze-thaw cycles accompanied by frost heave of up to 2 cm high and creep of the uppermost soil shallower than 15 cm. Such a shallow movement prevails on slopes with a thin debris mantle, resulting in small lobes and stripes. The annual freeze-thaw depth in rockwalls is calculated to be about 5 m, which delimits the maximum size of rock debris produced by frost weathering. The annual freeze-thaw activity reaches depths of 2 m or slightly more in debris slopes. The associated soil movement eventually develops solifluction lobes with a riser of 30 cm or higher. The growth and decay of permafrost, originating from long-term climatic change, cause freeze-thaw action reaching meter-to-decameter scale depths. Despite extremely low frequency, segregational freezing lasting many centuries or millennia may permit the accumulation of ice-rich layers near the top and bottom of the permafrost body. Permafrost melting can trigger cliff falls and debris flows in the thawing phase of millennial freeze-thaw cycles.

Type of freeze-thaw

Frequency (a-1)

Freeze-thaw depht (m)

Major processes and landforms







Debris fall, needle ice creep, frost creep, small stripes and lobes





Boulder and block falls, solifluction, active layer glide, large stripes and lobes





Cliff fall, deep landslide and debris flow

Table 1. The role of freeze-thaw cycles in controlling slope instability in the Swiss Alps

Références citées :

Aldrich, H. P. (1956). Frost penetration below highway and airfield pavement. Highway Research Board, 135, 124-149.

Barsch, D. (1996). Rockglaciers: Indicators for the Present and Former Geoecology in High Mountain Environments. Springer, Berlin (331 pp.).

Cheng, G. (1983). The mechanism of repeated-segregation for the formation of thick layered ground ice. Cold Regions Science and Technology, 8, 57-66.

Dramis, F., Govi, M., Gugliemin, M. and Mortara, G. (1995). Mountain permafrost and slope instability in the Italian Alps: the Val Pola landslide. Permafrost and Periglacial Processes, 6, 73-82.

Haeberli, W. (1996). On the morphodynamics of ice/debris-transport systems in cold mountain areas. Norsk Geograpisk Tidsskrift, 50, 3-9.

Haeberli, W., Cheng, G., Gorbunov, A. P. and Harris, S. A. (1993). Mountain permafrost and climatic change. Permafrost and Periglacial Processes, 4, 165-174.

Haeberli, W., Evin, M., Tenthorey, G., Keusen, H. R., Hoelzle, M., Keller, F., Vonder Mühll, D., Wagner, S., Pelfini, M. and Smiraglia, C. (1992). Field report, permafrost research sites in the Alps: excursions of the international workshop on permafrost and periglacial environment in mountain areas. Permafrost and Periglacial Processes, 3, 189-202.

Hallet, B., Walder, J. S. and Stubbs, C. W. (1991). Weathering by segregation ice growth in microcracks at sustained sub-zero temperatures: verification from an experimental study using acoustic emissions. Permafrost and Periglacial Processes, 2, 283-300.

Mackay, J. R. (1981). Active layer slope movement in a continuous permafrost environment, Garry Island, Northwest Territories, Canada. Canadian Journal of Earth Sciences, 18, 1666- 1680.

Matsuoka, N. (1991). A model of rate of frost shattering: application to field data from Japan, Svalbard and Antarctica.

Permafrost and Periglacial Processes, 2, 271-281.

Matsuoka, N. (1994). Diurnal freeze-thaw depth in rockwalls: field measurements and theoretical considerations. Earth Surface Processes and Landforms, 19, 423-435.

Matsuoka, N. (1997). Monitoring rockwall instability in the Murtèl-Corvatsch region, Upper Engadin. Mitteilungen der Versuchsanstalt für Wasserbau, Hydrologie und Glaziologie, ETH Zürich, in press.

Matsuoka, N., Hirakawa, K., Watanabe, T. and Moriwaki, K. (1997). Monitoring of periglacial slope processes in the Swiss Alps: the first two years of frost shattering, heave and creep. Permafrost and Periglacial Processes, 8, 155-177.

Prick, A. (1997). Critical degree of saturation as a threshold moisture level in frost weathering of limestones. Permafrost and Periglacial Processes, 8, 91-99.

Schindler, C., Mayer-Rosa, D., Keusen, H. R., Bezzola, G. R., Haeberli, W. and Zimmerman, M. (1993). Felsstürze Randa 1991. In Cartographie géomorphologique – Cartographie des risques, Institut de Géographie, Lausanne, Travaux et Recherches, 9, pp. 117-129.

Smith, D. J. (1987). Solifluction in the southern Canadian Rockies. Canadian Geographer, 31, 309-318.

Zimmerman, M. and Haeberli, W. (1992). Climatic change and debris flow activity in high- mountain areas: a case study in the Swiss Alps. Catena Supplement, 22, 59-72.