Recent studies oh periglacial processes and environments are reviewed on the basis of the mechanisms of frost action in rocks and soils. Frost action modifies the landform when the ground thaws after having been subject to frost heave or shattering. Freezing expansion, the cause of frost heave or shattering, begins just after the surface temperature has fallen below 0°C, proceeds rapidly with cooling until-2 or-3°C, and eventually attains a peak value usually at-5 to-7°C. This indicates that frost heave or shattering occurs just above the descending freezing front, and that lower temperature (e.g. less than-10°C) does not increase the power of frost action. Frozen soils are deformed as a result of frost creep, solifluction or active layer glide; the type of process is dependent on their grain size and moisture condition.
Diurnal and annual freeze-thaw cycles have different effects on periglacial landforms depending on their frequency and penetration depths. Diurnal freeze-thaw penetrates no more than 20 cm in soils and 50 cm in rocks, thus causing shallow soil movement of usually less than 10 cm and producing small debris less than 20 cm in diameter. Such shallow ground activities are believed to be predominant in most mid-latitude alpine environments where diurnal cycles occur frequently. The annual cycle controls the maximum depth of soil movement and the maximum size of fallen debris, which rarely exceed 200 cm and 500 cm, respectively, in Japanese alpine environments, as indicated by the records of ground temper.ature and theoretical considerations.
Recent laboratory and theoretical works, combined with field measurements, have enabled us to construct quantitative models of frost action as a function of environmental and geological factors. For example, the rate of bedrock frost shattering was expressed as a function of freeze-thaw frequency, degree of saturation and tensile strength by a simple model which agreed well with field data from several periglacial environments.
There are many problems to be solved on frost action environments. The most important ones are the influence of permafrost on frost action, the origin of block slopes and the sensitivity of frost action to climatic change. Permafrost may intensify frost action in overlying active layers: by acting as an impermeable layer, by producing cryostatic pressure, or by causing the two-sided freezing. Field measurement data, however, do not necessarily indicate high magnitudes of frost action in permafrost regions. The influence of climatic change on frost action cannot be discussed unless this problem is solved. Block fields and slopes, usually regarded as periglacial landforms, have not yet been explained in terms of the mechanisms of frost action. We should carefully evaluate each factor controlling frost action to solve these problems.

View full abstract

Smooth debris-mantled slopes develop on the windward side of high altitude areas above timberline in the Tanigawa Mountains, central Japan. Most of these slopes occur on a gentle north-facing slope, as the Tanigawa Mountains run roughly in an east-west direction. The slopes investigated are vegetation-free or sparsely vegetated ones that are mantled with rubble layers around Mt. Sennokura (2, 026.2 m a. s. l.) and Mt. Tairappyo (1, 983.7 m).
Six experimental sites (A, B, ……, and F) with gradients ranging from 0° to 39° were chosen. Ground temperatures were measured at site A from early May 1976 to early June 1977. Ground conditions and transitions of positions of painted stones at sites A to F were observed 6 to 10 times in 1974-1978.
The conclusions are summarized as follows:
Downslope displacements of surface gravels recorded by various methods range from 2.1 to 3.1 cm per year, on average. These values are similar to values obtained in Iide Mountains (Higaki, 1990). There are some types of slow mass movement operating on debris-mantled slopes: needle-ice creep, frost creep, gelifluction and debris creep. Rapid mass movement processes such as debris flows and rolling down of gravels often occur, but only small amounts of the angular gravels are moved. Rainwash and snow meltwater in a warm period also carry the angular gravels on bare slopes. In this paper the author intends to quantitatively clarify the relative degree between periglacial and non-periglacial processes operating on a gentle rubble slope. In consequence, the rate of slow mass movement caused by the freeze-thaw process in late autumn and spring may cover 60 percent of the total annual mass transfer. Accordingly the amount of debris movement such as slope wash in a warm period above 0°C may represent 40 percent of total slope distance. The average distances of surface gravels during the spring thaw period are approximately 2 times longer than those during the late autumn freeze period. Needle-ice creep and gelifluction are dominant processes operating on the rubble slopes from mid or late March to late May in the spring thaw period. In the late autumn freeze-thaw period, needle-ice creep and frostcreep dominate on the rubble slopes, but these effects are weak. These conclusions are different from the results of an investigation in the Kitakami Mountains (Sawaguchi, 1987) because of differences in altitude and geomorphic and climatic environments.

View full abstract

In this study the author attempts to discuss slow mass-movement processes caused by freezing and thawing in relation to the compositions of slope materials and fluctuations of ground temperature, in Spitsbergen and in the Japanese high mountains.
Study sites in Spitsbergen are located in Adventdalen and Reindalen, Nordensiöldland, central Spitsbergen (Fig. 1). These sites are debris slopes with gradients ranging from 6° to 32°, mainly covered with Jurassic, Cretaceous, and Tertiary shale fragments.
The rates and deformation patterns of downslope movement of surface rubble are measured by eighteen painted stone-lines. Many grass fibre-tubes, furthermore, were inserted vertically into the ground in order to detect downslope movement in the active layer.
Year-round ground temperature measurement was achieved at a flat surface (480 m a.s.l.) of the Adventdalen site using data logged at 3-hour intervals in 1988-1989. Sensors were installed at 5, 15, 30, 50 and 80 cm depths.
Studies to slow mass-movement in the Japanese high mountains were performed in Mt. Shirouma, the Kitakami Mountains, and Mt. Hoh-o. Long-term ground temperature records were observed on the Kitakami Mountains, Mt. Hoh-o and Mt. Akaishi.
The results of this study can be summarized as follows.
The average movement rate of painted stones from 1988 to 1990 was 3.3 cm/yr. in Spitsbergen (Fig. 2). Deformation features of all painted stone-lines showed patterns parallel to the basement lines independently of the compositions of slope materials (Figs. 3. 4. 5. 6). The movement rates at the tops of grass fibre-tubes were similar to the rate of movement measured by painted stone-lines (Fig. 8). These imply that deformations of painted stone-lines and grass fibre-tubes occurred by the same processes, that is, frost creep and gelifluction. Furthermore, movement of slope materials by diurnal freeze-thaw cycles is negligible in Spitsbergen, because of the very low frequency of this cycle (Fig. 10; Table 3). Therefore it is considered that annual freeze-thaw cycles mostly contribute to the movement of the slope materials in this area.
Average movement rates of painted stones measured in the Japanese high mountains are 26.1 cm/yr. in Mt. Shirouma, 21.7 cm/yr. in the Kitakami Mountains, and 39.7 cm/yr. in Mt. Hoh-o (Fig. 2). These rates are one figure faster than those of Spitsbergen. Deformation features of painted stone-lines are classified mostly into two patterns, one a festoon type and the other a parallel type (Fig. 7). The parallel type occurs on slopes mantled with thick surface rubble of coarse size. The festoon type is recognized on slopes mantled with thin surface rubble and fine materials.
The vertical profiles of many flexible tubes from the above-mentioned mountains on which they were inserted to become almost negligible at about 30-40 cm in depth (Fig. 8). This depth is about one half as shallow as that of Spitsbergen.
Freeze-thaw cycles in the Japanese high mountains overlap with diurnal ones and annual ones, especially those occurring in very high frequency (Fig. 9; Table 3). Movements on the slopes covered with thin rubble and fine materials are caused by diurnal and annual cycles. On these slopes deformations of painted stone-lines show the festoon pattern, and the movement rates are faster than those on slopes consisting of coarse rubble, which almost all move in annualfreeze-thaw cycles.
Considering the above mentioned results, in Spitsbergen differences incompositions of slope materials do not reflect movement processes and the rate of movement, because movement of slope materials occurs in almost annual freeze-thaw cycles. On the other hand, in the Japanese high mountains movement of slope materials occurs in two different freeze-thaw cycles, which are annual and diurnal ones.

View full abstract

The authors discuss the development of block fields and sorted polygons on the flat surfaces at the summits of fells (monadnocks; tunturi in Finnish) in Finnish Lapland in relation to the retreat of the ice sheet during the last glaciation.
There are many fells in the study area (Fig. 1) with their summits elevated several hundred meters; the highest peak is 806 m. Block fields and inactive sorted polygons with erratic boulders in their forming materials were developed on the flat surfaces at the summits of fells above about 600 m. Only block fields with erratics were developed on the summits between about 500 m and 600 m in elevation. Furthermore, on the summits below about 500 m, neither block fields nor sorted polygons were developed but glacial striaes and erratics existed.
All of the summits in the study area were covered with ice sheet during the late glaciation. However, because the higher summits with block fields appeared above the ice sheet like nunatak forms in the early stages of deglaciation, it was cold enough to produce block field-forming materials. On the other hand, the lower summits without block fields were covered with ice sheet for a longer period than the higher ones. It was not cold enough to produce block field-forming materials on the lower summits (Fig. 3, Table 1). Therefore, in the present study area, the authors define a boundary zone, which is a transitive zone of freeze-thaw weathering with an altitudinal range below the linear weathering limit (Nesje et al., 1987).
Inactive sorted polygons are commonly developed on the summits of fells higher than 600 m in their elevation. On the other hand, active sorted polygons are developed on the river floors and lower river terraces at the bottoms of the shallow valleys, mostly about 100 m in elevation (Fig. 2). Very severe temperature inversion was observed at Kevo (Fig. 1) in one of the valleys where active sorted polygons were developed. This is a very common phenomenon in Finnish Lapland. The values for freezing index, thawing index, and freeze-thaw days and cycles (Fig. 4, Table 2) show that valley bottoms have a more suitable climatic environment for periglacial processes than mountain tops. The authors think that severe temperature inversion is one of the important factors for active sorted polygon forms.

View full abstract

Three-dimensional fabric analysis was done in order to characterize the fabric of slope deposits gathered from various sources. Although the distribution of periglacial slope deposits on a logarithmic ratio plot partly overlaps those of nonperiglacial deposits such as soil creep, sediment flow plots, debris flow plots, talus, outwash and stratified slopes (in openwork beds) indicate low values of both C and K compared with those of periglacial slope deposits (Fig. 1, Table 1). This result leads to the conclusion that (1) the periglacial slope deposits have higher C and K values than non-periglacial deposits except for glacial ones; (2) the periglacial slope deposits are plotted in a zone with C ranging over 2.5, K ranging over 0.5 on the logarithmic ratio plot; and (3) it is, therefore, possible to distinguish the periglacial slope deposits from the non-periglacial ones by means of logarithmic ratio plots.
Three-dimensional fabric analysis of 2, 350 clasts in the site located on the fossil periglacial slope deposits in the Hidaka Mountains has been done to evaluate the relations of macro fabrics to both length of a-axis (Table 2) and clast shape (Table 3). The main results are that (1) clasts with a longer a-axis have macro fabrics with little deviation from mean orientation (Figs. 2 and 3); and (2) clast shape has little effect on fabric in the longer a-axis range (Fig. 4, Table 4).
Taking these results into consideration, azimuths and dips of clasts with longer a-axis have to be measured in order to gain fabrics with little deviation from mean orientation in periglacial slope deposits.

View full abstract

The authors conducted a field survey related to the genesis and occurrence of permafrost in James Ross Island and Seymour Island, Antarctic Peninsula region, during the 1989-1990 Antarctic summer season. Mean annual air temperature in both islands is estimated at about-10°C.
Seymour Island is located in the Weddell Sea. There was no ice-sheet over the island during or since the last glaciation period. Marine terraces of three different levels are distributed in the is-land: an upper terrace at Meseta (about 200 m a. s. 1.), a middle terrace at Sub-Meseta (about 50 m a. s. 1.) and a lower terrace at Larsen (about 5 m a. s. 1.). In order to estimate the thickness of permafrost, geo-electric resistivities were surveyed on these terraces. In the lower terrace at Larsen, long-term monitoring of ground temperature profiles was carried out for two years. The annual mean ground temperature and the temperature gradient indicate that the permafrost base is 34 m deep. This coincides with the depth obtained by geo-electric resistivity measurements.
In James Ross Island, about 90% of the ground surface is covered with an ice sheet or glaciar. Ice-free ground spreads to the northwestern part of the island. Coastal terraces of three different levels develop around Santa Marta Point along Croft Bay: upper terrace(21-24 m, 32-35 m a. s. 1.), middle terrace (10-17 m a. s. 1.) and lower terrace (3-5 m a. s. 1.). The group of upper terraces is composed of glacial till or fluvio-glacial deposits. Shell samples were collected from the deltaic deposit which composes the middle terrace, and from marine sands and gravels which cover the surface of the lower terrace. Then ¹⁴C dating was done with the results of about 25, 000 y. B. P. and 3, 000 y. B. P. respectively. According to geo-electric resistivity measurements, the depths of permafrost bases on the upper and lower terraces are estimated at 40 m and 3-5 m, respectively. It is considered that permafrost on the upper terrace occurred prior to the last glacial maximum age. The comparison between Antarctic and Arctic permafrost suggests that the permafrost in Antarctic regions is shallower than that in Arctic regions under similar annual mean temperatures at the present time.

View full abstract

In Japan research studies on periglacial smooth slopes have increased in number since the early 1970s. This increment is considered to be reflected by the development of Quaternary research in the 1960's. The Research Group for Alpine Geomorphology, mainly composed of young geomorphologists, has played an important role in the development of these studies, which were mostly carried out in the alpine zone of the Japanese Alps. The research began with field investigations such as measurements of grain-size distribution of slope deposits and measurements of rates of periglacial debris movement. In succession, the processes of debris movement and slope forms were examined. These quantitative studies covered aspects insufficiently investigated in European studies performed by Spreitzer and his colleagues. It was revealed that the grain-size characteristics and the stability of slope deposits largely depended on differences in the lithologic features of each area, and the chacteristics of surface materials affected the vegetational distribution. Some chronological studies revealed that debris production from bedrock had occurred even in the Late Glacial age and the Neoglaciation age. The importance of fossil periglacial slope studies is increasing for environmental reconstruction of the Last Glacial Age. Studies of plant ecology were also conducted on periglacial smooth slopes. Based on recent experiments both in the laboratory and in the field, studies of the mechanism of frost shattering are under way. They have come to support the adsorption force theory. The mechanism of formation of periglacial block slopes and the convexity of crest slopes are themes which require investigation in the future.

View full abstract

The altitudinal zone affected by periglacial processes is classified into periglacial and transitional zones. The periglacial zone is the optimal zone of zonal periglacial processes and/or of periglacial denudation-smooth slopes, and is generally defined as the zone between snowline and lower boundary of solifluction. The transitional zone is the zone where extrazonal periglacial processes occur above and below the periglacial zone. There are a number of unclarified points regarding the altitudinal positions of the periglacial and the transitional zones, and their relations to the permafrost zone. The present paper examines the altitudinal positions of the periglacial, transitional, and permafrost zones in Japan, and tries to clarify the altitudinal relationships among them, as well as the genetic relationships between periglacial processes and permafrost.
In Japan, the alpine zone can be substituted for the periglacial zone whose altitudinal extent is ca. 1, 000 m. Its upper boundary is expressed by the snowline, while its lower boundary is represented by the scrubline which correlates well with the warmest month isotherm of ca. 10°C. On the other hand, the altitudinal extent of the transitional zone below the periglacial zone reaches ca. 2, 000 m. The lower boundary of this transitional zone is represented by the lower boundary of the Fagus crenata forest zone, i. e., the 21.1°C mean summer isotherm. The lower boundaries of continuous and discontinuous permafrost zones are represented by the-6°C and the 21.1°C annual mean isotherms, respectively. The 10°C isotherm for the warmest month (August) descends north ward with the gradient of ca.-70 m/°N, while the gradients of the above-noted-6°C and the-2°C annual mean isotherms attain ca.-160 m/°N. Because the 10°C August isotherm intersects the-6°C and the-2°C annual mean isotherms, the Japanese periglacial and the transitional zones are classified into the following types (I a to II c) based on their altitudinal relations to permafrost zone:
(Ia) The periglacial zone occupied by only continuous permafrost. This zone occurs on the Hokkaido mountains higher than ca. 2, 250 m, north of ca. 42°N.
(Ib) The periglacial zone occupied by both continuous and discontinuous permafrost. This zone occurs in the Japanese Alps and on Mt. Fuji above 2, 900 to 3, 100 m in altitude.
(IIa) The transitional zone whose upper part is characterized by the presence of both continuous and discontinuous permafrost. This zone includes the Hokkaido district north of ca. 42°N.
(IIb) The transitional zone whose upper part is occupied by only discontinuous permafrost. Its distributional range is ca. 33° to 42°N in latitude. Most of high mountains of the Tohoku and Chugoku districts are included in this zone.
(IIc) The transitional zone where no permafrost occurs. All the high mountains of the Kyushu district are included in this zone.
Periglacial processes are classified into the following two types (A and B), based on their relations to permafrost:
(A) The periglacial processes that form small-scale landforms and/or-phenomena (e. g., patterned ground, solifluction lobe, involution, etc). This type of periglacial process is not controlled by permafrost.
(B) The periglacial processes that form large-scale landforms (e. g., block stream, periglacial denudation-smooth slope). This type of periglacial process is controlled by permafrost.
These results obtained in this paper are indispensable for reconstructing change in the paleoenvironment in Japan.

View full abstract

This paper discusses the altitudinal boundary between periglacial and non-periglacial zones in northeastern Japan during the Last Glacial Age based on the lower limit of fossil periglacial slopes and the upper limit of distribution of Pleistocene tephra layers.
The characteristics of fossil periglacial slopes in Japan are summarized as follows:
Form: Convex-concave smooth slope on main ridges.
Lower limit of the distribution: ca. 2, 000 m a. s. l. in Central Japan (Table 1). and with height gradually reduced northward.
Slope material: Debris layers or block layers. At present, most of them are stable and covered with alpine and subalpine vegetation.
The lower limit of fossil periglacial slopes is bounded by the upper limit of dissected slopes. The boundary ranges from 100 m to 500 m in height in the upper reaches of the Tama River (Fig. 1). In this area, periglacial block streams in the past and debris flow deposited by heavy rainfall in the present are sporadically distributed around the lower limits of fossil periglacial slopes (Fig. 2). Based on the above-mentioned data, altitudinal morphgenetic zones in the upper reaches of the Tama River during the Last Glacial Age can be divided into four zones as follows: the continuous periglacial slope zone, the sporadic periglacial slope zone, the dissected slope zone, and the stable slope zone. The boundaries of these morphogenetic zones range in height between 2, 000 and 1, 200 m a. s. l. (Fig. 3). From this it may be inferred that the boundary between periglacial and nonperiglacial zones largely varied in altitudinal extent during the Last Glacial Age.
The existence of tephra layers on the mountain slopes and indicates that the slopes have been stabilized since the tephra was deposited. Figure 4 shows altitudinal distribution of marker-tephra layers (1st group: early-middle stage in the Last Glacial; 2nd group: late stage in the Last Glacial) in northeastern Japan. The upper boundaries of both the 1st and 2nd groups are at gradually lower altitudes from south to north. The upper boundary of the distribution of tephra layers rose from the early-middle Last Glacial (1st group) to the late Last Glacial (2nd group). The lower li mits of fossil periglacial slopes in northeastern Japan indicated in previous studies are located above the upper boundary of the distribution of the 1st group. On the other hand, the locations of the macrofossil plants and fossil pollen during the Last Glacial Age were below the upper boundary of the distribution of the 2nd group. These facts indicate the existence of stable slopes covered with vegetation during the Last Glacial Age.

View full abstract

Low-relief erosion surfaces on high ridges in the Akaishi Mountains have been considered to be “remnants of peneplains”. The purpose of this study is to reconsider the origin of those surfaces by using a slope process-response model based on the continuity equation and on the rates of slow mass-movement processes investigated on the periglacial smooth slopes.
In the Akaishi Mountains periglacial smooth slopes mainly formed during the last glacial age are preserved in a high altitudinal zone (Figs. 1, 2). Low-relief erosion surfaces on high ridges are the gentlest parts (slope angle is less than 20°) of those periglacial smooth slopes (Fig. 3).
There are several quantitative studies about the slow mass-movement due to ground frost in the alpine zone of the Akaishi Mountains (Koaze, 1964; Okazawa et al., 1975; Higuchi, 1987). According to these studies the annual rate of surface gravel movement is in proportion to the slope gradient (Fig. 5; equation (1)). In Daishyoji-daira (Fig. 4), the vertical profiles of the annual rate of downslope soil movement (Fig. 6) were also measured (Okazawa et al., 1975) on periglacial smooth slopes mantled with surface rubble layers. These data indicate that the potential rates of weathering is greater than the rate of transport; in other words, the transport limited condition on those slopes.
In this case, based on the above empirical process laws of debris transport, the following continuity equation is deduced:
dy/dt=-5.8×10^-2 {(d²y/dx²)+(dy/dx)/x}
where, x [m] is the distance from the divide, y [m] is the altitude and t is time elapsed [year].
The development of an initially straight slope under slow mass-movement (Fig. 7) obtained from this equation shows that, without flat interfluve as an initial landform, the erosion surfaces can be formed over a period of tens thousands of years under periglacial climatic conditions. Therefore the low-relief erosion surfaces observed on high ridges do not necessarily have to have been derived from flat interfluves such as “peneplains”. But they probably have been formed at about their present height by periglacial processes in the glacial ages during the late Quaternary.

View full abstract

The authors concluded a study of the depositional slopes distributed in the Chugoku Mountains. These depositional slopes are divided into four types: colluvial slopes, block streams, debris-man-tled smooth slopes, and rock fall talus. The colluvial slopes are distributed in the piedmont area, and are divided into six types: Surface I, II₁, II₂, III, IV and V, respectively. The block streams, which can be divided into the Upper and Lower Block Streams, and the debris-mantled smooth slopes are located around the summit. Rock fall talus is found at the flanks of the mountains. Several tephra, which are useful as time markers, are embedded in the sediments of these slopes.
The formative ages of the colluvial slopes were revealed to be as follows: Surfaces I, II₁, II₂, and III were formed prior to the last interglacial stage, during the first half of the last glacial age, during the second half of the last glacial age, and between the late glacial age andapproximately 6, 000 years ago, respectively. Then followed the formation of Surfaces IV and V. The Upper Block Streams took shape during the first half of the last glacial age, while the Lower Block Streams were formed during the second half of the last glacial age. The rock fall talus and the debris-man-tled smooth slopes were formed during the second half of the glacial age.
Each of the formative ages of Surfaces II₁ and II₂ the block streams, the rock fall talus and the debris-mantled smooth slopes is correlated with a cold period. This is because the rock fragments of the sediments were produced by frost shattering and then were shifted by various types of mass movements, including solifluction. This suggests that these formative processes are closely related to the periglacial environment.

View full abstract

A number of papers on the geomorphology of western Japan have revealed active formation of periglacial landforms during the Last Glacial. Traces of past freeze-thaw action can be found even in coastal lowlands, including those in south Korea (Fig. 1). In contrast, the altitude of the Last Glacial forest limit of western Japan, reconstructed from decreases in temperature, exceeds 1, 000 m (Fig. 2). This marked discrepancy in altitude has led Japanese geomorphologists to discuss whether the lowlands of western Japan experienced a periglacial environment or not. However, no rational conclusion has been reached so far. This paper examines the distribution of the periglacial landforms and physical factors such as climate, vegetation, and geology so as to offer an explanation for intense freeze-thaw action far below the forest limit.
Even under the present temperate climate, periglacial earth hummocks and stone banked steps can occur in the treeless mountains of western Japan (Fig. 3 and Photo 1). Active soil creep by freeze-thaw action during winter is also reported from coastal lowlands with sparse vegetation cover. Under the Last Glacial colder climate, more favorable conditions for periglacial processes undoubtedly prevailed so long as lands were free from thick vegetation. Accordingly, attention here is directed toward the Last Glacial vegetation rather than temperature depression at that time.
Reconstruction of the vegetation map of Japan at the Last Glacial maximum has been attempted by several researchers. They yielded different results, because data for map compilation were not identical: change in temperature, palynological evidence, and distribution of existing plants and insects were separately compiled using different criteria. The method based on existing life was adopted by Hiura (1980). He recognized that some kinds of life closely related to vegetation types have scarcely widened their territories since the Last Glacial maximum, and thus are useful for reconstructing vegetation at that time. According to Hiura's map, treeless and/or sparse forest areas broadly occurred in the lowlands of western Japan and south Korea (Fig. 4). These areas agree fairly well with distribution of fossil periglacial landforms (Fig. 5-a); consequently Hiura's map is the most applicable from the geomorphological viewpoint. In addition, these areas are thought to have low potential for tree growth, because 1) half of them correspond to the distribution of recent treeless mountains (Hageyama: Fig. 5-b), which were human-induced but have been maintained due to an unfavorable physical setting that includes slope form, geology and climate; 2) the other half are in the proximity of active volcanoes, and thus were subjected to frequent falls of volcanic ash during the Last Glacial (Fig. 5-d). These correlations also support the validity of Hiura's vegetation map. Moreover, most of the areas have small amounts of snow in winter (Fig. 5-c), giving weak protection against freeze-thaw action. The condition of little snow also applies to the Last Glacial, because the atomospheric pressure pattern of Japan in winter has been basically unchanged since then.
In conclusion, factors including slope form, geology, climate and frequent fall of volcanic ash worked together to lower the land potential for tree growth in the lowlands. This caused recession of thick forestation in spite of temperatures high enough for tree growth, resulting in strong freeze-thaw action. Thin snow cover during winter also facilitated the occurrence of periglacial landforms. The landforms are located below the Last Glacial treeline determined by temperature, and thus can be called “extrazonal”. However, such extrazonal landforms developed in broad zones of western Japan and south Korea.

View full abstract