SURVEY OF CRYOGENIC PROCESSES, PERIGLACIAL FORMS AND PERMAFROST CONDITIONS IN SOUTH AMERICA

This geocryological inventory contributes to the state of the art and to recent advan.ces of Geocryology in the countries of South America with perrnafrost occurrence or seasonally frozen ground, where this field of research is of scientific interest. The aim is to increase the knowledge about the are as of the present cryosphere being modified by man and/or suffering changes due to climatic factors. A brief climatic analysis helps to understand the main South American cryogenic regions still poorly known and lacking the corresponding geomorphological cartography. The survey also emphasizes the fact that very few data are available on altitudinal Andean perrnafrost degradation caused by global warming processes, which should be monitored for hydrological and other reasons. The main cryogenic processes observed such as cryometeorization, nivation, solifluction, cryoturbation and sorting are described in different lithologies, places, etc. Some quantifications are given, such as data on solifluction movements and sedimentological methods applied to detect cryogenic phenomena. The latest data of Andean perrnafrost are summarized. They have recently been obtained with the help of different methodologies, including geophysics and/or, through borehole with ground temperature measurements. The most common Andean cryogenic forms are presented: microforms, patterned ground, felsenmeer, cryoturbation structures, thufurs in moors and solifluction lobes. Characteristic mesoforrns of the Central Andes, rock glaciers, or important elements of a periglacial environment such as sedimentary cryogenic slopes, cryoplanation surfaces or asymmetrical valleys are also described.


lNTRODUCTION
Geocryology is the science that studies the environment and the ecology of cold regions, their natural, geological and physico-chemical processes in relation with cycles of freezing and thawing as welI as the relation between alI those phenomena and human life.
The original concept periglacial (introduced by Walery von Lozinski on the Eleventh lnternational Congress of Geology in Stockholm in 1910), originalIy applied to describe the climatic and geomorphologic processes in areas close to Pleistocene ice, was complemented by the Russian term Geocriologija (NAUK AKADEMIJA 1960) for its wider concept referring to low temperatures, with permanently or even short term, seasonal or daily frozen ground.
When the ground, like soil, rock and organic material, remains at or below O°C for at least two consecutive years, it is calIed permafrost ( VAN EVERDINGEN 1998).
High latitudes together with a low MAAT (mean annual air temperature) are conditioning factors for the occurrence of Antarctic permafrost and for permafrost of subantarctic islands.ln our case the altitude provides the basis for us to classify permafrost as mountain permafrost of the Andean type.However the variety of dry permafrost with temperatures below zero but without ice is also a frequent phenomenon, and should not be neglected when considering cryotic characteristics of the Central Andes and the region of the South American Dry Diagonal or the Desert Andes.
It is the aim of this work to present the state of the art and recent advances of Geocryology in the countries of South America with permafrost occurrence or seasonally frozen ground and where this field of research is of scientific interest.
The future role of Geocryology has to be emphasized because it may provide important insight to the paleoclimatic reconstruction in climate changes and hydrology in South America.

PERMAFROST
One of the first attempts to map cryogenic regions in South America was made by BARANOV in 1964.
The map of MAAT of South America by HOFFMANN (1975) (Fig. 1, adapted by Trombotto) and the lines for +5, O and -5°C may help to determine locations with possible permafrost as welI as the most important areas with seasonal freezing or periglacial phenomena.Some areas require more specificinformation (e.g.the coldest site in South America in Patagonia, ice covers).
From the > 13 million km2 that correspond to the permanently frozen ground in the Southern Hemisphere, the largest part is that of the Antarctica, whereas there is very few data about permafrost occurrence on the South American continent.
Some estimations held that mountain permafrost would cover approximately 30,000 km2 (GORBUNOV 1978), but recently it is assumed that it covers even :;70,000 km2 (HAEBERLI et. ai. 1993).They divided permafrost into maritime permafrost in the Southern and Tropical Andes and continental permafrost in the Central Andes.
ln Argentina the term of approximately continuous permafrost (as introduced by GARLEFF & STINGL 1986) is applied to the type of permafrost which is very much restricted by topography and is limited to the MAAT isotherms of _2°C and -4 °C (at the Central Andes, 33°S and approximately 4500 m ASL) with precipitations between 500 -900 mm/y and limited to the isotherms of -1°C and -2°C with a precipitation of 300 mm/y in the case of the Argentine Puna region (Table 1).
The lower limit of Andean permafrost is restricted by a type of discontinuous permafrost (Table 1, Fig. 2) which may be observed in some parts through rock glaciers (BARSCH 1977).At present the lower permafrost limit at the Cordón deI Plata (Central Andes, Mendoza) is found at an :; elevation of 3700 -3800 m. ln the Southern or Wet Andes, the so-called Patagonian Andes, the occurrence of permafrost was registered at 51 °30' S in Santa Cruz (ROIG 1986) at an elevation of 980 -1100 m in the years 1977 and 1978.ln the Lake Region between 35°and 45°3 0'S, in the province of Chubut however, permafrost near Lake Vintter (approximately 44 OS)had been registered at an elevation of 2060 m in recent years (C. Bianchi, pers. comm., 1998).
The term island permafrost (Table 1) was introduced for the phenomenon of isolated patches at 4000 m ASL in the Central Andes of Mendoza (TROMBOTTO 1991) and relict permafrost for frozen ground at 3400 m but with a thick detrital cover in rock glaciers.
Towards the Equator, in the region of the Chimborazo (6275 m ASL, Ecuador), a typical tropical environment but with dry conditions, HEINE (1994) reported the limit of continuous permafrost to be at 5250 -5300 m ASL.Island permafrost, also mentioned by the author, is found below 5000 m.In other parts of the Andes of Ecuador, with soil temperatures around O°C but under humid conditions, there is no permafrost, but a glacial environment.
At the same time, subterranean ice may be preserved as in the ca~e of inactive ice (Table I) found at the salt lakes in the Altiplano and Puna of Atacama (Bolivia and Chile) at 4300m and approximately 22°S (HURLBERT & CHANG 1984) who assigned this type of ice to the Little Ice Age.Thermic particularities and regional dry conditions seem to reinforce the effects of periglacial processes and the persistence of ground ice in the Central Andes.Between 1991 and 1993 Signs of activity in the rock glaciers of Mendoza for example establish lower limits of permafrost; and thus the designation island permafrost (TROMBOTTO 1991).
Additional investigations are required for the distribution of Andean permafrost.It is necessary to map it, to study its topography, its ice content and its behaviour regarding global climatic change as well as its importance to domestic water supply.

INVENTORY OF CRYOGENIC PROCESSES OBSERVED IN SOUTH AMERICA
The main cryogenic processes observed in the South American Andes are cryoweathering (Fig. 3), nival processes, solifluction, cryoturbation, sorting, frost cracking and permafrost creeping with different lithologies.Some sites are being monitored.The studies of the processes in the periglacial environment in the laboratory and in the field were principally guided by the following phenomena.
• Sorting: sorted materiais due to cryogenesis create a cryogenic law, represented by pebbles and cobbles above and sediments below.CORTE (1962 a, 1962 b)  < 0.02 mm.The presence of silt has an important granulometrical significance to generate frost heaving.With water flow and the growth of ice the transport of material in its open system is to be exp1ained.On the other hand it is assumed that the upward movement of b10cks leaves a space quick1y filled by fine matelial that stops the b10cks from settling during the thawing processo • Proceeding in a similar way AHUMADA (1987), in her doctoral thesis, proved the same (sorting) for heavy mineraIs in the active layer.
• I mportant phenomena related with silica deposition in periglacial environments were part of TROMBOTTO's (1988) research for his doctoral thesis .
• As to important cryogenic chemical phenomena VOGT & CORTE (1996) had observed for a long time the possible origin of Ca C03• • Desiccation cracks and thermal contractions were very much discussed among geologists in South America.The dehydration process of the Pleistocene active layer and therrnal contraction may often use the same cracks (TROMBOTTO 1996).
• In comparison to other sedimentary media there is no data abour fluvial phenomena.Under the impact of winds in Mendoza taffonis of cryogenic -petrologic origin could be observed on pyroc1astic rocks at 34°S.Loess is still discussed as an aeolian sediment, but a cryogenic component in its origin cannot be ruled out.As TROLL (1944) assumed, freezing is an essential phenomenon in order to explain the production of silt in the loess.It may well be that the Patagonian loess matelial (TROMBOTTO 1996) as well as the Andean silt or cold loess was influenced by cryogenesis.TROMBOTTO & REGAIRAZ (2000) have recently described a case of cold loess for Mendoza.• Among the cryogenic processes of the Andean cyc1es of freezing and thawing, the mechanisms of different South Amelican 1ithologies and their expelimental correlation are yet unknown (Fig. 4).The cyc1es have to be reproduced in order to simulate the observed natural conditions as in the Central Andes for example, with a low content of humidity and high amplitudes and temperature oscillations between + 30°C and -20°C within 24 hrs. :i7 Trollbrot is one example of an Andean cryogenesis, where the fragments remain together and resemble sliced bread.

SEASONAL FROZEN GROUND
The area affected by seasonal freezing in South America, and Argentina in particular, is much larger than that covered by permafrost.
The important processes are solifluction and the effects of needle ice.The action of the wind or the impact of solar radiation combined with cryogenesis have been studied by PÉREZ (1984PÉREZ ( , 1992)).
In the Central Andes of Mendoza, CORTE (1983) incorporated environments with seasonal freezing with small-size patterned ground and weathering pits (OpferkesseZ) into an area he called parageocryogenical.
On the Patagonian mesetas and in the Argentine Precardillera keppavond (HUML UM & CHRlSTIANSEN 1998) may be found.This is the technical term used on the Faeroe Islands to define active small-scale pattemed ground.ANGELlNI (1990) and BUK (1992) made maps delimiting regions with a maximum seasonal freezing index and mean annual freezing frequencies in Patagonia.Segregation ice ar pipkrake has a strong impact in these areas and influences works of civil engineering in road constmction (Patagonia).Although no permafrost has been detected in the Andes of Mérida and Sierra de Perija (Venezuela), the periglaciallevel of these processes may reach a zone of over 1000 m above 3600 m ASL.Sorted pattemed ground is characteristic of the Páramo.At the Páramo de Piedras Blancas (Venezuela) the limit of needle ice for example, lies at an elevation of 3600 m (SCHUBERT 1979).In the region of the Tropical Andes, the phenomena of seasonal freezing are concentrated on an area called Tierra ReZada which lies above the timberline, with a MAAT of 7°C (subpáramo) to a MAAT of -1,5 °C at the snowfalllevel (Pico Bolívar, 5008 m ASL) in the so-called superpáramo (LAUER 1979).
Two superficial miniature forms that may be observed in these areas affected by the growth of needle ice and its pressure are gaps and miniature stripes.The gaps, or gaps around stones (WASHBURN 1979), are little mounds produced by needle ice and characterized by rocks that were lifted by upheaval and are left in position.Posterior to and frequently after this phenomenon a certain discontinuity or a ring with an empty space between the rOck and the sediment is found.As WASHBURN (1979) pointed out needle-ice occurred even in Brazil.
On the other hand the needle ice leaves different microstructures.
They are usually cmmbly and therefare they are called lumpy surfaces.Their appearance resembles plowed lands (WASHBURN 1979).SAPPER (1915, see WEISE 1983) introduced the term Rasenabschalung; a related term referred not directly to forms but to processes, but applied to peat areas.

FLUVIAL AND AEOLlAN PROCESSES IN A PERIGLACIAL ENVIRONMENT
In a cryogenic environment rivers and streamlets remain frozen during some time of the year.When they are not frozen they play an active role in the denudation of the valleys, a1though this fact is generally ignored.
Streamlets help to increase erosion in areas with permafrost and contribute to the formation of subterranean ice.Streams and subterranean ice extmded by freezing of some parts of the ground also create the characteristic ice in various layers called icing ar Aufeis.It is formed by the successive flow of water that freezes and is then covered by a following flow that freezes and so on.This phenomenon may be frequently observed in the Andes.
The influence of the wind is also decisive for the shaping of either continental or coastal cryogenic forms.A great variety of aeolian forms are found in cryogenic environments.
Clasts and blocks with one ar more surfaces shaped by the wind are designated ventifacts of cold environments.Depending on how many sides are eroded by the wind they are called 1: Einkanter, 2: Zweikanter and 3: Dreikanter (terms of German origin).The eroded surfaces are used to determine the wind direction (GONZÁLEZ BONORlNO & TERUGGI 1952).These ventifacts are part of the desert pavement so typical for the Andean aeolian periglacial environments.Rock cavities caused by aeolian erosion and with some participation of chemical weathering in their genesis are called taffoni (see FRENCH 1988).These taffoni ar honeycomb forms are stones with concavities produced by aeolian erosion (FRENCH 1988), usually in granitic rocks.The erosion of sandstone in a periglacial en'vironment may produce mushroom shaped structures.AUER (1956) described mushrooms as relicts of sandstone for Tierra deI Fuego.In the protected area ofLaguna deI Diamante (34°S, Cordillera Principal) some excellent examples of erosion, particularly in tuff or pumice, may be observed.
Another sediment of aeolian origin or of periglacial genesis is the covering debris (Deckschutt) in the so-called dells or periglacial valleys, with a very important aeolian component.The covering debris represents a diagnosis of the geoforms of cryogenic origino Cold loess is mentioned by HAMELIN & COOK (1967) referring to the loess deposited in the surroundings of glaciers.Generally the loess consists mainly of 50 -60 % silt « than 0.06 mm), 5 -30% clay, and 5 -10% sand.These values may vary by about 10%.
The most outstanding loess deposits of the Southern Hemisphere are those of the Pampean Plains.In comparison to other loess they possess an important textural component of fine sand and volcanic glass.These deposits consist of 50% silt, 50% sand and reveal a high content of volcanic ashes and pIagioclases.Predominant heavy mineraIs are hornblende, hypersthene, and augite.These loess were assigned to an arid climate, but not as indicators of cold climates (TERUGGI 1957).
However, a loess profile located at Las Carreras Valley, on the foot of the Cordón deI Plata, an eastern range of the Central Andes, in the province of Mendoza, Argentina, was dated with therrnoluminescence.It was assigned to the Middle Pleistocene and would express a cryomere of almost 200 ka that may be correlated with the conesponding glacial episode in the N orthern Hemisphere.This fact identifies this profile as a key element for the Quaternary stratigraphy in South America.
On the other hand it cannot be ruled out that for the silt-Ioess material from Patagonia (Camwy Formation, TROMBOTTO 1996) during the cryomeres or cold episodes, cryogenesis was involved.As early as 1936 GROEBER pointed out an important cryogenic relation between cold Andean environments and mechanical weathering.This author however, was looking for an explanation of the old "Bonaerense" and the Pampean Plains as well as for the sands called "Médano Invasor" (Invading Dunes Mesas are fIat mounds referring to Chilean -Bolivian perrnafrost areas (about 1 km 2) in relation with shallow saline high mountain lakes, as Laguna Colorada and Salar de Pujsa (over 4000 m ASL) in the Desert Andes, which hide in their interior strata of pure ice, veinlets or frozen sediments.
They are found in areas with an active layer rich in clay and silt with calcite (eventual thenardite), which enhances the strong reflection in isolation.The ice content varies between 10-87 %, with a thickness between 1 mm and 1 m.The maximum thickness of the permafrost in this area is assumed to be 25 m (GORBUNOV 1993) and the perrnafrost table lies between 20 and 70 cm.While HURLBERT & CHANG (1984, 1988) believe them to be relicts forrn the Little Ice Age, GORBUNOV (1993) thinks they are relicts from the Late Pleistocene to Upper Holocene.-;

Patterned ground
Patterned ground is characterized by a combination of processes with the intervention of frost heaving caused by needle ice, sorting, soil creep, dehydration and thermal contraction caused by the action of freezing and thawing.
According to WASHBURWs (1970) important classification, the most important forrns of patterned ground in South America have been grouped into: nets, stripes, polygons and circles (Table 2).They may be distinguished by their respective geometries.They are designated sorted or unsorted according to their textural differences and the definition of shape.
Nets (Fig. 5) or inegular shapes (Fig. 6) without permafrost but closely linked to the problems of night frost according to TROLL (1944) appear in the Tropical Andes.As PÉREZ (1984) proved, the genesis of patterned ground in the mountains of Venezuela is influenced by solar radiation above all other factors.The nets are usually of small size, in the range of a few centimeters, but under certain favourable conditions and with a more complex genesis as in the Central Andes, they reach the ranges of meters (Fig. 6).These shapes may also be the result of cryoturbation.Desiccation fissures also favour the formation of nets and there are cases of stripes combined with thermal contraction cracks.
For some time we have records of schist stripes in Peru (TROLL 1944)   or with dehyclration of layers saturated with melting water provided by snow patches ar seasonal ice.Thermal contraction polygons (Fig. 9) are also frequently seen in the Central Andes.
The type of sorted stone circles (Fig. 10) with a depression in the centre are also called Amundsenrings (TROMBOTTO 1991).
Extrusion structures (frost boil, Frostbeule, ostiol) ar tundra volcanoes (Figs. 6 and 11) are conical farms of fine sediments expelled by cryoturbation and cryostatic pressure in the substratum.They are frequently involved in the genesis of patterned ground.They may build micronets within circ1es ar stone polygons or else contribute directly to the cre'!tion of stonenets.
During the thawing process these extrusion structures are flattened and they constitute isolated spots of fine material that resemble "cakes of sand or clay", which gave arigin to the German name Erdkuchen apart from the term Frostbeule.They are frequently part of the sorted forms on the surface and they do not need permafrost.They are found very frequently throughout the Andean Cordillera.
The term nubbin (from middle low German: knot on a tree, applied by WASHBURN 1979) reaches back to a medieval Germanic word for a round protuberance some times prolonged    1988).The mechanism of these arctic structures is called subduction.It is explained by DYKE & ZOLTAl (1980) who calculate its movement to be less than 1 mm per year.

Solifluction fonns
The sIm" creep downhill of saturated soi!, caused by freezing and thawing in cryogenic regions is defined as solifluction.With permafrost it is called gelifluction.Some characteristic Andean fonns are: • Solifluction sheets in the Central and Southem Andes (Fig. 12).They are observed on surfaces with little inclination.They create fonns interwoven or related with reliefs characterized by coarse sediments.They may be observed on slopes of> 2°.A type of solifjuction layer with a reduced inclination angle contributes to the genesis of pattemed ground or enhances solifluction contact.
• Solifluction lobes (Fig. 13).They appear as protuberances ar tongues and may be identified on flanks of the slopes.Solifluction lobes may have terraced surfaces, rocky fronts and may have terraced surfaces, rocky fronts and may be combined with fissures provoked by the discharge of melting water at their frontal basis.ln this last case the clasts are usually in vertical position.They are the classica1 product of macrosolifluction or slope solifluction as TROLL (1944) called it.The lobes may show a remarkable vertical sorting on their most superficial parts.At the Cordón deI Plata, the lobes display angles between 28°and 29°mostly, with blocks between 30 and 40 cm on their surface and they are related   1983).
with the fonnation of debris slopes.The rate of movement of solifluction lobes in greywacke at the Lagunita deI Plata ranged between 3 and 6 cm per year (1983 -1996).The maximum speed was registered at an elevation of 4300 m (TROMBOTTO et aI. 1984, TROMBOTTO 1991 ).
• Solifluction terraces frequent throughout the Andes.Solifluction terraces are relatively smallsized, stepped solifluction fonns.They may appear combined with solifluction lobes and they do not require permafrost.ln the entire Andean region these forms are most abundant.They reveal a vertical sorting with large quantities of fine sediments.They appear in a _wide range of petrographical classes and with different inclination angles on the slopes.The most frequent slopes with terracettes in the area of the Lagunita deI Plata, Cordón deI Plata vary between 12°and 25°.A particular type was identified in the Central AMes (TROMBOTTO 1991) because of its magnificence and because it is a geofarm that appears in combination with snow patches.These giant forms, to be observed at the Lagunita deI Plata, display solifluction processes in great steps in combination with the influence of snow patches, which persist for a long time in their frontal parts.
Garlands do not require pennafrost.It is a fonn of creep stopped or restricted by vegetation with a generally terraced surface.They were described extensively until approximately 5000 m where nets, stripes and small stone polygons appear, as in the Argentine Puna for example (lGARZÁBAL 1983).CORTE (1953) found garlands in Argentine, while trying to explain cryopedological phenomena at the Pampa de los Avestruces at an elevation of 3500 m (Mendoza).ln Peru GRAF (1971), found them at 4850 m at the Valle Viscas (11°35' S).In Chacaltaya, Cordillera Real in Bolivia, HASTENRATH (1971) registered them at 4500 m. ln the Southem Andes, GARLEFF (1977) reported solifluction farms combined with ar stopped by vegetation in the region of the cold woods of Araucaria and the subantarctic Magellan woods, above the timberline, ar in the area of transition to periglacial soil (at Tronadar at 1500 m and at the Fitz Roy at 1000 m approximately).These authars also documented garlands in the Patagonian Sierras and the Patagonides, Argentina, at about 1000 m in the Sierra de Tecka and in San Bemardo for example.At the same time they mentioned lobes of blocks even with sorting for these elevations at the Patagonian plateaus (mesetas), and pattemed ground at the Meseta de Ia Muerte.
Block tongues that move downhill by the effect of solifluction characterize the periglacial Andean leveI at different latitudes and elevations.Rock streams or stoneruns are a particular type of block tongues.
The term kurum used in the Transbaikal (ROMANOVSKIY et ai.1989) applies to rock streams ar stoneruns whenever well-defined flows and detrital slopes or Schutthange may be identified.These cover large extensions of the slope surfaces.These are regional forms and they depend on altitude and the humidity regime in the active layer, in other words on cycles of freezing and thawing.They move several mm per year.According to ROMANVOSKlY et aI.
(1989), they may be classified into: a) those created by basement rocks and b) those created by sedimentary sephitic rocks.
Wanderblocks ar ploughing blocks are blocks that had been uplifted by the ice which slowly moves downhill due to solifluction.Sediments cover the side that is pushed.
They are found in theAndean Cordillera even in areas with seasonal freezing.
BIGARELLA et a!.(1969) found out that certain fe1senmeer forms on the P1ateau of Itatiaia, Brazi1 wou1d correspond to a semiarid and co1d episode of the P1eistocene.Other authors described these as a resu1t of "cryoturbations" together with the shaping of pediments (see BIGARELLA et a!.1969).These relict forms ofItatiaia however rather seem to be re1ated with "bou1der streams" in the sense of MODENESI (1992) and CLAPPERTON (1993) and the indirectly invo1ved perig1acia1 process wou1d correspond to the solifluction of subjacent sedimentary 1eve1s.This former fact is very important and wou1d have to be analyzed with more detail, as it might represent the most northem pa1eoclimatic limit of perig1acia1 phenomena outside the Andes.

Nivation hollows
Whether snow patches really react as erosive agents in cryogenic environments and at the snow-covered ground in mountain areas, without the participation of other factors, is questioned by HALL (1998).Another possibi1ity to consider in re1ation with Andean nivation hollows is that they can indicate climate fluctuations as well as climatic changes.
Seasonal or perennial snow patches are characteristic of the Andean 1andscapes (Figs. 13  and 17).Snow patches seem to be strong1y influenced by the E1 Nino / Southem Oscillation or ENSO phenomena and by snowfalls, in the Cordillera as well as by the observed processes of a warming up climate.The 10wer limit of perennia1 snow patches in the Andes of Mendoza (33°S) 1ies at 4300 m on the south-facing slopes.In the Southem Andes (38°S) however, these snow patches may be found in Argentina be10w e1evations of 2200 m (GARLEFF 1977).The pressure and weight of the snow causes krummholz in Araucaria trees and Nothofagus that grow close to the tree1ine.

Cryop1anation surfaces
The term cryoplanation, or EAKIN (1916) altiplanation, is used to exp1ain the combination of phenomena, which create a cryogenic environrnent of p1anation, cryop1anation ar a p1ain surface on the summits of high mountain areas; cal1ed summit surface in this 1ast case, or a terrace in valleys and on slopes, called cryoplanation terraces (Fig. 17).They may reach a width of 250 m and a 1ength of 1 km, with a Quatemary cover of a thickness of up to 3 m (CZUDEK 1989).
Among the most important processes are: a) the activity of snow patches or niva1 phenomena, b) cryoweathering, and c) solifluction, reinforcing the remova1 of the cryosediment (DEMEK 1969).
Accurding to Suchodrovskiy's theory (1967( ,1979( , see CZUDEK 1989) ) there wou1d be no general change caused by cryogenesis on the slope, but rather an overprint processo This makes us think of a pre-existing p1anation which for geo10gica1 reasons increases the erosion action, whi1e the summit surface wou1d represent a final state in a cycle of erosive states (CZUDEK 1989).
It cannot be ruled out that a surface of a cryogenic environment has a polygenetic origino Summit cryoplanation surfaces are closely related to Andean continuous permafrost.In the Andes of San Juan, Argentina (30°S) cryoplanation surfaces were taken as indicators of continuous permafrost, with temperatures < -4°C, at elevations over 5000 m by SCHOLL (1992).In the Southem Andes they appear as low as 1590 m (approximately 46°S) in the area of Cerro Ap Iwan (GARLEFF 1977).Cryoplanation surfaces between slopes may favour the accumulation of cryoregolithe as a cryobasin (Fig. 16), as in the cases of Mendoza, at almost  4000 m (TROMBOTTO et aI. 1984).

Tor (felsburg) (HAMELIN & COOK, 1967)
These are areas of incomplete cryoweathering closely linked with the structures of the rocks and in particular with their degree of fissility .According to the classification by FRENCH (1988) the following tors may be distinguished: 1) slope tors: located on the valIey sides and surrounded by slopes with a detrital cover, with inclination angles between 20°and 30°, depending on the type of rock and 2) apical tors.The term Chalt (in aoniken, a Patagonian native language) is used to describe a group of tors (Fig. 4) with a shape of saw-tooth surface (R. Casamiquela, pers. comm., 1993), an example of this is the Cerro Fitz Roy.

Cryogenic slopes
Due to the Andean periglacial morphodynamics the folIowing are the prevailing geoforrns on slopes: • Detrital slopes (Schutthaenge) of periglacial origin (Fig. 13) are of enorrnous hydrological importance for theAndean oasis (TROMBOTTO 1991).These are slopes built by rocks and blocks of considerable thickness and with an important percentage of embedded fine material or cryoregolithe, displaying creep or even other attached geoforrns.These slopes may contain interstitial ice or "cement ice", that cements rocks together and is of important hydrological value in the summer season.The granitic detrital slopes in the Lagunita deI Plata, Mendoza, display inclination angles of predominantly 29°to 34°, while those of the greywacke vary between 24°and 31°.
• Richter denudation slopes are wide and gentle slopes (Fig. 17) with a relatively thin layer of cryofragmented rocks.Also glatthang (from German Glatthang, pl.Glatthaenge).These are usualIy cut by the scars of snow avalanches and snow -debris avalanches and mudflows, which are characteristic of semi-arid periglacial environments in the Central Andes.GARLEFF (1977) described Glatthaenge for theAndes of San Juan, ValIe de Agua Negra, above 4500 m and to the south.They have been reported in different parts of Mendoza, as for example in the area of the Laguna deI Diamante (>3300 m), where STINGL & GARLEFF (1983) mapped these slopes.They studied the cryodynamics in the region of Cuyo and part of the Northwest region (NOA) in Argentina in search of the generating factor for this kind of slopes in this semiarid mountain regions.The same authors emphasize the importance of the petrology of the Nevados of Famatina (29°S), with schists and quartzites, that help to form these slopes on the periglaciallevel above 4000 m. ln Chile on the border with Perú, ABELE (1982) analysed these phenomena and distinguished a type of gently inclined slope associated with the cycles of freezing and thawing of another slope in the area of neblina (a misty altitudinallevel), where the effect of the growth and hydration of the salt is observed.
• Talus or talus cones (Fig. 18).GeneralIy great accumulations of angular shaped rocks and cryoregolithe of different sizes may be observed on the sides and feet of the mountains in cryogenic environments.These are created by the dynamics of freezing, weathering and the contribution of nivodetritic avalanches and are denominated talus.They belong to the phase that FRANCOU (1982) calIed d' eboulisation, that is the disintegration caused by primary chutes or primary rockfalIs.Many different types and combinations of this phenomenon have been classified so far, according to the changes in their genesis and according to their respective shapes: rockfall talus (for the falI and accumulation of cryofragments), alluvial talus, avalanche talus, etc. (RAPP 1959, WHITE 1981).FRANCOU (1984) points out that in Huamparcocha, Peru (12°S), above 4900 m, the origin of the talus, apart from the exposition, is influenced by the petrographic characteristics of andesites and ryolithes.These rocks, which in that region are also affected by intrusive igneous material, are altered by tectonics and hydortherrnalism (epidotisation), and seem to be FIGURE 18-Protalus and talus roek glaeier at Strelkov glaeier valley, Cerro El Plata (6310 mASL), Cordillera Frontal, Mendoza, Argentina (Mareh 1996).
easily attacked by cryoweathering.Similar conditions, with intrusives, tectonics and metamorphism were also determined as important factors for the genesis of periglacial forms at the Lagunita deI Plata, Mendoza (TROMBOTTO 1991).
• Protalus ramparts.Protalus ramparts are particular sedimentary forms associated with snow patches in postfrontal depressions, attached to the lateral mountain slopes on which rocks slide down and accumulate at their fronts.In the Central Andes they are very well represented and are often interpreted as embryonic states of a type of rock glacier.
A transversal cut into periglacial slopes and valleys may frequently reveal grezes litées (Fig. 19) as in theAndes ofMendoza.Their genesis needs to be studied more thoroughly.They consist of altering sedimentary layers, granulometrically different from one another.The main axes of their clasts are orientated slope downwards.TROMBOTTO (1991) presented supranival and ennival (from the upper part or from the interior of a snow patch) as possible genesis of some structures on the periglacial slopes of the Andes of Mendoza.
Debris slopes and relict grezes litées are abundant in the Cuyo region of Argentina, below and c10se to the present periglaciallevel, but would still have to be mapped in detail.
Andes.For decades their enormous hydrological value for the Central Andes has been pointed out (CORTE 1976, BUK 1983).
Through the open system of the active layer a water storage is produced, which maintains the geocryological balance.The solid precipitation (snow) that penetrates the active layer and the freezing of the latter creates the system of water storage in high mountains areas.In the summer season the active layer is thawing and the discharge of the rivers increases.
Thawing, summer snowfalls and the lowering of the permafrost table to greater depths, caused by the warming process of the last decades, establish a direct relation between the behaviour of temperatures on rock glacier surfaces and the discharge of the Andean rivers.Frozen areas, with permafrost or debriscovered ice in the Central Andes, as well as in other SouthAmerican cryogenic regions constitute therefore more important water suppJies than glacier regions.
In the tropical Andes rock glaciers appear from 20°S onwards, associated with the periglacial levei located at 4500 m and above (FRANCOU 1984).Francou and his colleagues have been studying the rock glacier Cerro Caquella in Bolivia as indicator of climate variability (1999).In Argentina the first reference of rock glaciers would be the publication by CATALANO (1926), who called them litoglaciares and describes them for the Puna region.
The glaciers of the Central Andes often culminate in detrital or morainic tongues that actually represent debris-covered glaciers.Following the c1assification by BARSCH (1969) these forms were designated in general debris rock glaciers (Figs. 20 and 21) in order to distinguish them from the ones of exc1usively cryogenic origin or the so-called talus rock glaciers (Fig. 18).CORTE (1987) later presented a very complex taxonomy incIuding other geoforms.The embryonic forms of talus and cryogenic slopes (Fig. 18) fed with material from nivodetritic avalanches and indications of movement were termed protalus or embryonic rock glaciers (TROMBOTTO 1991) In the Central Andes and down to 3200 m the rock glaciers may have ice and be inactive, but below 3000 m they are considered fossil and dose to this elevation they were only active during the Würm (BARSCH & HAPPOLDT 1985).
It has to be mentioned again that the air temperature is usually in certain disharmony with the soil temperature profile, resulting from a very particular thermal balance according to the radiation, orientation, constitution of the rocks, and origin of the ice being among the most important factors.
High surface temperatures, between air and soil, as well as their evolution with depth, were registered in the periglaciallevel of Mendoza andSan Juan, Central Andes, Argentina. They surpass 30°C (TROMBOTTO 1991, SCHROTT 1994).At the same time a correlation between global radiation and daily soil temperature variations was observed to depths of 1 m (HAPPOLDT & SCHROTT 1989).Adetailed study by SCHROTT (1994) sustains that the appearance of permafrost is mainly related to the.energetic balance determined by high solar radiation (20,6 and 22,3 MJ m2 d•I).A shadowed spot with extremely low radiation would explain atypical permafrost (e.g. at low elevations, SCHROTT 1999).
In the northern Chilean Andes, and within the region of the South American Arid Diagonal between 24°and 27°S, Swiss colleagues (KAMMER et aI.1999) mapped rock glaciers with great precision resulting in diagnostic parameters for their activity and defined their disappearance as active geoforms at a precipitation regime of :S;175 mm.
Rock glaciers in the Central Andes, Argentina, show signs of activity even at MAAT of O°C at elevations of 3700-3800 m and mean annual precipitations around 500 -600 mm.
The rock glacier El Salto for example has permafrost at an elevation of 3600 m (BARSCH & KING 1989) and at Morenas Coloradas, a rock glacier of glacigenic origin, the permafrost limit reaches down to even 3400 m.These last conditions seem to be related to the dry cIimate and the global energetic balance of periglacial environments in the Central Andes, factors which increase the efficiency and the radius of influence of cryogenic phenomena to lower elevations.
Dryness, the global energy balal}ce and regional atypical phenomena coincide with what LLIBOUTRY (1986) observed for the major rock glaciers in Chile, towards 33°S with signs of activity down to 3500 m and temperatures of 1 to 2°C.
Among the few existing measurements of rock glacier movement in•South America, there are the ones made by MARANGUNIC (1976) who calculated the rate on the margins of the rock glacier El Pedregoso, in Chile, at 3700 m at approximately 32°S, to be llcm / year.This rock glacier has a surface of 0,3 km2 and a summer discharge of 4-9 l/sec.
The case of the Dos Lenguas rock glacier in San Juan, Argentina, at 30°S is similar.It has a similar surface and a discharge of 5-8 l/sec.(SCHROTT 1994).Schrott thinks that the upper basin of Agua Negra, where the Dos Lenguas is located, with 2 km2 of rock glaciers in total, may oause a discharge of up to 50 l/seco On the other hand the Morenas Coloradas basin with perrnafrost and rock glaciers generates the discharge of the Vallecitos river with 505 1/ sec., which is of vital importance to water supply of Mendoza.This study of periglacial hydrology is of strategic importance for the near future.

Asymmetrical valleys
Another typical characteristic of periglacial environments are the asymmetrical valleys.In the Southem Hemisphere north-facing periglacial slopes are warmer and less steep with a major frequency of freezing and thawing cycles and more solifluction than the south-facing slopes.The south-facing slopes are colder, steeper and have rougher surfaces.
Periglacial saucer-shaped valleys, dells, many of which are asymmetrical, are characteristic geoforms of an Andean periglacial environment.They may also be identified by their covering with solifluction sediments (see SEMMEL 1985).
GARLEFF et ai.(1989) made a detailed inventory of fossil dells for the geomorphological map: LaJunta-AguaNueva, where they interpreted morphodynamic and paleoclimatic changes in the area bordering the Andes (350 S).

PERIGLACIAL PRESENCE ON THE PEATLANDS
Many of the Andean peatlands of southem South America represent the environment of a moa r or anmoor, depending on their content of organic material.In the Central Andes the peatlands usually have a genesis similar to that of a minerotrophic feno The inhabitants of such peatland environments (Fig. 22) in certain Andean subtropical regions and Patagonian call their habitat mallín or vega (anmoors).
These mallines (plural) are also designated as humid or swampy prairies (BOELCKE 1957).They frequently border the basalt plateaus of Patagonia.These mallines or moors are of great importance in Geocryology as far as their present seasonal freezing is concemed and because many of these environments are remnant ecosystems that were created as the Pleistocene ice retreated.They are also old areas associated with paleopermafrost e.g. with a paleoperiglacial environment.An environment of the fen -ar minerotrophic peatland type ~is produced by the existence of an aquatic environment, for example close to mountain springs, streamlets or riverbeds ariginated by the thawing of the altitudinal cryogenic leveI soil, when the water has difficulties to percolate or to drain as it encounters imperrneable sedimentary beds (e.g.glacial sediments, volcanic ashes etc.).Geomorphologically speaking, two main types of anmoors or mallines are found: one in the valleys and one on mountain slopes.They occur at the leveI thatAMBROSETTI et ai. (1986) called altoandean levei, or Andean tundra (see TROMBOTTO 1991).
These environments favour the growth of diverse vegetation, saturated with water and accumulated vertically, as the layers are not completely degraded, thus creating peatlands or very important organic soils in this way.The cold temperatures have an essential role in slowing down decomposition processes of the organic material.Seasonal cryogenesis acts as a generator of thufurlike seasonal freezing forms, in the environments of the Central Andes as well as in the environments of the Argentine-Chilean South (e.g.continental peatlands with the genus Carex and Caltha, ROlG et ai. 1985).Thufurs (Fig. 23), which often appear in association with the periglacial phenomena, have been widely ignored by researchers.
In the area of Última Esperanza in the Magellan region in the south of Chile and in TieITa deI Fuego and the Andean woods in Argentina, the moors however, are closely linked with high precipitations and with characteristics cOITesponding to ombrotrophic moors.ROlVAINEN (1954) (1980,1997) ana1ysed thermokarsts (Fig. 24) or cavities caused by subsidence or thawing.Here the ice began to disappear and the isotherm of 0°C raised, or facies of thermokarst, in debris rock glaciers in Mendoza.A degradation phase was suggested by FRANCOU et ai.(1999) for the tropical rock glacier Caquella in Bolivia based on low apparent resistivity in the geoe1ectric soundings.
This topic is a1most unexplored in South America.According to TROMBOTTO et ai.(1997, 1999;compare BARSCH & KING 1989) not only has the active layer deepened in the last 20 years, but also the isotherm of 0°C ascended altitudinally up to 3860 m at 33°S in the last decade.A thermokarst area of the rock glacier Morenas Coloradas, Mendoza (at 4000 -41 00 m) does not reveal any visible permafrost occurrence but does reveal the original forms that today appear dry or with lakes which freeze or thaw according to meteorologica1 annual variations.Another phenomenon revealing permafrost degradation is the occurrence of mudflows, which represent sudden movements of soil, a form of natural disasters.

PAST CRYOGENIC ENVIRONMENT AND PALEOPERMAFROST
Outstanding fóssil cryogenic structures in southern South America during the Last Glacial (LG) and according to various authors may be observed on the accompanying map (Fig. 25).The influence of the periglacial environment during the Pleistocene seems to have reached even into Brazil, participating in the evolution of slopes and leaving behind relict talus, signs of solifluction and freezing.No signs of permafrost are observed, as at the p1ateau of ltiatiaia fnOS), a region with mountains over 2700 m (MODENESI 1992, MODENESI-GAUTTIERI & TOLEDO 1996, MODENESI-GAUTTIERI & NUNES 1998).KEIDEL (1922) described debris slopes and summit peniplains in the Puna region (Argentina) and made first assumptions about c01d episodes in the past.CZAJKA (1955), later stated that the rock glaciers reached down to approximately 2600 m in the Nevados deI Aconquija (NW Argentina) at approximately 27°S and 66°W.The same author (1957) also reported fossil rock glaciers in the Sierra de Ia Ventana (Province of Buenos Aires), at 38°S and 62°W and made one of the first paleoclimatic deductions about cryogenic forms and phenomena in Patagonia.He considered a sequence of climatic denudation cycles, with tundra -10ess phase during the than 50% of today's precipitation in the second case.These conditions must have considerably affected and stopped the cryogenic activity in the South American Arid DiagonaI.
The Patagonian epigenetic ice wedge casts were mentioned by AUER (1956AUER ( , 1970)), CZAJKA (1955CZAJKA ( , 1957) ) and CORTE (1967,1997).Syngenetic ice wedge casts were recently mentioned by TROMBOTTO (1996).Based on the epigenetic and syngenetic ice wedge casts the extent of permafrost during the Pleistocene Age up to the rivers Río Negro and Río Colorado could be estimated.The difference of the MAAT between present temperatures and those during the LG is about 140 C for South Patagonia (TROMBOTTO 1994).
The Falkland Islands express the characteristics of the past periglacial maritime climate.On the Falkland Islands ANDERSSON (1906) was one of the first scientists to mention solifluction, introducing the term intemationally.Fossil periglacial geomorphology of the maritime type was analysed by CLARK (1972).The islands offer a great variety of cryogenic fossil forms: cryoplanation surfaces, tors, nivation hollows, ice wedge casts, felsenmeer and the famous stoneruns or blockstroeme.
to explain the origin and the deposition of saridstone and loess sediments of the Pampean region up to the south of Brazil during the dry episodes from IS4 onwards, considered to be the coldest episode of the last cycle (?).It still is difficult to trace back the provenance of the silt, and the techniques for the exact identification of its genesis are somewhat limited.
).Recently this last theory was presented again by IRIONDO & KROHLING (1995).They used this theory