The use of fossil periglacial phenomena as an index of former climatic conditions is limited by two basic problems. First, dating periglacial features directly is often difficult, if not impossible; generally they are dated by reference to the deposits within which they are found, thereby obtaining only a maximum age for the features. Secondly, although regions of modern periglacial activity can be circumscribed by particular isotherms, the occurrence of similar activity in the past can only indicate an upper limit to temperatures at the time, not a lower limit (R. Williams, 1975). Thus, in general terms, permafrost today only occurs in areas where the mean annual air temperature is <-2 °C and it is virtually ubiquitous north of the -6 to -8 °C isotherm in the Northern Hemisphere (Ives, 1974). Evidence of more extensive permafrost in the past, however, only demonstrates that temperatures were below these levels, and provides little information on how much lower. Mapping the distribution of relict periglacial features may indicate how far the southernmost boundary of the permafrost zone was displaced, but within this zone only the limiting maximum paleo-temperature estimates are possible. Nevertheless, periglacial features are of particular interest because they provide information about the periods of extreme temperature depression during past glacial episodes. They also provide information about areas close to the ice-sheet margins, for which there are few other sources of proxy paleoclimate data.
It has already been mentioned that permafrost only occurs in areas with mean annual temperatures below a certain level, but permafrost itself may leave no morphological evidence of its former existence. Paleoclimatic inferences can only be based on features which develop in regions of permafrost and disturb the sediments in a characteristic manner. In this way fossil or relict features can be identified and their distribution mapped (Fig. 7.7). The most useful and easily identified features include fossil ice wedges, pingos, sorted polygons, stone stripes, and
periglacial involutions (Washburn, 1979a). The problem is to identify those climatic factors that are necessary for the formation of the features in question; commonly this can only be done in general terms (Table 7.1). Ice wedges, for example, result from thermal contraction at subfreezing temperatures. Winter temperatures of-15 to -20 °C (or less) are required before active frost cracking occurs, but the exact requirements depend on the material being considered. Cracking and ice-wedge formation will occur at higher temperatures in silts and finegrained material than in gravels where mean annual temperatures of -12 °C may be necessary. Furthermore, the amount of snowfall is a significant factor because
TABLE 7.1 Climatic Threshold Values for the Distribution of Periglacial Geomorphic Features (after Karte and Liedtke 1981)
Periglacial geomorphic features^
Climatic threshold values^ MAT(°C)§ MAP(mm)n
1 Periglacial geomorphic features whose formation requires permafrost
1.1 Features connected with continuous permafrost Ice-wedge polygons
Sand-wedge polygons Closed system pingos
1.2 Features connected with discontinuous permafrost Open system pingos
1.3 Features which occur in connection with continuous, discontinuous, and sporadic permafrost
Depergelation forms ("thermokarst" forms, active layer failures, detachment failures, ground ice slumps, permafrost depressions, alas, "baydjarakhs," "dujodas," alas thermokarst valleys, beaded drainage, thaw lakes, oriented lakes, thermo-erosional niches, thermo-abrasional niches, degradation polygons, thermokarst mounds) Seasonal frost mounds (frost blisters, hydrolaccoliths, bugor) Palsas
Features whose formation requires intense seasonally frozen ground but which also occur in connection with permafrost 2.1 Seasonal frost-crack polygons (ground wedges)
Frost mounds (thufurs)
Tundra hummocks (high latitude occurrences) Earth hummocks (high latitude occurrences) Earth hummocks (high altitude occurrences)
Non-sorted circles (mud boils, mud circles)
Sorted circles and stripes (>1 m)
Other climatic indication: rapid temperature drops in early winter
Other important indication: high ground-ice content
Other climatic indications: continental climates with high incoming radiation, sublimation, evaporation and little snowfall
Other climatic indication: mean temperature of coldest month
TABLE 7.1 (Continued)
2.5 Sorted circles and stripes (<1 m)
2.6 Gelisolifluction microforms (lobes, steps,
ploughing blocks) 2.7 Nivation and cryoplanation features
(nivation hollows, cryoplanation terraces, frost-riven cliffs)
3 Features which are linked to diurnally frozen ground and needle ice but which also occur in connection with seasonally frozen ground and permafrost
3.1 Miniature polygons
3.2 Miniature sorted forms and stripes <+1°C
3.3 Microhummocks t For a description of these features, see Washburn (1979b). t The thermal threshold values represent upper limits for development of features. § Mean annual air temperature. H Mean annual precipitation.
snow will insulate the ground surface from the effects of severe cold. This has been demonstrated in many areas where today active ice-wedge formation does not normally occur; where snow is artificially removed (e.g., from roads or airport runways), frost cracks and ice wedges will develop. Paleoclimatic reconstructions based on such phenomena are thus subject to a certain amount of uncertainty, and similar problems have to be faced when dealing with other types of periglacial features. Nevertheless, it is possible to make conservative estimates of temperature change based on the former distribution of different types of periglacial features (Fig. 7.8). Accuracy is really limited by our understanding of the climatic controls on similar contemporary features. From this preliminary map, it would appear that mean annual temperatures in Europe were at least 14-17 °C below recent averages during the maximum phase of the last (Wisconsin/Wiirm/Weichsel) glaciation (Washburn, 1979b). Although many features used to compile the map are not well-dated, and often are simply considered to reflect conditions during the maximum stage of the last glaciation, it is worth considering the point made by Dylik (1975) that maximum temperature depression was generally not coincident with the maximum extent of ice (maximum "glaciation"). Dylik considers the extent of glacier ice to be more of an index of snowfall (i.e., of cold and humid conditions) than simply of low temperatures. Periglacial features may thus achieve their maximum development during periods of minimum temperature prior to the maximum extent of major ice sheets; this may explain the large discrepancy between botanically derived paleotemperatures for the last glacial maximum (sometimes indicating mean annual temperatures only 3-6 °C below those of today) compared with paleotemperature estimates derived from periglacial phenomena.
Where a variety of periglacial features of varying ages can be identified in a limited area, it may be possible to reconstruct paleotemperature through time. This has been attempted by Maarleveld (1976) using observations of relict periglacial features in the Netherlands (Fig. 7.9). Maarleveld associated each type of feature with particular temperature constraints; pingo remnants, for example, indicated maximum mean annual temperatures of -2 °C, whereas "extensive coarse snow meltwater deposits" were indicative of a range in mean annual temperature of -5 to -7 °C. Unfortunately, Maarleveld does not identify which sections of his graph are based on estimates of maximum temperature and which are based on a defined temperature range, so the graph may be more precise in some sections than in others. Nevertheless, as a first approximation to paleotemperature reconstruction through time the results compare well with other proxy data series (Fig. 7.9) and indicate the potential value of periglacial studies for paleoclimatic analysis.
Was this article helpful?