The Birth and Death of Glaciers Glaciers are the products of climate, and they are born only in places where the climate and various other factors are found in just the right combination. Glaciers do not form on the lowlands of the middle latitudes, but they do form and thrive in mid-latitude hghlands such as the Alps and Himalayas. They exist close to sea-level on some islands of the Canadian Arctic and on certain sub-Antarctic islands. But not all of these islands support glaciers, and there are vast areas on the very edges of the Ardic Ocean which are thought never to have been glaciated. Clearly glaciers do not simply prefer high-altitude and high-latitude areas. Other factors also play an important part, and before we look at how glaciers grow and behave it will be worthwhile to spend a few moments considering the situations in whch they are born. BIRTH IN THE MOUNTAINS Most glaciers are born in the mountains because here conditions are especially suitable for snow to collect and survive from one year to the next. Once snow has collected and survived the summer melting season it will accumulate and a snowpatch will begin to grow. The rate of accumulation, and the rate at which a snowpatch grows into a glacier, depends upon the interaction of many different factors, including temperature, latitude, altitude and aspect. Snowfall exerts a very potent control over the occurrence of glaciers. In general the areas which receive the greatest amounts of snowfall are very favourable for the birth or growth of glaciers, but it should be remembered that a high snowfall at a locality is of little use if it is all melted away during the summer. Many parts of the Alps, the Scottish mountains and the Rockies receive high snowfall totals every year, but this snowfall is inadequate for the birth of glaciers Page 1 because of the typically warm summers experienced by those areas. In other words, the snow which falls is "ineffective" from a glaciological point of view. On the other hand if an area is high enough to reduce summer temperatures, for example in tropical mountains, high snowfalls may be "effective" and glaciers are formed. In some polar areas summer temperatures seldom rise above freezing, and here quite small snowfall amounts are sufficient for glacier birth and growth. For example, central Antarctica and the northern part of the Greenland ice sheet are more arid than many of the world's subtropical desert areas, but the occasional snowfall is still quite adequate to maintain the ice sheet surface. Highland icefields and valley glaciers flowing down towards the ice-covered Hare Fjord, Ellesmere Island, Canada. (Source: Canadian Dept of Energy, Mines and Resources) From the foregoing it will be clear that glaciers need relatively low temperatures in order to survive and thrive. Many glaciers exist in areas where mean annual air temperatures are as low as -10 degrees C or even -50 degrees C, but mean annual air temperature is probably not all that important; after all, so long as the air temperature is below zero snow can accumulate, and so long as air Page 2 temperature is above zero melting can occur. The most critical factor therefore seems to be the mean air temperature during the summer months, and so long as this is not much above zero glacier survival is possible. Where summer temperatures are substantially above zero glaciers can only survive where there is more winter snowfall than can be melted away. It might be supposed that most glaciers exist close to the poles. T h s is not strictly so. True, the regional snowline (i.e. the line at whch the amount of snow falling is equal to the amount of summer melting) does fall gradually from the tropics towards the poles, and it is true that the polar lands support more glaciers than the tropics or sub-tropics. But we have already mentioned the fact that many polar areas are extremely arid, and some authorities believe that the more humid middle latitudes (especially around 50 degrees - 60 degrees N and S) are more suitable for the birth of glaciers than the Arctic and Antarctic. Altitude has an important effect upon the location of glaciers, and most of the smaller glaciers of today are mountain glaciers. They occupy collecting grounds in the hghlands, and they often flow down spectacular and beautiful mountain valleys. The glaciers of the tropics, in particular, would not exist without the presence of high mountains which have the effect of lowering local air temperatures and ensuring that most of the annual precipitation falls as snow rather than rain. For example, Mexico has only three mountains which support glaciers, and these are all above 5000 m. In East Africa there are glaciers on Mount Kenya, Mount Kilimanjaro and in the Ruwenzori, but none of the ice extends below an altitude of 4400 metres. The shape of the land surface (or its relief) determines the precise location of glaciers withn those areas which are climatically suitable. Hollows or depressions in the mountains are particularly suitable for the accumulation of snow, for they collect not only "direct"precipitation but also windblown snow from exposed plateau surfaces and mountain sides. If the slopes around a hollow are steep enough more snow may be added by the mechanism of avalanching, which is a perfectly normal process in all steep mountain country where snowfall totals are high. The highest peaks in a particular upland area may never support glaciers because their slopes are too steep, but gentler mountain summits perhaps 300 m lower in altitude Page 3 may be able to retain the snow which falls, causing glaciers to form. Certain types of glaciers can only occur where there are certain types of land surface w h c h favour their formation. Cirque glaciers and mountain valley glaciers are formed in steeply undulating mountain country, but ice caps normally form in gently undulating plateau areas, as for example in parts of Norway and Iceland. Quite closely connected with relief is the factor of aspect or orientation. Those of us who live in the northern hemisphere know that south-facing slopes tend to be sunny and warm, w h l e northfacing slopes tend to be shady and cool. It follows that glaciers, which are products of cold climates, will tend to concentrate on the northern sides of mountains wherever possible. Another factor is prevailing wind direction. Since many of the northern hemisphere glaciated areas experience prevailing south-westerlies, it turns out that most snow collects on the lee side of mountains and in hollows whch face north-east. Hence many of the small glaciers of the midlatitude hghlands have this orientation. A final factor in glacier formation is distance from the nearest ocean. Most people are familiar with the idea that as one travels inland from the coast, rainfall (or snowfall) totals tend to diminish. Thus, all other things being equal, glaciers have a greater chance of survival close to a coast where there is adequate moisture. Further inland only the highest mountain areas will receive heavy snow-falls. If the prevailing winds have to cross mountain barriers close to the coast they may lose most of their moisture before reachng the inland mountains, and the latter may remain quite arid (and unglaciated) because of the rain-shadow effect. From the foregoing we can see that the birth of glaciers depends on many different factors that interact with one another to create conditions suitable for glaciation. Each glacier takes so long to form that a study of its present-day environment gives us only part of the answer to our question "Why is it there?" Really we have to know about the glacier's past environment as well; we now know that small glaciers may exist today because of the climatic conditions which prevailed hundreds of years ago, while large glaciers may exist because there happened to be suitable climatic conditions several thousands of years ago. Page 4 SNOW INTO ICE Glaciers grow through the accumulation of snow and ice to considerable thicknesses. As mentioned above, most of the snow which feeds glaciers comes from direct precipitation, wind drifting (called deflation), and avalanching from steep mountain slopes. It is worth looking briefly at the process by which snow is converted into ice, and also at the characteristics of the ice which makes up the bulk of a glacier. (dl Compact granular rmw firn or ndvd (f * 0 0 0 0 0 0 How snow crystals are turned into glacier ice as they are buried deeper and deeper byfieshfalls ofsnow. This process may take more than 5000 years in Antarctica, but on mid-latitude glaciers a decade may be sufficient. When fresh snowflakes fall onto a glacier surface they have a delicate and beautiful hexagonal form. Snow crystal forms come in an almost infinite variety, and their complexity and size varies according to the temperature and other conditions prevailing at the time of the snowfall. The largest and most delicate crystals fall during calm weather when the air temperature is close to zero. Under Page 5 colder conditions and when there is a strong wind blowing, crystals tend to be smaller and more granular, partly because they are damaged in collisions with one another. Once on the glacier surface the crystals undergo a gradual conversion to glacier ice. As soon as they are buried by other snowflakes they lose some of their delicate tracery and become simpler in outline. Eventually the six "arms" are lost by melting and the snowflakes become snow granules. These may be less than 1 mm in diameter. As more snow layers build up on the glacier surface the pressure on the buried granules increases. Most air is expelled and the granules are packed closer and closer together. At sub-zero temperatures the changes in granule shapes occur simply because of pressure and physical damage. Where it is warmer, melting and refreezing occurs. There is more melting at the points of contact between granules, and the water so produced is recrystallised, causing them to grow. When this process has continued for a year or more the enlarged crystals are called firn (neve in French). By now the crystals are several millimetres in diameter, and there is very little air trapped between them. As they continue to grow they become welded together and impermeable to air or infiltrating meltwater; the trapped air from the original snow-pack can only exist as small air bubbles in the ice. This is now true glacier ice, and the individual crystals sometimes reach the diameter of a football. Quite commonly the thick annual snow layers are transformed into narrow ice layers (or foliations) deep withn the glacier. Not all of the ice which makes u p the bulk of a glacier comes from this slow process of conversion. In cool humid environments such as the Antarctic Peninsula, the sub-Antarctic islands and even western Iceland, glaciers may receive great additions of a special type of ice called rime ice. This is formed by the direct freezing of water droplets from the atmosphere when they come into contact with a cold surface. Generally rime ice forms in great cauliflower-shaped masses on rock summits and valley sides, but when these masses break off and fall down they help to build up the glacier surfaces below. Another type of ice is called superimposed ice; this is formed at relatively high temperatures (above or at freezing-point) when meltwater from the glacier surface infiltrates into the snow and firn beneath and freezes when it reaches the underlying ice. Some ice caps in the Canadian Arctic are made almost entirely of superimposed ice, whch forms in a new layer every year. Page 6 Foliated or banded ice exposed in the snout of the Cook Glacier, South Georgia. The bands are mostly -- but not a l w a y s -- annual accumulation layers. (Source: C. Clapperton) We should not be too earth-bound in our ideas concerning the formation of ice. There is perfectly respectable ice on the planet Mars, differing mainly in that it is made of solid carbon dioxide. It Page 7 blankets large parts of the Martian polar areas, and glacial astronomers (or should they be called "astronomical glaciologists"?) are now studying the characteristics of the two ice caps. These ice caps show up remarkably clearly on the photographs transmitted back to Earth from spacecraft. On earth, the time taken for the conversion of fresh snow to glacier ice varies greatly from place to place. On some mid-latitude glaciers where annual snowfall totals may be as high as 5-10 m, the process may take less than five years. On the cold glacier surfaces of the Greenland and Antarctic ice sheets the process may take more than 5000 years. Also the depth of the firn-ice transformation varies greatly, from 20 m or less on mid-latitude glaciers to 60 m or more on true polar glaciers. These differences help us to understand the ways in which different types of glacier behave, and they also provide us with guidelines to the dynamism or activity of glaciers in different environments. Clearly those glaciers which have the highest rates of accumulation and which convert snow to glacier ice at the greatest speed need to evacuate this surplus material at the highest possible rates if they are to remain in any sort of equilibrium with their environment. This means that the ice has to move away downslope relatively rapidly. Conversely glaciers whose surfaces build up but slowly tend to behave in a sluggsh fashion. GLACIERS LARGE AND SMALL It follows that if there are many different types of glacier ice, so there must be many different types of glacier. Glaciologists have spent a great deal of time over the years in attempting to classify the subjects of their study, and for our part we cannot understand the events of the Ice Age unless we know a little about the types of glaciers which have come and gone over the centuries. One of the most popular ways of distinguishing glaciers has been through the use of their temperature characteristics. It has been realised for many years that the temperature of glacier ice about 10 metres below the surface can give a guide to the mean annual air temperature of a locality. For example, the mean annual air temperature at the South Pole is about -51 degrees C, and this figure Page 8 was arrived at originally not by prolonged meteorological observations but by measuring the "10 m ice temperature" in a deep pit. Ice which is at such low temperatures tends to behave in a different way from ice in temperate latitudes, and glaciologists commonly refer to the two types of ice as cold ice and temperate ice. The major difference between these is explaii~edby reference to the pressure melting point of ice. The melting point of snow and ice is zero centigrade at sea-level, but on glaciers it varies with altitude and with depth beneath the surface. In general, the thicker the ice the lower the pressure melting point, and beneath thick glaciers such as the Antarctic ice-sheet meltwater can exist at a depth of 2500m around -1.6 degrees C. An image of the north polar ice cap of Mars, taken on the Mariner 9 voyage in 1972. The ice cap is made of carbon dioxide ice. (Source: NASA) Ice which is always below the pressure melting point is referred to as cold ice, while ice which is sufficiently close to the pressure melting point to contain water for some of the time is called temperate ice. The ice in the centre of the Greenland ice sheet is cold; at a depth of 500 m it has a temperature of about -23 degrees C, but although the temperature rises with increasing depth it is still no Page 9 hgher that - 10 degrees C when the ice sheet bed is reached at a depth of c.1300 m. Many of the Alpine and North American glaciers are made largely of temperate ice,allowing a layer of water to exist on the glacier bed throughout most of the year and allowing substantial amounts of water to exist in the body of the glacier itself. A vertical aerial photograph of a polar glacier snout on Ellesmere Island, Canada. This is the Lower Gilrnan Glacier, near Lake Hazen. (Source: Royal Canadian Air Force) It is no easy matter to define a glacier as simply "cold1'or "temperate" for it may be made u p of different types of ice in its upper and lower parts. In Antarctica the ice is cold close to the surface and temperate close to the bed; in Svalbard glaciers are warm beneath the collecting grounds and cold close to their snouts. It is therefore wiser to refer to the different types of glacier by using labels Page 1 0 which describe their location in terms of their latitude. Quite commonly nowadays glaciologists talk about temperate, sub-polar and polar glaciers; these terms tell us something about the temperature characteristics of the glacier ice involved in each case, and they are less open to misinterpretation. Another, quite different, way of referring to glaciers is through the use of descriptive terms whch tell us somethng about their shape or size. Using these criteria the simplest scheme is one in which we recognize only two different types of glacier: those w h c h assume their forms and dimensions largely independently of the shape of the underlying bedrock topography, and those which are controlled or constrained by topography. Under the first heading we may include ice caps and ice sheets, whch build themselves up to massive ice domes and which may completely submerge hills and valleys and even complete mountain ranges. Ice caps are usually less than 50,000 sq km in area. Ice sheets are quite different in scale, and the only two surviving at the present day are in Greenland and Antarctica. The highest point on the Antarctic ice sheet is more than 4200 m above sea-level, and there Radio-echo sounding profile of part of the Antarctic ice sheet. The profile shows a complex landscape of mountains, ridges and valleys, with all but the highest summits buried beneath the ice. (Source: Scott Polar Research Institute) are vast areas where the ice is more than 3000 m thck. The so-called East Antarctic section of the ice sheet measures some 4500 km x 2500 km, and the area of the Antarctic ice sheet as a whole is about 12.5 million sq km. Ice sheets and ice caps tend to build themselves up Page 1 1 to a characteristic cross-sectional form w h c h is called an equilibrium profile. The profile has been calculated from theoretical work on the physical properties of ice, and it is matched very closely by the surveyed surface profiles of the Antarctic and Greenland ice sheets and many smaller ice caps. The main features of an equilibrium profile are similar to those of a parabola, where the gradient is steep at the glacier edge and lessens progressively with distance from the margin towards the centre. Glaciers which are constrained by the relief of the land are different in many respects. They are generally much smaller than ice sheets and ice caps, and they d o not have dome-shaped upper surfaces. They occupy bedrock depressions in the mountains, and these act as glacier nourishment areas called icefields. Sometimes these icefields d o submerge bedrock ridges, but always the direction of ice movement is controlled by the slope of the bedrock surface in the valleys between the ridges. A highland icefield w i t h valley glaciers and piedmont lobes, north of Surprise Fjord, Axel Heiberg Island, Canada. (Source: Canadian Dept of Energy, Mines and Resources) Page 1 2 Highland icefields and valley glaciers near the east coast of Greenland. The iceljields "feed" ice via steep and crevassed icefalls into the valley glaciers, which in turn flow towards the lowlands. (Source: Geodetic Institute, Denmark) T w i n valley glaciers on the Blosseville Kyst, Greenland. Strictly these should be called "outlet glaciers" since they are fed from a part of the Greenland ice sheet. (Source: Geodetic Institute, Denmark) Page 13 The glaciers which leave the icefields are called valley glaciers. Exceptionally they may be over 100 km in length, but more commonly they are less than 30 km long. Regardless of their length, they are generally joined by tributary glaciers from upland cirques or ice caps on either side of the glacial trough. Cirque glaciers are small ice masses which occupy armchair-shaped hollows in the mountains. These glaciers, if they are very active, may flow downhill for considerable distances, thus becoming indistinguishable from normal valley glaciers. However, if the glacier snout does not extend beyond its rock basin it can be referred to as a true cirque glacier which is typically very broad in proportion to its length. These small glaciers are found in many mid-latitude upland areas in distinct groups; by their presence they indicate that climatic and other conditions are only just adequate for the continuing existence of glacier ice. GROWTH IN THE LOWLANDS Most glaciers grow in the uplands which give them birth, because here snowfall totals are hghest and summer temperatures are low enough for glacier ice to survive. Those glaciers which reach the lowlands are usually subject to the processes of ablation, which tend to reduce both glacier volumes and dimensions. We will look at these processes in a moment, but before doing so we should consider briefly some exceptional circumstances which can lead to glacier growth at low altitudes and in "unexpected"situations. One peculiar type of glacier is the piedmont lobe, in which a glacier spreads out as a broad, flattish sheet of ice as soon as it is freed from the constraints of its upland valley. Such glaciers are common in Antarctica, North Greenland and Arctic Canada, where they occur in broad valleys which are otherwise free of ice, or else on coastal lowlands. Piedmont lobes in relatively arid environments demonstrate the amazingly viscous character of sub-polar and polar ice. One author has likened the behaviour of the ice to white molasses poured over the landscape, and many glaciology students have enjoyed making models of piedmont glacier lobes by pouring plaster-of-paris down narrow wooden chutes. The Commonwealth Glacier in the McMurdo area of Antarctica is a famous example of a Page 1 4 glacier of this type, and there are many other examples in North Greenland and on Ellesmere Island. Perhaps the most famous of all piedmont lobes is the vast Malaspina Glacier in Alaska. After leaving the mountains the glacier spreads laterally into a lobe about 40 km across, with a surface area of 2400 sq. km. With its great extent and its characteristic pattern of striped moraines it is a prominent feature on satellite photographs oi the Alaskan coast. Temperate ice is even more viscous or deformable than polar and sub-polar ice, but because temperate glacier snouts tend to exist in areas of relatively high melting, spectacular piedmont lobes have little chance of survival. An idea which has caused much controversy among glaciologists and climatologists is the theory that ice sheets and ice caps may sometimes grow outside the more normal process of spreading outwards and downwards from collecting-grounds in the mountains. Long ago it was suggested that the Greenland, Scandinavian and Laurentide (North American) ice sheets grew gradually from glaciers which originated in the mountains and then filled the lowlands, allowing the ice surface to build up over many thousands of years of climatic cooling. The newer theory, loosely and spectacularly labelled the "snowblitz theory", involves the growth of ice sheets or ice caps in the lowlands. The suggested mechanism involves a sequence of cold winters during which the ground surface becomes snow-covered to a considerable thickness. If large expanses of snow survive from one winter to the next, the warming effect of the summer sun will be reduced because large amounts of solar energy will be radiated back into space by the bright snow surface. Thus, over time, the summer melting seasons will become less and less effective and the winter snowfalls more and more effective. This is a situation referred to as "positive feedback". Eventually, after but a few centuries, so much of the surface will be snow-covered that a point of no return is reached, and an ice cap or ice sheet will develop. This is a highly speculative and somewhat alarmist theory, and it is doubtful whether a snowblitz of this type could occur in the middle latitudes. It could, however, occur in high latitude areas such as the Canadian Arctic, and there were signs during the early 1970s of an increased snowcover in certain areas which had the apparent effect of lowering local summer temperatures. Also, it has been suggested that the Laurentide ice sheet grew by the coalescence of a Page 1 5 number of broad snow domes or ice caps which developed on the plateau areas of Baffin Island and Labrador-Ungava. It may well be that if climatic conditions are sensitive enough, glaciers can be born and nurtured in both lowland and upland situations, greatly speeding up the onset of a glacial episode. THE BALANCE EQUATION So far we have only examined the birth and growth of glaciers. We have looked at the ways in whch they form and the situations in which they can best exist. But glaciers cannot increase in size indefinitely, for as they grow bigger they trespass into areas which are environmentally unsuitable. Once this happens they melt away, and the rate of melting conditions glacier size just as much as the rate of snow and ice accumulation. If a glacier maintains its dimensions over the years it is said to be in balance; this means that the amount of material being added to the glacier is balanced by the amount of material being lost. Most of the material which builds up the glacier falls onto its surface on its highest parts in a zone called the accumulation zone. Conversely, most of the material which leaves the glacier is removed from its lower parts. The area of the glacier involved is called the ablation zone. Sometimes the two zones occupy more or less equal areas of the glacier surface, but it is more common for the accumulation zone to be the larger of the two. The figure on the next page explains, in very simplified form, the relations between the two parts of a glacier. Generally accumulation is at its greatest near the head of a glacier and falls off gradually towards the snout. Conversely ablation (or melting) is at its greatest near the snout, and decreases gradually up-glacier. Hence somewhere in the middle of a glacier we should expect a line where accumulation more or less balances ablation. This imaginary line, much beloved of glaciologists, is called the equilibriumline, for here the glacier surface is neither being raised by accumulation nor lowered by ablation. As we can see from the diagram, a glacier whch is more or less in a state of balance receives a "wedge" of snow each year in its accumulation zone and loses a "wedge" of ice each year in its ablation Page 1 6 A model glacier. This diagram shows in a very simplified form how a small glacier gains material in its accumulation zone and loses material in its ablation zone. Because a glacier moves, i t can achieve an approximate state of balance between "input" and "output". zone. The state of balance between the two zones is maintained by the mechanism of ice movement. As ice moves down glacier from the accumulation zone to the ablation zone it ensures that the glacier surface maintains a more or less constant altitude from one year to the next. If a glacier is not in equilibrium because it is gaining more ice than is lost through ablation, it will advance. On the other hand, if the rate of ablation is faster than the rate of accumulation, the glacier must retreat. The rate of advance or retreat will vary according to how strongly negative or positive its annual budget happens to be. In the real world it is often extremely difficult to determine the budget or mass balance of a glacier. On small glaciers the budget may be positive one year and negative the next, but because glaciers take a long time to respond to climatic events the behaviour of the snout is sometimes apparently independent of the short-term changes in snowfall and melting rates. For example, a glacier snout may continue to retreat during several years of positive budget, for the surplus snow and ice whch it receives will take time before it works its way through to the glacier snout. Paae 1 7 GENERAL CLIMATE __+ LOCAL GLACIER CLIMATE GLACIER MASS A N D ENERGY EXCHANGE N E T MASS BALANCE GLACIER RESPONSE ADVANCE OR RETREAT A "flow diagram" showing the relationships between climate and glacier snout behaviour. The "response time" which elapses between an environmental change and a snout advance or retreat varies according to the size and "dynamism" o f a glacier. HOW GLACIERS DIE The removal of material from the ablation zone of a glacier takes place by a variety of different mechanisms. If the lower part of a glacier is located in an area of high summer air temperatures, the effect will be to remove ice from the surface by direct melting. Generally no more than a few metres are removed each year, but some Norwegian, Icelandic and Alaskan glaciers lose up to 15 m of ice fro~n the vicinity of their snouts each summer. Mostly the product of this melting is meltwater, and the work which it can do is often spectacular. In cold and windy environments glaciers may lose a great deal of surface snow by deflation, and on the Antarctic ice sheet snow is constantly shifted from one area to another by high winds, giving rise to characteristic wind-eroded forms called sastrugi. These may be more than a metre in height, and they are notorious for the problems wluch they present to ice sheet travellers. Page 1 8 A "honeycomb" texture on the lower part of a glacier surface. The pits are caused by enhanced ablation around dark-coloured rock fragments and even around grains of sand. If the air is dry the melted ice may be evaporated directly into the atmosphere. If a polar environment is very arid, glacier surfaces may be lowered by evaporation - that is, by the direct conversion of ice into water-vapour. In some areas this mechanism may account for up to 5 per cent of the total annual loss of ice from the ablation zone. Of much greater importance in h g h latitudes is the mechanism of calving. This occurs only at the snouts or on the edges of glaciers which are in contact with lakes or the sea; the most efficient calving of all occurs where glaciers are actually afloat, as in some of the fjords of Greenland and around parts of the Antarctic continent. Most of the ablation around Antarctica is achieved by calving, leading to the presence of vast quantities of icebergs and smaller ice fragments in the surrounding seas. Some of the calving glaciers of Greenland produce more than 100 million tons of icebergs per day. Another type of ablation which is important beneath temperate glaciers is basal melting. Because of the heat transmitted from the centre of the earth the temperature on a glacier bed is raised as a result of the contact between ice and rock. Certain types of rock Paae 1 9 Large "tabular" icebergs being calved from the snout of a n outlet glacier in North- West Greenland. In the photograph w e can also see large amounts of "brash iceff, some old icebergs, and both old and new sea ice. The floating glacier snout is a t the bottom of the photograph. (Source: Geodetic Institute, Copenhagen) Paoe 2 0 transmit more heat than others; for example glacier beds in volcanic districts receive about three times as much heat as glacier beds on the old Precambrian Sheld areas of the world. When t h s heat is added to the heat created by a glacier sliding on its bed, and when we consider that thick glaciers may have their bottom layers close to the pressure melting point in any case, it is not difficult to understand that bottom melting can occur. Beneath some glaciers several centimetres can be removed from the bottom ice layers each year by t h s type of ablation. Chaotic wastage or melting of a glacier snout, Svartisen, Norway. Here the ice i s breaking up into detached blocks, and much of the "dead ice" is covered w i t h rock debris. (Source: M. Alexander) Generally glacier wastage close to the snout is not particularly spectacular, but sometimes catastrophic wastage can occur. If the climate around a glacier snout becomes warmer, or if a glacier advances strongly into an area where it cannot possibly survive, very rapid melting may affect vast areas of the glacier surface. The ice of the ablation zone may be broken u p into separate blocks, deep channels will be cut by meltwater streams, and the glacier surface may be converted into a wilderness of ponds, masses of rock debris, Paae 2 1 The debris-covered lower portion o f t h e Khumbu Glacier, notfar from Mount Everest. Here the disintegrating glacier i s bounded by huge ridges of lateral moraine. (Source: E. Schneider) Rapid melting on the snout of a glacier on Spitzbergen. A s the ice melts, debris slides down the glacier face, creating a striped effect. (Source: E. Kosiba) Page 2 2 hummocks and hollows, lakes and quagmires. This type of catastrophic melting is not by any means restricted to small glaciers, for it is known to have occurred during the decline and death of the Laurentide ice sheet in North America. The mid-latitude ice sheets seem to behave in a consistent way; they grow to vast size, eventually pushing so far southwards that they trespass into areas where the climate is too warm for their continued survival. At the same time they become so large that the snow-bearing winds which nourish them can no longer carry their loads into the ice sheet interiors; in other words, their source of supply is cut off or drastically reduced. When there is a high rate of melting near the ice sheet edge and an inadequate renewal of ice from the interior, ice sheet decline is only a matter of time. Catastrophic melting sets in, and the ice sheet simply disintegrates. ACKNOWLEDGEMENTS I am grateful to the original publishers of "The Ice Age" for allowing the use of a chapter from the book in a new format. I thank the members of the Editorial Board of the Norwegian Glacier Centre for their constructive comments and suggestions during the preparation of this text. I also thank the copyright owners of the illustrations for their original consents. I have been unable to trace the owners of some of the illustrations, and will be grateful for any information which will enable me to make due acknowledgement in any future editions of this booklet. Page 23