Humlum, Ole: Glacier behaviour and the influence of upper-air conditions during the Little Ice Age in Disko, central West Greenland. Geografisk Tidsskrift 87: 1-12. Copenhagen, June 1987.

An inventory of 205 glaciers and small firn areas in
southwestern Disko, central West Greenland, displaylarge
differences as to the amount of ice recession since
the end of the Little Ice Age. Especially high-situated ice
bodies facing N and NW have experienced a substantial
retreat, while other ice bodies have not. The cause for this
difference is investigated by considering the amount of
vertical equilibrium line displacement since the end of the
Littel Ice Age. From this, it is argued that especially
variations in upper-air wind conditions -as measured at
the 850 mb level- are responsible for the observed
heterogeneous glacial behaviour.

Ole Humlum, ass. prof, Department of Geomorphology, Geographical
Institute, University of Copenhagen, Øster Voldgade
10, DK-1350 Copenhagen K.

Keywords: Glacier behaviour, snow drift, Disko, West Greenland.

The existence and importance of past widespread glaciation were first recognized by Agassiz (1840) and the direct connection between glacier behaviour and climate was refined by Matthes (1942) and Ahlmann (1948, 1953). Later work by Nye (1960, 1963), Meier (1965), and Hoinkes (1968) demonstrates that a definite relationship exists between climate, energy exchange, mass balance, and the dynamic response of a glacier, and that the connecting links between these factors may explain the current behaviour of glaciers in response to the climate of the present and recent past. From this, reconstructions of past climates has been attempted from various indications on former merglacier extensions and snow-lines. Typically, the distribution of moraines and mean cique floor elevations have been the object for investigation (see e.g. Flint 1971, Miller 1961, Porter 1964, Ito and Vorndran 1983).

Palaeoclimatic inferences drawn from geomorphologic data should be based on an understanding of the relationship between climate and glacier fluctuations. Complex interactions exist between prevailing regional and local climatic conditions, regional and local topography, energy exchange at the glacier surface, mass-balance pertubations, and glacial dynamics (Meier 1965, Andrews et. al. 1970). Glacier fluctuations represent integrated responses to these factors, and due to differences as to the above factors, even glaciers within a geographically restricted area may well display different responses on identical overall climatic changes. This obvious represents a potential pitfall for palaeoclimatic studies using geomorphologic data on moraines, cirque floors and former snow lines, especially when the data are obtained from areas at present without glaciers, so that local variations as to glacial dynamics may not be studied.

The purpose of the present paper is to present an inventory on a group of glaciers in Disko, central West Greenland, to all appearances representing a rather homogeneous group, but who nevertheless since the close of the Little Ice Age has displayed a very uneven response to climatic changes. The reason for this heterogeneous behaviour will be discussed. In all likelihood a controlling factor is differences as to altitude above sea level and thereby different influence of upper-air meteorological conditions.

STUDY AREA

The study area is the southwestern part of the island Disko, central West Greenland (about 70°N, Fig. 1). The area consist of two peninsulas, 465 km² and 640 km², situated northwest and south of Diskofjord, respectively, and the 30 km² island Qeqertaq in Diskofjord (Fig. 2). This part of Disko consist of Tertiary plateau basalts, and the landscape is dominated by cirque carved lava plateaus and U-shaped walleys. Extensive top plateaus characterize the mountains in the southern peninsula (Fig. 3), while plateaus are less extensive in the northern peninsula. In the central and northern part of this peninsula the landscape approaches an alpine landscape (Fig. 4). The upper land surface rises from about 500-600 m asl. in the western part of the two peninsulas to above 1050 m asl. in the eastern part of the peninsulas. The island Qeqertaq reaches an altitude of about 600 m asl.

Climatic conditions within the study area can only be inferred in overall outlines, as only one meteorological station is presently being operated on Disko, in the town Godhavn/Qeqertarsuaq at the southern coast (Fig. 1). In Godhavn the mean annual air temperature was -3,5° during the period 1950-69, coolest month was March (-14.8° C), and warmest month July (7.4°C). Mean annual precipitation was 406 mm (water equivalent), of which about 75% usually falls during the period June-December, while the remainig part of the year is comparatively dry. Most precipitation falls in connection with advection of moist, maritime air masses along the Davis Strait from south and southwest. In Godhavn, at sea level, snow is the dominant type of precipitaion from late September to late May. The prevailing wind is from east and northeast, with the exception of the period May-August, where southwesterly and westerly winds dominate. Mean wind speed at sea level has a maximum during autumn and early winter, while it has a minimum in the period February-April. In general, the weather pattern in coastal central West Greenland is dominated by advection of moist air masses from S along the Davis Strait during the summer and autumn, while dry and cold air flowing off the Greenland Ice Sheet dominates the weather pattern during the remaining part of the year.

In Godhavn the sun is below horizon from late November to middle January, and the period with midnight sun is from last half of May to late in July. At summer solstice, solar radiation available daily (in the absence of atmosphere) exceeds the maximum radiation ever available at the equator.

According to inhabitants in Godhavn and local fishermen, also the southwestern part of Disko is characterized by a polar, maritime climate; in general much of the same nature as that experienced at Godhavn. This applies espcially for the exposed outer coast, where the orographic precipitation is at maximum. Moving eastwards along Diskofjord, climate rapidly becomes more continental. This overall picture is supported by the authors personal experience gained during several journeys within the study area in connection with a three-year stay in Godhavn

At present Disko (about 8575 km²) supports an extensive local glaciation. A total of almost 1000 ice caps, valley glaciers, cirque glaciers and isolated firn areas have been mapped by the author from aerial photographs recorded in the years 1953 and 1964. Together, almost 20°/( Ü 610 km2) of the island is oresentlv covered by glaciers, Maps showing the present glaciation level (GL) in Diske have been prepared by the author (Humlum 1985 and 1986).

Within the study area 205 individual glaciers and isolatedfirn
areas presently exist. Figure 2 display the distributionof
these ice bodies as well as the present glaciatior

level in the area. Glaciers and firn areas were mapped in
the field or from the above aerial photographs from 1964.

The largest coherent ice cover within the study area is the 85 km² ice cap Lyngemarkens Iskappe N of Godhavn (Fig. 2). Ten outlet glaciers radiate from this ice body. Almost all other glaciers within the study area are small cirque- or valley glaciers, usually not exeeding a few km² in size. Most glaciers face N or NW, but with the exception of the sector around SSE glaciers are found with almost any aspect (Fig. 5).

Since about 1895-90 AD, glaciers in Disko have been receeding, only interrupted by a small readvance about 1930-35 (according to the authors own lichenometric observations). Today, the terminus of most glaciers within the study area are still melting back, although several glaciers appear to be thickening in their upper reaches.

Talus production is intensive in Disko, and many glaciers therefore carry a substantial supra- and englacial load. Moraines dating from the Little Ice Age (ended about 1900 AD, Weidick 1968) are therefore dominated by talus-like material and are often very conspicious (Fig. 6). Due to the intensive talus production, also rock glaciers are frequent within the area (Humlum 1983).

During an inventory of the Littel Ice Age (LIA) extension of glaciers in Disko, a difference as to the LIA extension of glaciers facing N-NW and otherwise became apparant. Several glaciers facing N or NW (as indicated by the orientation of the equilibrium line) has experienced a considerable retreat since the end of the LIA, while few glaciers facing otherwise then reached significantly beyond their present terminus. Below, the meteorological background for this dissimilarity as to glacial response will be investigated.

RECENT EQUILIBRIUM LINE ALTITUDES

As variations in the equilibrium line altitude is a sensitive measure of mass balance changes for a glacier, the above problem is investigated by considering equilibrium line altitudes within the study area.

No systematic mass balance measurements have been undertaken on glaciers in Disko, and the equilibrium line altitude (ELA) for individual glaciers thus had to be determined by way of an indirect approach. Several techniques are available, and for a comprehensive review the reader is referred to the paper by Gross et al. (1977). In most cases the following visual technique was adopted for estimation of the local ELA: Where possible, the ELA for a glacier considered was determined as the level, below which medial moraines or other englacial debris are cropping out on the glacier surface (Lichtenecker 1938, Visser 1938). As most glaciers within the study area are situated in cirques encompassed by prominent headwalls, usually yielding large amounts of fresh talus, the ELA could be estimated by this procedure at the majority of glaciers. Observations were carried out in the field in a few cases, but most cases were decided by use of aerial photographs recorded shortly before the close of the ablation season 1964 (August 27th). Usually, the first lasting winter snow appears on the glaciers in Disko early September.

The official weather reports from the Danish MeteorologicalInstitute show precipitation at Godhavn in July- August 1964 to be somewhat above normal; 124.9 mm water equivalent against a mean of 98.1 mm for the period 1950-69. The mean air temperature July-August 1964 was 5.5°C, against a mean of 7.3°C for the period 1950-69. Provided the record from Godhavn is fairly representativefor the study area as a whole, any discrepancy

between the current ELA_{mea n} and the ELAi₉₆₄ as determinedby the above procedure is likely to be that the ELA,964 would be somewhat below the current ELA_mear Furthermore, because debris comprising medial moraines and other supraglacial load originally may have been incorporatedin the upper part of the accumulation area, close to the headwall, it will usually crop out on the glacier surface some distance below the equilibrium line. The above visual inspection thus only provides a minimum altitude for the equilibrium line. As most glaciers within the study area however are rather short, less than 2 km. this potential error is considered unlikely to disturb the overall pattern very much.

Summing up, the visual estimation of the ELA_i964 may result in a value somewhat lower than the current ELA_mean This should be carried in mind in what follows. As a firsl approximation, the results of this visual procedure will however be applied in the present context.

For glaciers facing E or W, the equilibrium line os often at higher altitude near the northern flank of the glacier than at the southern flank. This is especially the case where the glacier considered is situated shortly north of a

prominent mountain ridge. In cases like these the arithmetric mean of the two extreme values was adopted as representing the ELA for the glacier. In cases of doubt, the distribution of snow-filled versus empty chutes in the headwall above a glacier was studied in some detail. In my experience from Disko, in general, chutes terminating above the local equilibrium line and facing north, west or east, usually are filled with snow at their lower end at the close of the ablation season, while chutes ending below are not. Finally, if still in doubt, the overall surface form of the glacier was considered. The ELA was then approximated as the level where the surface changed from concave to convex (see also Andrews and Miller 1972).

For five glaciers, neither of the above tecniques appeared adequate. The ELA was then estimated using an assumed steady-state accumulation areas ration (AAR) of 0.65. Several studies indicate that for steady-state glaciers the accumulation area/total area varies between 0.6 and 0.7 (see e.g. Meier and Post 1962, Glen 1963, Grosswald and Kotlyakov 1969, Gross et al. 1977).

In order to obtain supplementary observations on the
current ELA, the aerial photographs were also used to investigatethe

vestigatethedistribution of large snow banks (isolated fira areas), known to or supposed to represent permanent features. As no significant movement is expected to occur within these snow accumulatins, the lower altitude of these features at the end of the ablation season was adoptedas an approximation of the local ELA. Only snow banks/fira areas situated in terrain not due to its topographyexcluding a lower extension than the actual observed - e.g. by way of a steep cliff - were considered. In case of doubt, the presence of a frontal moraine ridge or a bergschrund (indicative of significant deformation) was used to disciminate between true glaciers and permanent snow banks/firn areas.

By the above techniques, the current (1964) local ELA was estimated for 106 glaciers and 93 isolated firn areas within the study area. For 6 firn areas the character of the adjoining terrain made an estimation impossible. The resulting 199 ELA_{I96 4} estimates are plotted in figure sa. Height determinations were mainly derived from unpublished, preliminary 1:100000 map sheets with 50 m contour intervals (Danish Geodetic Institute). Elevations between contours were interpolated and based on the topographic form between contours. In general, ELAs are estimated to be accurate within +/- 25 m. The aspect of the equilibrium line was defined as being parallel to the general glacier surface slope direction at the equilibrium line.

The pattern emerging in figure 5a appears to justify the incorporation of ELA estimates derived from large snow banks/firn areas, as both ELA^er and ELA_firn values fall within the same overall zone of the diagram. Lowest ELA estimate is about 250 m asl. (05°), highest about 930 m (190°). A clear tendency appears to be present: The zone of ELA estimates rises gradually from N and NW towards S and SE. Notwithstanding aspect, maximum ELA estimates are however mostly derived from firn areas and ice caps situated on mountain plateaus (shown as open and filled squares in figure sa, respectively). At these sites terrain surface gradients are small and the ice bodies are virtually unprotected against wind action. If these highlying ELA estimates on firn areas are ignored due to the special topographic setting, the above azimuth/ELA relation becomes even more pronunced. This is emphasized by figure sb, where the two solid lines are envelopes for all observations except for the above exposed firn areas. Within the two solid envelopes, zone »1« contains all ELA estimates derived from ice caps (filled squares). Also ice caps are usually exposed features, and ELA estimates from these features are thus all within the upper part of the overall ELA zone. Below zone »1«, zone »2« represents the observational spread of ELAs derived from glaciers and firn areas in shielded sites - e.g. cirque valleys - only. However, also some ELA values derived from ice bodies in shielded settings are found in the upper zone »1« of the overall ELA zone.

The vertical spread of the ELA zone depicted in figure 5b is assumed to represent the influenze of topoclimate within the study area, whereas the gradational rise of ELA values toward a more southerly or southeasterly aspect is assumed to be caused mainly by the overall difference as regards net radiation as a function of aspect.

LITTLE ICE AGE EQUILIBRIUM LINE ALTITUDES

In order to investigate the background for the above mentioned different glacial behaviour since the end of the Little Ice Age, the distribution of Little Ice Age equilibrium line altitudes (ELA_UA) has been investigated.

The mapping of Little Ice Age ELAs was accomplished by considering the uppermost occurence of Little Ice Age moraines at the individual glaciers. These moraines are usually very conspicious in Disko (see figure 6), and, as a main rule, due to the ice movement direction, they are assumed to be deposited below the contemporary ELA only. The moraines are dated to the Little Ice Age period on account of 1) unpublished photos obtained by professor K.J.V. Steenstrup in 1898 and dr. L. Koch in 1913 (kindly placed at my disposal by dr. A. Weidick, GGU), 2) lichenometry, and 3) weathering of surface boulders on moraine crests. Naturally, not all sites have been visited in the field, but LIA-moraines in Disko are notoriously easy to identify in aerial photos due to their large size and the absence of any significant vegetation.

At 84 glaciers within the study area an estimate of the ELA_{L iA} could be obtained by this procedure, while faulty occurence and/or preservation of moraines made this impossible at the remaining 22 glaciers. Distinct marginal forms are mostly absent adjoining to the small firn areas, and the ELA_UA thus could only be estimated at 14 of these features. In these 14 cases the present small ice bodies have probably seen transitional to real glaciers during the Little Ice Age, as a moraine-like ridge today is seen beyond their lower part.

EQUILIBRIUM LINE DISPLACEMENT

The ELAuA estimates were compared with the estimates on current ELAs, and the vertaical displacement of the equilibrium line at the individual sites since the Little ice Age was calculated from this. The results are depicted in figure sc. No estimates were obtained from small firn areas situated on mountain plateaus.

At first sight the resulting plot appears to be without any clear overall pattern; most displacement values (AELA) are within a range of 20 to 40 m, but also values above 200 m are encountered. Moreover, even the maximum AELA values appear to be dispersed as to azimuth without any clear tendency.

Differentiating the AELA estimates according to the
type of ice body leads to a much more clear pattern. As ice
caps all are in a extremely exposed setting on top of the

mountain plateaus, AELA estimates (14) derived from these ice bodies are initially excluded from the analysis. The distribution of estimates derived from the remaining 84 glaciers and firn areas can then be deliminated by the envelope as shown in figure sc. Estimates on AELA valuesless than 40 m may be encountered at any azimuth except for the sector from SE to S (no observations, see figure sa). AELA values above 40 m are however only met with between WNW and NE, and the distribution appears to be symmetrical about NNW. This direction is identical to that designated as being of special interest from experience gained by an inventory on the glacier recession since the Little Ice Age (see above).

FACTORS CONTROLLING EQUILIBRIUM LINE DISPLACEMENT

The marked difference as to vertical displacement of equilibrium lines since the Little Ice Age is probably a feature derived from glacial responses on different changes in local climate at the individual ice bodies. Local climate is to a certain degree controlled by local topographic factors, and an investigation on factors controlling this heterogeneous response will therefore be attempted by considering the isolated influence of various topographic parameters on AELA values.

Consider figure 7a for a moment. This diagram shows a plot of AELA values (vertical) against the present ELA (horizontal) at the individual ice bodies. For ice bodies with the equilibrium line at comparable low altitude, the spred of AELA values is rather small, while it is considerable for higher ELA values. Of special interest is the fact that although the maximum values of AELA tend to grow as ELA increases, small values of AELA are nevertheless encountered for all values of ELA. As to ice body types, small firn areas (open circles) are characterized by comparatively small AELA values, while ice caps (black squares) dominate among high ELAs with large AELA spread. Glaciers in shielded topographic sites display both small and large AELA values.

From this one may conclude that the greater the current ELAmean, the greater are the changes to encounter a very great ELA displacement since the Little Ice Age. And if one excludes the very exposed ice caps from the analysis, and also includes information from figure sc, the preliminary conclusion is that provided an ice body considered has a northerly or northwesterly aspect (at the equilibrium line), and furthermore has a high current mean ELA, this ice body may well have experienced an above normal vertical displacement of the equilibrium line since the Little Ice Age. Only smaller vertical ELA displacements are to be expected for the remaining ice bodies within the study area.

Figure 7b elaborates further on this hypothesis. This diagram shows AELAs against the general altitude of the terrain surface at the top of the headwall above the glacier/ice body considered. For ice caps the altitude of the ice-free terrain adjoining to the ice body is considered. Especially in the southern part of the study area, this terrain surface may take the form of a rather smooth plane, and is not seldom of considerable horizontal extent. In the northern part of the study area, terrain is in general more alpine, and smooth top plateaus are of lesser extent.

Also from this point of view, altitude appears to represent a controlling factor on the glacial behaviour since the Little Ice Age. For ice bodies where the terrain surface above the glacier or adjoining to the glacier (ice caps only) is at relative low altitude, the corresponding AELA is small, while AELA_max,_mum values tend to increase as greater absolute altitudes of the general terrain surface are encountered. Low AELAs are however still met with for some sites characterized by great terrain surface altitudes also.

HIGH-ALTITUDE METEOROLOGICAL CONDITIONS IN DISKO

The observation that ice bodies at high sites tend to display greater mean vertical equilibrium line displacement since the Little Ice Age than ice bodies at lower sites may at first sight appears to be somewhat surprising. In general, ice bodies at low sites (close to the regional GL) would more likely be expected to experience close to critical climatic conditions than would ice bodies at higher sites (high above the regional GL) within the same area. Adding to the complexity is furthermore the fact that only ice bodies facing N and NW (and at high sites) display large AELAs, while all other ice bodies does not. Whatever the precise cause, it must apparantly act very selectively.

Interpretations of observed ELAs and AELAs are affected by the nature of the ELA, being a time-transgressive feature. Climatic and topographic controls are integrated over a period of considerabel lenght, the relevant period probably being decades or even centuries for high-arctic glaciers. Earlier investigations do not agree on the nature of pertinent controls on glacierization and glacial behaviour, and, actually, the importance may well vary from region to region.

In the case of Disko, some observations on snow drift appear relevant in the present context. Large-scale eolian accumulation and deflation forms exist within the accumulationareas of many glaciers in Disko (Fig. 8). These features display a local relief of 10-70 m, and are probably not short-lived forms. Rather they are the result of a stable dominant air flow acting during several years. These air-flow forms has been mapped from aerial photographs(1953 and 1964), and the reconstructed air flow pattern responsible for the observed niveo-eolian forms is depicted in figure 9. Short arrows in eastern Disko indicatethe dominant wind direction according to coastal sand dunes. The general pattern is clearly that of a dominantair flow towards N and NW, although local deviations up to 90° may occur where air masses are channeled into prominent valleys or through topographic saddles in a

range crest. Within the present study area the air flow thus
reconstructed is from SE towards NW.

Possibly with the exception of the air flow observations derived from coastal sand dunes, all air flow directions in figure 9 are thought to depict the dominant air flow during at least authumn, winter and spring, as air temperatures at the glaciers then are well below O°C and new snow is at hand in large quantity.

Observations on upper-air meteorological conditions (DMI 1967, 1977) at Egedesminde/Aasiaat, 65 km SE of southern Disko (Fig. 1), lend support to this supposition. Figure 10 shows radiosonde data on resultat wind directions and -speeds for the period 1956-70 at Egedesminde for surface level (27m asl.) and the 850 mb level (usually 1150-1450 m asl.) soundings. At the ground surface, the resultant wind is from NE and E in October-February, while winds with westerly component dominate during the summer. This corresponds well with the observations from Godhavn mentioned earlier in the paper. At the 850 mb level, however, the air flow is primarily meridional from S during the whole year, with maximum southerly component from January to June. Mean monthly velocities are often in excess of 20 knots. The zonal flow from E close to sea level is probably the result of the katabatic flow of cold air off the Greenland Ice Sheet from the almost permanent surface inversion, caused by strong radiational effects. During summer, this outflow is interrupted in the coastal areas by advection of warm and moist air masses from S and SW along the Davis Strait. High above sea level, the meridional flow from S at the 850 mb level however persist on a year-round basis, which is in accordance with the recurrent suggestion that the atmospheric circulation in the Greenland area includes a substantial meridional component (see e.g. Putnins 1970).

From this I would therefore suggest that many glacierzed mountains on Disko penetrates into the upper-air circulation system, and are for most of the year therefor exposed to fairly strong meridional winds from southerly directions. Below, terrain is mostly exposed to zonal winds from easterly and westerly directins as outlined above.

The air flow pattern displayed by figure 9 is not in perfect alignment to that observed at the 850 mb level at Egedesminde (Fig. 10), but this may well be the result of the meridional air flow being deflected somewhat against NW by the 1500-2100 m high mountains in interior Disko and in the peninsula Nugssuaq NE of Disko (Fig. 1). Within the present study area, this deflection around Disko result in the dominant wind direction being towards NW across the mountain tops. On the other hand, wind directions as indicated by the coastal sand dunes (Fig. 9) are in accordance with the zonal air flow pattern near sea level.

Summing up, the above considerations clearly point toward
the potential importance of high-level snow drift

from SE towards NW for the distribution of snow in the uppermost mountain areas in Disko. Not the least, this must obviously be of some influence for the mass-balance characteristics for high-lying glaciers within the study area, also. The very existence of large-scale niveo-eolian surface forms on most high-lying glaciers in Disko (Fig. 8) bears testimony to this conjecture.

SNOW DRIFT AND DIFFERENTIAL EQUILIBRIUM LINE DISPLACEMENT

Snow drift from SE appears to be frequent across the high areas within the study area (Fig. 9). To investigate the potential control of this feature on the observed equilibrium line displacements since the Little Ice Age, AELA values have been plotted against the potential snow drift upland size, measured towards SE from the top of the headvall or (for ice caps) from the ice margin, across onbroken, unglacierized and smooth terrain liable to experience extensive snow drift from SE.

Extensive high-altitude surfaces suitable for snow drift from SE are frequently found within the southern part of the study area, while the more alpine northern subarea is characterized by lesser values, only. Within this region high-level potential SE snow drift surfaces usually measure less than 300-500 m across. The results are depicted in figure 7c.

Figure 7c displays a rather interesting pattern. For potential high-level snow drift from SE across unglacierized surfaces measuring less than 500 m NW-SE, all AELAs are 50 m or less. For snow drift surfaces measuring more than 500 m across, AELA values are very scattered with a spread from 20 m to 220 m. In general, however, mean AELAs grow as the potential snow drift upland area increases. This lend support to the above hypothesis advocating the importance of snow drift as an explanation of the very different AELA values.

Ignoring all ice bodies with a potential SE snow drift upland area measuring less than 500 m NW-SE, and drawing isolines for the present minimum ELA (Fig. 7c), a consistent pattern emerges within the dispersed part of the diagram. High AELA values are seen to be associated both with high present ELAs and large potential snow drift upland areas. For potential snow drift uplands exceeding a certain critical value - about 1000 m in length - the present ELA however alone apears to be in control of the AELA.

A potential high-level SE snow drift surface measuring at least 500 m NW-SE thus appears to represent a controllingfeature for ice bodies within the study area. Ice bodies characterized by a potential snow drift upland less than 500 m in length all display ELAs below 50 m since the Littel Ice Age, even though there may still be recognized a weak tendency towards greater AELAs for succesively greater potential snow drift distances for this group also. In other words, for this group of glaciers and firn areas the

rise of the equilibrium line altitude appears to be relativelyconstant
regardless of the general altitude of the
glacier (as indicated by the current ELA). This is probably
because snow drift for small snow drift uplands as these is
of comparativly little importance only, and that the
A ELA in these cases primarily is controlled by other factorsas
e.g. changing air temperatures.
For potential snow drift uplands measuring more than
500 m SE-NW the spread of AELAs is considerable. But
in general, the higher above sea level the glacier consideredis
sideredissituated, the greater is the corresponding rise in
equilibrium line since the Little Ice Age.

Together, these observations point toward the general importance of snow drift from SE, and especially so for glaciers situated high above sea level. Only glaciers with a present mean ELA above 750-800 m asl. display above »normal« (50 m) AELA values. This is in good accordance with the information gained from radio-sonde data on upper-air meteorological conditions as presented above, and lends support to the supposition that the mountains on Disko penetrates into the meridional air flow system recorded at the 850 mb level (Fig. 10). It thus appears that the mass-balance for several of the high-level ice bodies within the study area extensively is controlled by snow dntt set up by this meridional air tlow. The above considerations also offers an explanation of a somewhat curious feature appearing in figure sc. It is immediately intelligible why only cirque- and valleyglaciers with a northwesterly aspect are associated large AELAs due to the snow drift from SE, but less so why several ice cap glaciers display high ELAs regardless of aspect? In this context it is important to bear in mind that ice caps typically are located on exposed mountain plateaus, and that snow drift from SE would supply snow to all radiating sectors of an ice cap, regardless of their individual aspect. Actually, the overall surface geometry of most ice caps within Disko tend to be elongated parallel to the air flow direction shown in figure 9, and they should probably be perceived as extraordinary large snow drift accumulation forms. Figure 5c thus is in accordance with what should be expected, provided the existence of imDortant hieh-altitude snow drift from SE.

PALAEOCLIMATIC IMPLICATIONS

Provided that the mass-balance of high-lying ice bodies with northwesterly aspect among other things is controlled by high-level snow drift from SE, certain palaeoclimatic implications may be attempted.

Ignoring ice caps, high-lying ice bodies with northwesterlyaspect
have experienced an above »normal« (in excessof

cessof50 m) rise of the equilibrium line since the end of the Little Ice Age. Several glaciers within this group displayAELAs of 100 m or even more. Meteorological observationscarried out at Jakobshavn/Illulissat since 1873 (at sea level) indicate a mean air temperature rise of about 2°C and a rise in the yearly precipitation of about 20% since the end of the Little Ice Age (ending about 1895 AD.). This climatic change must influenze upon all ice bodies within the area, regardless of the size of their potential snow drift upland. Very great AELAs (reduced by a normal/background AELA of 20-50 m) must then in all likelihood be ascribed primarily to a reduced importanceof high-level snow drift from SE since then. As the mean precipitation has gone up during the intervening period, a reduced snow drift influence would most likely be caused by a less intensive meridional high-level air flow over Disko nowadays, compared to what was normal during the Little Ice Age. On the other hand, the »background«AELA of 20 to 50 m characterizing ice bodies without significant high-level snow drift upland is then probably the result of changes in other meteorological factorsas temperature, precipitaion, mean cloud cover/net radiation since the end of the Little Ice Age.

Therefore, if anything, the observations described in the present paper calls for great care to be exersized when attempting reconstructions of past climate (typical temperatures) from observations on past and present glacier cover, ELAs and cique floor levels. Especially so in regions, where glaciers at present are absent, and where the present and past glaciological characteristics for the individual sites therefore are difficult to investigate.

DANSK SAMMENDRAG

Artiklen redegør for en undersøgelse af gletscherne i den sydvestlige
del af øen Disko, i det centrale Vestgrønland (fig. l og 2).

Som i de fleste områder på den nordlige halvkugle er også gletscherne i dette område smeltet tilbage siden slutningen af den såkaldte »Lille Istid«, der sluttede omkring århundredeskiftet. I undersøgelsesområdet er det dog bemærkelsesværdigt, at det især er højtliggende gletschere exponeret med N og NV, som er smeltet mest tilbage (fig. 5). Umiddelbart ville det forventes, at denne gruppe gletschere mere end andre havde været beskyttet mod et betydeligt massetab i den mellemliggende periode, dels på grund af deres store højde over havet, dels på grund af deres orientering bort fra den maximale indstrålingsretning.

En undersøgelse af ligevægtliniens vertikale forskydning (fig. 5) siden Den lille Istids afslutning bekræfter det ovennævnte forhold, men samtidigt åbnes herved mulighed for en analyse af årsagerne hertil. Denne analyse gennemføres ved at sammenholde den påviste vertikale forskydning med ligevægtsliniens nuværende beliggenhed, med bagvæggens højde, samt med dimensionen af det potentielle snefygningsopland ved vindretninger fra sydøst (fig. 7). Årsagen til netop at betragte denne vindretning er forekomsten af meget store snedrive-former i mange højtliggende gletscheres akkumulationsområder (fig. 8), konsekvent vidnende om dominerende vindretning fra sydøst (fig. 9), samt upper-air radiosonde data fra 850 mb niveauet over Egedesminde (60 km syd for Disko, fig. 10).

Analysen godtgør at årsagen til den ovennævnte uens gletscherreduktion skal søges i de specielle meridionale vindforhold i 850 mb niveauet. Her hersker året rundt sydlige vinde (fig. 10), hvilket ved snefygning netop er af særlig betydning for højtliggende gletschere der er exponeret i nordlig retnig. Den ekstraordinært store tilbagesmeltning af netop disse gletschere tilskrives derfor det forhold, at klimaændringen siden Den lille Istids ophør inden for undersøgelsesområdet formodentligt har været ledsaget af reducerede vindhastigheder i 850 mb niveauet, hvorved de højtliggende og nordvendte gletschere ikke i samme omfang som tidligere får tilført sne ved snefygning. Derfor den særligt store gletscherreduktion for disse gletschere.

ACKOWLEDGMENTS

This paper is based on observations obtained while I was serving as scientific director of the Arctic Station (University of Copenhagen) in Qeqertarsuaq/Godhavn, Disko. My sincere thanks goes to the staff at the station, and especially to Jakob Broberg, skipper of Arctic Stations research cutter »Porsild«, for many joint trawels and for invaluable informations on typical weather pattern in the Disko region during the last 50 years. The paper benefitted from comments by professor Harald Svensson, University of Coperhagen.

REFERENCES

Agassiz, L., 1840: Etude sur les glaciers. Neuchatel. English translation:
studies on glaciers. Translated by A.V. Carozzi. New
York and London, Hafner Publishing Co., 1967.

Ahlmann, H.W., 1948: Glaciological research on the North Atlantic
coasts. Royal Geographical Society, Research Series, No.
1, 83 pp.

Ahlmann, H.W., 1953: Glacier variationsand climatic fluctuations.
American Geographical Society, (Bowman Memorial
Lectures, Series 3), New York.

Andrews, J.T.,m Barry, R.G. and Drapier, L., 1970: An inventory of the present and past glacierization of Home Bay and Okoa Bay, east Baffin Island, N.W.T., Canada, and some climatic and palaeoclimatic considerations. Journal of Glaciology, 9 (57): 337-362.

Andrews, J.T. and Miller, G.H., 1972: Quarternary history of northern Cumberland peninsula, Baffin Island, N.W.T., Canada: Part IV: Maps of the present glaciation limits and lowest equilibrium line altitude for north and south Baffin Island. Arctic and Alpine Research, 4:45-59.

DMI (Danish Meteorological Institute), 1967: Meteorological upper-air
observations 1951-1960, Egedesminde, Greenland, 41
pp.

DMI (Danish Meteorological Institute), 1977: Meteorological upper-air
observations 1961-1970, Egedesminde, Greenland, 41
pp.

Flint, R.F., 1971: Glacial and Quarternary geology. New York,
John Wiley and Sons, Inc., 892 pp.

Glen, J.W., 1963: Bulletin of the International Association of
Scientific Hydrologists (lASH), 8(2):68 (contribution to discussion).

Gross, G., Kerschner, H. and Patzelt, G., 1977: Melodische Untersuchungen über die Schneegrenze in Alpinen Gletschergebieten. Zeitschrift für Gletscherkunde und Glazialgeologie, X11(2):233-251.

Grosswald, M.G. andKotlyakov, V.M., 1969: Present-day glaciers
in the U.S.S.R. and some data on their mass balance. Journal
of Glaciology, 8:23-50.

Hoinkes, H.C., 1968: Glacier varation and weather. Journal of
Glaciology, 7(49):3-19.

Humlum, 0., 1983: Rock glacier types on Disko, central West
Greenland. Geografisk Tidsskrift, 82:59-66.

Humlum, 0., 1985: The glaciation level in West Greenland. Arctic
and Alpine Research, 17(3):311-

Humlum, 0., 1986: Mapping of glaciation levels: comments on
the effect of sampling area size. Arctic and Alpine Research,
18(4):407-

Ito, M. and Vorndran, G., 1983: Glacial geomorphology and snow-lines of the younger Quaternary around the Yari-Hotaka mountain range, Northern Alps, central Japan. Polarforschung, 53(1):75-89.

Lichtenecker, N., 1938: Die gegenwärtige und die eiszeitliche
Schneegrenze in den Ostalpen. Verhandl. d. 111. Intern. Quartär
Konferenz, Wien 1936: 141-147.

Meier, M.F., 1965: Glaciers and Climate. In »The Quarternary of
the United States«, eds. H.E. Wright jr. and D.G. Frey, Princeton
University Press, Princeton, N.J., 795-805.

Meier, M.F. and Post, A., 1962: Recent variations in mass budgets of glaciers in western North America. 1.U.G.G./I. A.S.H., Comm. Snow Ice, Obergurgl, Int. Ass. Sei. Hydrol. Publ., 58:63-77.

Matthes, F.E., 1942: Glaciers. In »Hydrology«, ed. O.E. Meinzer,
New York, McGraw-Hill Book Co., 149-219.

Miller, M.M., 1961: A distribution study of abandoned cirques in
the Alaska-Canada Boundary Range. In »Geology of the Arctic«,
ed. G.O.Raasch, University of Toronto Press, 2:833-847.

Nye, J.F., 1960: The response of glaciers and ice-sheets to seasonal
and climatic changes. Proceedings of the Royal Society of
London, Ser.A., 256 (1287):559-584.

Nye, J.F., 1963: The response of a glacier to changes in the rate
of nourishment and wastage. Proceedings of the Royal Society
of London, Ser.A., 275 (1360):87-

Porter, S. C., 1964: Composite Pleistocene snow line of Olympic
Mountains and Cascade Range, Washington. Bull. Geol. Soc.
Amer., 75(5):477-481.

Visser, P.C., 1938: Wissenschaftliche Ergebnisse der Niederländischen Expeditionen in den Karakorum und die angreenzenden Gebiete in den Jahren 1922-1935. Bd. 11, Glaziologie, Leiden, 216 pp.

Weidick, A., 1968: Observations on some Holocene glacier fluctiations
in West Greenland. Meddelelser om Grønland, 165(6):
202 pp.