Introduction

Greenhouse warming of the atmosphere should affect the Arctic regions first and most strongly. The large fluctuations of annual weather make it difficult to detect an ongoing climate change from meteorological records alone. In this paper we present recent changes in surface elevation and volume of McCall Glacier in Arctic Alaska.

Reference Kelly, Jones, Scar, Cherry and Tavakol.Kelly and others (1982) give a good summary of the changes in seasonal and annual air temperatures in the Arctic from 1881 to 1980. Warming after 1890 culminated in rite 1930s with winters and summers being warmer by about 2.5 and 1.3"C, respectively, than during the 1880s. The warming started in the Barents and Kara Seas and became most pronounced in the northwest Greenland region. The Arctic cooled in the 1950s and stayed cold during the 1960s, with annual temperatures about 0.85 C less than in the 1930s. Warming began to affect the Arctic anew in the 1970s. again starting in the Barents and Kara Seas and spreading westward. In contrast with the 1930s, the warming is most pronounced in the Alaskan regions where annual temperatures rose bv about I C (Fig. 1) . The overall warming trend since the end of the 19th century is clearly exhibited by the glaciers in Arctic Alaska; they have retreated by 150-700 m from their “Little Ice Age” moraines : Reference Hamilton.Hamilton, 1965; Reference CalkinCalkin, 1988). Our goal is to identify more recent changes in climate against this background, by using high-resolution data from McCall Glacier.

Fig. 1. Annual mean temperature, average of Anchorage, Barrow, Fairbanks and Aome (adapted from Reference Bowling, Weller., Wilson and Scverin.Bowling (1991)).

To link climate changes to glacier changes two things have to lie considered:

• (i) Changes in climatic variables such as precipitation, solar radiation and temperature must not cancel in their combined effect on the mass balance of a glacier. On McCall Glacier, a general warming throughout the year would increase summer melt, but winter climate is too cold and dry for winter precipitation to increase significantly. This makes McCall Glacier a sensitive indicator of climate wanning.

• (ii) Advance/retreat or thinning/thickening of the glacier snout is greatly amplified by ice flux but lags the original mass-balance change in an intricate way Reference Nye.Nye, 1965; Reference Johannesson, Raymond and Waddingfon.Johannesson and others, 1989). Measuring the total volume change of a glacier, by photogrammetry or surveying, is more difficult than observations near the terminus. The advantage is that a mass-balance change can be readily detected by a corresponding volume change without time lag.

McCall Glacier is located at 69 18'N, 145 48'W in the northernmost chain of the Romanzof Mountains, northeastern Brooks Range, Alaska {Fig. 2 inset). The glacier occupies a north-facing valley. Ii is about 8 km long, has an average width of 640 m and covers an area of 7.4km". The ice originates in three cirques (referred to, from east to west, as upper, middle and lower, cirque) and extends from more than 2700 m on the north face of Mount llublcy to the terminus at 1350m. The glacier surface forms a series of bulges and treads with slopes of up to 15'^: and as low' as 3^C; the average ice slope is 7.5'. Figure 3 is a view of the lower ablation area, about 1.5 km up-glacier of the terminus, looking towards the confluence of the three cirques.

Fig. 2. Map of McCall Glacier, showing 1972 pole positions resurveyed in 1993 (black dots); control monuments (triangles); 1993 mass-balance stakes (open circles); upper and lower detailed-transverse profiles; and position of 1993 camp ( cross).

Fig. 3. Looking up-glacier from the upper detailed-transverse profile. Xote “Hanging" (Racier to the left and stum-capped Mount McCall in the right background.

Benson, 1986). Maximum ablation at the terminus was 2.1 ma ¹ in 1957/58 and 1.6-2 m a ¹ in the 1970s. All mass-balance values are given as water equivalent unless noted.) On McCall Glacier all the ablation and about 75% of the accumulation occur during June September. Internal accumulation accounted for up to 54%, and superimposed ice for up to 5%, of the total accumulation Reference Trabant and Benson.Trabant and Benson, 1986). McCall Glacier has a low- mass-exchangc rate; annual accumulation and ablation, averaged from 1970 to 1972, were +0.16 and -0.3 m a respectively. The mean equilibriuindine altitude from

Measurements Prior to 1993

McCall Glacier is tlie only glacier in ¡he U.S. Arctic with a glacioiogical record of several decades. Shorter records exist for some: of the small cirque glaciers of the central Brooks Range ¡Reference Calkin, Ellis, Haworth and Burns.Galkin and others, 1985). McCall Glacier was studied during the International Geophysical Year (IGY) in 1957)58, and from 1969 to 1975 a.s a contribution to the International Hydrological Decade (IHD). As part of the IGY fleld program, a photogram- metric map of scale I : 10 000 was produced (Reference BrandenbergerBrandenberger, 1959). Observations of the IGY team on the equilibrium and firn-line position arc also compiled in this map. Another photogrammctric mission was carried out in 1971 by Reference Dorrer and Wendlcr.Dorrer and Wcndlcr : 1976).

Reference KeelerKeeler, 1958 mapped mass balance on lower McCall Glacier during the IGY. Mass-balance maps of the whole glacier exist for the balance years 1968/69. 1969/70, 1970/ 71 and 1971/72 , Reference Weller., Fahl and Corbin.Weller and others, 1972; Trabant and 1969 to 1972 was 2050 m; mountain shadowing and preferred snow deposition on the west side of the glacier make the equilibrium line run normal to topographic contours over an elevation range of 350 m (Reference Weller. and Ishikawa.Weller and Ishikawa, 1974).

The extensive mass-balance network of the 1970s was surveyed each year, giving good definition of the ice surface and the surface velocity field at that time (unpublished information from Trabant, 1969-72), Maximum annual velocity was about 16 m a¹. Two detailed transects of the ice surface in the lower ablation area (Fig. 2' were surveyed for several years starting in 1969 '(unpublished information from Benson, 1969-75). The upper profile was resurveyed in 1987, Maps of a perennial a.ufeis field below McCall Glacier show the outline and elevation contours of the terminus in 1970. These data sets give accurate base lines for measuring changes in glacier volume and terminus position.

Meteorological observations were carried out in the upper cirque at about 2300m during 1957/50 (Reference Orvig.Orvig, 1961). An automated weather station operated from 1969 to 1972 on the rock ridge between upper and middle cirque. Sparse data were obtained during the winters. Comparison with weather stations of interior Alaska and the Arctic coast revealed a distinct mountain climate dominated by the proximity of the Arctic front in the region of McCall Glacier (Wendler and others, 1974). Approximate mean annual temperatures were about 12 C at 1700 m (Reference Trabant, Harrison, Benson., Weller and BowlingTrabant and others, 1975). Annual precipitation averaged over the glacier was estimated as 500 mm a"¹. During summer, winds arc mainly from the southwest (Wendler and others, 1974).

A polythermal temperature regime is suggested for McCall Glacier. In the accumulation zone the glacier is at about — 1⁰G throughout its thickness (Reference Orvig. and Mason.Orvig and Mason, 1963), while the ablation zone is cold at the surface !-8°G at 10 m depth: but is probably underlain by a layer of temperate ice at the bed (Reference Trabant, Harrison, Benson., Weller and BowlingTrabant and others, 1975).

1993 Field Work

The data presented in this paper Were gathered from 26 June to 10 August 1993 as part of a 3 year study of recent changes of McCall Glacier. Surveying the glacier surface, using optical methods and airborne and ground-based GPS methods, was the key tool for defining the glacier surface in 1993. The new data set allows comparison with past surveys in 1958 and the 1970s. Most of the control monuments used for the 1970s survey were recovered, and the more stable ones were selected for the 1993 survey. A GPS base line about 15 km long lies the glacier-control network to a survey benchmark that is fixed to the United States Coastal Geodetic Survey network of northern Alaska.

From the 1972 survey of the IHD mass-balance network 55 reliable pole coordinates were selected (Fig. 2). Oil the glacier a person was iteratively directed into the known horizontal position using a theodolite with an electronic distance measurer. Most horizontal coordinates were recovered to within 0.1 m radius, and the changes in vertical coordinate AZ since 1972 were recorded. At some positions, stakes were drilled into the ice for mass-balance and velocity measurements. The upper and lower detailed cross-glacier profiles \ Fig. 2) were resurveyed in 1993 using the bedrock markers of the previous surveys.

A complete center-line profile was obtained using ground-based stop-and-go kinematic GPS from near the pass of the upper cirque down to the terminus. A detailed outline and elevation profile of the 1993 terminus up to about 150 tn up-glacier were determined both optically and by kinematic GPS methods.

The most important correction to the elcvation- change data came from the different dates of the surveys in 1972 and 1993. From readings of all 1993 mass-balance stakes (Fig. 2), we interpolated ice and snow ablation for the period 29 June-5 August 1993 as a function of elevation (see Fig. 4a inset). Lower and upper cirque have somewhat different ablation, possibly due to nonuniform mountain screening. The maximum value ol the ablation correction was 1.38 m at stake 2 about 550 m from the terminus; the average for all points was about 1.0 m. Repetitive surveys about a month apart of several 1993 mass-balance stakes gave horizontal and vertical velocities. The vertical movement between 29 June and 5 August is smaller than + 0.01 m and can be neglected.

Fig. 4. a. Elevation change, 1972-93: a. vs elevation, b. cross-glacier variation, c. contour map.

Fig. 4.

Fig. 4.

The 1993 vertical coordinates were corrected for refraction and earth curvature. This correction increased the 1993 elevations by less than 0.2 m for most base lines ( ˂ 1.7 km). A few exceptionally long base lines had larger corrections ;0.63 m for the longest base line of 2.8 km). A similar correction was also calculated for the 1972 v ertical coordinates- Dite to shorter base lines, however, it was smaller than the inherent error of the 1972 survey (about 0,3m). The final error in the 1993 elevation data is + 0.05 m; similar accuracy exists in the horizontal coordinates. The error in the elevation change from August 1972 to August 1993 is dominated by the accuracy of the 1972 survey which was 0.3 m in vertical and 2 m in horizontal.

In the context of a future long-term monitoring program that includes glaciers from all over Alaska, center-line elevation profiles ol’ McCall Glacier and two other glaciers in the northeastern Brooks Range Esetuk and Okpilak Glaciers, about 20 km southeast and 15 km south, respectively were obtained using a lightweight, airborne-laser ranging system. These data will be presented elsewhere.

Changes in Elevation, Volume and Terminus Position

Figure 4a-c illustrate elevation change from 1972 to 1993. The surface dropped everywhere from a minimum of 1 m at the head of middle cirque to over 42 m near the terminus. The 1 m minimum is exceptional; the average drop in the accumulation areas of the three cirques is more like 3-3.5 m. This can best be seen from Figure 4a, where the elevation change, AZ, for all available positions is plotted against 1972 elevation, Z, regardless of the horizontal coordinates. The scatter represents variations in AZ both across the glacier and between different cirques. Figure 4b shows elevation change for selected transverse profiles. Cross-glacier variation of AZ ranges from about 0.5 m at the transects near the confluence of the cirques (Fig, 2: to about 5 m at the lower detailed-transverse profile. This cross-glacier variation is about 10% of the mean surface drop for each profile. No clear large-scale pattern, such as the glacier's sides dropping more than the center, is evident. .Meandering surface streams which incise channels a few meters deep into the ice are a ubiquitous: feature on McCall Glacier. Some of the recovered pole locations from 1972 were in or near such streams in 1993, indicating that a shift of the surface drainage network may be responsible for cross-glacier variations in elevation change.

The ice surface along the two detailed transects on the lower glacier is shown in Figure 5 for a number of years. The surface topography is almost uniformly lowered between successive dates, making transverse variation of elevation change comparatively small. The characteristic indentation in the middle of the upper profile is preserved from year to year. A second interesting detail in Figure 5 is the constant angle of the side-moraine slope as thé apparent width of the glacier diminishes over time. Direct observations show that the side moraines are ice-cored. The overlying till slides down at some angle of repose when the glacier thins, protecting more glacier ice from ablation. 1 he bedrock valley walls arc much steeper than the constant angles of repose revealed by Figure 5.

Fig. 5. Time evolution of ice surface along a. upper and b. lower detailed-transverse profile.

Average elevation drop of the upper profile as a function of time is plotted in the inset of Figure 5a. Around 1975 the drop rate increased from about 0.3 to about 1 ma . The elevation drop of the upper detailed transect from 1972 to 1993 provides an independent cheek of the elevation change obtained at the 1972 pole positions. A pole position about 20 m up-glacier from the upper detailed transect has an elevation drop of 21.1m while the nearby part of the upper transect dropped 20,9 m.

A map ol elevation change from 1972 to 1993 has been constructed by smoothing and interpolation Fig. 4c:. Total volume change was 5.5 + 0.2 x 1.0 ⁷ m^:i of ice. The error was estimated from the interpolation statistics arid the 0.3 m error in elevation change of the 1972 locations. Mean mass balance from 1972 to 1993, i.e. volume loss divided by mean glacier area, was. 0.33 + 0.01 ma^-1

Comparison Of The Periods 1972-93 and 1958-72

Dorrer directly compared stereo-photo pairs from 1971 and 1958 to deduce volume loss during this period (Reference Dorrer.Dorrer, 1975: Dorrer and Wend In. 1976'i, Due to the different illumination, scales, and snow coverage prevailing during the two photogfammctric missions, elevation change could be obtained for only three regions in the ablation area, around 1500. 1715 and 1900 m elevation. Mean elevation changes in these regions were 4.5 ±0.4, 2.9 + 0.6 and 2.0±0.4m, respectively. In Figure 6, mean annual thinning rates for the periods 1958-71 and 197293 are compared. In both cases a suitable exponential was fitted to the data (cf. Reference eidiek.Weidick, 1968; Dorrer and Wendler, 1976.. Two features are immediately obvious: (i. in the earlier period the accumulation area is only very slightly affected while in the later period it is characterized by a noticeable surface drop; and (it) the difference in mean annua! elevation change between the two periods increases drastically down-glacier. Extrapolation of Dorrer’s curve to the region close to the terminus gives rise to rather large uncertainties. According to his error bars, annual thinning of the ice there was between -0.5 and -0.85 ma ¹ during 1958 71, while from 1972 to 1993 the terminal region thinned by almost 2.4 m a ¹ (2.1 in a w.c.), a value in excess of the maximum ablation at the terminus during the 1970s. Mean mass balance during 1958 71 was ? 0.13ma ¹ as compared to 0.33m a ¹ lor the period 1972 93.

Fig. 6. Mean annual thinning vs elevation: 1972-93 (this paper), 1958 72 based on the Id map. and 1958-71 from Doner and Wend!er (1976). Positions along the center line of the glacier are connected by lines.

Elevation change between 1958 and 1972 can also he calculated by locating the 1972 pole positions on the 1958 IGY map. The result is shown in Figure 6; surface drop from 1958 to 1972 based on the IGY map differs greatly from Dorrer’s estimate. Mean annual balance would be 0.45 m a^-1, which is more than three times Dorrer’s value.?Strong thinning would occur not only close to the terminus but also about 1.5 km up-glacier, This irregular pattern of elevation change and two further arguments make us believe that there are systematic errors in the 1958 map considerably exceeding the expected error of +2.5m.

The first argument is based on the time history of elevation along the upper detailed traverse ; Fig, 5a inset),. Elevation can be extrapolated for 1958 using the 1GY map. and the thinning rate there from 1958 to 1972 is then about 1 m a just as from 1975 to 1993. This implies the research period in the 1970s was characterized by exceptionally low ablation (thinning rate 0.3 m a ^l). In contrast, mass balance calculated by Dorrer and YVencUer (1976) front weather records of Barter Island, 110 km north at the Arctic coast, indicates that during the period 1969-75 the glacier thinned at a higher annual rate than the mean rate during the 1960s. A thinning rate of 0,3 m a on the other hand, would give a surface drop of 3.9 m for the 13 year period 195871. This is very close to what Dorrer’s curve shows for an elevation of 1600 m, the 1972 elevation of the upper detailed traverse. Secondly, Reference Trabant and Benson.Trabant and Benson (1986) mention several IGY mass-balance stakes that reappeared in 1969 and ¡972 after they had apparently crossed the mean equilibrium line. The mean equilibrium line reconstructed in this way coincided best with the 1972 equilibrium line, making the 1972 balance of 0.19 m a ¹ a crude estimate of the mean balance in tire period 1958-72. Dorrer’s value of -0,13 m a ¹ for the mean balance is within the error range of this estimate while the value of -0.45 m a ¹ deduced from the IGY map is not.

Terminus Retreat

Figure 7 shows reconstructions of the 1958, 1970 and 1993 termini together with the youngest end moraine that outlines the maximum extent of the glacier around 1890 (lichen date). Terminus altitudes were 1354m in 1993, 1327 m in ¡970 and about 1320m in 1958. Changes ol retreat rate and surface elevation at the terminus were derived from direct observation and geometric considerations, treating the terminus up to the lower detailed profile as a wedge with constant surface and bed slope (see Table 1 . Implications are: (i) Terminus-re treat rate more than doubled between the time periods 1958' 70 and 1970 93. (ii) Thinning from 1958 to 1970 is faster than Dorrer’s curve suggests for the terminus Fig. Gi but still within the expected error range, iii The thinning rate from 1970 to 1993 is less than the 2.4 m a ¹ interpolated from Figure 6; it is almost constant down-glaeier of the lower detailed transverse, suggesting a stagnant, passively melting terminus for most of that period, (iv.i Surveys from 1970 to 1972 of the lowermost mass-balance stake (about 40 m from the 1970 terminus) imply that thinning of the terminus accelerated more gradually compared to the step-like increase in thinning rate around 1975 documented at the upper detailed- transverse profile.

Fig. 7. The terminus of Mccall Glacier for the years 1993, 1970 and 1958, and before the retreat from its "Little Ice Age" maximum(persumably(around 1890).

Interpretation —Evidence for Recent Climate Change ?

A significant climate change within the total observation period 19511-93 is strongly implied by the available data. Mean mass balance calculated from volume change became more negative by a factor of about two. between the periods 1958- 71 and 1972 93. IfDorrer’s estimate of 0.13 ma ^-1 for the 1958-71 mean mass balance is accurate, this factor would be 2.5; a value of 0.2 m a \ which is considered a minimum, would reduce the factor to about 1.7. The difference in mean mass balance between these two periods corresponds to the stepwise change in temperature in the 1970s (Fig. 1),

The terminus has doubled its retreat rate from 1970 93 as compared to 1958-70. Data from the early 1970s [last column of Table 1) favor a steadier increase in retreat rate beginning around 1970 rather than a step change. The similar relative increases of terminus-retreat rate, thinning rate of the lower ablation area and mean mass balance imply that all three are indicative of the same climate change.

Table 1. Ion contint of wedge ice and precipitation.

Conclusions

• (1) The- increase in the rate of volume wastage by a factor of more than two from the earlier to the later time period indicates a severe change in mass balance, most likely near the border of the periods,i.e. in the early 1970s.

• (2) The observed mass-balance change is probably mainly due to higher summer temperatures.

• (3) The step-like increase in surfacc-drop rate of the lower ablation area between about 1975 and 1980, and the increase in terminus-retreat rate, appear to be correlated with the recent climate change. This suggests a fast reaction time of McCall Glacier of less than a decade.

• (4) How representative the detected climate change is for the region cannot be decided at the moment. Limited data from other Brooks Range glaciers and an approach to modeling the mass balance of McCall Glacier from atmospheric data will be employed to answer this question in the future.