As a continuation of Professor Ahlmann’s studies of the glaciers along the North Atlantic coasts, a glaciological research programme was initiated in Kebnekajse in 194.6; its main object was a detailed study of the mass balance of Storglaciären. This study has been going on for 16 consecutive years and a reasonably accurate record of accumulation and ablation since the winter of 1945–46 is now available.

Before going into any details it may be useful to state some basic facts. The glacier area is now 3.1 km.², the average accumulation is 4.0 × 10⁶ m.³ of water equivalent or 130 g./cm.², the average ablation is 6.2 × 10⁶ m.³ or 200 g./cm.² and the average net loss is 2.2 × 10⁶ m.³ or 70 g./cm.², i.e. 55 per cent of the total income. The glacier is obviously retreating rapidly, which can also easily be observed at the front and along the sides and by observing the altitude to which the last winter’s snow recedes every autumn. It is thus easy to show the existence of a large deficit, but it is quite difficult to give accurate figures for this deficit expressed in, say million cubic metres per year.

Without putting a lot of effort into the field work, one can get very large errors. We have chosen a glacier which has a very simple geometry—a well defined accumulation area, parallel sides, no ice falls, etc.—but still, the accumulation is so unevenly distributed that a very great number of measurements have to be made in order to plot an accumulation map accurate enough for the computation of the total accumulation within to per cent of its true value—and, in fact, an accuracy of to per cent is far from satisfactory for a mass balance study.

Over the lower part of the glacier, where the snow lies on solid ice, a number of pits are dug at the middle or the end of May for a determination of the average snow density. At the same time the snow depth is measured with a steel rod at about 200 places in the ablation area. Depths and densities, which in fact vary little from one place to another but change considerably with time, give water equivalents. This method has always been easy to use and it is believed that the accumulation maps are good over the lower half or lower two-thirds of the glacier.

In the accumulation area the problem is rather different. It happens seldom that the autumn surface becomes solid enough to permit soundings with a steel rod; almost every observation has therefore had to be made in pits. Since it is in these higher reaches of the glacier that the maximum snow depth is found, at least the same absolute accuracy is needed as that which can be obtained in the ablation area, and with 4 to 6 m. of snow the digging of one or two hundred pits has been out of question. So, where the observation net ought to be very dense, we have few observations, and where we could have accepted fewer figures, we have plenty of them. This is unfortunate, but I assume that the same is true for most accumulation studies wherever they have been carried out.

During the last few years we have improved our methods. We still dig a number of deep pits for density measurements and for studies of stratification, but the majority of snow depth observations in the accumulation area are now made by core drilling with a coring auger of the SIPRE type. In this way we have managed to obtain the same accuracy over the whole glacier. The 1961 accumulation map (Fig. 1) is based on 370 observations, i.e. about 120 per km.².

Fig. 1.

Accumulation map for Storglaciären for 1961

Even though ablation does not vary as much as accumulation it also has to be measured at a great number of points. Variations can be caused by heat from nearby rock walls, by differences in ice structure and in amount of dirt on the surface, by surface drainage and, which is very important, by the thickness of the original snow cover. In the ablation area the net loss from the beginning to the end of the ablation season is easily measured with the usual stake method making use of the accumulation inventory at the end of May. In the accumulation area of Storglaciären stakes are used, but only combined with often repeated pit observations. The only satisfactory method which we have used for accurate ablation measurements in firn, is to compute it as the difference between two successive accumulation measurements. The density of our ablation observation net has been only 13 per km.² as compared with 120 for accumulation, but probably these densities give about the same accuracy (Fig. 2), since the ablation is more a function of altitude and thus varies less than accumulation and also because the greatest volume of melted water comes from the ablation area, where it is so much easier to measure (more than two-thirds come from there).

Fig. 2.

Ablation map for Storglaciären for 1961

I would welcome an initiative from the Glaciological Society or from this Symposium to work out some standards for mass balance studies, because so many people seem to think that a mass balance can be computed from any scattered observations. The classical examples are of course early publications on the mass balance of the Greenland Ice Sheet. Maybe we could agree upon desirable densities of the observation net (in points per km.²) for studies with different aim, for example 100 accumulation and 10 or 20 ablation observations per km.² for a long term study (of, let us say first order) associated with climatic problems or connected with a small glacier’s ice and/or water discharge. For other studies, when just a rough knowledge is needed, one may define mass balance studies of second and third order. Other observation procedures may also have to be defined. It is possible that fewer measurements per km.² may be sufficient on glaciers in the Alps than on the Scandinavian ones, because of the lower wind velocities in the Alps—but all this should first be looked into carefully, so that we can know how far we can trust published results.

We look upon our regime programme on Storglaciären as we look upon a meteorological station—one year’s observations are useful, a ten-year period is very valuable, but it is not until it covers some periods with glacier advance and glacier retreat that it appears to full advantage. Data should be published as soon as possible—and for the last two years ours have been delivered to the editor of Geografiska Annaler only a couple of months after the end of the ablation season—and when a long enough series of data has become available it should be analysed and one should try to establish which climatic conditions have caused the variations in accumulation, ablation and balance. We now have sixteen years of data, and we have good meteorological records from the glacier and from several meteorological stations in the neighbourhood, but at the moment we have no good climatologist and glaciologist available to do the work.

During the sixteen years of measurements the accumulation has varied between 2.1 and 6.9 × 10⁶ m.³ (68 and 220 g./cm.²) and averaged 4.0 × 10⁶ m.³ (130 g./cm.²) . The ablation has varied between 3.2 and 9.6 × 10⁶ m.³ (100 and 300 g./cm.²) and averaged 6.2 × 10⁶ m.³ (200 g./cm.²). We had a surplus in 1948–49 because of maximum accumulation followed by the next lowest ablation, we measured a balanced regime in 1947–48 and all the other fourteen years have given a deficit. The average thinning of the whole glacier has been 15 m. since 1946 (nearly 10 per cent of its volume has melted away). It is also of interest to know that the loss during the last three-year period has been greater than during any other three-year period under observation.

But how to explain this loss? Is accumulation or ablation responsible, or is it both? A thorough analysis has not yet been made. We know however, that the temperature climate has become considerably warmer during this century. This temperature rise began at the turn of the century and amounts to about 2° C. for the period June–August. Figure 4 shows the observed ablation as a function of the summer temperature anomaly found by taking the average of the mean summer temperature for five nearby meteorological stations—the anomaly being defined as the deviation from the mean value for the years 1901–30. In this case the “summer” is defined as 1 June-30 September since we often have a considerable amount of melting late in the season. The graph says that in order to get cquilibrium—with the present accumulation and areal extent unchanged—the summer temperature must fall to 0.7° C. below the 1901–30 average, i.e. to 1.2° C. below the average since 1946. We can also read that the accumulation should have to increase from 4.0 to about 5.5 × 10⁶ m.³ to cause a balanced regime with the 1901–30 summer temperatures and to about 6.5 × 10⁶ m.³ with the present rate of summer melt.

Fig. 3.

Accumulation, ablation and surplus (+) or deficit (—) on Storglaciären for 1945–61. (Data for the years 1950–59 from E. Woxnerud)

Fig. 4.

Deviation of mean summer temperature for five stations in Swedish Lappland frorn the average figure for the years 1901–30

We know that the summer temperatures have increased with an amount which corresponds well to the 1.2° C. required for equilibrium, but during only one exceptional year have we measured an accumulation of 6.5× 10⁶ m.³ or more. It is even less likely that accumulation has decreased from about 6.5 × 10⁶ m.³ a year to 4.0 if one notes that according to a paper by Reference ÅngströmAnders Ångström (1941) the annual precipitation during 1901–30 was at least as large or probably 5 per cent larger than during 1861–1900, and it can also be shown that the average winter precipitation as recorded by our five neighbouring meteorological stations was 11 per cent larger during 1945–60 than during 1901–30.

It is therefore most likely that, before the present retreat started, the accumulation on Storglaciären was approximately the same as it is today or 4.0 × 10⁶ m.³, possibly slightly more because of the larger area, and that the retreat is a result of a June—September temperature rise of just over 1° C.