Glaciers as water resource indicators of the glacial areas of the USSR

Snow and Ice-Symposium-Neiges et Glaces (Proceedings of the Moscow Symposium, August 1971; Actes du Colloque de Moscou, août 1971): IAHS-AISH Publ. No. 104, 1975. Glaciers as water resource indicators of the glacial areas of the USSR V. G. Khodakov Abstract. Information from direct measurement of precipitation and runoff of glacial territories is scant and not always reliable. At the same time such regions are marked as a rule by especially high water resources. The simple fact of glacier existence as well as some morphological characteristics of the glaciers enable us to calculate the estimated water resources of the regions where they occur. The following data are used: (1) Summer ablation of glacier surface A s, mean summer air temperature t and summer solar radiation balance sum B^, generalized in formula (2). (2) Altitude of nourishment limit and snow concentration coefficient K there (from the glacier inventory of the USSR). (3) Fields of air temperature and solar radiation received from meteorological stations situated in non-glacial areas. The values of t and B^ are adjusted to the nourishment limit. Value A s is calculated from f and Bfc. The value of is known to be equal to melt runoff and quality of solid precipitation for the whole glacier zone. Comparison of these calculations with the results of direct measurements in representative glacial basins of the IGY and the IHD produced good results. Macro- and mesogeographical regularities of water resources of glacial areas of the USSR are illustrated on a map of calculated values Y m averaged by districts. Résumé. Les données des mesures directes de précipitations atmosphériques et d'écoulement des régions englacées sont peu nombreuses et pas toujours sûres. En même temps les régions se caractérisent, en règle générale, par de grandes réserves en eau. L'existence des glaciers et quelques-unes de leurs caractéristiques morphologiques permettent d'évaluer les ressources en eau des régions où il y a les glaciers. Les données suivantes sont utilisées: (1) Ablation estivale de la surface du glacier A$, moyenne de la température de l'air estivale f et total estival du bilan des radiations solaires B^, généralisé par la formule (2). (2) Altitude de la ligne d'équilibre et coefficient de concentration de la neige à cette altitude (selon le Catalogue des glacier de l'urss). (3) Températures de l'air et radiation solaire au réseau des stations météorologiques non glaciaires. Les valeurs t et B/ç sont évaluées pour l'altitude de la ligne d'équilibre. Elles permettent d'évaluer la valeur dea s. On voit que Y m =A S /Kéga\e l'écoulement dû à la fusion et égale la quantité des précipitations solides pour toute la région glaciaire. La comparaison des valeurs ainsi estimées avec les résultats de mesures directes aux bassins glaciaires représentatifs de l'année Géophysique Internationale et de la Décennie Hydrologique Internationale a été satisfaisante. Les variations macro- et mésogéographiques des réserves en eau des régions glaciaires de l'urss sont illustrées par une carte des valeurs théoriques de Y m, moyennées pour chaque district. INTRODUCTION The climatological significance of the snow line was noted for the first time by Humboldt (1948). Since then, a number of scientists have tried to use its indicative role to determine the climatological and hydrological characteristics of high-latitude and high-altitude regions. Similar efforts have also been made in the field of palaeogeographic study. However, poor knowledge of the glacial areas restricted the

Glaciers as water resource indicators of the glacial areas of the USSR 23 possibility of working out correct quantitative correlations between water and heat characteristics of balances on the one hand, and the position of snow line and the glaciers in general on the other. The greatest success was achieved in Scandinavia (Ahlmann, 1948). In high-altitude and mountain glacier regions of the USSR the data on atmospheric precipitation and runoff were not sufficiently precise. Moreover the data available are not always reliable. This must be attributed, above all, to the totalizers in high-altitude areas where winter recordings were collected in spring. A number of causes of errors (the overgrowth of the inlet by hoar-frost and snow, the leaks in the tanks, etc.), as well as the 'error of inclination' (Khodakov, 1964) can increase the total error up to hundreds of per cent of the measured value. In some cases gross systematic errors of measurement of runoff in glacial areas are also known. In most cases they can be referred to the percolation capacity of moraine and alluvial deposits in the valleys and, consequently, lead to the underestimation of the runoff (Makarevitch et al, 1969). As high-latitude and high-altitude regions are more and more explored the information about water and snow resources is more and more needed for hydrological calculations as well as calculations of snow avalanches, snow drifts, pressure on constructions, accessibility of passes, possible use of pastures and so on. Here I attempt to estimate water and snow resources of all the glacial areas of the USSR. INITIAL INFORMATION AND METHOD OF CALCULATION Three groups of sources of mass information were used: (1) Value of summer snow and ice ablation A s and its factors. (2) The climatological data on air temperature and solar radiation in the USSR. (3) The altitude of nourishment limit (that is, snow line, zero balance line on the surface or the equilibrium line) and glacier morphology. The greatest part of the first group of initial information was received after IGY and IHD programmes had been carried out. As a rule, all the major elements of water-ice and heat balances were and are measured in selected glacial basins of the IGY (Koreisha, 1963; Krenke et ai, 1970; Makarevitch et al, 1969; Preobrazhensky, 1960; Troitsky et al, 1966; Tronov, 1968; Chizhov et al, 1968). I used these facts as a basis for further conclusions and for control of the results. Information from abroad on summer ablation of glacier surface as well as summer mean values of meteorological elements, which were generalized by Kotlyakov (1968); Krenke and Khodakov (1966); Moiseeva and Khodakov (1971) and Khodakov (1965), was also used. This enabled me to enlarge the range of values which were included in empirical formulae for further calculations. It seems that the glacial areas have the most snow and in a number of cases the greatest water production of all the territories of the USSR. Figure 1 shows the distribution of principal elements of water balance (annual precipitation X, runoff Y and spring maximum snow storage X c ) which was averaged to 10 km stretches in two essentially different regions. The section shown in Fig. 1A was worked out by mass measurements of X and X c ; the section shown in Fig. 1B was worked out by mass measurements of X and Y. The averaged scale makes it possible to compensate melt runoff surplus in the ablation area by its corresponding shortage in accumulation area of glaciers and to compare general water resources of glacial areas with those of adjoining high-altitude regions. In the Urals the maximum of all three curves is to be found in the centre of the glacial area. In the Tien-Shan the maximum of annual precipitation is displaced to the northwest of the glacial area. Near the 'water regime peaks' of both regions the XJY relation is 0.8. As the evaporation from ice and snow surfaces compared with the water resources is, generally speaking, unimportant (in the Urals condensation

24 V. G. Khodakov Precipitation X, runoff Y, and maximum snow storage,,n g/cm B -. 1 1 ' 1 ' 1 1 =r= 20 40 60 80 100 120 140 FIGURE 1. Average distribution of precipitation X, runoff Y and maximal snow storage X c on the profiles across the Polar Urals in the region of Vorkuta (A) and the Tien-Shan in the region of Alma-Ata (B). prevails), X c, as a matter of fact, is equal to Y m - the value of melt runoff of the glacial area and for the glacier surface is equal to summer ablation. Ablation from ice and snow surfaces (ignoring mechanical ablation) is ultimately determined by the heat balance of the melting surface. However, the solution of the heat balance equation to obtain the value of ablation requires a large amount of information which is practically impossible to obtain for most of the glacial areas. The dependence of summer ablation of the glaciers on mean summer (July-August) air temperature t at a height of two metres above the surface was investigated. Analysing 93 pairs of values by the least squares method the following formula was obtained: A s = 0.096 (r + 10) 2S3 (g/cm 2 ) (1) Practically it coincides with the earlier formula using a cubic parabola derived with less data (Khodakov, 1965). Nevertheless, the increase of data caused an increase of mean square deviation of calculated values compared with values measured up to ±89 g/cm 2. That is why another variable was included in the dependence - the shortwave solar radiation balance for three summer monthsb k,k kcal/cm 2, information for which we could get for 22 points (Fig. 2). Taking into consideration the dissimilarity of the variables and the complexity of the calculation, I preferred not to use the least squares method and derived the working formula by way of selection: A s = 0.1 (t + 13^B k + 4) 3 (g/cm 2 ) (2) It seems possible using formula (2) to calculate the value of summer ablation of the whole glacier surface. However, one has to contend with some difficulties arising both

Glaciers as water resource indicators of the glacial areas of the USSR 25 Ablation in g/crn 2 As, Temperature (June-August) t, in C FIGURE 2. Dependence of summer glacier and fiin surface ablation A s on mean summer air temperature t and shortwave radiation balance B^. from the enormous scale of calculation and the shortage of initial information. Moreover some melt water in glacier accumulation areas freezes. The value of A s according to the stratigraphie system {Combined heat and water balances of selected basins, 1969) and information used in practice characterizes the mass decrease of the layer of the current budget year only. The volume of glacier internal nourishment has not yet been calculated, so I have calculated only A s at the glacier nourishment limit. That line is interesting for several reasons. In particular, the value of total accumulation is exactly equal to that of total ablation. A detailed equation of the ice balance is as follows: X c +D+ V+8-E w =A s (3) where X c is the annual snowfall sum, D the snowdrift balance, V the avalanche snow accumulation, S the share of current year ice which was formed owing to the freezing of liquid precipitation (these values as a whole for the current budget year), E w evaporation from snow surface for the accumulation period. In some respect 8 and E w values compensate each other, and, besides, are small compared with A s. Thus it is sufficiently accurate to write equation (3) as: X C +D+V = A S (4) So, summer ablation on the nourishment limit quantitatively characterizes snow storage of that glacial region as well. During fairly normal years the nourishment limit is, in a physical sense, the line of contact between the current year's ice and the 'old' glacier ice. But the 'new' ice zone

26 V. G. Khodakov in mountain glaciers is narrow, so in most cases one can identify it with the firn line without any essential correction. This firn line can be perfectly interpreted on air photographs and even on photographs taken from satellites. The altitude above sea level for each glacier is to be found in the multi-volume The glacier inventory of the USSR. I also used corresponding information from the monograph of Kalesnik (1963) and from numerous articles in the collection The data of glaciological research. Chronicle, Discussions. These data are not homogeneous, sometimes referring to certain data of a certain year and sometime to mean annual data. Generally speaking, annual variations of firn line are small. On the central Tuyuksu glacier, a typical mountain-valley glacier, their range is ± 125 m (Makarevitch et al., 1969). At the same time, space variations of that value within the USSR approach 5000 m. Mean summer air temperature from a meteorological station, the nearest one to the glacier, was adjusted to the altitude of firn line by linear extrapolation. When there were two meteorological stations with a considerable difference of altitudes the vertical temperature gradient was computed. In most cases its values was 7 ± 1 degrees/km. This figure was used for calculation when it was impossible to select a respresentative pair of meteorological stations (mainly in Siberia and the Far East). The values computed that way were decreased by the value of 'the glacier temperature jump', At, defined as the difference of mean summer temperatures in glacial and nonglacial meteorological stations and adjusted to the second altitude. The analysis of the data of both ice sheet and mountain glaciers of the world made it possible to ascertain a clear dependence of A? on their characteristic size L(km) as follows: log At = 0.28 log L - 0.07 (5) The characteristic size was defined as the distance between ice divides and the edge of the ice sheet glacier, and to the start of the narrow tongue in the case of a mountain glacier. In most cases the value of L for the USSR ranges from 1 km to 10 km, and At correspondingly, ranges from 0.9 to 1.6. Calculation of the second parameter B k involves more difficulties. The basis for this calculation is the annual global shortwave radiation data for summer Q (kcal/cm 2 ) for nonglacial, mainly plain-situated stations (Radiation regime of the USSR, 1961). They reflect general climatic, mainly zonal changes of summer solar radiation in a range of 32 to 62 kcal/cm 2. Besides that, a general tendency for Q to increase while going up the mountain for unshaded areas, as well as a reverse effect caused by the increase of local orographic cloudiness, is known (Borzenkova, 1970). The nourishment limit is situated, as a rule, in the flattest part of the glacier and that is why the direct influence of slope exposition is considerably weaker than orographic cloudiness influence and when the horizon is closed. Undoubtedly the the strongest factor involved in B k is albedo. The reflecting surface of the nourishment limit is snow and firn and thus the mean summer value of albedo is similar in different glacial areas and is estimated by direct measurements 0.55 ± 0.1.1 could not consider all these factors separately, so a method of generalized calculation was chosen. On the basis of a few data of B k measurements in the area of the firn line the relation of B k to the mean value of global radiation B k /Q was calculated. It turned out to be on the average 0.3 ±0.1. This value was accepted when calculating B k of different glacier regions on the basis of the climatologie data of Q. As a result applying formula (2)A S values for glaciers in different regions were calculated. ANALYSIS OF RESULTS It appeared that the calculated value of A s varies within quite a wide range even on a comparatively small territory. On the one hand, the variations were doubtless caused by the errors in calculation due to the error in formula (2) and the errors of

Glaciers as water resource indicators of the glacial areas of the USSR 27 calculation of its arguments t and B k. On the other hand, they are also conditioned by different concentrations in drift and avalanche snow on glaciers. Including the snow concentration coefficient in the analysis, dividing both parts of equation (4) by the annual solid precipitation sum X c, we find: «- l * 5 ^ - ^ (6) Assuming now that all snow from the nonglacial slopes near the nourishment limit is transported to the glacier by avalanches and blizzards, then F B K- F f B f (F is the basin area, Ff the area of the glacier, B the width of the basin, Bf the width of the glacier near the firn line). Now, K can be easily measured on a map or can be approximately estimated by using The glacier inventory of the USSR. The calculation indicated that the larger the glacier, the smaller is K. Its characteristic value for small cirque glaciers is 2-3, for average mountain-valley glaciers about 1.5, for large mountain-valley glaciers about 1. It seems that on glacier domes, flat summit glaciers and volcanic cones the snow concentration coefficient can be less than 1 because of snow drift transport (V= 0,D < 0). Nevertheless, that fact lias not been taken into consideration because of the absence of necessary information. Ignoring the value of evaporation, which is small near the nourishment limit, it is possible to say: ^A s = X c =Y m (7) (Y m is the value of melt runoff of the basin at the nourishment limit altitude). To reduce the sporadic errors of the calculation the values of Y m obtained on specific glaciers were averaged for glacier regions. Each region included a compact concentration of glaciers in orographically homogeneous territory. When the value of Y m clearly tended to 'an abrupt change in the territory (on different large-scale slopes, towards the centre of the mountains) smaller regions were chosen in order to characterize quantitatively these changes (Fig. 3). It is difficult to estimate objectively the accuracy of the data received. An experiment made to compare the calculated values of Y m with melt runoff values from representative glacier basins proved that the difference was within 20%. However, such an experimental verification is not completely independent because the data received from those basins were used in formula (2). Even expecting far less accuracy one can, however, make certain conclusions which are based on Y m calculation. In most cases the value of Y m turns out to be considerably larger (in some cases few times larger) than the quantity of solid precipitation as well as annual precipitation sums according to the data received from the meteorological stations nearest to the glaciers. This fact is to be kept in mind while using that information in different calculations. The contrast between specific water resources of glacier and the adjoining regions (especially plains) is rather great. Large-scale geographical regularity of Y m distribution in North Eurasia can be clearly traced on a map (Fig. 3). The maximum points are situated near the Iceland-Kara Sea and Aleutian low pressure areas. A large-scale geographical regularity of water resource distribution of glacial regions is clear. Within a certain mountain country of maritime or continental climate Y m diminishes rapidly from the periphery to the internal part along the direction of prevailing flow of atmospheric moisture. In conclusion I should like to point out that the indicative role of glaciers still is not used completely in the present work. Further details of water resource distribution in

28 V. G. Khodakov FIGURE 3. Map of calculated water resources of the glacier regions of the USSR and bordering countries (in the numerator value Y m, g/cm 2, in the denominator height of firn line, km). mountainous and high-latitude areas would be advisable if there are few or insufficient direct hydrometeorological observations of precipitation and runoff. The consideration of other elements of water balance like evaporation by 'penitent snow' distribution is apparently possible. REFERENCES A Guide of Glacier Inventory of the USSR (1966) Gidrometeoizdat, Leningrad. Ahlmann, H. W. (1948) Glaciological research on the North Atlantic coasts. London. Borzenkova, I. I. (1970) On the peculiarities of radiation regime of mountain regions. Works GCO, issue 263, Gidrometeoizdat, Leningrad. Chizhov, O. P. et al. (1968) The glaciation of Novaya Zemlya. Nauka, Moscow. Combined Heat and Water Balances of Selected Basins (1969) UNESCO/IAHS. Humboldt, A. F. (1948) Kosmos, P. 1-5. Kalesnik, S. V. (1963) Issues on glaciology. GIGL, Moscow. Khodakov, V. G. (1964) On possible error of precipitation measurement. Meteor, and hydrol, 6. Khodakov, V. G. (1965) Dependence of summed ablation of the glacier surface on air temperature. Meteor, and hydrol, 6. Koreisha, M. M. (1963) Present-day glaciation of the Suntar-khayata range. A. S. USSR, Moscow. Kotlyakov, V. M. (1968) The Earth snow cover and glaciers. Gidrometeoizdat, Leningrad. Krenke, A. N. and Khodakov, V. G. (1966) On relation of glacier surface melting and air temperature, MGI, 12. Krenke, A. N., Borovik, Z. S. and Rototayev, K. P. (1970) Snow accumulation on the glaciers of the Caucasus. NIGMI, 45, Gidrometeoizdat, Leningrad. Makarevitch, K. G. et al. (1969) The glaciation of the Zailii Alatau. Nauka, Moscow. Moiseeva, G. P. and Khodakov, V. G. (1971) On the calculation of annual surface ablation of glaciers and firn banks. MGI, 18. Preobrazhensky, V. S. (1960) Kodar glacier region (Transbaikalj. A. S. USSR, Moscow. Radiation Regime of the USSR (1961) Gidrometeoizdat, Leningrad.

Glaciers as water resource indicators of the glacial areas of the USSR 29 The Fedchenko Glacier (1962) A. S. Uzbek SSR, Tashkent. Ttte Glaciation of Elbrus (1968) Moscow University. Troitsky, L. S. et al. (1966) The glaciation of the Urals. Nauka, Moscow. Tronov, M. V. (1968) A representative mountain glacier basin of Aktrru in the Altai. MGI, 14.