Bashkara glacial lakes (North Caucasus, Russia) and outburst hazard

The Bashkara group of glacial lakes with a total area of 93,000 m², and a total volume over 900,000 m³ is located in the upper part of Adyl-Su basin at the margin of Bashkara glacier (Fig.1).

Figure 1. Bashkara lake group. 1 � Bashkara Glacier, 2- Lake Bashkara, 3 � Lake Lapa, 4 � Lake Mizinchik. Photo by V.V. Krylenko.

This valley-type glacier is roughly 4 km in length and has an area of about 3 km². The main branch of the glacier flows from Mt. Ullukara (4302 m a.s.l.) to the north and turns to the northwest at the snout. A secondary branch issues from a mountain pass between Mt. Bashkara (4164 m a.s.l.) and Mt. Dzhantugan (4012 m a.s.l.). This branch descends towards the northwest and flows into the main branch in the upper part of the snout. During the Little Ice Age (LIA) the glacier snout was divided in two parts as a result of the main Ullukara branch pushing the Bashkara branch out to the right side of the valley. As a result, a moraine ridge loop was created. In late 1940s - early 1950s Lake Bashkara started to form inside the loop due to the stagnant ice melt. The lake was not shown on a schematic map of the Bashkara glacier in 1933 (Oreshnikova 1936), and did not yet exist in 1946 (E.A. Zolotarev, pers. comm.). However, it was shown on a 1957 schematic map (Dubinsky & Snegur 1961). At that time the lake area was two times less than today. A second smaller lake downstream of the glacier snout is also mentioned in the literature. In August 1958 and 1959 Lake Bashkara burst through its ice dam. After these GLOFs the level of the lake was lowered by two meters. The smaller downstream lake was also involved in the flood. As a result catastrophic debris flows totalling 2 million m³ were formed from the moraine deposits and travelled 12 km downstream. The initial water impulse was only 60,000 m³ (Seinova & Zolotarev 2001).

In the end of the 1980s new small lakes formed south of the drained downstream lake by the edge of the retreating Bashkara glacier (Seinova & Zolotarev 2001). In 1991-2001 these lakes grew in area and volume (Chernomorets et al. 2003) but, in our opinion, their outburst was unlikely due to small volume of the lakes and relatively good surface drainage. The larger eastern lake was named Lake Lapa and the smaller western lake was named Lake Mizinchik (Fig. 1). Lake Bashkara has subglacial/englacial drainage which goes towards these downstream lakes. Lakes Lapa and Mizinchik are dammed by a low ice-cored moraine. Should the drainage channel of Lake Bashkara become cluttered, a catastrophic GLOF involving all three lakes may occur. It is likely to erode downstream moraine and debris flow deposits and transform into a debris flow. The situation becomes more dangerous each year. Buildings, bridges and camping sites may be damaged.

Lake level dynamics

Water level of a glacial lake is one of the most important indicators of short-term outburst hazard. Lake Lapa is drained by a well-developed runoff channel, so the water level variations are not large. During summer the range of variations is about 10-15 cm, while the seasonal range does not exceed 30 cm. The interannual lake level is also stable.

For Lake Bashkara the situation is different. We conducted detailed field investigations to identify the influence of ablation and rainfall on lake level fluctuations. The seasonal range of levels is up to 150-200 cm (Fig. 3), with the highest level usually observed in the beginning of the warm period. In late June-early July the level goes slightly down and remains quasi-stable until late summer. In the last week of August to first week of September lake level sinks rapidly and we observe its minima at the end of the ablation period. There is no surface drainage channel from Lake Bashkara. Its water filters through englacial channels and cavities in the stagnant part of Bashkara glacier snout, which is covered by abundant surface debris. For this reason, in our opinion, lake level fluctuations are driven both by meteo-glaciological factors (air temperature, precipitation, snow and ice melting) and by water exchange in the lake-glacier system, which includes such processes as water accumulation in the glacier, filling of englacial cavities with water at the beginning of the ablation season and water release from the glacier in a later period. This may be a mechanism similar to that on South Cascade Glacier (USA) (Tangborn et al. 1975), which, like Bashkara, is a warm glacier, and where meltwater accumulated within glacier early during the ablation season and later released.�

In 1999-2005 we observed a clear tendency for an interannual rise of Lake Bashkara (Fig. 2). Up to 2002 the level rose insignificantly, but in the cool and rainy summer of 2003 the lake level was 50 cm higher than in 2002. This trend continued and in 2005 the level was about 100 cm higher than five years earlier. We hypothesize that this tendency is a result of changes in the englacial and subglacial drainage system.

Figure 2. Lake Bashkara water level dynamics during warm periods in 1999 �2005. Zero level refers to the zero mark on the permanent level gauge.

Changes in lake areas and volumes

We observe that Lake Bashkara is quasi-stable whereas Lake Lapa is drastically expanding. Area and volume of Lake Bashkara are driven mainly by level fluctuations and vary within 60,000-70,000 m² and 675,000-800,000 m³. Below the zero mark on the permanent level gauge, the volume varies annually within a range of 750,000-790,000�m³ which, we note, is within the measurement accuracy. The maximum measured depth (34 m) in the centre of the lake (Fig. 3). is also stable, so the lake bed relief is assumed stable during the observation period.

Figure 3. Bathymetric map of Lake Bashkara.

Dynamics of Lake Lapa area and volume is presented in Fig. 4a. In 2001-2006 the lake area increased three-fold and the volume multiplied by a factor of five. The process is driven by the glacier terminus retreating from the eastern lake shore. The eastern part of the lake has maximum depths and features the most rapid shore recession (Fig. 4b & 5). The most intensive volume expansion was observed in 2002-2003 due to a depth increase and in 2003-2004 due to a notable enlargement of deep areas. In 2005 volumetric increase slowed due to depth stabilization, but in 2006 it accelerated as the lake merged with a very large thermokarst hole in the glacier. Average lake depth has been reduced due to sedimentation on the lake bed. The maximum depth increased two-fold in 2001-2002, was quasi-stable from 2003-2005 and decreased by 1 m in 2005-2006.�

a������ ����������������������������������������������������������� ����������b

Figure 4. Lake Lapa: a - lake area and volume in 2001-2006; b - bathymetric map in 2006.

In 2001 Lake Mizinchik was comparable to Lake Lapa in area and volume. In summer 2001 transformation of the Bashkara river outlet initiated rapid sedimentation in the lake. Part of the glacier terminus adjacent to Lake Mizinchik has been stable in recent years. As a result the lake in 2004 was only 25% of its size in 2001. In 2005 we could not perform depth sounding of Lake Mizinchik because it was too shallow for measurements. By 2006 Lake Mizinchik had practically disappeared (Fig. 5).

In our opinion, lakes Lapa and Mizinchik are a striking instance of proglacial lake evolution. In the case of stable glacier termini we often observe reduction in lake depths and debris flow hazard. Retreating glacier termini cause lakes to expand (similar cases are presented by Kattelmann 2003, Richardson & Reynolds 2000, O�Connor & Costa 1993) and possibly link to subglacial relief features. In this case lake volume and outburst hazard increase, but at some stage lake depths stop increasing. After this two possible scenarios may occur: GLOF or filling of the lake by sediments.

Development of the Bashkara lake group is closely connected to the Bashkara glacier dynamics. The right part of the snout experiences rapid degradation, terminus retreat, and decrease of ice flow velocities down to zero. As a result active thermokarst processes are observed here. According to our repeat survey data, the surface of the glacier snout lowered by about 10 m on average from 1999-2005. In early summer 2004 a giant thermokarst hole about 40m deep formed between Lake Bashkara and Lake Lapa about 100 m from Lake Lapa shore. In 2005 this hole was about 50 m deep and about 80 m in diameter. In 2006 it became a gulf of Lake Lapa (Fig. 5f). There are many thermokarst depressions between the lakes, so we predict subsequent narrowing of the ice dam between Lake Bashkara and Lake Lapa. In 2004 the dam width was about 500 m, by 2006 it reduced to about 250 m. This signifies a significant increase of the debris-flow hazard for downstream areas. Bridges in the Adyl-Su River valley, two hotels, and a large camping site hosting up to 500 tourists in summer are all in the danger area. In case of a particularly disastrous outburst, part of the Elbrus settlement with several thousand inhabitants, as well as two bridges on a federal road may be damaged.

Figure 5. Changes of Lake Lapa and Lake Mizinchik as recorded by repeat geodetic surveys. Lake coastlines: a - in 2001; b - in 2001-2002; c - in 2002-2003; d - in 2003-2004; e - in 2004-2005; f � overall in 2001-2006.� Coastlines are shown with solid lines for a later date and with dashed lines for an earlier date.�1 � Lake Mizinchik, 2 � Lake Lapa. Coordinates are given in metres on a local grid.

Possible scenario and consequences of lake outburst

The condition of the ice and moraine dams in the Bashkara area is constantly changing. This makes it difficult to define the scenario of a possible outburst and to calculate its parameters. To assess features of a possible outburst flood, as well as the discharge and volume of the resultant debris flow we applied the method of geographical analogy. We reviewed descriptions of ten GLOFs in Northern Tian Shan and Dzhungarskiy Alatau mountains published in Russian journals Selevye potoki (Debris flows) and Meteorologia i Gidrologiya (Meteorology and Hydrology). The case which is likely similar to the Bashkara lake group was an outburst of a lake group in the vicinity of the Tushinskiy Glacier of the Dzhungarskiy Alatau range on 8 September 1982 (Tikhomirov & Shevyrtalov 1985). The volume ratio between the large upper lake and the two small lower lakes, and the longitudinal profile of the valley were similar to the Bashkara group. The event also happened on the background of glacier recession. In late August � early September the lake drainage channel was blocked due to ice failure. On 8 September 1982 a break-up of the ice plug led to the lake outburst. The level of the upper lake dropped by 3.5 m, the lower lakes were overrun and their dams rapidly eroded. The outburst had a peak discharge of about 150 m³/s, which continued for 1.5 hours. Within 1100 m downstream the flood transformed into a debris flow having a peak discharge of about 290 m³/s. When the debris flow reached a steep moraine terrace (up to 12.5 km path distance from the lakes) it transformed into a disastrous mud-debris flow having a discharge of about 2400 m³/s and a volume of 2.7 million m³.

A similar outburst mechanism is likely for the Bashkara lake group. Subglacial drainage channels are in an unstable condition due to thermokarst processes. In case of a channel blockage water will fill englacial cavities in stagnant ice which will then burst. Water release may rapidly lower Lake Bashkara by 5 to 10 m and Lake Lapa may be involved in the resulting GLOF. As a result up to 500,000 m³ of water may form a flood downstream. Even if the debris dam of Lake Lapa is overrun, flood discharge is not likely to exceed 150 m³/s. Most likely the GLOF will transform into a debris flow in the area of steeply sloping LIA moraines, as observed during previous bursts in late 1950s.