Observational Results and Discussion

Fig. 4 shows longitudinal distributions of water temperature, SSC, and surplus density a in Tsho Rolpa, Imja and Lugge. It should be noted that the water of 4 °C is not located in the bottom layer, but at ca. 20 m depth of Tsho Rolpa, at ca. 6 m depth of Imja, and at the surface of Lugge. This looks very abnormal, since, if the water is clear, its density is maximum at 4°C. Most of the lake water exhibits water temperature of 1.5 to 5°C, which, if clear, gives the water density of 999.92 to 1000.00 kg/m under the air pressure of 1 atm. This means that, if SSC is more than 0.1 g/L, the bulk density pC of lake water is always more than 1000 kg/m3

for the particle density of 2,730 to 2,760 kg/m3. The SSC values in the three lakes range from 0.08 to 0.9 g/L. It is thus seen that the lake water density depends on SSC rather than water temperature. In fact, the SSC patterns in Fig. 4 are similar to the a patterns, though it is relatively unclear in Lugge, due to the rough isoplethic intervals of SSC and a. Hence, the density structure of the three lakes could be controlled by behaviors of suspended sediment supplied by glacier-melt or landslide from on the steep slope of side moraine. The suspended sediment in the lakes consists of more than 70 % inorganic clay particles (less than 4 ^m in diameter) [16]. Such fine particles have very small settling velocity of the 10- m/s order at water temperature of 1.5 to 5°C. The sediment rich water could thus induce the long stagnation of water mass at more than 0 °C, which is effective for ice-melt below the lake bottom over the year [2].

The isotherms' pattern in the upper layer of Tsho Rolpa (top of Fig. 4A) suggests that the surface water heated by solar radiation is transported toward the glacier terminus (glacier front), and then, is moved in the opposite direction after cooling by the glacial ice. Meanwhile, the distributions of SSC and temperature in the lower layer indicate that turbid and cold water intrudes into the bottom layer. This intrusion was probably produced by turbid meltwater inflow through an englacial tunnel at the base of the glacier terminus, followed by sediment-laden underflows [21, 22]. The surface layer of Tsho Rolpa is almost isopycnal (i.e., similar a values) at 0 to 25 m depths. This means the active vertical mixing by strong winds over the lake surface. The isotherms in Imja (top of Fig. 4B) indicate that warm water produced by solar radiation concentrates in the surface layer, and that cold water such as in Tsho Rolpa does not exist in the lower layer. This suggests weak wind mixing in the surface layer and no turbid meltwater inflow at the base of the glacier terminus. The low SSC in the whole Imja (middle of Fig. 4B) evidences no inflow of sediment rich meltwater. In Lugge, the heating of surface water is not prevalent, reflecting a thermal condition in the post-monsoon season. The patterns of isotherms and a in the surface layer suggests the weak wind mixing as in Imja. However, some of turbid meltwater inflow from the glacier terminus appears to occur both in the surface and bottom layers, since the SSC isopleths concentrate in the layers. The a patterns in Tsho Rolpa and Imja indicate that the lakes are pycnally stable by the sedimentladen underflow in the lower layer. However, Imja is almost neutral because of no turbid water inflow. The partly high a or SSC values at the water surfaces of Imja and Lugge mean that the sediment supply to the lake surface by meltwater input or landslide is active.

20

£

40

i

«0

tj

i

11X1

120

o

20

1

40

1

40

i

80

*

100

120

0

20

I

40

f

60

£

b

$

00

20

Suspended Sediment Concenlraljon ($!}

TTD4 TTD3 ülacicr Fn

1

4

» - s

W»tcr Temewratur® <°C)

-I

Suspended Sediment Concenlraljon ($!}

- --

Surplus Density o

Figure 4. Longitudinal distributions of water temperature, suspended sediment concentration (SSC) and surplus density a in (A) Tsho Rolpa (2 June 1996), (B) Imja (15 July 1997) and (C) Lugge (26 September 2002).

Fig. 5 shows time series of flow vectors and water temperature observed at 1 hr intervals at 25.4 m depth of site A in Tsho Rolpa, compared with wind vectors at site AWS (Fig. 1) [16]. As soon as valley winds prevail on daytime along the elongated lake surface, the flow velocity increases in the opposite direction with decreasing water temperature. This indicates that the countercurrents occurred near the pycnocline or the thermocline at ca. 25 m depth in order to compensate for the upper wind-driven currents flowing toward the glacier terminus (Fig. 4A). Meanwhile, the vertical water circulation built up by wind-driven currents and compensation currents is accompanied by the uplift of the lower cold water, thus decreasing water temperature [23]. In addition, the monitoring of flow velocity, water temperature and SSC at 1.5 m above the bottom at site G evidenced the generation of sediment-laden underflows from the base of the glacier terminus [16].

Figure 5. Hourly records of (A) flow vectors and (B) water temperature at 25.4 m depth of site A, compared with leeward wind vectors at site AWS [16].

The monitored results and the longitudinal structures in Fig .4A allowed us to introduce a conceptual model of lake currents related to heat transport (Fig. 6). The hydrodynamics of Tsho Rolpa consists of vertical water circulation in the upper layer above the thermocline, sediment-laden underflows in the bottom layer, and the intrusion in the middle layer from the interaction between the two currents near the glacier terminus. The calving at the glacier terminus could be facilitated by the active dispersal of cooled water in the downlake direction, which is produced by the interaction between sediment-laden underflows and wind-driven vertical circulation. The strong southeastern valley wind produces the setup by wind-driven currents toward the glacier terminus. During the setup, the thermocline is lifted up at the downlake of a nodal line, and lowered at the uplake of the nodal line.

Figure 6. Conceptual model of the hydrodynamics of Tsho Rolpa related to the heat dispersal [16]. The white circle shows the location of the monitoring at site A (Fig. 5). The black circle indicates a nodal line with no oscillation of thermocline or pycnocline during the setup toward the glacier terminus.

Figure 6. Conceptual model of the hydrodynamics of Tsho Rolpa related to the heat dispersal [16]. The white circle shows the location of the monitoring at site A (Fig. 5). The black circle indicates a nodal line with no oscillation of thermocline or pycnocline during the setup toward the glacier terminus.

Fig. 7 shows temporal variations of wind vectors observed at site AWS (Fig. 1). A wind system is common to the three lakes, i.e., being independent of rainfall or non-rainfall, valley winds prevail on daytime along the elongated lake surface (daily maximums of 3.6 to 8.6 m/s), but, at night, no wind or weak mountain winds occur. The wind at Lugge was observed in the post-monsoon season, and at Imja and Tsho Rolpa, in the pre-monsoon season. However, the wind system at Lugge is consistent in the pre-monsoon, monsoon and post-monsoon seasons (personal communication with Dr. N. Naito in 2004), though the wind at Lugge in Fig. 7 is relatively weak. As a feature of the terrain around the meteorological station of Tsho Rolpa, the crest of the end moraine near the station is only at 2 to 3 m above the lake level (Fig. 2B). Thus, the wind record directly exhibits wind speed and direction at near the water surface. In Imja and Lugge, the dead-ice zone is still developed at downlake at 10 to 25 m above the water level of Imja and at ca. 28 m above the water surface of Lugge (Fig. 1). The meteorological stations at the two lakes are set at near the level of the dead-ice zone. Thus, the wind records probably do not represent wind speed and direction at near the lake surface, since the relatively high dead-ice zone at downlake could obstruct the ventilation of the valley wind. The weak wind mixing in Imja and Lugge (Fig. 4B and 4C) is probably due to the attenuation of valley winds by the topographic screening effect of the dead-ice zone. The topographic effect possibly depends on the lake surface area in the downwind direction. Such a topographic effect is demonstrated by numerical simulation in the next section.

Wind Vector (leeward)

Wind Vector (leeward)

II 12 13 14 15 16 17 IS 19 20 21

September 2002 October 2002

Figure 7. Leeward wind vectors recorded at site AWS of (A) Tsho Rolpa, (B) Imja and (C) Lugge (Fig. 1). Outlines of the lake surface and location (black circles) of meteorological stations are shown.

September 2002 October 2002

Figure 7. Leeward wind vectors recorded at site AWS of (A) Tsho Rolpa, (B) Imja and (C) Lugge (Fig. 1). Outlines of the lake surface and location (black circles) of meteorological stations are shown.

Was this article helpful?

0 0

Post a comment