GROUP D FINAL PROJECT

Group D Final Project

THE EFFECT OF SNOW MELT BLOCKS ON SEDIMENT ENTRAINMENT IN A SUPRAGLACIAL STREAM

Sarah Holah

Andrew Hudson

Miriam Lehmann

Adam Mellor

Alex Miles

Deborah Southwell

Richard Taylor

Georgina Turner

Rhys Williams

with assistance from

Andy Evans

Aim and Objectives

To assess the effect that snow block erosion has on the entrainment of river sediment in a supraglacial stream on the Ödenwinkelkees glacier.

  • To assess the rate and nature of snow melt
  • Use theoretical shear stress to calculate theoretical sediment size able to be entrained and compare this value to actual sediment sizes found
  • Estimate the sediment entrainment occurring due to snow block melt throughout the total melt season

Hypothesis

1. Sediment entrainment will increase with the addition of snow blocks.

2. The theoretical sediment size able to be entrained, and the actual sediment found to be entrained, will be comparable.

3. An estimation of annual entrainment resulting from snow block erosion will be possible

Location

Located 1.5 kilometres from the Rudolfshütte in Hohe Tauern, Austria, the Ödenwinkelkees (see figure one), a corrie glacier, is approximately 2 kilometres long. A large proportion of the ablation zone is covered by sediment from the surrounding topography. The supraglacial stream is located 500 metres away from the glacier terminus.

Figure 1: Map of study area

Background and Context

Glacial river systems are an important resource in Alpine environments, providing inhabitants with water for everyday use such as drinking, and also electricity through hydroelectric power (HEP) schemes, making them valuable not only domestically but also commercially. However, a major problem associated with HEP schemes is sedimentation within the dams, which can restrict water storage capacity and movement through generator systems. Supraglacial streams contribute to powering the HEP schemes, but also act as a sediment source as they collect and transport material in their flow. In order to address the problem of dam sedimentation it is thus essential to understand the behaviour of, and influences on, sediment entrainment within supraglacial streams.

Adenlof & Wohl (1994) found that in the lower regions of their alpine study area that woody debris acted as ‘a major local control on temporal and spatial patterns of bedload movement’. In higher, glacial regions, where vegetation becomes increasingly scarce, large snow blocks could potentially act as a substitute to woody debris, having similar effects on sediment distribution patterns. Snow blocks entering the stream as a result of ablation could alter entrainment patterns, making it important to gain an increased understanding of the relationship between snow block erosion and stream sediment to enable researchers to plan strategies to combat dam sedimentation.

There are two mechanisms by which a snow filled channel will clear throughout the melt season. These are melting and snow blocks falling into the stream. Benn and Evans (1998) explains the importance and impact of snow blocks stating that supraglacial streams will ‘enlarge rapidly by the entrainment of rafts of snow from the banks’.

Supraglacial streams form through the process of percolation of meltwater through the snow. As the water drains through it will form rills that join to form a drainage network. The low permeability of glacial ice (Benn & Evans, 1998) coupled with a high melt rate leads to a large accumulation of surface water that if located on a slope will flow forming surface channels by entraining surrounding snow banks. As more of the surrounding snow banks are eroded into the channel they will melt, increasing the flow discharge. This process will be more prevalent during peak times of ablation such as just after midday when insolation is greatest. As the discharge increases, the stream will have a higher capacity to carry bed load as well as the ability to entrain larger sediment. This is provided by greater stream energy which is able to overcome the frictional forces preventing the sediment from being picked up by the flow. A greater frequency of impact events helps to dislodge more sediment leading to an overall increase in bed load. This will mean more sediment will be carried down the glacier into proglacial streams causing braiding and further implications on sedimentation in water supply and Hydro Electric Power stations.

Snow is added to the channel either as meltwater or by snow blocks directly falling in. Snowmelt enters the channel by downward percolation through the air spaces in the snow. Deeper layers of snow are more compacted, their internal air spaces reduce making percolation of water more difficult. This results in less water reaching the channel through englacial flow, typical of late in the ablation season when the overlying fresh snow from the accumulation season will have melted leaving the underlying firn and ice from previous accumulation seasons. Typically the density of fresh snow is 50-200kg/m-3 and old compacted snow is 400-830kg/m-3 (Paterson, 1994, p9). The greater density for old snow means that when left unsupported, such as when underlying ice is undercut by the flow of the stream, it is more likely to fall as one large block. The addition of a large block of snow to the channel will cause huge turbulence in the flow, this will dissipate large forces across the different boundary layers. These forces will result in a larger shear stress being applied to the sediment found on the wetted perimeter than during normal flow conditions (“normal” being when no large blocks of snow are suddenly added). With greater shear stress it will mean larger particles with a heavier mass will be entrained by the stream as their critical shear stress value (the minimum stress required to overcome the resisting forces that hold the material together) (Benn & Evans, 1998, p152) would have been reached.

The addition of large blocks of snow to channels can cause pulses in suspended sediment, known as “slugs”. These occur independently of changes in discharge and reflect a sudden release of sediment previously stored in the now collapsed snow/ice channel banks.

The density of streams decreases up-glacier due to decreased melting and water production with increasing altitude as air temperatures are lower. This would imply that the ability for the supraglacial stream to entrain sediment would increase downstream.

Methodology

To successfully test the hypotheses the following data collection methods were employed.

1. Channel Survey

The length and depth of the supraglacial stream was measured to put the entire system into context and to extrapolate our results for the entire channel. Pools and riffles were identified and mapped (see figure two).

Figure 2: Annotated Diagram of the Channel Profile

2. From the channel survey we selected a suitable riffle-pool sequence, based on its accessibility and measured the entire length of the sequence and the pool riffle sections.

Cross sections of both were taken to give a more detailed insight into the sampling area and to allow calculation of the discharge in the supraglacial river channel. Every hour the velocity and depth of the flow were measured which would allow further analysis of the stream discharge and changes over time.

3. To analyse snow block erosion, it was decided to compare the sediment movement in the river in a “normal” state whereby the river was flowing without snow blocks entering the channel. This could then be compared with a measured block of snow being added to the flow at the top of the riffle-pool sequence.

  • Every hour, four “blank” (runs without a snow block added) and four snow block runs were completed.
  • Before the snow block was added the dimensions were measured as can be seen in figure 3a.
  • To allow for a more thorough comparison the blank and snow runs were alternated.
  • For each run the sediment was caught in a suber net, located at the end of the riffle-pool sequence, for 30 seconds.
  • For the snow run the time that it took the snow block to travel was also recorded
  • The mass of the sediment collected from all the runs was recorded.
  • Any sediment above 20mm in diameter was identified and the dimensions measured.

Figure 3a: Measuring the dimensions of snow blocks

4. Observations were made of the snow melt characteristics both along the river channel and at the snow to water transition upstream (see figure 3b). This included the nature of the snow melt, the rate of melting and the size of snow blocks entering the stream through natural processes, this would give us an estimation of the amount of snow melt entering the supra-glacial stream throughout the day. This information could then be used to analyse the general trend of the ablation season.

Figure 3b: Observing snow melt characteristics

In addition to this the average snow density was measured for the entire channel. A “snow block run” was carried out where the snow block was added to a measured section of the channel and its size measured every 9.5 metres to assess how quickly the ice melted.

5. The sediment in a downstream pool was measured, a different site was selected to avoid disrupting the riffle-pool sampling area, to assess the range of sediment size and amount.

Limitations

There were several limitations involved in this study. The suber-net used to collect the eroded sediment was rectangular in shape. The channel, however, was u-shaped and therefore some sediment would have flown underneath the net instead of into the net.

Also, it was difficult to measure little cornices without breaking them. As the block of snow was first measured and then thrown downstream, hands touching it could have melted the snow, thus affecting the results.

Another limitation was the spring-balance because it only measured up to 500 grams. On one of the snow runs a rock was caught in the net that was heavier than 500 grams and therefore had to be classified as 500+.

Walking and standing close to the channel caused smaller blocks of snow to fall into the water, this may have affected the amount of sediment caught in the net.

Due to time restrictions it was not viable to collect a whole data set throughout the full melt period of the day. As a result of the blocks of snow flowing downstream a lot of sediment was transported in this direction and collected. There was, however, only a limited amount of sediment in the pool. It has been assumed that it would be replaced from upstream locations, but the disturbance caused would not see an instant recovery to natural levels.

Results

Figure 4 shows the mass of sediment collected every 30 seconds from the supraglacial stream in normal flow (black line), and with a block of snow added (red line). We clearly see that the mass of sediment is much higher when the snow block has been added.

p value = 0.005

Figure 5 shows that as average discharge rises the amount of sediment in the blank run increases. This is due to the fact that higher discharges carry larger amounts of sediment. A P value of 0.005 shows that the results are significant and we can accept that sediment levels increase with discharge.

p value = 0.080

Figure 6 shows an inverse relationship with discharge and mass of sediment during the snow run (i.e. snow blocks added). This indicates that the rising sediment during the snow run is not due to discharge but due to the addition of the snow. However, the p value of 0.080 shows that we can’t accept the results statistically, although the trend appears obvious.

Figure 7 shows the mass of sediment VS the mass of the block. A weak positive correlation can be seen, as block mass increases so does the mass of sediment.

On average it can be seen that a 1kg snow block will cause 0.565kg of sediment entrainment. It was calculated that 10.156kg mass of blocks fell in the melt period of one day and therefore it could be suggested that there is 5.763kg sediment entrained in one day (0.565 x 10.156 = 5.763). Although some snow blocks will pass through more than one pool it was found that snow blocks either break up or become lodged after approximately 20m. A snow block will therefore not pass through more than two pools and sediment entrainment in subsequent pools will not occur to any significant level. As a snow block is passing through two pools, the sediment entrained in one day should be doubled, due to the calculated value being for one block passing through a single pool. The total entrained in one day will be 11.526kg

The melt season typically occurs from May to July (60 days). There is little opportunity at the start of the ablation season for snow blocks to fall as there is no defined channel as it is snow filled. Therefore to calculate the sediment entrainment caused by snow blocks entering the channel a period of 45 days will be used as this is when the appropriate channel conditions for snow block occurring exist towards the latter half of the ablation season.

Therefore:

45 x 11.526 = 518.67kg in a melt season.

Theory and Discussion

The theoretical shear stress can be calculated and this can then be compared to the actual shear stress found in the supraglacial stream

Shear stress (τw) is found by multiplying the mass of the object acting, by the angle of slope that it is travelling down. The force exerted by the front of the snow block is demonstrated by the diagram and equation below.

Note- Buoyancy effects have been taken into account by using a reduced mass – the mass under water only

There is a shear stress between the stream bed and the bottom of the ice block. This is calculated by…

The contact area between the snow block and any sediment in the channel needs to be known in order to calculate the shear stress being applied by the block.

hu=Height of ice under water

hs=Height of bed rock

hw=Height of water

hsi=Height of sediment hit

hsw=Height of sediment not hit

The shear stress applied by the snow block is calculated by the force of the snow block divided by the area that it hits. The following equation is used to calculate the total shear stress applied by the blocks.

height of water / 0.3
height of ice block / 0.25
density of water / 1000
density of ice / 800
gravity / 9.8
slope / 7
thickness of sediment / 0.25
density of rock / 2500
Height of ice in water / 0.2
height of ice not under water / 0.05
height of water under ice / 0.1
height of sediment not hit / 0.1
height of sediment hit / 0.15
force from block / 1071.875
stress from block on hit area / 7145.833
shear stress from block / 2744

Therefore τi = 2744

The amount of sediment that is entrained by the snow block depends on its angle of friction.

Shear stress = mass x tangent of angle of friction. As the shear stress is known, a range of friction angles can be used to calculate the range of masses, and volumes, of sediment expected to be entrained due to the snow block. See table 1

Angle of friction / volume of sediment grain that can be entrained
10 / 6078.9052
20 / 2944.952359
30 / 1856.541959
40 / 1277.410882
50 / 899.4099172
60 / 618.8473198
70 / 390.1305949
80 / 189.0004824

Table 1

The volume of actual sediment grains entrained from various runs is displayed below.

Actual Entrainment volumes
656
875
450
1250
480
600
294
240

The average volume of the actual entrained sediment was 605.6cm3. This fits in well in the range of predicted entrained sediment volumes and it therefore can be said that actual and predicted entrained sediment volumes are comparable.

Snow block melt characteristics and observations

To date, the upper region of the supraglacial channel is still snow filled, although approximately 10% of the stream as a whole is at this state. Where melting of snow occurs the surface edge is smooth and pointed. Where snow block detachment has occurred surface edge is vertical and sheared. It was observed that where the channel was snow filled, melting occurred. Snow block detachment occurred in the rest of the channel almost entirely in the outside bend of a meander. It was observed that the snow blocks that became detached were on average 1/5 – ¼ of the length of the snow shelf. It could also be proposed that after a snow block detaches the succession of similar snow blocks is quicker than if no block had detached initially. Once a detached snow block enters the stream approximately 60% break within 1-2 metres. It was also very common for them to become caught on sediment banks, limiting their entrainment potential.

Conclusion

The aim of this investigation was to assess the effect that snow block erosion has on the entrainment of sediment in supraglacial streams. After analysing the rate and nature of snowmelt sediment entrainment, we have been able to successfully prove the first hypothesis.This states that sediment entrainment will increase with the addition of snow blocks.

The theoretical shear stress calculation supported the results of the actual shear stress values which were obtained.

The conclusions of the previous two hypotheses allowed our third hypothesis to be tested, and the results obtained showed that the actual entrainment resulting from the erosion of snow blocks was able to be realistically estimated.

In conclusion, it can be said that the aims and objectives of this investigation were suitable in answering our hypotheses. Whilst there were certain limitations in our study, our results do appear robust enough to be used in future studies and further research, which will be essential to increase our knowledge of the subject area.

Bibliography & References

Adenlof, K.A. & Wohl, E.E. (1994). Controls on bedload movement in a sub-alpine stream of the Colorado rocky-mountains, USA. Arctic & Alpine Research. 26(1), pp.77-85