Origins and effects of sediment on the Ödelwinkelkees Glacier
Austrian Fieldtrip
Höhe Tauern Alps
5th – 13th July 2004
Group A
Laura Adams, Elizabeth Angood, Ian Atkin, Michael Brown, David Fouracre, Timothy Henman, Benjamin King, Alexander Maddocks, Cathy Magnus, Andrew McGarity
Contents
Abstract
Introduction
Background research
Aims and hypotheses
Methodology
Results
Analysis
Discussion
Limitations
Further study
Conclusion
Appendices
References
Abstract
This investigation aimed to find the origin of the sediment on the surface of the Odelwinkelkees Glacier and the differential ablation rates related to this sediment. Fieldwork was carried out in order to assess the source of the sediment as well as to find the critical angle of dirt cone failure and to map the glacial surface for use in an ablation model. It was found that the source of the sediment was from the valley sides which had later been re-worked by fluvial processes. The critical angle was found to be 43º and using the ablation model the effects of sediment on ablation rates were found to be inconclusive.
Introduction
This investigation was carried out on the Ödelwinkelkees Glacier located in the Höhe Tauern Alps, Austria (See Fig. 1).
Figure 1: Aerial photograph showing the Ödelwinkelkees glacier and surrounding area
This is a valley glacier which is currently experiencing ablation i.e. a net loss in mass due to melt and is typical of this type of glaciers in that there is a large moraine accumulation at the snout (approximately 50% of the ice surface is covered in sediment). The snout of the glacier is approximately 2170m above sea level. It has steep valley walls and a braided proglacial stream network. The sediment distribution varies widely across the glacier with visible tiger stripes of fine sediment in places, as well as large, thick piles of debris.
The combination of moraine and ice provides a useful insight into the differential melt rates on the glacier surface and therefore it is well adapted for modelling. Therefore this glacier is an ideal study site to investigate the effects of sediment on ablation rates and to enable future predictions to be made.
In order to investigate ablation rates relating to the sediment on the glacial snout, it is necessary to determine the current glacial surface to enable expected changes to be modelled.
Aims
1. To investigate from where the sediment on the glacier surface originated.
2. To discover if a critical angle for dirt cone failure exists and if so, determine its value.
3. To determine the effect of sediment on melt rate using the Evans model.
Hypotheses
1. There is a critical angle for dirt cone failure; this angle remains the same irrespective of other factors.
2. Areas covered in thick layers of sediment will experience less ablation than bare ice/snow areas, whereas areas with thin layers of sediment will experience heightened ablation.
Background Information
Sediment on the glacial surface, and more specifically the snout of the glacier has a significant affect upon glacial melt rates. A key factor is the thickness of the sediment. With thin sediment layers, due to sediment’s lower albedo, more insolation is absorbed than the surrounding snow and ice because its thermal conductivity is greater. Therefore, there is more energy for sediment to emit than ice and this causes a greater rate of melting (Benn and Evans, 1998). However, if the sediment is too thick then solar energy is unable to reach the underlying ice and thus the sediment acts as an insulator.
This results in enhanced ablation with respect to the bare snow and ice surfaces because long wave radiation is emitted by the sediment throughout the night. With thick sediment layers, this effect is negated by the insulating effect of the debris layer. The result is lower ablation in areas covered by significant amounts of sediment.
Therefore the behaviour of sediment is determined by its thickness, where it can either cause or prevent melting of the ice below it. At a certain value, known as the ‘critical thickness’, this change in behaviour occurs where the sediment prevents the ice from melting and begins to insulate it. According to Benn and Evans (1998), it is at a value located between 5 and 10mm, which is the critical threshold at which melting is reduced. For instance, small particles will sink into the ice due to the melting around them, whereas large particles will simply rest on the glacier surface.
Drewry (1972, cited Benn and Evans, 1998) observed that melt rates of clean ice were considerably greater than melt rates of debris covered ice over a 26 day period. Pelto (2004) also observed that ablation rates are significantly reduced under debris cover. Studying the Lymann and Columbia glacier, he observed that melting was reduced by 25-30%. On the Columbia glacier 3.3m w.e. (water equivalent) of melt was observed on the clean ice, which is greater than 2.3m w.e. of melt for the debris covered areas.
Dirt cones are cones of ice covered in a layer of sediment, and can be used as an indicator of the critical angle of sediment failure as they provide easily measurable failure faces. The critical angle is the point at which a sediment layer can no longer support itself and therefore slips down slope. Their method of formation lends itself to propagation of cones across the glacier surface. A thin layer of dirt warms the underlying ice, as a result of the high albedo of the debris in relation to bare ice. This warming forms a hollow on the ice surface, into which more debris is transported by wind. When the debris layer reaches the critical thickness, it begins to insulate the ice from insolation, rather than enhancing melting. This means that the debris covered area ablates at a lesser rate than the surrounding ice, forming a cone above the ice surface.
The cone will become progressively steeper as the surrounding ice melts. Eventually the cone will reach a point at which the debris covering will slide off. This debris collects at the bottom of the cone, serving to protect a new area of ice, propagating cones across the ice surface. The process is repeated as long as ablation is the dominant process on the glacier.
It is also useful to understand from where the sediment originates. It would be expected that sediment on the surface of the glacier would be from either the valley sides due to frost shatter erosion, streams flowing onto the glacier, subglacial material brought to the surface by ice movements within the glacier, or fine wind-blown sediment.
Once on the glacier, the sediment tends to accumulate at the snout due to a number of processes within the glacier. Sediment can be carried to the front of the glacier and revealed on the surface by melting. The movement of the glacier itself can also bring sediment to the surface by creating shear zones. This can lead to fractures in the glacier, allowing ice to ride up over dead ice at the front, depositing the subglacial and englacial sediment on the surface. Englacial streams also act to bring sediment to the surface.
These processes all result in the sediment accumulating at the snout due to the fact that at the snout, the ice is moving slower due to often lying in a basin, being thinner (having less momentum), and being buttressed by moraines. All these factors lead to the accumulation of sediment in the ablation zone. This results in varied melt rates which are to be investigated and modelled. Varying transportation paths of the sediment affect the wear on the particles, and hence the sediment morphology. This is useful in identifying the origin of the sediment (Boulton, 1978).
In order to model ablation, it is essential to have an idea of the ice surface height and the depth of sediment on the surface. These measurements were taken and used as input data in an ablation model. The model was run for the Ödelwinkelkees glacier, which was chosen as a good example of a valley glacier in an accessible area.
Methodology
In order to carry out the survey of an area on the Ödelwinkelkees glacier we first had to choose our area, size and location, and then mark this area out so that it was easily visible. After a brief discussion a decision was made to create a survey area of 75 metres square because we believed that this was a sufficient area to provide us with an accurate representation of the glacier surface. This was subsequently divided into 5 metre squares. The location of the survey area itself was chosen as it covered areas of snow, sediment and dirt cones, which helped us to achieve our aims. With the size and location of the survey area determined the next task was to actually mark out the area to ensure the accuracy of our readings and was therefore useful when running the model. The technique to mark out this area was simple; a starting point was positioned at the centre of the baseline and, using a 30 metre tape measure, 40 metres was mapped out to the west and a subsequent 35 metres to the east. In addition to this every 5 metres a stone cairn was constructed to delineate the survey grid.
Due to the orientation of the glacier a decision was made to map the sides running parallel to the glacier at 160° and perpendicular to the glacier sides at 240° thus giving the square area desired.
For the edges of this area we ensured that the compass bearings were correct and started to map the remaining sides. We again built cairns every 5 metres. However larger cairns were built at 20, 40 and 60 metres. The area was then completed in the same way as the sides and base were established. To make measuring easier a line of cairns was plotted down the centre of the grid as well as one across the grid at 40 metres. Once completed we had the appropriate grid needed to conduct our survey.
- Sediment origins
Ten sample sites were chosen, at regular 5m intervals along the pre-set grid. At each site the top 3 inches of material were examined (approximately 20 stones). This involved taking the following measurements for each stone:
- Longest length
- Width
- Depth
- Sharpest angle
- Roughness
This was done using a tape measure, compass and pre-defined roughness scale (see Appendix 1). The percentage cover of fine sediment, e.g. clay, sand or silt was also estimated at each of the sites.
These field measurements were later used to calculate:
- Roundness = width/length
- Shape = height/length
- Sphericity = depth/length
These results were then compared to graphs of known sediment origin (taken from Benn and Evans – See Appendices 2 & 3) to determine possible origins of our data.
2. Dirt Cones
The dimensions of a number of cones were measured, including their width, length, and steepest angle. The length was taken as the axis aligned with glacier flow direction, and their width as perpendicular to this. The angle was measured by lying an ice axe on the steepest slope, and measuring this angle with a clinometer. The height was calculated using trigonometry. The sediment depth was measured at the apex of each cone, by removing debris until the ice core was revealed. A qualitative measurement of sediment type was taken, with two categories: mixed and fine sediment. It was also noted whether the debris cover on each cone had failed, or whether it was judged to be near failing. This was possible as some shifting of the sediment was visible, but complete failure had not yet occurred.
Our sampling strategy measured every available dirt cone. There were some which were located in highly crevassed and unstable snow areas, which were not accessible, and some which were too large to effectively measure. The steepest angle of the larger cones was still measured, where it was safe to do so. In total twelve cones were measured, with the first seven being small enough to complete full length, width and height measurements.
3. Glacial Survey
For the survey, two pairs of students of similar height surveyed opposite sides of the grid. A member from each of these pairs stood at the centre of the baseline with a clinometer and took a reading of the slope angle of the glacier. This was achieved with the help from the second student in each pair as they in turn sighted the position of every point within the grid using a compass to ensure they were located at the correct point. The student at the baseline then read the angle from the clinometer by lining it up level with the sighter’s eyes. This was repeated for every 5 metre point within the grid going up the glacier in a uniform fashion.
In addition to the slope angle readings, sediment or snow depth were recorded depending on the type of cover at each point in the grid. In areas of sediment an estimation of depth was made. The deep central areas were assumed to be greater than 50cm with the edge areas being given a depth value of 40cm. In snow covered areas an ice axe was pushed into the snow pack until it came into contact with the glacier surface. The depth was then estimated off the handle of the ice axe, except in areas where the ice axe didn’t come into contact with the glacier surface. Here a depth of greater than 50cm was assumed.
Model Method
The glacier surface model designed by Dr Andy Evans uses melt rate, critical angle and sediment thickness to determine melt in relation to sediment cover. Ice thickness and sediment thickness are inputted into the model as well melt rate and the pre-determined critical angle. It then determines ice thickness after the given melt rate and selects cells in a random manor and deducts the appropriate amount of ice depending on the melt rate for this cell. The model then assesses the critical angle and any sediment above this value is shifted to the adjacent cell. This is repeated time over time until all the cells have been assessed. An ice surface image is then displayed showing different areas of melting in relation to sediment cover. Lighter areas represent high melt where as darker areas represent lower levels of melt.
Prior to inputting data into the glacier surface model it was adjusted using various equations in Microsoft Excel. Firstly the diagonal distance to every point in the grid was ascertained from the fixed sampling point on the baseline. In addition to this, some trigonometry was used to determine the sediment height. We rearranged the SINE rule to calculate this.
Initially, the model was adjusted to our preference (critical angle 43°, one day, 140mm daily melt, 5000mm square width), and for future predictions, the relative number of days is also included. The first ice surface was then loaded on to the existing mud surface. The model was run under our given conditions and a glacial ice surface was produced. The model was then ran for Group E’s data producing another surface. A prediction from our data was created using the model and we then compared it to that of Group E’s actual data for that day. Both data sets were transferred into Microsoft Excel where the difference between the two days was calculated. Negative figures were produced, and in order to avoid this, the data was squared and used to create a fourth ice surface. However this then required the data to be rooted and produced a further surface. The original data was then entered into Minitab and a graph was plotted. Finally a regression of the data sets was performed to test the significance of any relationship between the two data sets.
Using field observations, it was ascertained that the average daily melt is 140mm. From this we can quantify the maximum human error for areas covered by snow. The difference in melt rates between the two data sets was then subtracted which gave us the human error. By finding the highest number in the data set we can predict how many days it will be before the human error will become insignificant and the real melt rate will be seen. A predicted ice surface was produced for the number of days calculated for the human error to become insignificant from our original data set.
Results and Analysis
- Sediment Origins
After calculating roundness, sphericity and shape for each of the given sites, the results were plotted in graphs comparable to those of Benn and Evans (see Appendices 2 & 3).
The plotted results are shown in Figures 2 and 3 (compare respectively to Appendices 2 & 3).
Figure 2:
Figure 3: (See Appendix 2 for data)