1
References Benn, D.I. and Hulton, N.J.R. (2010). Computers and Geosciences, 36, p605-610; Benn, D.I. and Ballantyne, C.K. (2005), Journal of Quaternary Science, 20, p577-592; N.R.J. Nye, J.F. (1952), Nature, Vol 169, p 529-30; Osmaston, H. (2005), Quaternary International, 138-139, p22-31; Rea, B.R. and Evans, D.J. A. (2007), Palaeogreography, Palaeoclimatology, Palaeoecology, 246, p307-330; Schilling, D.H. and Hollin, J.T. (1981), In. The last great ice sheets. p207 -220; Vieira, G. (2008), Geomorphology, 97, p190-207; Acknowledgements DP would like to thank UW for funding the research project, the on-going support and advice from Tom Bradwell and Andrew Finlayson and sponsorship of the coring equipment from Van Walt Ltd. David Ashmore, James Lea, Will Hughes and Benedict Reinardy are thanked for their time and advice in the field. Danni Pearce *1 , Brice Rea 2 , Iestyn Barr 3 and Des McDougall 1 *1 [email protected] University of Worcester, Institute for Science and the Environment, Henwick Grove, Worcester, WR2 6AJ, U.K. 2 University of Aberdeen, School of Geosciences, Elphinstone Road, Aberdeen , AB24 3UF, Scotland, U.K. 3 Queen's University Belfast, School of Geography, Archaeology and Palaeoecology, Belfast, Ireland, U.K. The traditional geomorphological approach provides information on the size of the glacier and ice surface but little other information is gained. The numerical model produces a holistic reconstruction, providing glaciological insights to likely boundary conditions, including basal shear stress, the presence and thickness of plateau ice into an otherwise static reconstruction. It is suggested when reconstructing plateau icefields the modelling technique should be used. 5. Conclusion Reconstructing plateau icefields: Evaluating empirical and modelled approaches Glacial landforms are widely utilised to reconstruct former glacier geometries with a common aim to estimate the Equilibrium Line Altitudes (ELAs) and from these, infer palaeoclimatic conditions. Such inferences may be studied on a regional scale and used to correlate climatic gradients across large distances (e.g., Europe). To approximate palaeo-ELAs the reconstructed three-dimensional shape is required. Published reconstructions usually take one of two approaches: 1) The traditional approach uses geomorphological mapping with hand contouring and intuitions gained from the morphology of contemporary ice-masses to derive the palaeo-ice surface (e.g., Benn and Ballantyne, 2005). 2) Numerical models formulated from physics to simulate theoretical glacier surface profiles, whilst respecting the geomorphological evidence (e.g., Rea and Evans, 2007). No study has compared the two methods for the same ice-mass. This is important because either approach may result in differences in glacier limits, ELAs and palaeo-climate. This research uses both methods to reconstruct a Younger Dryas (YD; 12.9 -11.7 cal. ka BP) plateau icefield in the Tweedsmuir Hills, Scotland and quantifies the results from a cartographic and geometrical aspect. 1. Introduction The mapped landsystem can be correlated across the study area to delimit a YD plateau icefield c. 42 km 2 (Figure 2). The age is supported by new 14 C dating of basal stratigraphies and Terrestrial Cosmogenic Nuclide Analysis of in situ boulders. 2. Methods – Geomorphological approach In order to reconstruct the ice-surface profiles a valley centre-line flow model is used. Based upon Nye’s (1952) equation for mechanical equilibrium, for the case of an infinitely wide glacier which does not slide over its bed. In the valleys lateral drag is calculated using a shape factor. The following step-by-step calculation is used: ℎ +1 =ℎ + Eq. 1 (Schilling and Hollin, 1981) where: h ice surface elevation τ av basal shear stress s shape factor ρ ice density g acceleration due to gravity Δx step length t ice thickness i iteration number Figure 1. Symbols used in the numerical reconstruction (adapted after Schilling and Hollin, 1981). Where: r = bedrock elevation; t = centre-line ice thickness; h = ice surface elevation; Β = bed slope; α = ice surface slope; Ƭ b = basal shear stress; i = denotes the first iteration, with n+ denoting the iteration number; Δx = step length. i+1 i+2 i+3 x h 0 Bedrock i h i h i+1 h i+2 h i+3 Ice r i r i+1 r i+2 i+4 r i+3 r i+4 h i+4 Δx Δx Δx β t i t i+1 Δx t i+2 t i+3 t i+4 Ƭ b Δh α When the glacier long profile is constrained by geomorphological evidence, the Ƭ av is varied until the reconstruction matches the geomorphology. If absent, an arbitrary Ƭ av of 100 kPa is used (Rea and Evans, 2007; Vieira, 2008). Modelling approach Glacial landforms were digitally mapped using a combination of aerial photos, NEXTMap TM and mapping in the field used a ruggedized tablet PC. The three-dimensional ice-mass geometry was contoured through extrapolation and interpolation of the mapped evidence following established approaches (e.g., Sissons, 1980; Lukas and Bradwell, 2010). ELAs are calculated using the Area Altitude Balance Ratio method (BR = 1.67 to 2) and BR = 1.9 is used to calculate the area weighted mean ELA. Both techniques produce encouragingly similar geometrical configurations with both reconstructions covering c. 42 km 2 . However, important differences occur, which influence the ELA calculation. When landforms are absent or fragmentary (e.g., trimlines), as in the accumulation zones on plateau icefields, the geomorphological approach increasingly relies on extrapolation between lines of evidence and on the individual’s perception of how the ice-mass ought to look. This can result in an under/overestimation of the ice surface compared to the model most likely due to reworking and paraglacial modification. The numerical approach addresses this issue by using an iterative procedure to calculate the likely ice height in the valleys and on the plateau. The ELA is also influenced by the hypsometry. The modelled ice surface is more robust since they are derived from glacier physics and the geomorphological evidence, rather than the traditional approach which (when landforms are absent) follows the underlying topographic contours (Osmaston, 2005). The ELA results suggest caution is required when comparing values from differing methods Deriving ELAs would benefit from a standardised methodology to permit more accurate regional /large scale comparisons. 4. Discussion Geomorphological Reconstruction Modelled Reconstruction The model delimits a plateau icefield c. 42 km 2 (Figure 3). Calculated s range (0.41 to 0.69). τ av are c. 50 – 100 kPa although, two valleys with good geomorphic control produce low τ av of 25 – 45 kPa most likely due to basal slip, which is not explicitly incorporated in the model. Ice thickness on the summits is c. 100 m. The final reconstruction is processed in GIS to provide a more realistic ice surface. Figure 3. Three-Dimensional glacier reconstruction derived from the Schilling and Hollin (1981) model. DASHED line represent the flowlines used to calculate the surface profiles. YELLOW represents an increase in ice compared to the geomorphological reconstruction and BLUE is a decrease. Ice- surface contours are reconstructed at 50 m intervals. Overlaid onto shaded DTM, Scale 1:75,000. INSET are examples of the ice surface profiles the model produces when calibrated with geomorphological evidence. Scale and location is the same as Figure 2 Figure 2. 3-Dimensional glacier reconstruction derived from the geomorphological mapping. Ice-surface contours are reconstructed at 50 m intervals. Overlaid onto shaded DTM, Scale 1:75,000 Mean ELA c. 527 m 400 0 100 200 300 400 500 600 700 800 900 1000 0 1000 2000 3000 4000 5000 6000 7000 8000 Elevation (m) Distance from terminus (m) Reconstructed ice surface Bedrock Target elevation Reconstructed ice surface - Winterhope Mean ELA c. 562 m 0 100 200 300 400 500 600 700 800 900 1000 0 1000 2000 3000 4000 5000 6000 7000 8000 Elevation (m) Distance from terminus (m) Reconstructed ice surface Bedrock Target Elevation (spot heights) Reconstructed ice surface – Gameshope Winterhope Gameshope τ av 43 kPa τ av 25 - 50 kPa 3. Results
