EGU24-6076, updated on 08 Mar 2024
https://doi.org/10.5194/egusphere-egu24-6076
EGU General Assembly 2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

Ribbed Moraines and Generalised Linear Modelling: Can modelled ice conditions be used to predict the presence of specific landforms?  

Thomas James Barnes, Karianne Staalesen Lilleøren, Thomas Vikhamar Schuler, and Louise Steffensen Schmidt
Thomas James Barnes et al.
  • Universitetet i Oslo, Matematisk-Naturvitenskap fakultetet, Geofag, GeoHyd, Oslo, Norway

In this study, we investigate the relationship between subglacial conditions and the presence of ribbed moraines in Norway. Ribbed moraines are low-lying subglacially formed ridges, transverse to glacial flow and numerous processes have been proposed to explain their formation. So far there is no agreement about the formation process but most of them are linked to the presence of subglacial water. We therefore hypothesise that there is a relationship between hydrological conditions at the bed of the Fennoscandian Ice Sheet, and the presence of ribbed moraines. To test this, we extract subglacial conditions from a numerical model of the Fennoscandian Ice Sheet and derive further modelled hydrological conditions using a MATLAB-based hydrological toolbox. Our conditions include: (i) subglacial hydrological sinks, (ii) subglacial hydraulic head, (iii) flow accumulation, (iv) ice thickness, (v) ice-flow velocity, and (vi) basal temperature. We use these data in a presence-absence generalised linear modelling approach, to evaluate the coexistence of ribbed moraines and specific conditions. From this we can infer whether they have a consistent series of conditions which determine their presence. We focus on two areas, a training dataset in Vinstre, South-Central Norway, and a validation dataset in Femunden, Central-Eastern Norway. These sites cover known and well mapped areas of ribbed moraines, which are used as ground truth data. Comparison is possible through superimposing presence-absence predictions on the ground truth data in GIS as a pair of gridded, spatially referenced datasets. In comparing the model output to ground truth data, we aim to provide new assessments of the validity of the many ribbed moraine formation theories. For example, if hydrological conditions prove a poor predictor, then we can consider the presence of water as less likely a prerequisite for the formation of ribbed moraines. 

How to cite: Barnes, T. J., Lilleøren, K. S., Schuler, T. V., and Schmidt, L. S.: Ribbed Moraines and Generalised Linear Modelling: Can modelled ice conditions be used to predict the presence of specific landforms?  , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6076, https://doi.org/10.5194/egusphere-egu24-6076, 2024.