Surface crevasses on Svalbard: Spatial Distribution Analysis with Focus on the Lomonosovfonna Ice Cap

• Main Author: Hawrylak, Monika
• Format: Others
• Language: English
• Published: Uppsala universitet, Institutionen för geovetenskaper 2021
• Subjects: crevasses; spatial distribution; model; glacier velocity; Lomonosovfonna; Svalbard; Physical Geography; Naturgeografi
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-446531

Understanding the formation mechanics of glacial crevasses is crucial in a variety of glacial applications. Besides being a serious safety hazard during field campaigns, crevasses influence calving rates, mass balance, and the hydrological network of glaciers. Therefore, knowledge about their spatial distribution and potential development zones is highly beneficial. In this project, spatial distribution of surface crevasses on the Svalbard archipelago is investigated using a simple crevasse depth model, so called Nye’s zero-stress model, and a set of surface ice velocity data. The model is run for various ice temperatures as it is one of the parameters affecting crevasse development. The crevasse occurrences are also mapped and digitised manually using high resolution remote-sensed images. A special emphasis is put on the Lomonosovfonna ice cap, central Spitsbergen, where the crevasse distribution is studied in detail. The results indicate the greatest density of crevasses in regions of high ice surface velocity and concurrently high strain rates. These areas encompass mainly fast-flowing outlet valley glaciers, while the more stagnant ice caps and ice fields are characterised by little to no crevassing except for their margins. The modelled spatial distribution is concurrent with the observations, particularly for certain ice temperatures, highlighting the importance of accurate ice temperature measurements in modelling and a need of separation of the accumulation and ablation zones. Overall, the Nye’s zero-stress model yields accurate results and proves to be a suitable tool for the task. With room for fine-tuning, it is a promising tool that is easy to incorporate in other models.