A 16-year record of supraglacial lake extent in West Greenland: analysis of inland expansion and rapid drainage events

• Main Author: Cooley, Sarah W.
• Format: Thesis
• Language: English
• Published: Apollo - University of Cambridge Repository 2016
• Subjects: Greenland; Ice Sheet
• Online Access: https://dx.doi.org/10.17863/cam.8752; https://www.repository.cam.ac.uk/handle/1810/263410

Supraglacial lakes forming and draining on the Greenland are particularly important for understanding the hydrologic control on flow due to their ability to rapidly transport water to the base of the ice sheet and induce basal sliding (Das et al., 2008). In the past decade, numerous studies have mapped supraglacial lakes using remotely sensed imagery and noted a steady inland progression of lakes coincident with increasing melt (Fitzpatrick et al., 2014; Howat et al., 2013). However, there remains significant uncertainty in this progression, particularly in regards to the impact of observational bias on lake detection (Leeson et al., 2013) and possible physical altitudinal limits to inland expansion of lakes and rapid drainage events (Poinar et al., 2015). The aim of this dissertation is to examine patterns in inland expansion and rapid drainage frequency through an analysis of supraglacial lake processes in central West Greenland from 2000-2015. Using a fully automated lake detection approach adapted from Liang et al. (2012), ~515 supraglacial lakes per year are mapped over a 63,000 km2 study area. In high melt years, lakes form and disappear earlier and a greater number of lakes appear on the ice sheet as lakes extend to higher elevations. Strong positive trends in the number of lakes, cumulative surface water loss and maximum lake elevation are also identified over the 16-year record. Comparison of different methods of rapid drainage detection from previous studies reveals the percent of lakes detected as rapidly draining varies from 3% to 38% depending on how rapid drainage is defined and indicates that the number of rapid drainage events is strongly dependent on the frequency of cloud-free lake observations. Statistical analysis of this observation bias in rapid drainage detection suggests a true rapid drainage probability between 0.36 and 0.45, which is considerably higher than previous studies have reported. Finally, no evidence is found to support a physical altitudinal limit on inland expansion of rapid drainage events, suggesting rapid drainage events will continue to move inland as the climate warms and thus may propagate basal sliding-induced mass loss to higher elevations.