Disko Bay sea-ice variability: Oceanic and Atmospheric drivers

• Main Author: King, Dana Margareta
• Format: Master Thesis
• Language: English
• Published: The University of Bergen 2023
• Subjects: 756213; Arctic; Greenland; Climate change; Disko Bay; North Atlantic; North Atlantic oscillation; Sea ice
• Online Access: https://hdl.handle.net/11250/3109662

Climate change impacts the Arctic sea-ice cover and thickness with a related Arctic Amplification and an increased ocean heat transport towards the Arctic. There has been a profound decline of Arctic winter sea-ice extent, exemplified by a gradual reduction in maximum sea-ice extent during March. Disko Bay on the west coast of Greenland is an excellent case study to investigate atmosphere, ocean, and sea-ice variability. The inflowing West Greenland Current influences Disko Bay, and the variability of the inflow appears to be related to the North Atlantic Oscillation. In this thesis, I have investigated the variability of the winter sea-ice concentration in January, February, and March (JFM) from 2003 to 2023 in connection with atmospheric and oceanic forcing. Atmospheric forcing is provided by ERA 5, while hydrographic observations are available for nine years. Two Sea-Ice Concentration (SIC) products of different spatial resolutions are evaluated and used to create a mean SIC time series for Disko Bay. This revealed a pronounced year-to-year variability and a small positive trend in SIC. A random forest model is used to relate daily atmospheric variables to SIC variability. I consider the variables which increase the mean square error of the model performance the most if they were to be removed (\%IncMSE). In addition, a correlation of atmospheric and oceanic parameters to SIC was applied. Correlations between SIC and oceanic forcing were related to determining the presence of cold and fresh Polar Water (PW) ($\theta$ < 1 °C and S $\leq$ 33.7), its thickness and the temperature of the top 50 dbar, T$_{\text{50 dbar}}$, in the water column. For atmospheric forcing, the three most important variables are Sea Surface Temperature (SST), the 2 m Temperature (T2M), and Surface net Latent Heat Flux (SLHF) with an \%IncMSE of $\geq$ 40\% each. Events of high SIC (> 75\%, 2008 and 2012) and low (< 30\%, 2010 and 2011) years were selected, and their variability was further investigated. This revealed that SST, ...