Measurements and Modelling of Arctic Coastal Environments

Abstract

The most pronounced effects of climate change are observed in the polar regions, especially in the Arctic, which is currently warming more rapidly than the rest of the world. The Arctic is losing sea ice, and the oceans are changing. The consequences of this polar transition affect the whole planet. The 13th goal of the United Nations Sustainable Development Goals is to take urgent action to combat climate change and its impacts. This Climate Action goal includes targets such as strengthening the resilience against climate-related hazards and improve climate change impact reduction, adaptation, mitigation and early warning systems. The main objective of this PhD was to improve our understanding of the physical environment of the Arctic using insitu measured data, remote sensing techniques and numerical modelling. Such an understanding is crucial to fully comprehend the ongoing changes and predict the effects due to climate change, and, therefore, indirectly contributes to the targets of the 13th United Nation Sustainability Goal. In this PhD thesis, three Arctic phenomena/topics were studied, including: 1) Wave propagation through ice-covered oceans 2) The physical oceanographic environment of the Arctic fjord Kangerlussuaq 3) Iceberg drift modelling in the Barents Sea First, ocean waves travelling through icy waters were studied using satellite remote sensing data. Synthetic aperture radar (SAR) data were found to be the most valuable source of satellite information to study waveice interactions. This type of data was used to study the change of wave parameters (peak wavelength and dominant wave direction) as waves enter the ice-covered waters of the Barents Sea, where accompanying in-situ data were available from the Barents Sea Metocean and Ice Network (BaSMIN) programme. A change in dominant wave direction was found towards the normal, relative to the ice edge, which is partly due to refraction and due to a SAR imaging artefact. The peak wavelength slightly increases as waves travel within the sea ice, which is due to the spatial dispersion of waves and is possibly enhanced by wave-ice interactions. Furthermore, the wave dispersion relation within sea ice was estimated from Sentinel-1 SAR Interferometric Wide swath data by using an innovative new implementation of a known method. In-situ data of sea currents from the BaSMIN campaign allowed the quantification of these data’s influence on the wave dispersion relation, which is a significant improvement over previous studies. The dispersion relation was derived for long waves (peak wavelengths between 100 m – 350 m) within thin ice (ice thickness less than 40 cm). The derived relation does not deviate from the theoretical open-water dispersion relation, which agrees with previous findings. Presently, however, the resolution of the SAR data is too coarse to study the wave dispersion relation of short waves within sea ice. Secondly, the seasonal variations of the physical oceanographic conditions and the wave climate in the Arctic fjord Kangerlussuaq were studied. Two high-fidelity numerical models were set up and calibrated against insitu data. Having such models allows us to fill in the spatial and temporal gaps left by in-situ data. The fjord consists of a deep, inner part with very slow currents and a shallow, outer part that is characterised by very strong tidal currents. These strong currents are most likely the cause of the absence of sea ice during winter in the outer part. During summertime, the inner part of the fjord is strongly stratified, and three water masses are present: water from the West Greenland Current, a brackish surface layer and a homogeneous deeplying water mass. The brackish surface layer is heavily affected by the meltwater from the inland ice, and this layer has a net outflow towards the open ocean. The deep-lying water mass is hardly subject to renewal during summer and appears dynamically decoupled from the open ocean. Furthermore, the 50-year return period significant wave height and peak wave period were estimated, which have a value of 1.8 m and 5 s, respectively. Finally, iceberg occurrence in the Barents Sea was studied using a numerical iceberg drift and deterioration model taken from the literature. Estimating annual iceberg encounter frequencies is vital for designing offshore structures and planning ice management operations. It was found that considerable uncertainties exist in two parameters of the model input at the iceberg sources, namely 1) the annual number of icebergs released from the sources that drift freely into the Barents Sea, and 2) the initial size characteristics of the released icebergs. Satellite remote sensing data collected by the Sentinel-2 constellation were utilised to derive iceberg size characteristics at the major iceberg sources. More than 22,000 icebergs were manually identified and provided statistics of the initial iceberg lengths and iceberg widths. Furthermore, a methodology is proposed using the Copernicus iceberg number density dataset, which primarily consists of Sentinel-1 data, to estimate the annual number of icebergs released. This results in approximately 2600 icebergs released per year. The model is forced with the newly obtained data to produce maps of the annual iceberg encounter frequencies and the annual expected number of icebergs in the Barents Sea.