Characteristics of snow structure along Kongsvegen glacier (Svalbard)

Snow cover data in Arctic regions like Svalbard are scarce while interest in distribution, properties and development of Arctic snow is evident (in context of climate change, too). Contributing to that issue this work is based on comprehensive observations of physical snow parameters recorded in snow pits and using special instruments, respectively. The data were collected in spring 2011 and at several locations in a height transect along Kongsvegen glacier. Analysis mainly concerns description of gross profile characteristics, the identification and physical understanding of major structures and their spatial and temporal characteristics. Three of them are considered in detail in which context snow hardness data are particularly important. Analysis was supported by data from automatic weather stations operated at the sites of snow investigations. These data also served to drive a numerical snow model. Model output was compared to observations and effectively served development of a process-based interpretation and reconstruction of the history of the investigated key layers. Different depth scaling method were developed to make the snow profiles comparable because snow depth strongly depends on altitude. The layer-wise depth scaling method yielded significant improvements in correlation between neighbouring snow hardness profiles along the glacier and at local scale. Three layers characterized by clear maxima in snow hardness, high density and mostly rounded grain types could thus be traced along the major part of the glacier. Attributions were comparatively uncertain if not impossible in the lower section of the glacier due to the higher temperatures and associated melt influences. Snow profiles were significantly correlated across decameter distances. The evolution history of the selected
layers was investigated using relevant model output data. The latter were overall successfully validated by comparison with measured key parameters like snowheight, surface temperature or end-winter vertical profiles. Some deficits in simulation skill were reflected in e.g. biases towards too low snow temperature and failure to reproduce fine structures in the vertical density profiles. This points to a need to improve related aspects of the model setup and parameterizations. There are indications that some problems may lead back to inappropriate parameterization of the density of fresh snow. Concerning the investigated examplary layers, however, the simulations largely reproduce the observed characteristics. Uncerainties are generally larger at lower sites again, and a general tendency towards too rapidly evolving snow metamorphosis may be noted. Backtracing individual layers proved feasible and allows estimating the approximate date of deposition and associated meteorological conditions. Follow up work may concern even more detailed analysis of the observed profiles, including parameters and sites which have not yet been considered in this work. Advancing methods for more sophisticated depth attribution of profile data is desirable as well as developing enhanced understanding of processes governing the development of observed physical properties (grain types in particular). Related simulations may mainly be improved concerning their setup and parameterizations (of fresh snow density for example). On the other hand, existing output already hints on interesting new research topics related to e.g. the development of a firn aquifer in the upper regions of the glacier and its dependencies on near- surface snow developments
by Maik Brakemeier
Masterarbeit University of Innsbruck 2017