Permafrost modelling over different scales in arctic and high-mountain environments

Permafrost is present in several parts of Norway and contributes to binding mountain rock walls together, as well as storing potential greenhouse gasses as organic carbon in mires. Global climate changes influence the ground temperatures, and as a result the permafrost will thaw in some areas. This thesis provides new insight in the decisive role of the snow distribution in high mountain areas for correct estimates of ground temperatures in climate models. Large parts of the high mountain areas and the mires in Scandinavia are frozen all year around. This phenomenon is called permafrost. Permafrost acts as a glue in mountain areas, and if the permafrost warms or thaws entirely, resulting destabilization of the rock walls could threaten both infrastructure and lives. In permafrost areas with a high organic content, thawing permafrost will release the carbon storage in the ground as greenhouse gasses, and act as a feedback mechanism to the global climate change. It is therefore crucial to understand how existing permafrost will react to the future changes in the climate. In order to know how the permafrost in mountains and lowland areas will respond to the future climate, precise and detailed models for ground temperatures are required. The low thermal conductivity of snow cover makes it an efficient insulator to cold winter temperatures, and it therefore has profound implications for the thermal regime of the ground. Because of drifting snow, snow depths can be highly variable – from bare-blow ridges to deep troughs, sometimes only meters apart. Climate- and weather models traditionally operate on coarse resolutions, and use snow depths averaged over lager areas (> 1 km). In this thesis Gisnås demonstrates the errors introduced into permafrost models by this simplified representation of snow. Particularly bare-blown areas, where we find the coldest permafrost, are not well represented in current models. Statistical methods for representation of small-scale variation in snow depths in regional weather- and permafrost models have therefore been developed in this study. These methods highly improve the representation of the ground temperatures as well as the response in ground temperatures to climate changes in permafrost models. The improved permafrost model is implemented for Scandinavia, and a new map of permafrost in this region is produced. The map provides new knowledge and a much higher level of detail of the total distribution of permafrost, and gives new insight in which areas that are vulnerable to future climate changes