Remote Sensing of Snow Dynamics over the Vissátvuopmi Palsa Mire, Northern Sweden

Abstract

Palsas are peat hills made of a permafrost core with ice lenses and are very important areas for biodiversity. Today, palsas are also climate indicators for global warming. Due to their sensitive location at the edge of the permafrost zone, they are very vulnerable to a changed climate. Palsas are disappearing faster and faster, as the climate has become increasingly warmer for the past decades. In order to observe these periglacial landforms before they disappear, it is of importance to investigate how a warmer, wetter and more snow-covered climate affects the palsa mires. Above all, snow dynamics is a climate variable that affects the entire permafrost zone. With the help of technological developments, new remote sensing methods have been used in recent decades to map snow dynamics such as snow depth and snow cover, and these can also be applied over palsa mires. There are yet few studies that use several different types of remote sensing methods to map snow dynamics over the same palsa. Even fewer have combined various remote sensing methods with measurements taken in the field over a palsa mire. The purpose of this study was therefore to investigate how the different remote sensing methods LiDAR (light detection and ranging) and optical satellite data can be used to measure snow dynamics and how their measurements agree to those taken in situ. Weather data was also processed in order to relate climate and topographic parameters with snow distribution over the palsa mire. Measurements of snow depth and snow cover were performed on two palsa types, a domeshaped and a ridge-shaped palsa in Vissátvuopmi, Sweden's largest continuous palsa mire. Both UAV (unmanned aerial vehicle) LiDAR and in situ measurements were performed in the field, where Sentinel-2 optical satellite data were collected over the palsas during the time period 2017-2022. Weather data over the study area was also collected and processed. The results of the snow depth from in situ and LiDAR measurements showed that some measuring points had relatively the same snow depth between the methods, while other points showed greater differences in snow depth between them. A correlation analysis between the different snow depth measurements resulted only in an intermediate correlation of 0.41 for the ridge-shaped palsa and 0.28 for the dome-shaped palsa. The P-values for each correlation was found to be lower than 0.05, indicating that the correlations were statistically significant. The study also shows how much impact meteorological variables such as air temperature and wind have on how snow is distributed over the ridge-shaped and dome-shaped palsas, especially in relation to the palsa's topography. The analysis of Sentinel-2 optical satellite data showed that the ridge-shaped and dome-shaped palsas received an earlier first snowmelt date during the year and a later first snow-free date, which in 2022 resulted in an increase in the snowmelt period by 15 days for the ridge-shaped palsa and 14 days for the dome-shaped palsa since 2019.