EGU24-559, updated on 08 Mar 2024
EGU General Assembly 2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

Detecting lowland thermokarst development by UAV remote sensing in the Stordalen mire, Abisko, Sweden 

Maxime Thomas1, Thomas Moenaert1, Éléonore du Bois d’Aische1, Maëlle Villani1, Catherine Hirst1,2, Erik Lundin3, François Jonard1,4, Sébastien Lambot1, Kristof Van Oost1, Veerle Vanacker1, Reiner Giesler5, Carl-Magnus Mörth6, and Sophie Opfergelt1
Maxime Thomas et al.
  • 1Earth and Life Institute, UCLouvain, Louvain-la-Neuve, Belgium (
  • 2Department of Earth Sciences, Durham University, Durham, United Kingdom
  • 3Abisko Scientific Research Station, Swedish Polar Research Secretariat, Abisko, Sweden
  • 4Earth Observation and Ecosystem Modelling Laboratory, Université de Liège, Liège, Belgium
  • 5Climate Impacts Research Centre, Department of Ecology and Environmental Science, Umeå University, Umeå, Sweden
  • 6Department of Geological Sciences, Stockholm University, Stockholm, Sweden

In situ field studies in thawing permafrost regions have shown that C emissions resulting from organic carbon (OC) decomposition depend among others on the variability in soil water content, which can be directly related to microtopography. A more precise assessment of the evolution of permafrost C emissions as a function of thermokarst development requires high-resolution quantification of thermokarst-affected areas, as lowland thermokarst development induces fine-scale spatial variability (~ 50 – 100 cm). Here, we investigate a gradient of lowland thermokarst development at Stordalen mire, Abisko, Sweden, from well-drained undisturbed palsas to inundated fens, which have undergone ground subsidence. We produced orthomosaics and digital elevation models from very-high resolution (10 cm) UAV photogrammetry as well as a spatially continuous map of soil electrical conductivity (EC) based on Electromagnetic Induction (EMI) measurements performed in September 2021. In conjunction, we measured in situ the soil water content from the different stages of thermokarst development at the same period. The soil EC values are contrasted along the gradient in line with contrasts observed in the landscape classification derived from the orthomosaics and digital elevation models: palsas are flat areas with low soil EC (drier), whereas fens are subsided areas with higher EC (water-saturated). Areas in the course of degradation (transition zones) are well identified based on their higher slope, and broad range of EC. Importantly, these transition zones are only detected using a very fine spatial scale (i.e., 10 cm) coupled to information on the microtopography. Compared to a set of previously collected orthomosaics and digital elevation models, our results show an acceleration of thermokarst development in this area with a rate of palsa decline 4 to 10 times greater in 2019-2021 than in 2000-2014.

How to cite: Thomas, M., Moenaert, T., du Bois d’Aische, É., Villani, M., Hirst, C., Lundin, E., Jonard, F., Lambot, S., Van Oost, K., Vanacker, V., Giesler, R., Mörth, C.-M., and Opfergelt, S.: Detecting lowland thermokarst development by UAV remote sensing in the Stordalen mire, Abisko, Sweden , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-559,, 2024.