Full text: Download
Abstract. Mountain permafrost is sensitive to climate change and is expected to gradually degrade in response to the ongoing atmospheric warming trend. Long-term monitoring of the permafrost thermal state is a key task, but problematic where temperatures are close to 0 ∘C because the energy exchange is then dominantly related to latent heat effects associated with phase change (ice–water), rather than ground warming or cooling. Consequently, it is difficult to detect significant spatio-temporal variations in ground properties (e.g. ice–water ratio) that occur during the freezing–thawing process with point scale temperature monitoring alone. Hence, electrical methods have become popular in permafrost investigations as the resistivities of ice and water differ by several orders of magnitude, theoretically allowing a clear distinction between frozen and unfrozen ground. In this study we present an assessment of mountain permafrost evolution using long-term electrical resistivity tomography monitoring (ERTM) from a network of permanent sites in the central Alps. The time series consist of more than 1000 datasets from six sites, where resistivities have been measured on a regular basis for up to 20 years. We identify systematic sources of error and apply automatic filtering procedures during data processing. In order to constrain the interpretation of the results, we analyse inversion results and long-term resistivity changes in comparison with existing borehole temperature time series. Our results show that the resistivity dataset provides valuable insights at the melting point, where temperature changes stagnate due to latent heat effects. The longest time series (19 years) demonstrates a prominent permafrost degradation trend, but degradation is also detectable in shorter time series (about a decade) at most sites. In spite of the wide range of morphological, climatological, and geological differences between the sites, the observed inter-annual resistivity changes and long-term tendencies are similar for all sites of the network.