ИСТИНА

Ice wedges are a prominent phenomena of permafrost landscapes, which cover almost a quarter of the northern hemisphere‘s landmass. These ice bodies are typically several cubic meters big and build a characteristic polygonal micro topography, thus determining indirectly the distribution of moisture, vegetation, and elements within the seasonally unfrozen active layer. However, the existence of ice bodies in the subsurface is not always delineable based on surface data. Furthermore, the increased potential of subsidence poses a hazard to any infrastructure nearby in case of a temperature increase. Also, the identification of hidden ice bodies is relevant for scientific field work such as scientific drilling. Ground-penetrating radar has been proven to be a suitable geophysical tool in permafrost areas for imaging sediments of the active layer at high resolution as well as determining the location of ice wedges in a non-invasive manner. However, the success of imaging based on widely used acquisition strategies (common-offset geometry, 2D data acquisition and processing flow) remains limited, mainly because the subsurface is heterogeneous (e.g. due to cryoturbation or scattered occurrence of ice lenses) and ice bodies exhibit a complex shape. We examine the influence of subsurface heterogeneity and ice wedge geometry on imaging these structures, based on synthetic data for a 2D polygon scenario of successively increasing complexity. Subsequently, we apply our interpretation strategies for identifying ice wedges to field data from Siberia.