ИСТИНА

Permafrost environments cover a fifth of the land area: 60% of Russia, 25% of China, 80% of Alaska and about 50% of Canada. The permafrost studies often take place in hard conditions  in wetlands and in impassable areas like taiga with dense vegetation cover. Permafrost structure is very sensitive to various mechanical and thermal influences. Consequently, permafrost geophysical methods should be non-invasive, possibly remote and quick. The ground penetrating radar (GPR) is promising in such conditions. The high dielectric contrast between ice and water determines the possibility of identifying the boundaries between melt and frozen grounds on the GPR data (radargrams), as well as local objects such as ice wedges. In August 2015 we carried out the fieldwork in the Chara Depression (Zabaykalsky Krai, Eastern Siberia, Russia). Modern and relict ice wedges were observed in almost all parts of the relief. We used GPR to detect ice wedges with invisible polygons, which was seen in the river outcrop. We wanted not only to determine the location of ice wedges, but also its shapes and sizes without drilling. This information is essential for thermokarst risk assessment, for paleogeographic studies, for the estimation of the amount of buried ice, etc. The geological section of fieldwork site was represented by frozen sandy loam covered with peat in the active layer. We used GPR system ZOND-12e (Radar Systems Inc., Latvia) with shielded 300 MHz antenna. The study was conducted using the funds of RFBR “My first grant” №14–05–31510. An integral part of the study was numerical modelling in the free software tool gprMax (http://www.gprmax.com). We made several models of subsurface based on photographs of outcrop, where the GPR survey was conducted. The comparison of real and simulated data is found to be extremely useful to interpret field data and understand the location and the size of ice within the permafrost. We analyzed also the influence of different factors, like shape of ice wedge, permafrost table boundary and deposit conductivity, on synthetic radargrams. Here, we present both field and modelling results.