Hazards of Activation of Cryogenic Processes in the Arctic Community: A Geopenetrating Radar Study in Lorino, Chukotka, Russia

The subsurface structure of permafrost is of high significance to forecast landscape dynamics and the engineering stability of infrastructure under human impacts and climate warming, which is a modern challenge for Arctic communities. Application of the non-destructive method of geo-penetrating radar (GPR) survey is a promising way to study it. The study program, which could be used for planning and monitoring of measures of adaptation of Arctic communities to environmental changes is provided in this paper. The main principle was to use etalons of coupled radargrams and archive geological data to interpret changes in the permafrost structure from a grid of 5–10 m deep GPR transects. Here, we show the application of GPR to reconstruct and predict hazards of activation of cryogenic processes from the spatial variability in the structure of permafrost. The cumulative effects of the village and climate change on permafrost were manifested in changes in the active layer thickness from 0.5−1.0 m to up to 3.5 m. Despite that the permafrost degradation has declined due to the improved maintenance of infrastructure and the effects of ground filling application, the hazards of heaving and thermokarst remain for the built-up area in Lorino.


Introduction
Reports on the growing number of destruction and deformation of infrastructure has recently emerged from different inhabited locations in the Arctic [1][2][3][4][5][6][7]. Locally, up to 80% of buildings in large industrial centers have become deformed [4]. Recurrent deformations of infrastructure in spite of the application of different geotechnical solutions are common [8]. The reasons for this not only include the extreme conditions to which the construction materials are exposed and high costs of the maintenance. They are also caused by alterations of permafrost, which often acts as a basement for foundations. Its high vulnerability to increasing air temperature in the Arctic, which is predicted to continue over the 21st century [9], is widely recognized to be a reason for current or future negative consequences to society [10,11] and represents a natural hazard.
Despite of the long history of regulated construction in the permafrost zone dating back to the mid-1950s, changes of heat and moisture conditions driven by climate tendencies are not accounted for Geosciences 2020, 10, 57 2 of 13 in building standards [12,13]. Another issue relates to uncertainties of subsurface conditions, which are higher in the permafrost zone than outside it because of extreme variability in patterns and volumes of ground ice. Ground ice is not only represented by massive ice (a conventional term [14]), but also by relatively thinner cryostructures. Together, they form hardly predictable ice content along and below an engineering object, causing unequal responses of permafrost to changes in local heat and moisture exchange during construction and exploitation. The growth of ground ice inclusions results in the frost heaving of soils and erected infrastructure. The thawing results in subsidence, and in the worst cases, in thermokarst. The differentiated behavior of permafrost leads to deformations and destructions, which cause evolving disastrous consequences, affecting social security [4].
Unlike any other cryospheric object, changes to permafrost cannot be easily observed [9]. The application of a geo-penetrating radar (GPR) to permafrost studies has allowed researchers to reliably recognize thawed and frozen layers, massive ice, and rocks in continuous and discontinuous permafrost [15][16][17][18][19][20]. The permafrost sections, comprising of layers with different iciness [21,22] and salinity [23], have been studied by combining GPR survey with drilling. Several case studies have mapped aureoles of thawing in degrading permafrost and findings of massive ice and other permafrost features under engineering structures [24,25]. The efficacy of the method is based on the contrasting electrophysical properties of thawed and frozen sediments, ice, and aquifers. A limitations of the method is fading of the signal in waterlogged peaty soils of tundra, or thick covering loams or saline sediments. The need to calibrate the instrument at every study site to unique geological sections is another constraint in the wide application of GPR.
The development of methods of subsurface sounding is a priority task under predicted shrinkage of permafrost area, increasing permafrost temperature and active layer thickness [9]. They do not only non-destructively complement the spatial structure of permafrost deduced from "snapshots" of a sampled active layer thickness [19], an exposed parts of geological sections, and survey drilling, which commute disturbance to the thermal regime of soils [26,27]. They are also the most effective way to monitor permafrost evolution under human impacts and environmental changes [28][29][30][31]. The combination of a GPR survey with available geological data is a powerful technique of estimation of the dynamics of cryogenic processes, and the evolution of structure of permafrost [16,32].
With this case study, we share our experience of the application of GPR in permafrost zone to reconstruct and predict cryogenic processes affecting the life of an Arctic community. Our goal was to assess the vulnerability of permafrost. To do so, we put forward the following objectives:

•
To collect the available geological data from point sources (boreholes and exposure); • To track the interfaces of a thawed layer and permafrost layers having contrasting electrophysical properties using GPR; • To reconstruct recent geological processes, including hazardous thermokarst, thermal erosion, and frost heaving.  [36,37]. A description of sedimentary units is given in the text.

Study Site
The geological section of the terrace is disclosed in a well-described exposure of the coastal bluff. It consists of deposits widely distributed on the Bering sea coasts [36]. A detailed description of the sediments can be found in our open access study [37]. Dark grey clayey loams and clays in the base of the bluff (Unit 1) are covered with sands (Unit 2a) banded with pebble and gravel-enriched layers (Unit 2b). The inlet of ice-and organic-rich saline fine sands (Unit 3) is disclosed in the center of the terrace, intersected by a deep gully. They are covered with grey saline sandy loams with gravel inclusions (Unit 5, 2-4 m thick), and peat dissected with ice wedges (Unit 4, up to 4 m thick). The icerich transient layer just below the active layer [38] was observed in the exposure. The territory of the settlement is regularly leveled with ground filling mainly comprised of waste of coal burned on local steam shops.
Permafrost is continuous, mostly with low ice content, and with sparse occurrence of massive ice bodies [39]. The permafrost temperature measured in the boreholes by Chukotka Engineering Survey Enterprise (ChESE) was −3.7…−4.5 °C in the 1990s. The mean annual temperatures of soil by depths in 1992 were 0 °C at 1.5 m, −3 °C at 4.5 m. The Circumpolar Active Layer Monitoring plot in pristine tundra 2 km away from Lorino has been showing a gradual increase of the active layer thickness from 0.5 to 0.7 m over the last decade [40,41]. Among the hazardous cryogenic processes and phenomena observed in Lorino are thermokarst, heaving, thermal erosion, and thermal abrasion, causing deformations of buildings and destruction of infrastructure [37,42]. Furthermore, solifluction, kurums, and icings are common in the surroundings.
The population of Lorino was 998 in 2018 [43]. The settlement occupies 9.43 km 2 of land, of which 0.27 km 2 is the built-up area, with 37 residential low-rise (up to four floors) buildings of the 1960s-2000s, with the modern ones being built on pile supports; two steam shops; a diesel power station; a first-aid post; several stores; two schools; and a playground [44]. Roads are unpaved. The population mainly consists of indigenous Chukchi people involved in traditional marine mammal hunting, fishing, reindeer herding, fur farming, community services, and trade.

Methods and Instruments
Widely available GPRs with antennas ranging from 10 to 1000 MHz exhibit little difference in terms of their technical structure, principles of operation, and capabilities. GPR uses broadband pulses of electromagnetic radiation in the wavelength range from 1 to 0.1 m and detects reflected signals from subsurface interfaces of layers differing by electrophysical properties. Such interfaces may correspond to differences in the composition of sediment horizons, boundaries of thawed sediments and permafrost, massive ice bodies, differences in saturation, or mineralization.
We used an OKO-2 (LogiS, Russia) GPR with a 400 MHz shielded antenna block because 50-400 MHz is the optimal frequency range for surveying permafrost laid by 0.5-2.5 m thick layers down to a 10 m depth [16,17,21,23,25,26,28,[45][46][47]. The GPR survey was performed by the steady motion of an The geological section of the terrace is disclosed in a well-described exposure of the coastal bluff. It consists of deposits widely distributed on the Bering sea coasts [36]. A detailed description of the sediments can be found in our open access study [37]. Dark grey clayey loams and clays in the base of the bluff (Unit 1) are covered with sands (Unit 2a) banded with pebble and gravel-enriched layers (Unit 2b). The inlet of ice-and organic-rich saline fine sands (Unit 3) is disclosed in the center of the terrace, intersected by a deep gully. They are covered with grey saline sandy loams with gravel inclusions (Unit 5, 2-4 m thick), and peat dissected with ice wedges (Unit 4, up to 4 m thick). The ice-rich transient layer just below the active layer [38] was observed in the exposure. The territory of the settlement is regularly leveled with ground filling mainly comprised of waste of coal burned on local steam shops.
Permafrost is continuous, mostly with low ice content, and with sparse occurrence of massive ice bodies [39]. The permafrost temperature measured in the boreholes by Chukotka Engineering Survey Enterprise (ChESE) was −3.7 . . . −4.5 • C in the 1990s. The mean annual temperatures of soil by depths in 1992 were 0 • C at 1.5 m, −3 • C at 4.5 m. The Circumpolar Active Layer Monitoring plot in pristine tundra 2 km away from Lorino has been showing a gradual increase of the active layer thickness from 0.5 to 0.7 m over the last decade [40,41]. Among the hazardous cryogenic processes and phenomena observed in Lorino are thermokarst, heaving, thermal erosion, and thermal abrasion, causing deformations of buildings and destruction of infrastructure [37,42]. Furthermore, solifluction, kurums, and icings are common in the surroundings.
The population of Lorino was 998 in 2018 [43]. The settlement occupies 9.43 km 2 of land, of which 0.27 km 2 is the built-up area, with 37 residential low-rise (up to four floors) buildings of the 1960s-2000s, with the modern ones being built on pile supports; two steam shops; a diesel power station; a first-aid post; several stores; two schools; and a playground [44]. Roads are unpaved. The population mainly consists of indigenous Chukchi people involved in traditional marine mammal hunting, fishing, reindeer herding, fur farming, community services, and trade.

Methods and Instruments
Widely available GPRs with antennas ranging from 10 to 1000 MHz exhibit little difference in terms of their technical structure, principles of operation, and capabilities. GPR uses broadband pulses of electromagnetic radiation in the wavelength range from 1 to 0.1 m and detects reflected signals from subsurface interfaces of layers differing by electrophysical properties. Such interfaces may correspond to differences in the composition of sediment horizons, boundaries of thawed sediments and permafrost, massive ice bodies, differences in saturation, or mineralization.
We used an OKO-2 (LogiS, Russia) GPR with a 400 MHz shielded antenna block because 50-400 MHz is the optimal frequency range for surveying permafrost laid by 0.5-2.5 m thick layers down to a 10 m depth [16,17,21,23,25,26,28,[45][46][47]. The GPR survey was performed by the steady motion of an Geosciences 2020, 10, 57 4 of 13 interconnected generator of electromagnetic waves and a detector. The motion along the predefined transects was supported by sledges.
The program of the survey consisted of four steps detailed in the Sections 2.2.1-2.2.4, followed by the geocryological and geoenvironmental interpretation described in the Section 3.3.

Reconnaissance Survey
We started with a visual description of the study area, using community design documents and Google Earth images, paying attention to building types and non-conformities of the surface. The test GPR survey was performed to adjust the intensity of radiation and mode of detection to a depth of penetration of 5-10 m, resolution of 100-200 ns, and acceptable noise level. Then, the etalon transects and 50-150 m long transects were set up with an interval of 50-100 m, along the streets and sidewalks of the built-up area ( Figure 2).
We started with a visual description of the study area, using community design documents and Google Earth images, paying attention to building types and non-conformities of the surface. The test GPR survey was performed to adjust the intensity of radiation and mode of detection to a depth of penetration of 5-10 m, resolution of 100-200 ns, and acceptable noise level. Then, the etalon transects and 50-150 m long transects were set up with an interval of 50-100 m, along the streets and sidewalks of the built-up area ( Figure 2).

Etalon Geological and Geophysical Sections
Geological data were available from 10-15 m deep boreholes, which had been drilled in the 1990s by ChESE for projects of a new school and the diesel electric station. Mechanical, physical, and chemical features of sediments along the cores and the geological sections had been documented. Boreholes were selected for the etalons in order to cover the maximum diversity of geological sections typical for Lorino. Previous geophysical data for the study area were not found. GPR transects were collected along the mouths of the boreholes, corrected for the ground filling, which had buried the borehole mouths 0.5-1 m below the surface. Etalons were made by the coupling of geological sections and geophysical sections (radargrams) and, thus, the electrophysical properties of the boundaries in the geological section were found. The most important one was the dielectric permittivity of medium (ε), which defines the velocity of wave propagation (V), the damping ratio, and delay of the reflected signal.

Etalon Geological and Geophysical Sections
Geological data were available from 10-15 m deep boreholes, which had been drilled in the 1990s by ChESE for projects of a new school and the diesel electric station. Mechanical, physical, and chemical features of sediments along the cores and the geological sections had been documented. Boreholes were selected for the etalons in order to cover the maximum diversity of geological sections typical for Lorino. Previous geophysical data for the study area were not found. GPR transects were collected along the mouths of the boreholes, corrected for the ground filling, which had buried the borehole mouths 0.5-1 m below the surface. Etalons were made by the coupling of geological sections and geophysical sections (radargrams) and, thus, the electrophysical properties of the boundaries in the geological section were found. The most important one was the dielectric permittivity of medium (ε), which defines the velocity of wave propagation (V), the damping ratio, and delay of the reflected signal.

Collection and Processing of Geo-Penetrating Radar (GPR) Transects
Radargrams of GPR transects were collected in forward and return strokes with different resolutions to have the overall depth of penetration of 10 m and allow for a more detailed survey of the top 5 m. Visually deformed sites were examined in more detail. A total of 36 GPR transects 50-150 m long were collected during 3 days in late autumn, when the active layer had just started to freeze under mean daily temperatures of +5 . . . −1 • C (Figure 2).
The techniques of processing of radargrams are mostly analogous to those of seismograms because the laws of wave propagation and reflection are similar for electromagnetic and acoustic waves [23]. Radargrams were processed and graphically analyzed with GeoScan31 (Geotech, Russia) in order to amplify useful signals and to attenuate the extraneous noise. Interpretation was based on the recognition of continuous reflection events corresponding to interfaces of media with different electric permittivity. Condensation or termination of the reflections gave images of the form of layers and lens-like objects in the section. Based on the ε and V estimated for each interface on the etalons, the boundaries on radargrams were interpreted as geological boundaries, and the distances between them as thicknesses of layers.

Spatial Modeling Based on the Grid of Transects
Thicknesses of layers and depths of interfaces were sampled from the sections at every 5 m, producing the grid of 565 × 7 data points superimposed on the village plan. Maps of depths (isochores) and thicknesses (isopachs) of the recognized layers were generated based on the grid, using Surfer 11.0 (Golden Software, Golden, CO, USA) to study spatial variabilities of the near-surface structure of permafrost.

Etalons
The following interfaces were reliably recognized on radargrams of etalons: active layer bottom, bottom of ice-rich ground filling, massive ice beds, boundaries of buried organic-rich deposits 1 m thick at depths of 3-4 m, and the table of loams (Figure 3a,b). The layer of thawed sediments was recognized owing to the microplicated structure formed by multiple thin layers and folds originating from recurrent freezing-thawing cycles. Another distinctive feature was the interface of the ice-rich transient layer on the bottom. The ground-filling bottom was less contrasting, but still recognizable. Characteristic features of massive ice beds were a lens-like form, reaching a diameter of 30-50 m, and no secondary (parallel) reflections. Boundaries of organic-rich sediments were the most contrasting because of the large difference in dielectric permittivity compared to covering and embedding layers. Recognizing the table of loams at 6-8 m was the most difficult task.
Based on the sequence of layers, two 10 m-deep etalon sections were collected in eastern and western parts of the built-up area of the settlement (Figure 2). Etalon no. 1 was mainly laid with the sandy sediments of Units 2 and 3 ( Figure 3a). Observed reflection interfaces were attributed to the frozen/thawed sediment interface, and the boundaries of alternating layers of different ice contents and sediment compositions. Four layers recognized on the radargram corresponded to the following geological layers:

GPR Performance in Built-Up Areas of the Permafrost Zone
The study took place among buildings and infrastructure of the village. Wasted metal and construction materials were often met within the section. Non-stationery high-frequency noise occurred from passing cars, work of the power station, etc. A GPR survey has at least several advantages in such conditions compared to microseismic sounding and different types of electric survey: compact and mobile instruments, controlled sensitivity of the instrument to technogenic noise, and short-term survey production.

GPR Performance in Built-Up Areas of the Permafrost Zone
The study took place among buildings and infrastructure of the village. Wasted metal and construction materials were often met within the section. Non-stationery high-frequency noise occurred from passing cars, work of the power station, etc. A GPR survey has at least several advantages in such conditions compared to microseismic sounding and different types of electric survey: compact and mobile instruments, controlled sensitivity of the instrument to technogenic noise, and short-term survey production.
Our application did not allow the features of permafrost, such as the grain size, mechanical properties, salinity, moisture, ice, or organic matter content, to be measured for each layer. These features had been studied by ChESE from the cores collected on several drilling sites in the village and the layers were physically characterized. On the other hand, we found spatial variations of thicknesses of the layers and the section in the whole built-up area.
These values might be used as a reference in similar conditions when there is no other data for calibration. We demonstrate its use in the reconstruction of geological processes and prediction of the geohazards in the following section.

Permafrost Structure: Reconstruction and Predicting of Cryogenic Processes
The typical sections of permafrost, as interpreted from radargrams, are shown in Figure 4. Spatial patterns of the thickness and deformations of layers allowed the dynamics of cryogenic processes to be interpreted and vulnerability of lands of Lorino to different geohazards to be classified. Our application did not allow the features of permafrost, such as the grain size, mechanical properties, salinity, moisture, ice, or organic matter content, to be measured for each layer. These features had been studied by ChESE from the cores collected on several drilling sites in the village and the layers were physically characterized. On the other hand, we found spatial variations of thicknesses of the layers and the section in the whole built-up area.
These values might be used as a reference in similar conditions when there is no other data for calibration. We demonstrate its use in the reconstruction of geological processes and prediction of the geohazards in the following section.

Permafrost Structure: Reconstruction and Predicting of Cryogenic Processes
The typical sections of permafrost, as interpreted from radargrams, are shown in Figure 4. Spatial patterns of the thickness and deformations of layers allowed the dynamics of cryogenic processes to be interpreted and vulnerability of lands of Lorino to different geohazards to be classified.  Sites with the uniform bottom of Layer A at 0.5-1 m had background values of seasonal thawing (Figure 4a), slightly deeper than that in pristine tundra nearby, likely due to different soil composition and drainage [41]. The thickening of Layer A up to 2-3 m accompanied by the thinning of Layer B marked taliks and, thus, indicated the active thermokarst (Figure 4a,b). The thickening of ice-rich Layer B to 1-1.5 m was associated with the aggradation of permafrost by the freezing of ground filling (Figure 4b,c). The local thickening of Layer F indicated heaving, observed at sites where the active layer reached the table of Layer F (Figure 4b). On sites where Layer B thickened to 2-2.5 m, pinching out Layers E and F (Figure 4c), the aggradation of permafrost occurred by the refreezing of a talik formed by the relic thermokarst of the marine terrace.
The thickness of layer A within the built-up area varied from 1.0 to 3.5 m, with an average of 2.2 m. The bottom of thawed layer was crucial to recognizing the distribution and spatial structure of taliks. Concentric taliks with a thickness exceeding 1.8 m (Figure 5a) originated from the infiltration of atmospheric precipitation. Linear taliks either occurred along streets or tributaries of gullies and were formed by thermoerosion. Areas with a thawing depth close to that of natural, undisturbed landscapes Based on the interpretation of layers and reconstruction of cryogenic processes, the built-up area of Lorino was classified by the hazard of cryogenic processes into four types ( Figure 6).
The highest hazard is expected in residential areas, where recent thermokarst or frost heaving were induced by changes of the active layer, especially where Layers E and F lay close to the surface (Figure 4b). They are located along the boundary of ground filling with less disturbed lands surrounding the built-up area. Both processes could be expected to intensify following the tendencies of increasing variability and temperature.
Overall, at the time of study in 2013, the aggradation of permafrost and formation of the ice-rich layer were observed in the majority of taliks. The intensity of permafrost degradation from thermokarst and erosion was diminished by the new rural design of the 2000s, reconstruction of utilities, and demolition or reconstruction of old buildings. Ground filling supported the rise of the permafrost table and deepened frost-susceptible sediments and ice-rich layers, thus decreasing the hazards of frost heaving and thermokarst.
taliks. Concentric taliks with a thickness exceeding 1.8 m (Figure 5a) originated from the infiltration of atmospheric precipitation. Linear taliks either occurred along streets or tributaries of gullies and were formed by thermoerosion. Areas with a thawing depth close to that of natural, undisturbed landscapes (0.5-1.0 m) were either found on the margins of the built-up area under patches of vegetation cover or along the roads paved with ground filling.  Apart from its practical value, the scheme ( Figure 6) illustrates stages of permafrost transition from a natural landscape to a social-ecological system. Active layer increased from 0.5-1.0 m on the margins of the built-up area to 1.0-3.5 m in the village under similarly changed climate and growing human impact. Permafrost was degraded by the thawing of near-surface ice lenses, but at the same time, it aggraded with an accumulation of the layer of ground filling.
up area have formed chains of initiating heaving mounds with ice 1.2-1.5 m thick and with a diameter of 2-3 m, as well as larger heaving areas with ice lenses 8-12 m in diameter (Figure 5d). The bottom of Layer F laid uniformly at 2.7-3.0 m, supporting the heaving origin of the anomalies of thickness, rather than filling of the hollows in the underlying sediments.
Based on the interpretation of layers and reconstruction of cryogenic processes, the built-up area of Lorino was classified by the hazard of cryogenic processes into four types ( Figure 6).

Conclusions
Our case shows the application of GPR to an in-depth study of geological (cryogenic) processes, phenomena, and hazards in social-ecological systems. To interpret GPR transects as geophysical data, the descriptions of geological sections are required. The study of etalons prior to spatial surveying helps focus on the variability of a section, rather than on documenting the electrophysical properties. The structure of geophysical layers allows the history of cryogenic processes to be reconstructed. Reconstructing the stages of aggradation and degradation of permafrost and activation of frost heaving extends the application of GPR surveys in cryology and climatology as an effective instrument for palaeoreconstructions.
Current tendencies in the Arctic are likely to cause a growth of hazards of activation of cryogenic processes. In turn, this could increase the deformation of buildings and infrastructure and decrease the quality of life, unless specific measures to control them are undertaken. The application of a GPR survey is a fast and effective method for creating the baseline for planning, and monitoring the efficiency of the measures of adaptation of the Arctic communities to climate change.
An example of an adaptation measure which could be listed among the best practices, when the ice-rich sediments lay close to the surface, is the application of ground filling in Lorino. It is also necessary to keep in mind that the amelioration of permafrost starts with the control of human impacts. Improved maintenance of the buildings and infrastructure, the demolition or reconstruction of old buildings, and repair works of the communication infrastructure in Lorino in the 2000s have created an environment favorable for permafrost aggradation.
Positive lessons from this study might not be applicable at another location due to differences in the geological structure, human impact, and tendencies in climate and environmental change. This is why more case studies addressing the planning and monitoring of adaptation measures are necessary. Additionally, this is why a detailed program of the study was included in this paper. The authors hope it is worth using for improving other permafrost locations.
Author Contributions: O.T. developed the methodology; conducted the investigation; completed software processing; and curated, validated, and analyzed data. G.K. secured funding, supervised the study, and administered the project. O.T. and G.K. conceptualized the study, allocated resources, and prepared the original draft. All co-authors prepared the visualization, reviewed and edited the article. All authors have read and agreed to the published version of the manuscript.