Time-Series InSAR Monitoring of Permafrost-Related Surface Deformation at Tiksi Airport: Impacts of Climate Warming and Coastal Erosion on the Northernmost Siberian Mainland

Abstract

The Arctic is the fastest-warming region on Earth, exhibiting a pronounced “amplifying effect”, which has triggered widespread permafrost thaw and increased the risk of surface deformation. In the Arctic coastal lowlands, permafrost is also affected by shoreline retreat. The impact of these dual stressors on surface deformation processes in the Arctic coastal lowlands remains poorly understood, particularly in terms of how permafrost thaw and shoreline retreat interact to influence surface stability. To address this gap, we employed PS-InSAR technology to monitor surface deformation from 2017 to 2021 at Tiksi Airport, the northernmost airport on the Siberian mainland, situated adjacent to the Laptev Sea. The results show that Tiksi Airport experiences localized significant surface subsidence, with deformation velocity ranging from −42 to 39 mm/yr. The near-coastal area of Tiksi Airport is strongly influenced by the ocean. Specifically, for extreme subsidence deformation (around –40 mm/yr), the surface subsidence velocity increases by 0.2 mm/yr for every 100 m closer to the coastline. Analysis of these deformation characteristics suggests that the primary causes of subsidence are land surface temperature (LST) warming and erosion by the Laptev Sea, which together lead to increased permafrost thaw. By revealing the combined effects of climate warming and coastal erosion on permafrost stability, this study contributes to enhancing the understanding of infrastructure safety and quality of life for residents in Arctic coastal subsidence areas.

1. Introduction

The Arctic is experiencing amplified global warming compared to other regions [1], leading to widespread permafrost degradation that poses significant risks to Arctic infrastructure by mid-century [2]. Permafrost, defined as soil or rock that remains below 0 °C for at least two consecutive years, plays a critical role in shaping Arctic geomorphology, ecological processes, and hydrological systems [3,4]. More importantly, permafrost degradation directly impacts human life, as rising ground temperatures reduce its load-bearing capacity, threatening the stability of infrastructure [5]. The thawing of ice-rich permafrost can cause ground subsidence and uneven surface deformation, further compromising engineered structures [6]. Model projections indicate that by 2050, approximately 33% of Arctic infrastructure could be damaged due to ground subsidence and loss of structural capacity, affecting up to 3.6 million people, nearly three-quarters of the Arctic population [2,5]. These findings underscore that permafrost degradation not only destabilizes buildings but also incurs substantial economic costs and threatens the livelihoods of Arctic residents. Consequently, there is an urgent need to assess the surface stability of Arctic infrastructure to mitigate these risks.

The Arctic coast, shaped by the interplay of sea ice and permafrost, is highly vulnerable to climate change, with coastal erosion being a primary driver of coastline retreat [7]. Coastal erosion rates have increased throughout the Arctic over the past several decades, often by a factor of two or more [8]. For instance, the coastline of the Tiksi Bykovsky Peninsula receded at an average rate of 0.59 m/yr between 1951 and 2006, with total erosion ranging from 92 to 434 m over these 55 years [9]. Such erosion has already led to the destruction of private homes and public infrastructure, including collective farm buildings [10]. In Alaska, by 2100, coastal erosion and flooding could damage 40–65% of the current coastal infrastructure and 10–20% of oil field infrastructure [11]. Compounding this issue, the increasing frequency of high-energy storms and extended ice-free seasons may further expose the Arctic coast to erosion, threatening critical infrastructure essential for the survival of Arctic communities. These challenges not only hinder the long-term development of Arctic coastal regions but also highlight their heightened sensitivity to climate change. Therefore, investigating surface deformation in Arctic coastal infrastructure is of paramount importance.

Interferometric Synthetic Aperture Radar (InSAR) is a remote sensing technology that detects surface deformation by analyzing phase differences in radar satellite images acquired at different times. Known for its high precision (millimeter-level accuracy) and ability to cover large areas under all weather conditions, InSAR is particularly well-suited for monitoring vast and inaccessible regions [12,13]. Permanent Scatterer InSAR (PS-InSAR), an advanced InSAR technique, focuses on analyzing points with stable scattering characteristics, enabling reliable deformation monitoring even in areas with complex terrain or uneven scatterer distribution [14,15]. Airports, with their stable scattering properties, are ideal targets for PS-InSAR monitoring, as this method can accurately detect subtle ground movements critical for maintaining the structural integrity of runways, taxiways, and other essential infrastructure.

In Arctic coastal lowlands, permafrost stability is threatened by the dual impacts of climate warming and coastline retreat, making the monitoring of surface deformation in these regions essential for sustainable development and environmental preservation. This study focuses on Tiksi Airport as a critical case to address three key research questions: (1) the extent of surface deformation at Tiksi Airport, (2) the influence of oceanic processes on surface deformation, and (3) the relationship between climate warming and surface deformation. By addressing these questions, this study aims to provide new insights into the development of Arctic coastal infrastructure and the planning of Arctic shipping routes, contributing to the resilience and sustainability of Arctic communities.

2. Materials and Methods

2.1. Research Area

Tiksi Airport is situated in Northern Yakutia, adjacent to the Arctic Laptev Sea, with geographical coordinates of 71°35′N and 128°55′E (Figure 1). As the highest-latitude airport on the Siberian mainland, it lies in a region underlain by Carboniferous and Permian rocks, with relatively thin Quaternary sediments not exceeding 1.5 to 2 m in thickness. The local climate is characterized by extreme seasonal variations, with January temperatures ranging from −25.6 °C to −34 °C and July temperatures ranging from 6.3 °C to 12.6 °C, based on data from the Tiksi meteorological station between 2010 and 2021. The annual average temperature during this period was approximately −12 °C [16]. The surface is dominated by tundra landscapes, including ice-wedge polygons [17]. Thermokarst lakes have developed around the airport runway, and the area is currently experiencing coastline retreat and intensified thermal erosion, both of which indicate ongoing degradation of the underlying permafrost [18].

2.2. Data

2.2.1. SAR Data

The Copernicus program of the European Space Agency (ESA) is an extremely important Earth observation initiative jointly developed by ESA and the European Union. The program aims to provide freely accessible and sustainable Earth observation data to support various observational activities related to Earth. Sentinel-1 is a radar imaging satellite of the Copernicus program. It provides C-band all-weather continuous imagery for prioritizing real-time dynamic monitoring of the world’s oceans and land masses, with greater reliability, faster re-entry cycles, and wider coverage now being used in a growing range of applications. To monitor surface deformation at Tiksi Airport, this study utilized 120 Sentinel-1A images taken from January 2017 to December 2021. Sentinel-1A data are downloaded at the Alaska Satellite Facility (https://search.asf.alaska.edu/#/, accessed on 11 January 2025). The specific information is provided in

Table 1. Sentinel-1A information in this study.

Data Information	Description
Start of the monitoring period	2 January 2017
End of the monitoring period	19 December 2021
Super master image	4 May 2019
Path/frame	149/353
Flight direction	Descending
Total number of images	120

.

2.2.2. DEM Data

In the InSAR interferometric process, ArcticDEM is utilized for terrain correction to eliminate phase variations caused by topography, thereby improving the accuracy of deformation monitoring. This dataset represents a significant advancement, offering a resolution of 2 m, compared to the 30 m resolution of typical open-source DEM datasets [20]. (Data sources: https://www.pgc.umn.edu/data/arcticdem/, accessed on 13 January 2025).

2.2.3. Landsat Data

Landsat 8 is a satellite launched by the National Aeronautics and Space Administration (NASA) in cooperation with the United States Geological Survey (USGS), continuing the tradition of providing critical Earth observation data since 1972. Landsat 8 is equipped with two main instruments: the Operational Land Imager (OLI) and the Thermal Infrared Sensor (TIRS), which can capture high-resolution multispectral and thermal infrared images of the Earth’s surface [21]. In Google Earth Engine (GEE), Landsat 8 data can be accessed for free. In this study, the thermal infrared bands of Landsat 8 are used to retrieve the land surface temperature (LST) of Tiksi Airport.

2.2.4. ALT Data

The CALM active layer thickness (ALT) data at the Tiksi site is limited to the period from 1997 to 2000, making it insufficient for analyzing long-term trends. To address this limitation, we utilized ALT data from the ESA Permafrost Climate Change Initiative (CCI) project. The latest version of the CCI ALT dataset (V4.0) provides Northern Hemisphere ALT data at a 1000 m resolution, spanning from 1997 to 2021 [19]. Notably, the CCI ALT data shows strong consistency with the ALT measurements from the CALM program [22]. (Data sources: https://climate.esa.int/en/projects/permafrost/, accessed on 18 January 2025). Since the resolution of the ALT data is 1 km, we calculated the average of the deformation data for all PS points within each ALT pixel to reduce the error in the data.

2.2.5. Meteorological Data

To analyze the factors influencing the seasonal changes in surface deformation at Tiksi Airport, we used measured air temperature and precipitation data from the Tiksi meteorological station (data downloaded from: http://meteo.ru/data/, accessed on 20 January 2025). The types of temperature and precipitation are daily data, and we have corresponded to the daily average temperature and precipitation data corresponding to the date of the Sentinel data collection, respectively.

2.3. Methods

2.3.1. PS-InSAR

Interferometric Synthetic Aperture Radar (InSAR) is a technique that utilizes the phase information of SAR data at different times in the same region for interferometry to extract deformation information [23]. Differential-Interferometric Synthetic Aperture Radar (D-InSAR) technology has many examples of applications in permafrost regions [24,25]. PS-InSAR technology is developed based on D-InSAR, because D-InSAR technology has some drawbacks, such as the length of the time baseline brings some images to the size of the SAR data processing volume, there will be out-of-coherence phenomenon in the area of high vegetation cover, too much reliance on precision orbit to align the data, and there will be some atmospheric effects affecting the accuracy of the deformation [26,27]. PS-InSAR experiments are conducted using Sarscape software in ENVI (https://www.sarmap.ch/index.php/software/sarscape/, accessed on 12 January 2025). The main processing steps of PS-InSAR include generating connection maps, image interferometry processing, first and second deformation inversions, and geocoding. The specific workflow is shown in Figure 2.

The main process step is as follows:

1.

Generate connection graphs: The process of generating a connection graph in PS-InSAR involves selecting suitable image pairs for interferometric processing. To minimize the impact of temporal and spatial baselines on coherence, PS-InSAR selects image pairs that are temporally close and spatially proximate (Figure 3). The connection graph typically uses thresholds for temporal and spatial baselines to restrict the selection of image pairs, ensuring that the generated interferograms have high coherence.

2.

Interference process: The interferogram generation process involves calculating the phase difference between SAR images acquired at different times to create interferograms. The specific steps are as follows: first, precisely align the SAR images acquired at different times (Registration); then, calculate the phase difference of the aligned SAR images to generate interferograms (Interferogram Generation); next, remove the phase variations caused by Earth’s curvature (Flat-Earth Removal); finally, use an external Digital Elevation Model (DEM) to remove the phase variations caused by topography (Topographic Phase Removal).

3.

First step of inversion: The purpose of the first inversion is to extract the phase time series for each pixel and identify coherent points (PS points). The specific steps are as follows: first, perform time series analysis for each pixel to calculate the phase values at various time points; then, through coherence analysis and phase stability detection, select stable coherent points, with results generally considered reliable if the coherence is greater than 0.5.

4.

Second step of inversion: The purpose of this step is to extract surface deformation information from the phase time series of PS points. The specific steps are as follows: first, perform phase unwrapping to obtain continuous phase change values; then, remove the atmospheric effects on the phase; finally, extract surface deformation information from the phase time series with the atmospheric effects removed (Deformation Inversion).

5.

Geocoding: Geocoding is the process of converting interferograms and inversion results from the radar coordinate system to the geographic coordinate system. The specific steps are as follows: first, use external DEM and satellite orbit information for coordinate transformation to convert radar coordinates to geographic coordinates; then, perform geometric correction on the converted images to ensure they align with the actual geographic locations; finally, overlay the corrected deformation map with geographic reference images to facilitate further analysis and interpretation. Because the deformation process of permafrost is a freezing and thawing process, we transform the deformation in the LOS direction to the vertical direction [28]:

$$d_v={\displaystyle \frac{d_{LOS}}{cos\theta }}$$

(1)

where $\theta$ is the angle of incidence of the satellite.

According to the references [29], ground surface deformation in permafrost regions can be decomposed into two components: long-term deformation, primarily caused by thawing of permafrost due to climate warming, and seasonal deformation, mainly caused by frost heave and thaw subsidence of the ALT within the annual freeze–thaw cycles [30]. Therefore, the ground deformation in permafrost regions can be represented as [31]:

$$y\left(t\right)=\mathit{\upsilon t}+c+a\mathit{sin}\left(\omega t\right)+bcos\left(\omega t\right)$$

(2)

where $t$ represents the time interval relative to the first period’s data, $y$ represents the surface deformation relative to the time, $a\mathit{sin}\left(\omega t\right)+bcos\left(\omega t\right)$ describes the seasonal component of deformation, $\omega$ is the angular frequency, defined as: $\omega ={\displaystyle \frac{2\pi }{T}}$, with the period T assumed to be 1 year, $\upsilon$ represents the velocity of deformation (mm/yr), $c$ is the constant bias term (fitting residual), and the seasonal deformation amplitude can be represented as $A=\sqrt{a^2+b^2}$.

2.3.2. Evaluation of Theoretical Precision

Coherence coefficients and phase variance are often used to represent theoretical errors when obtaining field validation data is difficult. The coherence coefficient measures the similarity of phase information between two SAR images. Its value ranges from 0 to 1, with higher values indicating greater similarity in phase information at corresponding pixels between the two images. Specifically, a high coherence coefficient (close to 1) indicates high similarity in corresponding pixels in the two SAR images, which usually suggests good data quality and suitability for further interferometric analysis. Conversely, a low coherence coefficient (close to 0) indicates significant phase differences at corresponding pixels in the two SAR images, which usually means poor data quality and potential unsuitability for precise surface deformation analysis.

Phase variance measures the degree of variation in phase values across a set of SAR images. It reflects the phase stability of a specific pixel at different time points. Specifically, low phase variance indicates that the phase values show little change over different time points, suggesting high surface stability at that location. Conversely, high phase variance indicates that the phase values show significant changes over different time points, suggesting low surface stability at that location, possibly due to interference or noise. This implies that the data accuracy at that location is lower. The formulas for the coherence factor and phase variance are available in the reference [32].

2.3.3. LST Inversion

Ermida et al. proposed a method for calculating Land Surface Temperature (LST) using Landsat satellite data, applicable to Landsat 4, 5, 7, and 8 satellites [33]. The LST retrieval uses the Statistical Mono-Window (SMW) algorithm, which is based on a linear relationship between land surface temperature and top-of-atmosphere brightness temperature, while also considering corrections for surface emissivity and atmospheric water vapor. On the Google Earth Engine (GEE) platform, we uploaded the vector boundary of Tiksi Airport and applied a cloud masking function to remove cloudy pixels from the images, ensuring that only clear-sky observations were retained for further analysis. To reduce noise and the impact of anomalies caused by temporary atmospheric conditions or sensor errors, we calculated the average value of the images from 2017 to 2021 to minimize the influence of outliers. The brightness temperature from the thermal infrared band was then converted to land surface temperature (LST). The corresponding method for this conversion can be found in the reference [33].

3. Results

3.1. PS-InSAR Accuracy Verification

The calculation results for the coherence coefficient and phase variance are shown in Figure 4. The coherence coefficient of Tiksi Airport primarily ranges between 0.5 and 0.94, with coherence coefficients greater than 0.5. The phase variance ranges from 0.09 to 0.32 mm (Figure 4). This indicates that Tiksi Airport, as a permanent scatterer, has maintained its scattering characteristics over a long period, exhibiting high coherence and stability. However, it should be noted that factors such as snow cover and surface water can affect the coherence and phase variance, introducing some uncertainties in the measurements. These factors may reduce coherence or increase phase variance, particularly during seasonal transitions.

3.2. Spatial Variation of Ground Surface Deformation

Figure 5 shows the deformation rate and distribution histogram for Tiksi Airport, where positive values represent uplift and negative values indicate subsidence. Between 2017 and 2021, the surface deformation velocity at Tiksi Airport ranged from −42 to 39 mm/yr. The distribution histogram reveals the frequency of deformation velocity intervals, with the highest frequency occurring between −10 and 10 mm/yr, suggesting relatively stable surface conditions. However, 3.8% of the PS points show a deformation velocity of less than −30 mm/yr, indicating areas of very rapid subsidence.

Figure 6 displays a 3D frontal view of the surface deformation velocity at Tiksi Airport. The number of PS points representing extreme subsidence (−40 mm/yr, shown in blue) is significantly greater than the number of PS points representing extreme uplift (40 mm/yr, shown in red), indicating an overall dominance of subsidence. The subsiding points are predominantly distributed in areas closer to the Laptev Sea, where higher deformation velocities are observed. This spatial pattern suggests a gradient of deformation intensity decreasing from the near-coastal zone toward far-coastal areas.

To clearly illustrate the distribution of uplift and subsidence at Tiksi Airport, we classify points with a surface deformation velocity < 0 mm/yr as subsidence points and those with a velocity > 0 mm/yr as uplift points (Figure 7). On the right runway, three representative points (P1–P3) were selected: P1 and P3 are classified as subsidence points, while P2 is classified as an uplift point. To qualitatively support the PS-InSAR results, we examined high-resolution historical Google Earth images. Although Google Earth imagery does not provide precise deformation measurements, it offers visual cues of surface conditions. For example, P1 and P3 are located in areas where surface irregularities or depressions are visually noticeable, while P2 lies in a relatively elevated and stable section. These visual observations qualitatively align with the deformation trends derived from the PS-InSAR analysis.

3.3. Temporal Variation of Ground Surface Deformation

During the summer, ground temperatures in Tiksi peak, and the active layer is fully thawed. Figure 8 shows the deformation data for August from 2017 to 2021, highlighting the annual deformation trends at the airport. In 2017, surface deformation was relatively stable, ranging from −50 to 37 mm, with only a few points subsiding more than 30 mm. By 2018, deformation ranged from −82 to 70 mm, and 3% of the points showed subsidence exceeding 50 mm. In 2019, the range expanded to −119 to 112 mm. By 2020, deformation increased further, ranging from −166 to 143 mm, and by 2021, it reached −221 to 196 mm. Compared to 2017, significant subsidence was observed on the runway near the Laptev Sea in 2020 and 2021. As a result, we identify the runway closest to the sea as the most hazardous area.

4. Discussion

4.1. The Relationship Between Ground Surface Deformation and Climatic Factors

To analyze the relationship between surface deformation and climate at Tiksi Airport, we calculated the mean land surface temperature (MLST) from 2017 to 2021 and examined its connection with surface deformation velocity. This spatial analysis helps clarify how LST influences surface deformation across different areas. In the temporal analysis, we collected mean daily temperature and precipitation data corresponding to the surface deformation measurement dates. This approach enables us to investigate the dynamic relationship between surface deformation and climate factors.

From 2017 to 2021, a negative correlation was observed between the mean land surface temperature (MLST) and surface deformation velocity at Tiksi Airport (Figure 9). As MLST increases, surface deformation tends to shift toward greater subsidence. Based on the fitted curve, a 1 °C rise in MLST corresponds to an increase in surface subsidence velocity of 0.46 mm/yr. LST serves as a critical link between the near-surface layer and the underlying permafrost. Higher surface temperatures facilitate greater heat transfer to the permafrost, leading to rising permafrost temperatures and gradual thawing. As permafrost thaws, surface deformation typically intensifies, resulting in increased subsidence.

To better understand the relationship between seasonal surface deformation and climate change at Tiksi Airport, we divide the year into two seasons: the warm season (June to September) and the cold season (October to May). By analyzing the time series deformation curves for Points 1–2 and comparing them with mean daily temperature and precipitation data (Figure 10), we observe a clear seasonal pattern in surface deformation. Subsidence occurs during the warm season, while uplift dominates the cold season. This pattern is closely tied to the freeze–thaw cycle of the soil. During the warm season, rising air and surface temperatures cause the soil to absorb more heat than it releases. In October, as temperatures drop, the soil begins to lose more heat than it absorbs, leading to the freezing of soil moisture. As temperatures continue to decrease, the surface layer also freezes, contributing to the observed uplift.

It is noteworthy that the subsidence velocity at Point 3, located near the coast, is exceptionally high, with a velocity of −41 mm/yr and cumulative subsidence reaching nearly 200 mm by the end of 2021. Even during the cold season, the subsidence trend remains significant, deviating from typical surface deformation patterns observed in permafrost regions. Research indicates that Arctic coastal permafrost is highly sensitive to climate change, with thawing not only releasing substantial amounts of greenhouse gases but also posing a serious threat to coastal infrastructure.

4.2. Impact of the Oceans on Surface Deformation

The PS points near Tiksi Airport show significant subsidence, highlighting the strong influence of coastal erosion on the airport. To explore the relationship between surface deformation and the ocean, this study calculates the shortest distance from each PS point to the coastline, using this as a proxy for proximity to the ocean (Figure 11a). Based on this metric, we analyze the correlation between surface deformation velocity and distance to the ocean. Through this analysis, we aim to uncover the mechanisms driving the interaction between surface deformation and oceanic proximity.

Most of the PS points at Tiksi Airport have deformation velocities ranging from −10 to 10 mm/yr, while some points exhibit extreme subsidence with velocities of around −40 mm/yr. We fitted the surface deformation velocities in these two intervals separately against the distance from the coastline. The results show that both intervals exhibit a similar deformation pattern, where surface subsidence becomes more significant as the distance to the coastline decreases. In the −10 to 10 mm/yr interval, the surface subsidence velocity increases by 0.1 mm/yr for every 100 m closer to the coastline. In the extreme subsidence interval near −40 mm/yr, the velocity increases by 0.2 mm/yr per 100 m (Figure 11b). It is noteworthy that the correlation coefficient between deformation velocity and distance to the coastline is 0.12 in the −10 to 10 mm/yr interval, while in the extreme subsidence region (around −40 mm/yr), the correlation increases to 0.37, indicating that oceanic factors have a more significant influence on extreme subsidence deformation.

To compare surface deformation differences between near-coastal and far-coastal areas at Tiksi Airport, we calculated the average deformation for all PS points in these regions. Figure 12b shows the time-series deformation curves from 2017 to 2021. The results indicate that both areas experienced subsidence during this period. By the end of 2021, subsidence in the far-coastal area was 5 mm, while in the near-coastal area, it reached 20 mm, four times greater than in the far-away coastal area.

In addition, the time series deformation curves reveal distinct patterns between the far-coastal and near-coastal areas (Figure 13b). In the far-coastal area, surface deformation follows the typical seasonal pattern of permafrost regions, with significant fluctuations: uplift occurs in winter due to freezing, and subsidence occurs in summer due to thawing, peaking between August and September each year. In contrast, the near-coastal area shows minimal seasonal variation, instead exhibiting a nearly linear subsidence trend. This suggests that surface deformation in the near-coastal area is influenced by the interaction between permafrost and the marine environment. Ocean temperatures fluctuate less than land temperatures, particularly in winter, when sea ice insulates the water, keeping it warmer than the land [34,35]. This continuous heat transfer likely accelerates the melting of underground ice in the permafrost, causing rapid subsidence. Additionally, the salt content in seawater lowers the freezing temperature of near-coastal permafrost [36,37], potentially preventing the soil from fully freezing during the cold season and further intensifying subsidence.

4.3. Relationship Between ALT and Surface Deformation

In permafrost regions, the active layer thickness (ALT) refers to the surface layer above the permafrost that undergoes seasonal freezing and thawing, also known as the seasonal thaw layer [38,39]. The ALT indicates the maximum thaw depth reached during this period and is a key metric in permafrost research. We calculated the mean active layer thickness (MALT) at Tiksi Airport from 2017 to 2021 and conducted a detailed statistical analysis of its relationship with the surface deformation velocity during the same period (Figure 14). The results show a correlation coefficient of −0.96 between the two variables, indicating a very strong negative correlation. Specifically, for every 1 m increase in the ALT, the corresponding surface subsidence velocity increases by 30 mm.

This strong negative correlation suggests that as the ALT increases, the surface subsidence velocity significantly accelerates. An increase in ALT means a greater portion of the ground ice is involved in the annual freeze–thaw cycle. The thawing of this ice weakens the soil structure, making the surface more prone to subsidence. As more ground ice participates in the thawing process, the surface becomes increasingly unstable, leading to higher subsidence rates. Therefore, the observed relationship between ALT and subsidence velocity reflects the amplified effect of permafrost thaw and its contribution to surface deformation.

Compared to existing InSAR studies in permafrost regions [29,32,40], our results highlight the relationship between ALT and surface deformation. Specifically, InSAR technology reflects the characteristics of surface deformation, which is fundamentally caused by the thawing of permafrost beneath the surface. The thawed permafrost becomes part of the active layer, leading to surface subsidence that can be detected through surface deformation (Figure 15). However, current research on surface deformation in permafrost areas has not deeply explored the intrinsic connection between surface deformation and permafrost deformation. In the Arctic region, the ALT is often difficult to measure directly. However, surface deformation information obtained through InSAR technology can reflect the close relationship between surface subsidence and permafrost thawing. Our study further reveals the impact mechanism of ALT on surface deformation, providing valuable insights for permafrost assessment and engineering monitoring in the Arctic region.

4.4. Significance and Limitations

The Arctic region, notably susceptible to global warming, is experiencing pronounced changes, with permafrost melting leading to significant challenges for infrastructure stability and safety. In the high-latitude Arctic coastal areas, there is a crucial gap in research regarding the interplay between global warming and coastal erosion. This study contributes to bridging that gap, enhancing our comprehension of how Arctic coastal built facilities respond to climatic changes and offering valuable data for their maintenance.

The challenge of acquiring field ground measurement data for validating surface deformation presents a limitation to current research. In addition, even though we chose a stabilized scatterer, we still did not remove the error caused by surface deformation due to snow in winter, which is a challenge for InSAR applications in permafrost regions. Future studies are recommended to incorporate precise ground measurements alongside supplementary data to thoroughly analyze surface deformation patterns in the Tiksi region. Future monitoring efforts should integrate diverse datasets to refine the accuracy of observations, ensuring a comprehensive understanding of the permafrost dynamics and their implications for Arctic infrastructure. In addition, in the Tiksi region, we can only obtain surface deformation data from 2017 to 2021 because of missing data, interference from weather factors, and limitations on the frequency of satellite observations. There are still many challenges to obtaining surface deformation data over long-time spans using satellite data.

This study, using the Tiksi airport as a case study, provides a preliminary analysis of the differences in surface deformation between near-shore and far-shore areas. However, the Arctic coastal region is a complex system shaped by the interplay of permafrost and marine environments, and isolating the analysis to near-shore and far-shore areas has inherent limitations. Future research should integrate precise ground measurements and supplementary datasets to comprehensively analyze surface deformation patterns in the Tiksi region, while also developing coupled models that incorporate key parameters such as coastline position changes, sea temperature, and sea ice extent. This approach will more effectively elucidate the interaction mechanisms between marine environments and permafrost, advancing our understanding of Arctic permafrost dynamics and their implications for infrastructure.

In addition, in order to improve the applicability of the findings, this study should be extended to other Arctic coastal regions in future studies (e.g., Alaska, Canadian Arctic, Greenland, and other coastal regions of Russia) would allow for a broader understanding of the different impacts of climate change on permafrost and infrastructure stability in different Arctic environments. Revealing the differences and similarities in permafrost response between regions will provide specific recommendations for infrastructure management in different regions.

5. Conclusions

This study focuses on Tiksi Airport and investigates the stability of permafrost in Arctic coastal lowlands under the combined effects of climate warming and coastal erosion using InSAR technology. The PS-InSAR method was used to obtain surface deformation data from 2017 to 2021. Data accuracy was ensured by controlling the coherence coefficient and phase variance.

The results show that surface deformation at Tiksi Airport is significantly uneven. The number of subsiding PS points is greater than the number of uplifting points. The deformation velocity ranges from −42 mm/yr to 39 mm/yr. Notably, the near-coastal airport runway area has subsidence approximately four times higher than that of the far-coastal runway area. By 2021, the maximum cumulative subsidence exceeded 200 mm. The near-coastal PS points show an almost linear subsidence trend with minimal seasonal variation, indicating a strong influence of the ocean on the near-coastal area of the airport. Further analysis reveals a positive correlation between surface deformation rate and ground temperature. As ground temperature increases, the subsidence velocity becomes higher. This suggests that the permafrost at Tiksi Airport is responding significantly to climate warming. In addition, the active layer thickness (ALT) is closely related to the deformation velocity, with areas of thicker ALT experiencing faster subsidence. These findings indicate that Arctic coastal permafrost is affected not only by climate warming but also by the intensifying effects of coastal erosion, leading to further instability.

This study highlights the multiple instability risks of Arctic coastal permafrost and emphasizes the importance of long-term surface deformation monitoring. Given the significant impact on infrastructure, it is crucial for policymakers to consider incorporating adaptive strategies for infrastructure management in these vulnerable regions. These strategies could include the strengthening of foundations, the use of insulation materials to delay thawing, and the relocation of critical infrastructure away from high-risk areas. Additionally, continuous monitoring of surface deformation and permafrost conditions should be prioritized to inform timely interventions. Future research should also explore innovative techniques for mitigating the impacts of permafrost degradation and coastal erosion, ensuring the resilience of infrastructure in Arctic coastal regions.