sensors

Article Monitoring the Degradation of Island Permafrost Using Time-Series InSAR Technique: A Case Study of ,

Sai Wang 1, Bing Xu 1,*, Wei Shan 2 , Jiancun Shi 1, Zhiwei Li 1 and Guangcai Feng 1 1 School of Geosciences and Info-Physics, Central South University, 410083, China; [email protected] (S.W.); [email protected] (J.S.); [email protected] (Z.L.); [email protected] (G.F.) 2 Engineering Consulting and Design Institute, Northeast Forestry University, 150040, China; [email protected] * Correspondence: [email protected]; Tel.: +86-0731-88830573

 Received: 9 February 2019; Accepted: 15 March 2019; Published: 19 March 2019 

Abstract: In the context of global warming, the air temperature of the Heihe basin in Northeast China has increased significantly, resulting in the degradation of the island permafrost. In this paper, we used an elaborated time-series Interferometric Synthetic Aperture Radar (InSAR) strategy to monitor the ground deformation in the Heihe area ( Province, China) and then analyzed the permafrost deformation characteristics from June 2007 to December 2010. The results showed that the region presented island permafrost surface deformation, and the deformation rate along the line of sight mainly varied from –70 to 70 mm/a. Based on the analysis of remote sensing and topological measurements, we found that the deformation area generally occurred at lower altitudes and on shady slopes, which is consistent with the distribution characteristics of permafrost islands. Additionally, the deformation of permafrost is highly correlated with the increase of annual minimum temperature, with an average correlation value of –0.80. The accelerated degradation of permafrost in the study area led to the settlement, threatening the infrastructure safety. Our results reveal accelerated degradation characteristics for the island permafrost under the background of rising air temperature, and provide a reference for future relevant research.

Keywords: island permafrost; deformation; time-series InSAR; global warming; geohazards

1. Introduction Seasonal permafrost and permafrost are broadly distributed in China, covering 54% and 25% of the land area, respectively. The distribution area of all permafrost is about 1.59 × 106 km2, ranking third in the world. The permafrost distributed in the northern part of Northeast China represents the second largest area of permafrost in China [1]. In recent years, global warming has caused a rise in air temperatures and increasing precipitation, which have accelerated the degradation of permafrost. Due to the widespread degeneration of permafrost, forests are dramatically transforming into swamps, further breaking the ecosystem equilibrium. This situation should receive attention [2]. As a sensitive indicator of climate change, the change in permafrost has attracted special attention around the world. Northeast China is the southern margin of the Eurasian permafrost region. The permafrost of these low-elevation islands is more sensitive to changes in climate and environmental, and has shown obvious signs of degradation in recent years. The surface deformation caused by island permafrost degradation is localized, which is different from the surface deformation (usually the overall uplift or subsidence) caused by permafrost degradation. Additionally, the degradation of island permafrost has a great impact on the local environment, leading to phenomena such as landslides [3], solifluction, and partial expressway collapse and destruction. With the increase of construction

Sensors 2019, 19, 1364; doi:10.3390/s19061364 www.mdpi.com/journal/sensors Sensors 2019, 19, x FOR PEER REVIEW 2 of 14

Sensorsor subsidence)2019, 19, 1364 caused by permafrost degradation. Additionally, the degradation of island permafrost2 of 14 has a great impact on the local environment, leading to phenomena such as landslides [3], solifluction, and partial expressway collapse and destruction. With the increase of construction projects, projects, understanding the distribution of surface deformation in the island permafrost area is the key understanding the distribution of surface deformation in the island permafrost area is the key to to disaster warning and prevention. However, compared with the permafrost in the Qinghai–Tibet disaster warning and prevention. However, compared with the permafrost in the Qinghai–Tibet plateau, the permafrost monitoring research in this region is insufficiently studied. More attention plateau, the permafrost monitoring research in this region is insufficiently studied. More attention needs to be paid to the permafrost degradation in the future. needs to be paid to the permafrost degradation in the future. Due to its wide spatial coverage, Interferometric Synthetic Aperture Radar (InSAR), as an all-weather Due to its wide spatial coverage, Interferometric Synthetic Aperture Radar (InSAR), as an all- monitoring method with competitive accuracy, has been widely used to obtain ground deformation weather monitoring method with competitive accuracy, has been widely used to obtain ground measurements. The traditional differential InSAR technique is inevitably affected by spatiotemporal deformation measurements. The traditional differential InSAR technique is inevitably affected by decorrelation,spatiotemporal atmospheric decorrelation, delay, atmospheric and random delay, noise, and which random limits noise, the measurementwhich limits the accuracy measurement to some extentaccuracy [4]. Toto some overcome extent these [4]. To limitations, overcome time-series these limitations, InSAR techniquestime-series suchInSAR as techniques persistent scatterersuch as interferometrypersistent scatterer (PS-InSAR) interferometry [5] and (PS-InSAR) small-baseline [5] and subset small-baseline interferometry subset (SBAS-InSAR) interferometry have (SBAS- been proposedInSAR) have [6]. Thesebeen proposed time-series [6]. InSAR These techniquestime-series canInSAR monitor techniques small deformationcan monitor small over adeformation long period, areover suitable a long for period, surface are deformation suitable for inversionsurface deformation in the study inversion area, and in have the beenstudy successfully area, and have applied been in manysuccessfully cases. Theapplied InSAR in techniquemany cases. has The been InSAR widely te usedchnique for monitoringhas been widely surface used deformation for monitoring caused bysurface earthquakes deformation [7], underground caused by earthquakes mining [8], [7], landslides underground [9], oil mining and gas [8], extraction landslides [10 [9],], groundwater oil and gas extractionextraction [ 11[10],], and groundwater thawing and extraction freezing [11], of permafrostand thawing [12 and,13 ].freezing of permafrost [12,13]. InIn this this study, study, we focuswe focus on the on degradation the degradation monitoring monitoring of island of permafrost island permafrost around the around expressway the fromexpressway Bei’an to from Heihe Bei’an in Heilongjiang to Heihe in Province,Heilongjiang China. Province, First, a setChina. of 20 First, scenes a set L-band of 20 ALOS/PALSARscenes L-band stripmapALOS/PALSAR images stripmap acquired images from Juneacquired 2006 from to December June 2006 to 2010 December were processed 2010 were with processed an elaborated with an time-serieselaborated InSARtime-series strategy InSAR to obtainstrategy deformation to obtain deformation of the study of area. the Subsequently,study area. Subsequently, the deformation the ofdeformation island permafrost of island was permafrost analyzed was in combination analyzed in withcombination elevation, with slope elevation, direction, slope and direction, temperature. and Moreover,temperature. we Moreover, analyzedthe we correlationanalyzed the between correlation thedeformation between the ofdeformation island permafrost of island and permafrost elevated temperature.and elevated Finally,temperature. we further Finally, analyzed we further the anal potentialyzed the geological potential hazards geological associated hazards withassociated island permafrostwith island degradation. permafrost degradation.

2.2. Study Study AreaArea TheThe study study area area is locatedis located in the in Sunwu–Aihuithe Sunwu–Aihui area, Heihearea, Heihe City, in City, Northern in Northern Heilongjiang Heilongjiang Province, ChinaProvince, (see China Figure 1(see). The Figure plains 1). ofThe the plains study of area the havestudy low area mountainous have low mountainous terrain, with terrain, an elevation with an of betweenelevation 140 of andbetween 425 m.140 The and elevation 425 m. The of theelevation study areaof the gradually study area decreases gradually from decreases the northwest from the to thenorthwest southeast. to the southeast.

FigureFigure 1. 1.The The studystudy areaarea andand itsits Digital Elevation Model (D (DEM).EM). The The red red rectangle rectangle on on the the left left outlines outlines thethe spatial spatial coverage coverage ofof thethe studystudy area,area, andand thethe rightright inset shows the the DEM DEM of of the the study study area, area, with with the the bold black dashed line denoting the expressway from Bei’an to Heihe and the white rectangle denoting the coverage of the ALOS/PALSAR data used in this study. Sensors 2019, 19, 1364 3 of 14 Sensors 2019, 19, x FOR PEER REVIEW 3 of 14

bold black dashed line denoting the expressway from Bei’an to Heihe and the white rectangle The central part of the image is intermountain depressions and valley terraces, distributed in marshes, denoting the coverage of the ALOS/PALSAR data used in this study. cultivated land, and other land types, while the south part is mainly cultivated land. The study area is located inThe the central transitional part of regionthe image between is intermountain the middle-temperate depressions and zone valley and terraces, the cool-temperate distributed in zone, and ismarshes, alternately cultivated affected land, by and low other pressure land andtypes, high while pressure the south from part the is surroundingmainly cultivated area, land. as well The as by the monsoonstudy area [14 is]. located in the transitional region between the middle-temperate zone and the cool- temperateThe climate zone, in theand studyis alternately area has affected strong by seasonal low pressure characteristics, and high pressure with cold, from dry, the andsurrounding long winters, and hot,area, humid,as well as and by the short monsoon summers. [14]. The annual average precipitation is about 540 mm, although The climate in the study area has strong seasonal characteristics, with cold, dry, and long winters, can sometimes reach up to 800 mm. The precipitation mainly occurs in the form of rainfall from July and hot, humid, and short summers. The annual average precipitation is about 540 mm, although can to September, accounting for 61% to 67% of the total annual precipitation [15]. The annual average sometimes reach up to 800 ◦mm. The precipitation mainly occurs in the form of rainfall from July to temperatureSeptember, is aboutaccounting−2 to for 2 61%C. Meteorological to 67% of the tota datal annual indicate precipitation that the air [15]. temperature The annual has average exhibited a ◦ significanttemperature variation is about [16 −].2 China’sto 2 °C. Meteorological air temperature data increase indicate that rate the of air 0.022 temperatureC/a in thehas pastexhibited 50 years a is highersignificant than the variation global or [16]. Northern China’s Hemisphereair temperature average increase temperature rate of 0.022 increase °C/a in ratethe past [17 ].50 The years northeast is regionhigher of China, than the with global an airor Northern temperature Hemisphere increase average rate of temperature 0.035 ◦C/a, increase is one of rate the [17]. areas The with northeast the fastest temperatureregion of riseChina, in with China. an air temperature increase rate of 0.035 °C/a, is one of the areas with the fastest temperatureThe rate of airrise temperaturein China. warming in the Sun Wu area is the highest in Northeast China [16]. AccordingThe to rate the of temperature air temperature data warming of the Sunwuin the Sun County Wu area Meteorological is the highest in Station, Northeast from China 1980 [16]. to 2018, According to the temperature data of the Meteorological Station, from 1980 to 2018, the annual average temperature, the annual maximum temperature, and the annual average minimum the annual average temperature, the annual maximum temperature, and the annual average temperature were plotted in Figure2. The temperature change rate during the 39 years from 1980 minimum temperature were plotted in Figure 2. The temperature change rate during the 39 years to 2018from was 1980 obtained to 2018 was by obtained linear fitting.by linear Thefitting. annual The annual average average minimum minimum temperature temperature increasedincreased at a ◦ significantat a significant rate of 0.1rate ofC 0.1 /a, °C which /a, which is the is largestthe largest and and about about 3.3 3.3 times times thethe annual average average maximum maximum air temperatureair temperature increases increases in Northeast in Northeast China. China.

Figure 2. 2. TheThe annual annual air air temperature temperature of Sunwu of Sunwu County, County, Heilongjiang Heilongjiang Province, Province, China, China, from 1980 from to 1980 2018. to 2018. The spatial distribution of island permafrost in the study area is highly variable and discrete. The spatial distribution of island permafrost in the study area is highly variable and discrete. The temperature, thickness, and thermal stability of permafrost, and the sensitivity of island The temperature, thickness, and thermal stability of permafrost, and the sensitivity of island permafrost permafrost to climatic and environmental changes have obvious local spatiotemporal differences to climatic[18]. In the and context environmental of climate change changes, a series have of obvious geological local disasters, spatiotemporal such as landslides, differences soil erosion, [18]. In the context of climate change, a series of geological disasters, such as landslides, soil erosion, and debris flows are prone to occur, and the severe soil impoverishment caused by island-type permafrost degradation may occur in some areas. The Bei’an to Heihe Expressway, which was built between 1997 and 2000, was unstable due to the melting of permafrost in 1999. Subsequently, the road was repaved to a ridge with stable geological Sensors 2019, 19, 1364 4 of 14 conditions. Since 1994, the area of cultivated land in the study area has increased, and the forest area has decreased dramatically. The surface temperature and its amplitude has increased, the seasonal melting depth has increased, and the total area of island permafrost has gradually decreased. Since 2000, the permafrost in the study area has been extensively degraded. In general, the degraded permafrost in our study area is highly sensitive under the described climatic conditions [19]. Any sustained warming or human activity can accelerate permafrost degradation, further destabilizing slopes and increasing the damage associated with permafrost. The InSAR monitoring results presented in this paper are of great significance for environmental assessment and regional infrastructure security in the study area.

3. Data Processing and Results In our study, a total of 20 L-band ALOS/PALASAR stripmap images over the Heihe area were used to retrieve the ground deformation caused by permafrost degeneration. The dataset was acquired along the ALOS satellite track 423 from June 2007 to December 2010 with Fine Beam Single (FBS) polarization or Fine Beam Double (FBD) polarization mode. The incidence angle of the image center point was about 38.75◦. Table1 details the information of the dataset used. The data coverage is shown in Figure1. Due to the dense vegetation cover in our region of interest and the periodical change of water content in the permafrost active layer, the application of InSAR is mainly limited by temporal decorrelation [20]. Nevertheless, the long wavelength (λ ≈ 23.6 cm) of the L-band is less affected by this limitation and maintains good coherence in the study area [20,21].

Table 1. Information of data used.

No. Acq. Date Orbit Frame Mode No. Acq. Date Orbit Frame Mode 1 20070610 07330 980 FBS 11 20090731 18737 980 FBS 2 20070726 08001 980 FBS 12 20090915 19408 980 FBS 3 20070910 08672 980 FBS 13 20091216 20750 980 FBD 4 20071211 10014 980 FBD 14 20100131 21421 980 FBD 5 20080126 10685 980 FBD 15 20100318 22092 980 FBD 6 20080427 12027 980 FBS 16 20100503 22763 980 FBS 7 20080612 12698 980 FBS 17 20100618 23434 980 FBS 8 20081213 15382 980 FBD 18 20100803 24105 980 FBS 9 20090128 16053 980 FBD 19 20100918 24776 980 FBS 10 20090615 18066 980 FBD 20 20101219 26118 980 FBD

In order to obtain highly coherent interferometric results, an elaborated data processing strategy combining the SBAS-InSAR and PS-InSAR methods is used to process a long dataset. Figure3 shows an elaborated data processing strategy for retrieving the deformation time-series and the corresponding mean deformation velocity. The processing strategy consists of three major parts, namely pre-processing, point targets selection, and deformation time-series retrieval. Data processing starts with a stack of focus Single Look Complex (SLC) images. Firstly, the GAMMA software (v20180704) [22] was used to import the dataset. Then, all SLC images were co-registered in GAMMA software. The image acquired on 31 July 2009 (No. 11, see Table1) was chosen as the reference image for SLC co-registration by minimizing the decorrelation. To reduce the topographic effect in image range offset, the SLC stack was co-registered with the geometrical SAR image co-registration method assisted by 1-arc-second (spatial resolution is about 30 m × 30 m) SRTM Digital Elevation Model (DEM). We only considered interferometric pairs with short perpendicular baseline and temporal baseline, i.e., 1.5 km, which is significantly smaller than the critical baseline [20]. Figure4 shows the network of selected interferometric pairs. Since the ALOS orbit was adjusted in mid-2008 [23], the long time-series images were divided into two subsets due to the large separation of baseline (i.e., 6–7). To connect the two subsets, several pairs with a relatively long baseline, namely 4–9, 4–10, 5–9, and 5–10, were selected manually. The coherences of these pairs remained at an acceptable level (coherence ≥ 0.4). SensorsSensors 20192019,, 1919,, xx FORFOR PEERPEER REVIEWREVIEW 55 ofof 1414

registeredregistered inin GAMMAGAMMA software.software. TheThe imageimage acquiredacquired onon 3131 JulyJuly 20092009 (No.(No. 11,11, seesee TableTable 1)1) waswas chosenchosen asas thethe referencereference imageimage forfor SLCSLC co-registrationco-registration byby minimizingminimizing thethe decorrelation.decorrelation. ToTo reducereduce thethe topographictopographic effecteffect inin imageimage rangerange offset,offset, thethe SLCSLC stackstack waswas co-registeredco-registered withwith thethe geometricalgeometrical SARSAR imageimage co-registrationco-registration methodmethod assistedassisted byby 1-arc-second1-arc-second (spatial(spatial resolutionresolution isis aboutabout 3030 mm ×× 3030 m)m) SRTMSRTM Digital Elevation Model (DEM). We only considered interferometricinterferometric pairspairs withwith shortshort perpendicularperpendicular baseline and temporal baseline, i.e., 1.5 km, which is significantlysignificantly smallersmaller thanthan thethe criticalcritical baselinebaseline [20].[20]. FigureFigure 44 showsshows thethe networknetwork ofof selectedselected interfinterferometricerometric pairs.pairs. SinceSince thethe ALOSALOS orbitorbit waswas adjustedadjusted inin mid-2008mid-2008 [23],[23], thethe longlong time-seriestime-series imagesimages werewere divideddivided intointo twotwo subsetssubsets duedue toto thethe largelarge separationseparation ofof baselinebaseline (i.e.,(i.e., 6–7).6–7). ToTo connectconnect thethe twotwo subsets,subsets, severalseveral pairspairs with a relatively long Sensorsbaseline,2019, 19 namely, 1364 4–9, 4–10, 5–9, and 5–10, were selectedlected manually.manually. TheThe coherencescoherences ofof thesethese pairspairs5 of 14 remainedremained atat anan acceptableacceptable levellevel (coherence(coherence ≥≥ 0.4).0.4).

Original SLC StackStack

DEM-assisted CoregisteredCoregistered SLCSLC ExternalExternal DEMDEM CoregistrationCoregistration StackStack

InterferogramInterferogram InterferogramInterferogram PhasePhase SimulationSimulation FormationFormation StackStack

DA Method

PTsPTs Phase,Phase, PTsPTs MergeMerge PTsPTs SelectionSelection Baseline...Baseline... SDSD MethodMethod

PhasePhase SmoothingSmoothing ResidualResidual Space-TemporalSpace-Temporal & Unwrapping Unwrapped Phase APS Filtering

Time Series LOS Deformation Model Time Series LOS Deformations & Modeling Deformation Deformations & Deformation Rate

FigureFigureFigure 3. 3.3. The The elaboratedelaborated data processing processing strategy. strategy.

Figure 4. The network of the selected interferometric pairs. Blue dots denote a SAR image acquisition. Lines between blue dots are interferometric pairs, the colors of which represent the average coherence (coh.) of the corresponding interferogram.

Subsequently, the original interferograms were generated by conjugate multiplication between the co-registered master SLC image and slave SLC image. Thus, a DEM was interpolated from 1-arc-second SRTM DEM, which had the same regular grid spacing as the reference SLC image. We utilized the DEM to simulate and remove the topographic phase contribution using the GAMMA software during the interferogram generation process. Then, we geocoded InSAR results from range-Doppler coordinates into map geometry corresponding to the WGS84 coordinate system using the GAMMA software. The selection of highly coherent point targets (PTs) is one of the key steps for our processing strategy and analysis. Currently, spatial coherence is usually used as a criterion for PTs selection. Sensors 2019, 19, 1364 6 of 14

However, the coherence would be overestimated in a lowly coherent area [24], as in the case of our study area. As a consequence, there would be many unreliable PTs selected. To select out reliable PTs, the Spectral Diversity (SD) [25] (implemented in GAMMA software) and Amplitude Dispersion (DA) method [26] derived from PS-InSAR [5] were used to select PTs with the coregistered SLC images, respectively. The spectral correlation minimum threshold was set to 0.30 and the DA threshold was set to 0.25. The final PTs list was the union of the PTs selected by the SD method and the PTs selected by the DA method. The interferometric phases and baseline information for each PT were extracted from the data grid. In order to alleviate the phase random noise, the interferometric phase was smoothed by a low-pass adaptive phase filter prior to phase unwrapping. The reference point was set to a stable area near the image center (see the black triangle in Figure5). The smoothed phases of PTs for each interferometric pair were unwrapped by a network programming-based phase unwrapping method [27]. Subsequently, the SBAS-InSAR method [6] was implemented to obtain the deformation time-series and the corresponding mean deformation velocity in the study area. A triangular filter was used for temporal and spatial atmospheric screen estimation [5,26]. Sensors 2019, 19, x FOR PEER REVIEW 7 of 14

FigureFigure 5. 5.Average Average annualannual deformationdeformation raterate alongalong the line of sight direction. The The background background is is DEM DEM shadedshaded relief, relief, and and thethe blackblack dasheddashed polygonspolygons denote the island island permafrost permafrost ar areas,eas, which which were were derived derived fromfrom Landsat Landsat images images in in our our previous previous work work [28 [28].]. The Th blacke black triangle triangle denotes denotes the the reference reference point, point, the boldthe blackbold black dashed dashed line denotesline denotes the expresswaythe expressway from from Bei’an Bei’an to to Heihe, Heihe, and and the the whitewhite and red red rectangles rectangles denotedenote selected selected areas areas for for further further analysisanalysis inin thisthis study.study.

4. Results, Analysis, and Discussion

4.1. Deformation Statistics with Topographic Factors It can be seen from Figure 5 that the deformation in the study area mainly occurred in the low- altitude valleys or flat areas, while there was almost no deformation in the mountainous areas. To further explore the relationship between ground deformation and elevation, we conducted a statistical analysis of the altitude and the deformation rate. Firstly, the elevation of the study area was divided into nine groups with an equal interval of 30 m. Then, we counted the number of PTs with significant deformation (i.e., a deformation rate >20 mm/a or < –20 mm/a) in each elevation group and calculated corresponding proportion. The statistical result is shown in Figure 6a. It can be seen that the proportion of PTs decreases with the altitude rises. Most of the significant deformation was located at lower altitudes of 210–300 meters, and that the altitude of 250 m had the highest possibility to deform. The results imply that the island permafrost at these altitudes was more possible to deform than at other elevations. Our results are in accordance with previous work [28], which found that the island permafrost was distributed in low-lying areas in Heilongjiang Province. Sensors 2019, 19, 1364 7 of 14

Finally, we obtained the deformation rate and the deformation time-series along the line of sight from 2006 to 2010 (refer to 10 June 2006). As shown in Figure5, the result was plotted on a DEM shade relief. The deformation rate ranged from –70 to 70 mm/a. Approximately 530,000 PTs were obtained in the experiment, most of which were nearly stable or showed slight deformation (ranging from −35 to 35 mm/a). However, there were some island-shaped deformation regions having a significant deformation rate. In order to show more details of the deformation, the island permafrost area was superimposed on the deformation rate figure [14]. It can be seen that most of the island permafrost area underwent significant deformation, substantially greater than 35 mm/a. These results indicate that the island permafrost was degrading. In addition, we shall note that there were several sites with positive deformation rate, meaning that the ground settlement moved towards the satellite. However, the positive deformation only occurred on low altitude or slope facing to satellite, see Figure5. Such positive deformation is very interesting. The high-altitude slope deformed, and the material migrated to lower areas, leading to possible uplift. Besides, when a slope facing to satellite, the slope settlement would also cause positive deformation rate. For the sake for simplicity, we only focused on the negative deformation rate in this paper.

4. Results, Analysis, and Discussion

4.1. Deformation Statistics with Topographic Factors It can be seen from Figure5 that the deformation in the study area mainly occurred in the low-altitude valleys or flat areas, while there was almost no deformation in the mountainous areas. To further explore the relationship between ground deformation and elevation, we conducted a statistical analysis of the altitude and the deformation rate. Firstly, the elevation of the study area was divided into nine groups with an equal interval of 30 m. Then, we counted the number of PTs with significant deformation (i.e., a deformation rate >20 mm/a or < −20 mm/a) in each elevation group and calculated corresponding proportion. The statistical result is shown in Figure6a. It can be seen that the proportion of PTs decreases with the altitude rises. Most of the significant deformation was located at lower altitudes of 210–300 m, and that the altitude of 250 m had the highest possibility to deform. The results imply that the island permafrost at these altitudes was more possible to deform than at other elevations. Our results are in accordance with previous work [28], which found that the island permafrost was distributed in Sensors 2019, 19, x FOR PEER REVIEW 8 of 14 low-lying areas in Heilongjiang Province.

(a) (b)

FigureFigure 6. (a 6.) significant(a) significant deformation deformation distribution distribution at different different altitudes. altitudes. (b) ( baspect) aspect distribution distribution of study of study area and deformation. area and deformation. The solar irradiance is greatest on the southern slope, followed by the southeast slope, southwest slope, east slope, and west slope. The solar irradiance on the northeast slope, northwest slope, and north slope are the smallest. Figure 6b shows the slope distribution statistics of the study area and the slope distribution statistics of the PTs with significant deformation, obtained by superposition analysis of the deformation area and the slope direction obtained by 1 arc-second SRTM DEM. We found that the deformation area was distributed in shadow with less solar radiation, including the north slope, the northwest slope, and the northeast slope. This is consistent with the slope distribution of permafrost in study area.

4.2. Correlation Analysis with Air Temperature The island-shaped and sparse island-like permafrost in the study area is widely distributed, and is the transition zone between permafrost and seasonal permafrost. Permafrost is extremely sensitive to surface heat exchange conditions [29]. According to the temperature monitoring data of Sunwu County Meteorological Station, since 1981, the temperature in the study area had shown a significant rising trend. The average annual temperature of Sunwu County from 1981 to 2018 is shown in Figure 3. Regarding the trend of rising temperature, it can be found that the annual average minimum temperature growth in the region is the fastest, with a growth rate of 0.1 °C/a. The increase of the annual average minimum temperature in the study area is greater than the increase of the annual average maximum temperature, indicating that the main reason for the increase in temperature is not the increase in direct radiation but the decrease of soil heat demand and the increase of soil radiation to the atmosphere [16]. In order to show further details of the experimental results, several island permafrost deformation areas were selected for analysis. Selected area P1 (see the white rectangle in Figure 5) is located on the beach of a river valley, as shown in Figure 7, where the water content in soil is high, the altitude is low, and the terrain is flat. Compared with the optical images acquired in 2004 and 2011, it can be seen that the land surface features had changed. Sensors 2019, 19, 1364 8 of 14

The solar irradiance is greatest on the southern slope, followed by the southeast slope, southwest slope, east slope, and west slope. The solar irradiance on the northeast slope, northwest slope, and north slope are the smallest. Figure6b shows the slope distribution statistics of the study area and the slope distribution statistics of the PTs with significant deformation, obtained by superposition analysis of the deformation area and the slope direction obtained by 1 arc-second SRTM DEM. We found that the deformation area was distributed in shadow with less solar radiation, including the north slope, the northwest slope, and the northeast slope. This is consistent with the slope distribution of permafrost in study area.

4.2. Correlation Analysis with Air Temperature The island-shaped and sparse island-like permafrost in the study area is widely distributed, and is the transition zone between permafrost and seasonal permafrost. Permafrost is extremely sensitive to surface heat exchange conditions [29]. According to the temperature monitoring data of Sunwu County Meteorological Station, since 1981, the temperature in the study area had shown a significant rising trend. The average annual temperature of Sunwu County from 1981 to 2018 is shown in Figure3. Regarding the trend of rising temperature, it can be found that the annual average minimum temperature growth in the region is the fastest, with a growth rate of 0.1 ◦C/a. The increase of the annual average minimum temperature in the study area is greater than the increase of the annual average maximum temperature, indicating that the main reason for the increase in temperature is not the increase in direct radiation but the decrease of soil heat demand and the increase of soil radiation to the atmosphere [16]. In order to show further details of the experimental results, several island permafrost deformation areas were selected for analysis. Selected area P1 (see the white rectangle in Figure5) is located on the beach of a river valley, as shown in Figure7, where the water content in soil is high, the altitude is low, and the terrain is flat. Compared with the optical images acquired in 2004 and 2011, it can be seen that the land surface features had changed. Sensors 2019, 19, x FOR PEER REVIEW 9 of 14

FigureFigure 7. Optical7. Optical images images of of selected selected areaarea P1P1 andand its deformation. (a (a) )image image acquired acquired on on 15 15 June June 2004; 2004; (b)(b image) image acquired acquired on on 19 19 July July 2011; 2011; ( c(c)) magnificationmagnification of the white white dashed dashed rectangle rectangle in in (c) (c) with with the the deformationdeformation rate rate overlap; overlap; the the white white dashed dashed polygonpolygon denotesdenotes the island island permafrost permafrost area. area.

AccordingAccording to theto the land land use use dynamic dynamic change change of of Sunwu Sunwu County County from from 1996 1996 to to 2010, 2010, the the increase increase rate of arablerate of landarable and land garden and garden land was land about was about 118% 118% and727%, and 727%, and theand forest the forest land land decreased decreased by aboutby 22%about [30]. 22% The [30]. ever The forest ever vegetation forest vegetation was replaced was replaced by reclaimed by reclaimed farmland, farmland, development development of urban of constructionurban construction implying implying that the humanthat the activities human activities had aggravated had aggravated the degradation the degradation process of process permafrost of permafrost in this region. Due to the interference of human activities, the island permafrost (see in this region. Due to the interference of human activities, the island permafrost (see Figure7c) Figure 7c) underwent continued degradation, causing ground deformation. The deformation profile underwent continued degradation, causing ground deformation. The deformation profile of area P1 is of area P1 is shown in Figure 8. The result shows obvious ground deformation, indicating that island shown in Figure8. The result shows obvious ground deformation, indicating that island permafrost permafrost was undergoing degradation. According to the comparison of the optical images acquired in the same season (see Figure 7a and 7b), we can see that the ever-vegetated areas have been reclaimed for agricultural use. This change leads us to believe that the degradation of island permafrost will be accelerated by human activities.

Figure 8. Profile of deformation time-series along the red line in Figure 7c for the selected area P1.

Furthermore, in order to show the relationship between the air temperature and ground deformations, we plotted the deformations of 13 selected areas (see red dashed rectangles in Figure 5) and air temperature in Figure 9. In the warm season, from April to September, the temperature causes the soil to absorb more heat than it loses, especially in July and August, when the temperature rises dramatically. Consequently, the permafrost melting and surface subsidence Sensors 2019, 19, x FOR PEER REVIEW 9 of 14

Figure 7. Optical images of selected area P1 and its deformation. (a) image acquired on 15 June 2004; (b) image acquired on 19 July 2011; (c) magnification of the white dashed rectangle in (c) with the deformation rate overlap; the white dashed polygon denotes the island permafrost area.

According to the land use dynamic change of Sunwu County from 1996 to 2010, the increase rate of arable land and garden land was about 118% and 727%, and the forest land decreased by about 22% [30]. The ever forest vegetation was replaced by reclaimed farmland, development of urban construction implying that the human activities had aggravated the degradation process of

Sensorspermafrost2019, 19 in, 1364 this region. Due to the interference of human activities, the island permafrost (see9 of 14 Figure 7c) underwent continued degradation, causing ground deformation. The deformation profile of area P1 is shown in Figure 8. The result shows obvious ground deformation, indicating that island waspermafrost undergoing was degradation.undergoing Accordingdegradation. to theAccording comparison to ofthe the comparison optical images of the acquired optical in theimages same seasonacquired (see in Figure the same7a,b), season we can (see see Figure that the 7aever-vegetated and 7b), we can areas see havethat the been ever-vegetated reclaimed for areas agricultural have use.been This reclaimed change for leads agricultural us to believe use. thatThis thechange degradation leads us ofto islandbelieve permafrost that the degradation will be accelerated of island by humanpermafrost activities. will be accelerated by human activities.

FigureFigure 8. 8. ProfileProfile of of deformation deformation time-series time-series along along the the red red line line in inFigure Figure 7c7 forc for the the selected selected area area P1. P1.

Furthermore,Furthermore, in order to show the the relation relationshipship between between the the air air temperature temperature and and ground ground deformations,deformations, wewe plotted plotted the the deformations deformations of 13of selected13 selected areas areas (see red(see dashed red dashed rectangles rectangles in Figure in5 ) andFigure air temperature5) and air temperature in Figure9. Inin the Figure warm 9. season, In the from warm April season, to September, from April the to temperature September, causes the thetemperature soil to absorb causes more the heatsoil to than absorb it loses, more especially heat than init loses, July andespecially August, in whenJuly and the August, temperature when risesthe Sensors 2019, 19, x FOR PEER REVIEW 10 of 14 dramatically.temperature Consequently,rises dramatically. the permafrost Consequently, melting the and permafrost surface subsidence melting and peaked. surface The subsidence temperature startedpeaked. to The drop temperature around October, started and to thedrop heat around released October, from depthand the in theheat soil released was greater from depth than the in heatthe absorbedsoil was greater from the than atmosphere, the heat absorbed which caused from thethe wateratmosphere, in the soil which to freeze, caused resulting the water in in frost the heavingsoil to offreeze, the ground resulting surface. in frost heaving of the ground surface.

FigureFigure 9.9. TheThe deformation time-series for area P1 P1 and and the the temperature. temperature.

TheThe timetime toto reachreach thethe maximummaximum meltingmelting depth and the time time to to reach reach the the maximum maximum degree degree of of frostfrost heave heave will will bebe slightly slightly laterlater thanthan thethe timetime whenwhen thethe highest/lowesthighest/lowest temperature temperature peak peak appears, appears, whichwhich is is consistent consistent withwith thethe generalgeneral lawlaw ofof permafrostpermafrost deformation. Furthermore, Furthermore, we we calculated calculated the the deformation lag time to be about 25 days. The deformation fitting curves (see Figure 9) are composed of an annual periodic function and a linear function. Periodicity is due to seasonal variation. The linear subsidence is related to the degradation of permafrost. After the first winter (2007) in the study period, a freezing process was carried out, followed by an apparent settlement in the summer. In the next two years, the frost heave of permafrost in winter was lower than the deformation variable caused by melting settlement. The subsidence trend caused by permafrost degradation was obvious. It can also be clearly seen from the change of the positive and negative temperature duration of the annual average minimum temperature in Sunwu County during the superposition study period that the days with continuous negative temperature decreased and the days with continuous positive temperature increased. Based on the change of the annual air temperature, not only did the average minimum temperature rise, but the corresponding period below 0 °C also shortened in our study area (see Figure 10a). The change indirectly indicates that the permafrost melting period is prolonged and the corresponding frost heave period is shortened. Such climate change is likely to lead to the degradation of island permafrost in the study area. To further illustrate the correlation between air temperature change and deformation, we removed the annual periodicity from temperature change and deformation, respectively, and obtained the corresponding linear trend, as shown in Figure 10b. By analyzing the correlation between the deformation trend and the temperature increase trend, it was found that the correlation between the deformation trend and the temperature change varies from –0.78 to –0.85, with an average of –0.80. This indicates that air temperature increase is an important factor affecting island permafrost deformation. Sensors 2019, 19, 1364 10 of 14 deformation lag time to be about 25 days. The deformation fitting curves (see Figure9) are composed of an annual periodic function and a linear function. Periodicity is due to seasonal variation. The linear subsidence is related to the degradation of permafrost. After the first winter (2007) in the study period, a freezing process was carried out, followed by an apparent settlement in the summer. In the next two years, the frost heave of permafrost in winter was lower than the deformation variable caused by melting settlement. The subsidence trend caused by permafrost degradation was obvious. It can also be clearly seen from the change of the positive and negative temperature duration of the annual average minimum temperature in Sunwu County during the superposition study period that the days with continuous negative temperature decreased and the days with continuous positive temperature increased. Based on the change of the annual air temperature, not only did the average minimum temperature rise, but the corresponding period below 0 ◦C also shortened in our study area (see Figure 10a). The change indirectly indicates that the permafrost melting period is prolonged and the corresponding frost heave period is shortened. Such climate change is likely to lead to the degradationSensors 2019, 19 of, x islandFOR PEER permafrost REVIEW in the study area. 11 of 14

◦ FigureFigure 10.10.( a(a)) thethe deformationdeformation ofof areaarea P1P1 andand thethe periodperiod ofof below/abovebelow/above 0 0 °CC;; (b (b) )linear linear residual residual of of deformationdeformation andand fittingfitting linearlinear residualresidual ofof airair temperature.temperature.

4.3. ToPotential further Geological illustrate Hazards the correlation between air temperature change and deformation, we removed the annual periodicity from temperature change and deformation, respectively, and obtained the correspondingUnder global linearclimate trend, warming as shown conditions, in Figure the pe10rmafrostb. By analyzing environment the correlation in the Heihe between area theis deformationfacing a great trend threat. and The the physical temperature and mechanical increase trend, properties it was of found the original that the permafrost correlation have between changed the deformationgreatly [31]. trend As an and early the temperatureproject developed change and varies cons fromtructed−0.78 in to the−0.85, northernwith an permafrost average of region,−0.80. Thisthe indicatesBei’an–Heihe that air expressway temperature section increase is located is an importantin the perm factorafrost affecting degradation island region. permafrost Being deformation.in a state of vulnerability, permafrost is very sensitive to human activities, such as engineering construction. Even 4.3.weak Potential thermal Geological influence Hazards will lead to great changes that accelerate permafrost degradation. The changes related to the Bei’an–Heihe expressway, a selected area P2 (see the white rectangle in Under global climate warming conditions, the permafrost environment in the Heihe area is facing a Figure 5), are shown in Figure 11. The expressway, which was designed and constructed from 1999 great threat. The physical and mechanical properties of the original permafrost have changed greatly [31]. to 2000, was originally located on lower altitude slopes. However, the permafrost slope became As an early project developed and constructed in the northern permafrost region, the Bei’an–Heihe unstable after construction [32], and subsequently a realignment of the expressway was carried out expressway section is located in the permafrost degradation region. Being in a state of vulnerability, in 2003. permafrost is very sensitive to human activities, such as engineering construction. Even weak thermal influence will lead to great changes that accelerate permafrost degradation. The changes related to the Bei’an–Heihe expressway, a selected area P2 (see the white rectangle in Figure5), are shown in Figure 11. The expressway, which was designed and constructed from 1999 to 2000, was originally located on lower altitude slopes. However, the permafrost slope became unstable after construction [32], and subsequently a realignment of the expressway was carried out in 2003.

Figure 11. Historic changes of the Bei’an–Heihe expressway (Landsat images). Sensors 2019, 19, x FOR PEER REVIEW 11 of 14

Figure 10. (a) the deformation of area P1 and the period of below/above 0 °C; (b) linear residual of deformation and fitting linear residual of air temperature.

4.3. Potential Geological Hazards Under global climate warming conditions, the permafrost environment in the Heihe area is facing a great threat. The physical and mechanical properties of the original permafrost have changed greatly [31]. As an early project developed and constructed in the northern permafrost region, the Bei’an–Heihe expressway section is located in the permafrost degradation region. Being in a state of vulnerability, permafrost is very sensitive to human activities, such as engineering construction. Even weak thermal influence will lead to great changes that accelerate permafrost degradation. The changes related to the Bei’an–Heihe expressway, a selected area P2 (see the white rectangle in Figure 5), are shown in Figure 11. The expressway, which was designed and constructed from 1999 to 2000, was originally located on lower altitude slopes. However, the permafrost slope became Sensorsunstable2019, after19, 1364 construction [32], and subsequently a realignment of the expressway was carried11 out of 14 in 2003.

Sensors 2019, 19, x FigureFORFigure PEER 11. 11. REVIEWHistoric Historic changeschanges ofof thethe Bei’an–HeiheBei’an–Heihe expressway (Landsat(Landsat images). 12 of 14

TheThe InSARInSAR monitoringmonitoring resultsresults show that there was significant significant ground ground deformation deformation caused caused by by permafrostpermafrost melting, melting, as as shown shown in in Figure Figure 12 .12. It canIt can be seenbe seen that that the newthe new road road has beenhas been shifted shifted to a higher to a altitude,higher altitude, which is which out of is the out island of the permafrost island perm area.afrost However, area. thereHowever, is still there significant is still deformationsignificant threateningdeformation the threatening expressway. the The expressway. deformation The time-series deformation for thetime-series selected areafor the P2 presentsselected aarea stepped P2 shape,presents and a thestepped overall shape, trend and is downward.the overall Ittrend can beis downward. seen that this It regioncan be experiencedseen that this almost region no thawingexperienced deformation almost no in thawing summer deformation and permafrost in summer frost and heaving permafrost in winter. frost heaving With the in increase winter. With of air temperature,the increase of the air water temperature, content inthe the water soil content increased in the after soil the increased permafrost after degraded, the permafrost which degraded, decreased thewhich friction decreased of the the active friction layer. of the The active soil was layer. prone The soil to deform was prone with to thedeform factor wi ofth gravitythe factor on of the gravity slope ◦ (abouton the 15slope), and (about the 15°), deformation and the deformation rate of the permafrostrate of the permafrost increased. increased. This is a reflection This is a reflection of the rapid of degradationthe rapid degradation of the island of the permafrost island permafrost in the study in th areae study under area the under increasing the increasing air temperature. air temperature.

FigureFigure 12.12. ((aa)) deformationdeformation ofof thethe selectedselected areaarea P2.P2. The The white white dashed dashed polygon polygon denotes denotes the the island island permafrost.permafrost. ((bb)) thethe deformationdeformation time-seriestime-series inin the region indicated by the the red red arrow. arrow.

ItIt can can also also be be seen seen from from this casethis thatcase mechanicalthat mechanical properties properties such as such stress as and stress strain and of permafroststrain of arepermafrost affected are by thawingaffected by due thawing to the due influence to the ofinfl airuence temperature of air temperature increase. increase. Besides, Besides, farming farming and/or and/or deforestation activities are also factors which accelerate the permafrost degradation. As a consequence, geological disasters such as landslides tend to occur. All of these factors will bring difficulties to local engineering investigation, construction design, quality control, and maintenance. InSAR, as a practical technique, can be applied to monitor and analyze the ground deformation of permafrost areas that are challenging for conducting field investigations.

5. Conclusions In the context of global warming, the air temperature of Heihe, China, has increased significantly, resulting in the degradation of island permafrost. In this paper, a large-scale ground deformation survey was conducted in the Heihe area by using a time-series InSAR technique. The findings are concluded as follows: 1) The time deformation of the Bei’an–Heihe area was obtained. From 2007 to 2010, the area had a deformation rate of −70 to 70 mm/a. The ground deformation caused by the thawing of permafrost was greater than that caused by frost heaving. The settlement characteristics of the permafrost in the study area were closely related to the altitude, being mainly distributed in low-altitude areas. 2) The significant increase of temperature has intensified the development of warming and drying trends in the Heihe area. The island permafrost in the study area is sensitive to climate change. Based on a joint analysis with temperature, we found that the correlation between the deformation trend and temperature change varied from −0.78 to −0.85, with an average value of −0.80, indicating a significant correlation between degradation grade and temperature increase. 3) Permafrost degradation caused by rising temperature affects the safety of expressway construction in cold regions. Based on an analysis of the settlement on the side slope of a Sensors 2019, 19, 1364 12 of 14 deforestation activities are also factors which accelerate the permafrost degradation. As a consequence, geological disasters such as landslides tend to occur. All of these factors will bring difficulties to local engineering investigation, construction design, quality control, and maintenance. InSAR, as a practical technique, can be applied to monitor and analyze the ground deformation of permafrost areas that are challenging for conducting field investigations.

5. Conclusions In the context of global warming, the air temperature of Heihe, China, has increased significantly, resulting in the degradation of island permafrost. In this paper, a large-scale ground deformation survey was conducted in the Heihe area by using a time-series InSAR technique. The findings are concluded as follows:

(1) The time deformation of the Bei’an–Heihe area was obtained. From 2007 to 2010, the area had a deformation rate of −70 to 70 mm/a. The ground deformation caused by the thawing of permafrost was greater than that caused by frost heaving. The settlement characteristics of the permafrost in the study area were closely related to the altitude, being mainly distributed in low-altitude areas. (2) The significant increase of temperature has intensified the development of warming and drying trends in the Heihe area. The island permafrost in the study area is sensitive to climate change. Based on a joint analysis with temperature, we found that the correlation between the deformation trend and temperature change varied from −0.78 to −0.85, with an average value of −0.80, indicating a significant correlation between degradation grade and temperature increase. (3) Permafrost degradation caused by rising temperature affects the safety of expressway construction in cold regions. Based on an analysis of the settlement on the side slope of a selected section, we believe that the influence of permafrost degradation will have an increasingly significant effect on construction in cold regions.

Permafrost degradation is accompanied by soil moisture loss, land desertification, vegetation coverage reduction, groundwater pollution, and other problems, affecting the microbial permafrost, carbon cycle, and cold region ecological hydrology. Therefore, it is important to monitor the tundra in the Heihe area. However, there are many factors that influence the deformation process of permafrost. Therefore, future work should further improve deformation observation (in space and time) by combining more abundant data, and establish the interaction between climate change and permafrost degradation.

Author Contributions: S.W. and B.X. conceived the research work, and S.W. wrote the first draft of the paper. W.S., J.S., Z.L. and G.F. contributed to experimental implementation and result interpretation. All authors contributed to paper writing and revision. Funding: This work was supported in part by the National Natural Science Foundation of China (Nos. 41804008, 41474007 and 41574005), the National Key R&D Program of China (2018YFC1505101), and the Major Projects of High-Resolution Earth Observation System of China (Civil Part) (No. 03-Y20A11-9001-15/16). Acknowledgments: The authors would like to thank the Japan Aerospace Exploration Agency (JAXA) for providing the ALOS/PALSAR data under RA6 data category (Project ID: P3214002). The optical images were provided by the United States Geological Survey (USGS). General Mapping Tools and Gnuplot were used to produce several figures. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Ran, Y.H.; Xin, L.; Cheng, G.D.; Zhang, T.J.; Wu, Q.B.; Jin, H.J.; Rui, J. Distribution of permafrost in China: An overview of existing permafrost maps. Permafr. Periglac. Process. 2012, 23, 322–333. [CrossRef] 2. Jorgenson, M.T.; Racine, C.H.; Walters, J.C.; Osterkamp, T.E. Permafrost degradation and ecological changes associated with a warmingclimate in central alaska. Clim. Chang. 2001, 48, 551–579. [CrossRef] Sensors 2019, 19, 1364 13 of 14

3. Kharuk, V.I.; Shushpanov, A.S.; Im, S.T.; Ranson, K.J. Climate-induced landsliding within the larch dominant permafrost zone of central siberia. Environ. Res. Lett. 2016, 11, 045004. [CrossRef][PubMed] 4. Hooper, A.; Bekaert, D.; Spaans, K.; Arıkan, M. Recent advances in sar interferometry time series analysis for measuring crustal deformation. Tectonophysics 2011, 514, 1–13. [CrossRef] 5. Ferretti, A.; Prati, C.; Rocca, F. Permanent scatterers in sar interferometry. IEEE Trans. Geosci. Remote Sens. 2001, 39, 8–20. [CrossRef] 6. Berardino, P.; Fornaro, G.; Lanari, R.; Sansosti, E. A new algorithm for surface deformation monitoring based on small baseline differential sar interferograms. IEEE Trans. Geosci. Remote Sens. 2002, 40, 2375–2383. [CrossRef] 7. Massonnet, D.; Rossi, M.; Carmona, C.; Adragna, F.; Peltzer, G.; Feigl, K.; Rabaute, T. The displacement field of the landers earthquake mapped by radar interferometry. Nature 1993, 364, 138–142. [CrossRef] 8. Jiang, L.; Lin, H.; Jianwei, M.A.; Kong, B.; Wang, Y. Potential of small-baseline sar interferometry for monitoring land subsidence related to underground coal fires: Wuda (northern China) case study. Remote Sens. Environ. 2011, 115, 257–268. [CrossRef] 9. Singhroy, V.; Mattar, K.E.; Gray, A.L. Landslide characterisation in canada using interferometric sar and combined sar and tm images. Adv. Space Res. 1998, 21, 465–476. [CrossRef] 10. Deguchi, T.; Narita, T. Monitoring of Land Deformation Due to Oil Production by Insar Time Series Analysis. In Proceedings of the Fringe 2015 Workshop, Frascati, Italy, 23–27 March 2015. 11. Bell, J.W.; Amelung, F.; Ferretti, A.; Bianchi, M.; Novali, F. Permanent scatterer insar reveals seasonal and long-term aquifer-system response to groundwater pumping and artificial recharge. Water Resour. Res. 2008, 44.[CrossRef] 12. Rong, Z. Monitoring surface deformation over permafrost with an improved sbas-insar algorithm: With emphasis on climatic factors modeling. Remote Sens. Environ. 2016, 184, 276–287. 13. Short, N.; Brisco, B.; Couture, N.; Pollard, W.; Murnaghan, K.; Budkewitsch, P. A comparison of terrasar-x, radarsat-2 and alos-palsar interferometry for monitoring permafrost environments, case study from herschel island, canada. Remote Sens. Environ. 2011, 115, 3491–3506. [CrossRef] 14. Wang, C. Study on Surface Deformation in Degraded Permafrost Region of Northeast China Based on d-Insar. Ph.D. Thesis, Northeast Forestry University, Harbin, China, 2015. (In Chinese) 15. Shi, P.J.; Shao, S.; Ming, W.; Ning, L.; Wang, J.A.; Jin, Y.Y.; Gu, X.T.; Yin, W.X. Climate change regionalization in China (1961–2010). Sci. Chin. Earth Sci. 2014, 57, 2676–2689. [CrossRef] 16. Wei, S.; Hu, Z.; Ying, G.; Zhang, C.; Wang, C.; Hua, J.; Yao, L.; Xiao, J. The impact of climate change on landslides in southeastern of high-latitude permafrost regions of China. Front. Earth Sci. 2015, 3, 7. 17. Edenhofer, O.; Seyboth, K. Intergovernmental panel on climate change (ipcc). Encycl. Energy Nat. Res. Environ. Econ. 2013, 26, 48–56. 18. Zhou, Y.; Wang, Y.; Gao, X.; Yue, H. Ground temperature, permafrost distribution and climate warming in northeastern China. J. Glaciol. Geocryol. 1996, 139–146. 19. He, R.-X.; Jin, H.-J.; Chang, X.-L.; Lv, L.-Z.; Yu, S.-P.; Yang, S.-Z.; Wang, S.-L.; Sun, G.-Y. Degradation of permafrost in the northern part of northeastern China:Present state and causal analysis. J. Glaciol. Geocryol. 2009, 31, 829–834. 20. Zebker, H.A.; Villasenor, J. Decorrelation in interferometric radar echoes. IEEE Trans. Geosci. Remote Sens. 1992, 30, 950–959. [CrossRef] 21. Rosen, P.A.; Hensley, S.; Joughin, I.R.; Li, F.K.; Madsen, S.N.; Rodriguez, E.; Goldstein, R.M. Synthetic aperture radar interferometry. Proc. IEEE 2000, 88, 333–382. [CrossRef] 22. Werner, C.; Wegmüller, U.; Strozzi, T.; Wiesmann, A. Gamma Sar and Interferometric Processing Software; Citeseer: Gothenburg, Sweden, 2000; p. 1620. 23. Hashimoto, M.; Fukushima, Y. An attempt to detect secular deformation associated with the subduction of the philippine sea plate with alos/palsar. In Proceedings of the Agu Fall Meeting, San Francisco, CA, USA, 14–18 December 2009. 24. Jiang, M.; Ding, X.; Li, Z.; Jiang, M. Hybrid approach for unbiased coherence estimation for multitemporal insar. IEEE Trans. Geosci. Remote Sens. 2014, 52, 2459–2473. [CrossRef] 25. Werner, C.; Wegmuller, U.; Strozzi, T.; Wiesmann, A. Interferometric point target analysis for deformation mapping. In Proceedings of the 2003 IEEE International Geoscience and Remote Sensing Symposium. (IEEE Cat. No. 03CH37477), Toulouse, France, 21–25 July 2003; pp. 4362–4364. Sensors 2019, 19, 1364 14 of 14

26. Ferretti, A.; Prati, C.; Rocca, F. Nonlinear subsidence rate estimation using permanent scatterers in differential sar interferometry. IEEE Trans. Geosci. Remote Sens. 2000, 38, 2202–2212. [CrossRef] 27. Costantini, M. A novel phase unwrapping method based on network programming. IEEE Trans. Geosci. Remote Sens. 1998, 36, 813–821. [CrossRef] 28. Wang, C.; Wei, S.; Ying, G.; Hu, Z.; Hua, J. Permafrost Distribution Study Based on Landsat Etm+ Imagery of The Northwest Section of the Range, China; Springer: Berlin, Germany, 2014. 29. Jin, H.J.; Shu xun, L.I.; Wang, S.L.; Zhao, L. Impacts of climatic change on permafrost and cold regions environments in China. Acta Geogr. Sin. 2000, 55, 161–173. 30. Wang, J. Impact of Land Use Planning on Land Use Change in Sunwu County. Master’s Thesis, Northeast Forestry University, Harbin, China, 2012. (In Chinese) 31. Jin, H.; Yu, Q.; Lü, L.; Guo, D.; He, R.; Yu, S.; Sun, G.; Li, Y. Degradation of permafrost in the xing’anling mountains, northeastern China. Permafr. Perigl. Process. 2010, 18, 245–258. [CrossRef] 32. Shi, X.; Liao, M.; Teng, W.; Lu, Z.; Wei, S.; Wang, C. Expressway deformation mapping using high-resolution terrasar-x images. Remote Sens. Lett. 2014, 5, 194–203. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).