DOI QR코드

DOI QR Code

Monitoring Time-Series Subsidence Observation in Incheon Using X-Band COSMO-SkyMed Synthetic Aperture Radar

  • Sang-Hoon Hong (Department of Geological Sciences, Pusan National University)
  • Received : 2024.03.25
  • Accepted : 2024.04.30
  • Published : 2024.04.30

Abstract

Ground subsidence in urban areas is mainly caused by anthropogenic factors such as excessive groundwater extraction and underground infrastructure development in the subsurface composed of soft materials. Global Navigation Satellite System data with high temporal resolution have been widely used to measure surface displacements accurately. However, these point-based terrestrial measurements with the low spatial resolution are somewhat limited in observing two-dimensional continuous surface displacements over large areas. The synthetic aperture radar interferometry (InSAR) technique can construct relatively high spatial resolution surface displacement information with accuracy ranging from millimeters to centimeters. Although constellation operations of SAR satellites have improved the revisit cycle, the temporal resolution of space-based observations is still low compared to in-situ observations. In this study, we evaluate the extraction of a time-series of surface displacement in Incheon Metropolitan City, South Korea, using the small baseline subset technique implemented using the commercial software, Gamma. For this purpose, 24 COSMO-SkyMed X-band SAR observations were collected from July 12, 2011, to August 27, 2012. The time-series surface displacement results were improved by reducing random phase noise, correcting residual phase due to satellite orbit errors, and mitigating nonlinear atmospheric phase artifacts. The perpendicular baseline of the collected COSMO-SkyMed SAR images was set to approximately 2-300 m. The surface displacement related to the ground subsidence was detected approximately 1 cm annually around a few Incheon Subway Line 2 route stations. The sufficient coherence indicates that the satellite orbit has been precisely managed for the interferometric processing.

Keywords

1. Introduction

Ground subsidence refers to the gradual settlement or sudden sinking of the Earth’s surface owing to the compaction of materials constituting the subsurface consolidation processes within the Earth (Milliman and Haq, 1996). Natural ground subsidence is mainly caused by limestone dissolution induced by fluid movement, such as groundwater, in soil or bedrock, and it is primarily observed in karst environments. However, land reclamation projects in weak ground areas or excessive extraction of subsurface fluids like groundwater or oil are considered anthropogenic causes of ground subsidence. The ground subsidence rate depends on the type and morphology of the materials composing the ground and varies from several months to a few decades. In areas where cavities exist in the subsurface, sudden land collapse or sinkholes can occur when the surface load exceeds the threshold (Bruno, 1992). Estuarine areas, usually characterized by sedimentary terrains and beneficial for human habitation facilitating urban development, often experience frequent ground subsidence. Increased surface loading from building construction on soft ground materials and the construction of social underground structures are the leading causes of significant ground subsidence, a severe social issue in urban environments. Furthermore, the influx of seawater into subsiding ground environments exacerbates the potential for flooding disasters due to climate change with the effects of sea-level rise due to global warming (Park and Hong, 2021).

Optical satellites that observe the Earth using visible and infrared bands utilize sunlight as the primary light source in the daytime. However, synthetic aperture radar (SAR) satellites equipped with microwave bands can observe surface targets even under undesirable weather conditions, such as rain or clouds not observable by optical satellites or during night time when the sunlight is absent (Moreira et al., 2013). The SAR payloads directly transmit and receive microwave pulse signals generated by the satellite platform, acquiring complex data types that can be translated into amplitude and phase. The amplitude is the intensity of the electromagnetic microwave wave backscattered from the target, reflecting the physical characteristics of the scattering object. The phase, obtainable only from the SAR sensors, contains the distance information between the target and the sensor, allowing extraction of topographic elevation of the target surface and precise observation of surface displacement with accuracy ranging from millimeters to centimeters(Hanssen, 2001). Differential synthetic aperture radar interferometry (DInSAR) is an advanced technique that quantitatively extracts surface displacements caused by earthquakes, volcanoes, landslides, ground subsidence, glacier movements, and water level changes by utilizing phase differences between two or more SAR images (Gabriel et al., 1989; Kampes, 2006; Massonnet and Feigl, 1998; Zebker et al., 1994). The DInSAR technique has been widely applied worldwide to monitor geological hazards, assess damage restoration plans, and understand surface displacement’s geoscientific background. Unlike field survey observation methods such as point-based Global Navigation Satellite System (GNSS) measurements, which observe a limited area with low spatial resolution, the DInSAR applications can simultaneously observe surface displacements over vast areas, making it highly economical (Abidin et al., 2008; Hong et al., 2009; Wright et al., 2004).

The conventional DInSAR can help detect large-scale surface displacements that occur quickly due to geological hazards such as earthquakes or volcanoes. However, a large temporal baseline between the acquired SAR images makes it difficult to selectively separate pure surface displacement components because additional phase signals due to atmospheric artifacts from the troposphere or ionosphere and seasonal scattering variations in the Earth’s surface inevitably become included (Fattahi et al., 2017; Jolivet et al., 2011; Jung et al., 2012; Kim et al., 2013). If statistical modeling of the surface displacements using a time-series of differential interferogram in a polynomial enables the exclusion of the residual random phases, which can be considered as noise from the actual surface displacement. A least square or singular value decomposition (SVD) method is a popular numerical analysis to model the time-series of surface displacements.

An integrating DInSAR time-series analysis based on large amounts of SAR data was proposed to extract temporal trends in phase changes related to surface displacements over long periods (Berardino et al., 2002; Ferretti et al., 2001). The small baseline subset (SBAS) method utilizes high-density interferograms over distributed targets by limiting the spatial geometrical baseline distance. In contrast, the persistent scatterer interferometry (PSI) method selectively extracts multiple points showing stable phases to estimate surface displacement over a long period (Berardino et al., 2002; Ferretti et al., 2001). The time-series of multiple differential interferograms is more effective for monitoring gradual surface displacements such as ground subsidence, especially for rapidly investigating ground subsidence and mitigating damage in urban areas. However, a considerable computational load for the data processing is required due to the large amount of data used.

This study evaluates a time-series surface displacement due to ground subsidence in Incheon Metropolitan City, South Korea. The data utilized in the study are COSMO-SkyMed X-band SAR data still in operation by the Italian Space Agency (ASI). The SBAS processing technique included in the Interferometric Point Target Analysis Software (IPTA) module provided by Gamma, a commercial software, was utilized for data processing. An automated time-series differential DInSAR script developed by Pusan National University was applied to streamline data processing. This study aims to preliminary analyze possible ground subsidence using high-resolution X-band SAR data temporarily acquired in the early 2010s.

2. Study Area and Data

2.1. Study Area

The study area is located in Incheon Metropolitan City on the west coast of South Korea, partially encompassing the southern part of Ganghwa Island and the eastern part of Yeongjong Island. Although the swath width of the collected SAR observation also covers some areas of Seoul, Goyang City, Gimpo City, Bucheon City, Siheung City, and Gwangmyeong City, this study focuses primarily on discussing surface displacement occurred in Incheon Metropolitan City. Road collapse accidents occurred around the construction site of the Incheon Subway in February 2012, and localized ground subsidence was reported near the construction site of Yeongjong Sky City in July 2014 (Kim, 2018). In addition, evaluations of possible ground subsidence based on compaction of soft ground materials have been conducted for the Songdo and Cheongna areas of Incheon Metropolitan City (Kim, 2010). According to the literature review, most of the ground subsidence reported in the Incheon Metropolitan City was localized due to the consolidation of facilities resulting from underground structures, and no significant ground subsidence has been reported due to soil compaction in soft ground material.

According to the feasibility reassessment report of Incheon Subway Line 2 conducted by the Korea Development Institute in 2008, the construction of Incheon Subway Line 2 was pursued to alleviate urban traffic congestion caused by the Geomdan New Town, the Yeongjong-Cheongna Economic Free Zone development projects. Another purpose of the construction project is to alleviate traffic congestion by partially connecting the north-south and east-west axes of Incheon Metropolitan City, induce pollution reduction, and balance regional development. The initial planned construction period was from 2007 to 2013, but construction began in June 2009 and was not opened until July 30, 2016.

This paper conducts the SBAS data processing to detect where surface displacement occurs and further analyzes the characteristics and causes of ground subsidence using the collected COSMO-SkyMed SAR data. To illustrate the study area’s location, the Landsat-5 Thematic Mapper (TM) optical image around Incheon Metropolitan City is shown in Fig. 1(a). The red box area indicates the swath of the collected SAR observation, and Fig. 1(b) depicts the mean intensity image of the observed radar data.

OGCSBN_2024_v40n2_141_3_f0001.png 이미지

Fig. 1. Location maps of the study area. (a) Landsat-5 Thematic Mapper high-quality ortho-rectified L1T dataset, which was acquired on 27 Sep 2011, showing the swath of COSMO-SkyMed HIMAGE mode (marked by the red box) utilized in this study. (b) An average amplitude image shows Incheon City near Seoul (Courtesy of United States Geological Survey).

2.2. Data

Twenty-four COSMO-SkyMed HIMAGE SAR images were collected from July 12, 2011, to August 27, 2012, to observe a time-series surface displacement in the study area. The COSMO-SkyMed satellite uses a 9.6 GHz X-band radar carrier frequency corresponding to a wavelength of approximately 3.1 cm. The swath width of collected data covers an area of approximately 40 × 40 km2, indicated by the red box area in Fig. 1(a). The collected SAR data were observed in horizontal-horizontal co-polarization (HH). The incidence angle is 38.8° at the center range, and pixel spacing is approximately 1.33 m and 2.06 m in the slant range and the azimuth direction, respectively. Detailed characteristics are presented in Table 1.

Table 1. Parameters of the COSMO-SkyMed datasets used in this study

OGCSBN_2024_v40n2_141_4_t0001.png 이미지

The COSMO-SkyMed, developed and is being operated in Italy, was launched in June and December 2007, October 2008, and November 2010, with four satellite constellations. The planned mission’s duration was from 2007 to 2019, and they are jointly operated by the ASI and the Italian Ministry of Defence (MoD). The primary applications of the COSMO-SkyMed include defense/security, flood, drought, landslide, volcano, and earthquake, as well as agriculture, forestry, environment, geology, and cryosphere in Earth environmental science. Due to significantly shorter revisit cycles, the constellation of the COSMO-SkyMed is a valuable resource for monitoring surface displacements, such as volcanoes, earthquakes, and glaciers. Moreover, despite being a short-wavelength X-band SAR, it allows for generating high-resolution interferograms with relatively low temporal decorrelation. In the initial operation, three observation modes, Spotlight, Stripmap, and ScanSAR, were supported, and a wide swath mode was added later to enhance ship monitoring capabilities at sea. The swath covers from 10 to 2,000 km, and there are differences in spatial resolution depending on the observation mode.

The topographic phase needs to be subtracted using the digital elevation model (DEM) information to calculate the differential interferogram. In this study, global TanDEM-X (TerraSAR-X add-on for digital elevation measurements) DEM with a spatial resolution of 90 m, distributed by the German Aerospace Center, was collected and processed (Wessel et al., 2018). The global TanDEM-X DEM was produced using SAR observations collected from nearly identical two satellites, with perpendicular baselines ranging from approximately 120 m to 500 m. The global TanDEM-X SAR observations were collected entirely in 2015, and the global TanDEM-X DEM production was completed in September 2016 after precise data processing. The absolute height accuracy of the DEM is approximately 1 m, showing better performance than the mission requirement of 10 m.

3. Methodology

In the InSAR application, coherence is a quantitative measure evaluating the correlation between two SAR observations, ranging from 0 to 1(Rosen et al.,2000). It is necessary to utilize combinations of highly coherent interferograms to estimate the continuous time-series of the phases of SAR acquired at different times. The SBAS method is designed for a time-series analysis that selectively utilizes interferograms with short geometrical baselines and has been widely used for surface displacement observation on various Earth’s surfaces. The SBAS method relies on more reliable interferograms by limiting the perpendicular and temporal baseline to account for the decorrelation effect. Its purpose is to extract a linear or non-linear time-series of surface displacement trends of only distributed scatterers with coherence above a certain level from interferometric SAR data pair distribution (Berardino et al., 2002; Hong and Wdowinski, 2013; Hong et al., 2010). The SVD numerical analysis using a matrix of coherent interferometric phases with temporal evolution is applied to extract a time-series of surface displacement.

Repetitive removal of residual phases which is excluded with a primary trend of surface displacement, can suppress atmospheric artifacts (Berardino et al., 2002). Multi-looking processing to suppress the phase noise is accompanied by decreased spatial resolution. Hence, the spatial resolution of the estimated time-series of surface displacement is lower than that of the collected SAR observations. The correct unwrapping processing in differential interferograms is also essential in the SVD analysis because discontinuity of the interferometric SAR data pair distribution might misrepresent the time-series of the surface displacement. Numerous SAR observations in a golden age for spaceborne SAR systems increase the number of interferograms that should be processed, leading to a considerable computing load. Several clustering and parallel processing methods have been developed to process data faster than serial processing, with a benchmark of 16 central processing units (Hong, 2019).

Even though the small perpendicular baseline is preferred for the coherent interferogram, the large temporal baseline often significantly degrades interferometric coherence in distributed scatterers. In the wetlands, forests, and polar regions, where temporal decorrelation is significant, the temporal baseline can also be limited when a network of interferometric SAR data pair distribution is calculated to maintain higher coherence (Hong et al., 2010). The small temporal baseline subset (STBAS) method has been proposed to overcome coherence degradation from temporal decorrelation. It utilizes interferogram networks composed only of short temporal baselines without constraints on perpendicular baselines (Hong et al., 2010).

This study constructed a network consisting of 80 interferograms for the SBAS data processing by setting the range of perpendicular baselines from 0 to 300 m and temporal baselines from 1 to 240 days. Fig. 2 illustrates the interferometric SAR data pair distribution produced using the criteria of the baseline conditions. The characteristics of selected interferometric pairs used in the SBAS data processing are described in Table 2. Commercial software Gamma was used for data processing, and Synthetic Aperture Radar PYthon scripts for Retrieval of Earth’s surface displacements (SARPYRE), an automated SBAS data processing script designed and implemented at Pusan National University, was applied. Using this automated script of the SARPYRE ensures consistent and effective production of results compared to manual data processing. To subtract the topographic phase, a 90 m spatial resolution of TanDEM-X global DEM was resampled with a 4 × 4 oversampling factor, and multi-looking processing was applied using 4 × 4 multi-looking factors to remove phase noise in the differential interferograms. In addition, an adaptive phase filter was also applied for phase noise removal (Goldstein and Werner, 1998). The low frequency of residual phase caused by possible satellite orbital errors in the filtered differential interferograms was removed.

Table 2. The list of interferometric pairs used in the SBAS data processing

OGCSBN_2024_v40n2_141_5_t0001.png 이미지

OGCSBN_2024_v40n2_141_5_t0002.png 이미지

OGCSBN_2024_v40n2_141_6_f0001.png 이미지

Fig. 2. Interferometric SAR data pair distribution for the small baseline subset (SBAS) data processing. SAR data acquired on 12 July 2011, was selected as the reference image. The temporal baseline ranges from 16 to 240 days, and the perpendicular baseline is from 1.7 to 299.5 m.

The minimum cost flow (MCF) algorithm was utilized for the phase unwrapping of differential interferograms, and then the phase unwrapping correction procedure was applied. Since each differential interferogram utilizes SAR observations acquired on different dates, inevitable atmospheric artifacts due to water vapor occur. In this study, an effective reduction of atmospheric artifacts was achieved by applying spatial filters to suppress localized random phase components. The estimated surface displacement through the SVD solver was geocoded using DEM data and exported into Google Earth format for visual analysis. Fig. 3 depicts the schematic diagram for the SBAS processing steps utilized in this study.

OGCSBN_2024_v40n2_141_6_f0002.png 이미지

Fig. 3. Schematic diagram of COSMO-SkyMed small subset baseline (SBAS) data processing for surface displacement monitoring in time-series. All processing procedures have been implemented using Synthetic Aperture Radar PYthon scripts to Retrieve Earth’s surface displacements (SARPYRE) developed by Pusan National University.

4. Results

This section presents and analyzes the results of a time-series of surface displacements estimated by the SBAS data processing. Fig. 4(a) illustrates the cumulative surface displacement occurring from July 12, 2011, to August 27, 2012, with a displacement range of –1 to 1 cm. Areas such as water bodies, forests, and intertidal flats were masked in the time-series of surface displacement due to difficulties in maintaining high coherence. The masking process used a threshold of 0.6 coherence calculating the average coherence of all interferograms. Fig. 4(b) represents the annual surface displacement rate within the –1 to 1 cm range, the same as cumulative displacement in Fig. 4(a). At the same time, Fig. 4(c) depicts the standard deviation calculation between the surface displacement phase and a linear model of the estimated phases. The impact of temporal decorrelation due to the relatively large temporal baseline can be observed in areas such as intertidal flats and forest regions, where a time-series of surface displacement cannot be obtained. It suggests that the interferometric SAR observations of the shorter revisit cycle or longer wavelength are required for monitoring time-series of surface displacement on the intertidal flats and forest regions.

OGCSBN_2024_v40n2_141_7_f0001.png 이미지

Fig. 4. Time-series of surface displacement results. (a) The total surface displacement between 12 July 2011, and 27 August 2011. (b) Mean deformation velocity maps. (c) The standard deviation of residual phase for uncertainty analysis. The red box in panel (b) was selected to analyze the time-series of surface displacement in detail.

In the overall results, most surface displacement was observed locally around urban areas; however, significant subsidence was not observed over widespread study areas. Although localized ground uplift was detected in Fig. 4(a), they were not distinctly observed in the annual surface displacement rate map of Fig. 4(b). Moreover, the scatter plots of the time-series surface displacement analysis indicate no consistent uplift, suspecting the presence of localized atmospheric artifacts or residual phase noise in the short wavelength X-band SAR observations. This localized ground uplift will be investigated further through additional experiments with longer-period datasets in future studies. This study only focuses on analyzing ground subsidence to interpret regional characteristics. The overall phase standard deviation converging to zero indicates that the SBAS data processing with a high confidence level was achieved.

Fig. 5(a) was displayed by magnifying the red box inset area in Fig. 4(b) for further investigations to represent the surface displacement resulting from ground subsidence in more detail. Five regions representing relatively significant ground subsidence were selected and labeled with the characters of b-f in Fig. 5(a). Figs. 5(b-f) presents plots of time-series of surface displacement for the selected regions to examine the temporal evolution of the ground subsidence, along with linear regression equations for each plot. The subsidence rates observed in the five regions indicate gradual downward surface displacement, reflecting surface displacement velocities corresponding to an annual subsidence of approximately 1 cm. Statistical analysis of time series surface displacement reveals relatively small phase standard deviation levels compared to linear model regression, indicating relatively linear surface displacement trends.

OGCSBN_2024_v40n2_141_8_f0001.png 이미지

Fig. 5. Scatter plots of the estimated ground subsidence. (a) A zoomed image showing that surface displacement might be caused by ground subsidence. The white dots marked with (b-f) were selected to examine the pattern of ground subsidence in time-series. (b-f) Plots of time-series of surface displacement at (b-f) in Fig. 4(a). Solid lines represent the linear regression model used to estimate the surface displacement velocity.

Geocoded time-series surface displacement maps were converted into a Google Earth kmz file format to analyze the regional characteristics of the ground subsidence and compare them with existing topographic maps. The five suspected ground subsidence areas are located near Incheon Subway Line 2 or Line 1 stations, suggesting the influence of underground construction associated with subsurface excavation. The suspected subway stations include Gajeong Jungang Market Station, Gajaeul Station, Jooan National Industrial Complex Station, Inha University Station, and Bupyeong-gu Office Station.

Considering that Incheon Subway Line 2 opened in July 2016, the surface displacement observed in this study from July 2011 to August 2012 is interpreted as reflecting ground subsidence occurring during that subway construction period. In future work, further analysis is needed to determine if the observed ground subsidence is still ongoing. These preliminary results can serve as an initial analysis indicating ground subsidence patterns according to the distribution of Incheon subway routes. Furthermore, a comparison study of surface displacement changes before and after subway completion can be worth investigating using long-term multi-temporal SAR data such as the European Sentinel-1 satellite data.

5. Discussion and Conclusions

In this study, a time-series ground subsidence analysis has been successfully conducted for the Incheon Metropolitan City using high-resolution COSMO-SkyMed X-band SAR observations acquired over approximately one year, from July 2011 to August 2012. The SARPYRE software, developed by Busan National University to efficiently process numerous SAR datasets, was used for the SBAS data processing. The SARPYRE, based on the commercial Gamma SAR processing package, supports a distributed processing system like the Slurm workload manager.

The surface displacements detected around several Incheon Subway Line 2 stations might primarily be attributed to ground subsidence. The annual ground subsidence amounts during this period are estimated to be approximately 1 cm. Considering the completion of Incheon Subway Line 2 in July 2016 and the acquisition period of the data used in this study from 2011 to 2012, the extracted surface displacements are presumed to result from temporary ground subsidence due to weak ground conditions or underground consolidation associated with subway construction. Locally observed ground uplift is interpreted as undesirable noise due to atmospheric artifacts commonly found in interferograms with short SAR wavelength bands. More precise time-series ground subsidence analysis would be required in the near future, utilizing long-term European Sentinel-1 SAR data collected before and after the completion of Incheon Subway Line 2.

The COSMO-SkyMed satellites with four-satellite constellation operation can mitigate the coherence degradation due to temporal decorrelation for interferometric applications. The minimum temporal baseline selected for interferogram generation was 16 days for the SBAS processing in this study. Although interferometric pairs with longer temporal baselines were also used, sufficient coherence was observed within urban areas. This indicates that time-series surface displacement extraction with a 16-day interval is achievable through precise orbit adjustment and attitude control. Korean Multi-Purpose Satellite (KOMPSAT) 5, which is currently operational in South Korea, and KOMPSAT 6, which is nearing completion and scheduled for launch, are designed with the same microwave frequency as the COSMO-SkyMed X-band SAR data used. It implies the potential capability of interferometric SAR techniques in urban areas with frequent data acquisition due to the revisit cycle of KOMPSAT 6 every 11 days. However, coherence significantly degrades in densely forested areas or regions with frequent surface changes like intertidal flats. Thus, more frequent observations are required for the surface displacement analysis in such areas (Hong and Wdowinski, 2013). A shorter revisit cycle through the constellation enables the maximization of InSAR applications of KOMPSAT 6.

Acknowledgments

This work was supported by a 2-Year Research Grant of Pusan National University.

Conflict of Interest

No potential conflict of interest relevant to this article was reported.

References

  1. Abidin, H. Z.,Andreas, H., Djaja, R., Darmawan, D., and Gamal, M., 2008. Land subsidence characteristics of Jakarta between 1997 and 2005, as estimated using GPS surveys. Gps Solutions, 12, 23-32. https://doi.org/10.1007/s10291-007-0061-0 
  2. Berardino, P., Fornaro, G., Lanari, R., and Sansosti, E., 2002. A New algorithm for surface deformation monitoring based on small baseline differential SAR interferograms. IEEE Transactions on Geoscience and Remote Sensing, 40(11), 2375-2383. https://doi.org/10.1109/TGRS.2002.803792 
  3. Bruno, M., 1992. Subsidence-induced well failure. SPE Drilling Engineering, 7(2), 148-152. https://doi.org/10.2118/20058-PA 
  4. Fattahi, H., Simons, M., and Agram, P., 2017. InSAR time-series estimation of the ionospheric phase delay: An extension of the split range-spectrum technique. IEEE Transactions on Geoscience and Remote Sensing, 55(10), 5984-5996. https://doi.org/10.1109/TGRS.2017.2718566 
  5. Ferretti, A., Prati, C., and Rocca, F., 2001. Permanent scatterers in SAR interferometry. IEEE Transactions on Geoscience and Remote Sensing, 39(1), 8-20. https://doi.org/10.1109/36.898661 
  6. Gabriel, A. K., Goldstein, R. M., andZebker, H. A., 1989.Mapping small elevation changes overlarge areas: Differential radar interferometry. Journal of Geophysical Research: Solid Earth, 94(B7), 9183-9191. https://doi.org/10.1029/JB094iB07p09183 
  7. Goldstein, R. M., and Werner, C. L., 1998. Radar interferogram filtering for geophysical applications. Geophysical Research Letters, 25, 4035-4038. https://doi.org/10.1029/1998GL900033 
  8. Hanssen, R. F., 2001. Radar interferometry: Data interpretation and error analysis(Vol. 2). Springer Science and Business Media. https://doi.org/10.1007/0-306-47633-9 
  9. Hong, S.-H., 2019. Parallel computing on intensity offset tracking using synthetic aperture radar for retrieval of glacier velocity. Korean Journal of Remote Sensing, 35(1), 29-37. https://doi.org/10.7780/kjrs.2019.35.1.3 
  10. Hong, S.-H., and Wdowinski, S., 2013. Multitemporal multitrack monitoring of wetland water levels in the Florida Everglades usingALOSPALSARdatawith interferometric processing. IEEE Geoscience and Remote Sensing Letters, 11(8), 1355-1359. https://doi.org/10.1109/LGRS.2013.2293492 
  11. Hong, S.-H., Wdowinski, S., and Kim, S.-W., 2009. Evaluation of TerraSAR-X observations for wetland InSAR application. IEEE Transactions on Geoscience and Remote Sensing, 48(2), 864-873. https://doi.org/10.1109/TGRS.2009.2026895 
  12. Hong, S.-H., Wdowinski, S., Kim, S.-W., and Won, J.-S., 2010. Multi-temporal monitoring of wetland water levelsin the Florida Everglades using interferometric synthetic aperture radar (InSAR). Remote Sensing of Environment, 114(11), 2436-2447. https://doi.org/10.1016/j.rse.2010.05.019 
  13. Jolivet, R., Grandin, R., Lasserre, C., Doin, M. P., and Peltzer, G., 2011. Systematic InSAR tropospheric phase delay corrections from global meteorological reanalysis data. Geophysical Research Letters, 38(17). https://doi.org/10.1029/2011GL048757 
  14. Jung, H.-S., Lee, D.-T., Lu, Z., and Won, J.-S., 2012. Ionospheric correction of SAR interferograms by multiple-aperture interferometry. IEEE Transactions on Geoscience and Remote Sensing, 51(5), 3191-3199. https://doi.org/10.1109/TGRS.2012.2218660 
  15. Kampes, B. M., 2006. Radar interferometry (Vol. 12). Springer.
  16. Kim, D. H., Bae, K. D., Ko, S. K., and Lee, W. J., 2010. Evaluation of spatial distribution of secondary compression of Songdo marine clay by probabilistic method. Journal of the Korean Geotechnical Society, 26(9), 25-35. https://doi.org/10.7843/kgs.2010.26.9.25 
  17. Kim, E.K., 2018. Policy suggestions for the creation of a safe road environment in Incheon City: Focused on measures for the ground settlement management. Korean Journal of Urban Studies, 13, 79-106. https://doi.org/10.34165/urbanr.2018..13.79 
  18. Kim, S.-W., Wdowinski, S., Amelung, F., Dixon, T. H., and Won, J.-S., 2013. Interferometric coherence analysis of the Everglades wetlands, South Florida. IEEE Transactions on Geoscience and Remote Sensing, 51(12), 5210-5224. https://doi.org/10.1109/TGRS.2012.2231418 
  19. Massonnet, D., and Feigl, K. L., 1998. Radar interferometry and its application to changes in the Earth's surface. Reviews of Geophysics, 36(4), 441-500. https://doi.org/10.1029/97RG03139 
  20. Milliman, J. D., and Haq, B. U., 1996. Sea-level rise and coastal subsidence: Causes, consequences, and strategies (Vol. 2). Springer Science and Business Media. 
  21. Moreira, A., Prats-Iraola, P., Younis, M., Krieger, G., Hajnsek, I., and Papathanassiou, K. P., 2013. A tutorial on synthetic aperture radar. IEEE Geoscience and Remote Sensing Magazine, 1(1), 6-43. https://doi.org/10.1109/MGRS.2013.2248301 
  22. Park, S. W., and Hong, S. H., 2021. Nonlinear modeling of subsidence from a decade of InSAR time-series. Geophysical Research Letters, 48(3), e2020GL090970. https://doi.org/10.1029/2020GL090970 
  23. Rosen, P. A., Hensley, S., Joughin, I. R., Li, F. K., Madsen, S. N., Rodriguez, E., and Goldstein, R. M., 2000. Synthetic aperture radar interferometry. Proceedings of the IEEE, 88(3), 333-382. https://doi.org/10.1109/5.838084 
  24. Wessel, B., Huber, M.,Wohlfart, C., Marschalk, U., Kosmann, D., and Roth, A., 2018. Accuracy assessment of the global TanDEM-X Digital Elevation Model with GPS data. ISPRS Journal of Photogrammetry and Remote Sensing, 139, 171-182. https://doi.org/10.1016/j.isprsjprs.2018.02.017 
  25. Wright, T. J., Parsons, B. E., and Lu, Z., 2004. Toward mapping surface deformation in three dimensions using InSAR. Geophysical Research Letters, 31(1). https://doi.org/10.1029/2003GL018827 
  26. Zebker, H. A., Rosen, P. A., Goldstein, R. M., Gabriel, A., and Werner, C. L., 1994. On the derivation of coseismic displacement fields using differential radar interferometry: The Landers earthquake. Journal of Geophysical Research: Solid Earth, 99(B10), 19617-19634. https://doi.org/10.1029/94JB01179