Landslides result from the outward and downward behavior of slope materials combined with rock, soil, or various materials like this (Varnes, 1978). It is more dangerous than numerous natural hazards and may harm natural resources and engineering activities, causing the sacrifice of life and property (Saha et al., 2002). In order to decrease the devastating effects of the landslide to minimum levels, the mass movements caused by this disaster must be determined precisely and quickly. Obtaining the location information of a landslide in an area will facilitate taking precautions against a possible landslide disaster. In this context, precise determination of location and boundary information of mass movements has gained importance.
LIM is a kind of map that includes the location and boundary information of landslides. This map has been used as a base map in many scientific studies on landslide disasters (Ayalew et al., 2004; Pourghasemi et al., 2013; Eker & Aydın, 2014 Notti et al.,2014 ; Kadi et al., 2019). Therefore, supplying this kind of map, which is used as a base map in landslide-based scientific research, with more precision and detail, allows the outputs to be produced more efficiently and realistically. The production phase of this map is generally based on surface observations. Rarely some traditional terrestrial measurement methods are also used. However, detecting and measuring landslides that occur in areas where limited transportation and population density are not available is challenging. Therefore, detecting landslides in such areas and the LIM produced is insufficient.
Permanent Scatterer Interferometry (PSI) is an improved Synthetic Aperture Radar (SAR) technique employed in deformation works and can calculate movement with millimeter precision (Ciampalini et al., 2016). This method is widely used in monitoring landslides. ( Righini et al., 2011; Lu et al. 2012; Şanlı et al. 2012; Abdikan et al. 2013; Cigna et al. 2013; Tofani et al. 2013; Bianchini et al. 2014; Notti et al. 2014; Sara et al. 2014; Mohammed et al. 2013; Oliveira et al. 2015; Ciampalini et al. 2016; Bhattacharya & Mukherjee 2016; Bianchini et al. 2017; Necula et al. 2017; Akgül 2018; Darvishi et al. 2018; Altıntaş & Gökalp 2019; Chao et al. 2019; Meghanadh et al. 2019; Fárová et al. 2019; Mondini et al. 2019; Shirani & Pasandi 2019; Aslan et al. 2020; Jennifer et al. 2020; Jiaxuan et al. 2020; Yazici and Gormus 2020).
There are many studies in which the Differential SAR Interferometry (DInSAR) technique has been successfully applied to detect surface deformations in large areas to date. However, the traditional DInSAR method has some limitations, such as orbital, atmospheric, and errors due to temporal and spatial correlation distortion. The PSInSAR method (Ferretti et al., 2001), which uses a series of SAR images recorded at different dates, was developed to overcome these limitations (Abdikan et al., 2014; Yazici and Gormus, 2020). Multiple interferometric approaches, grouped under the name PS-InSAR, have some essential features compared to classical InSAR, such as eliminating atmospheric and orbital errors, detecting linear and nonlinear deformation patterns, and generating a time series variation for each PS point.
PSInSAR requires multiple SAR images to be taken over the same area. Typically, 15–20 images are needed to perform PSI analysis with a C-band. (Ferretti et al., 2000; Colesanti et al., 2003; Plank et al., 2013) Due to its higher resolution and shorter wavelength, it is possible to use fewer data sets for PSI analysis to be performed with X-band (Bovenga et al., 2012). In any case, the more images available, the better the quality of the PSI deformation rate and time series estimation (Crosetto et al., 2016). The PS-InSAR approach allows the investigation of natural or man-made objects that present a stable backscatter signal over time. Although the X-Band (short wavelength) may generate more PS points in urban areas than other bands, it often produces very low PS points in densely vegetated areas due to the high temporal correlation resulting from the short wavelength. Therefore, the use of C-band (Sentinel-1) and L-band (ALOS PALSAR) data may provide more advantages, especially in rural areas (Pawluszek-Filipiak et al., 2020). In addition, deformation measurement can be made more sensitively with the C band compared to the L band, especially in urban areas (Tantianuparp et al., 2013). Data from the Sentinel-1 satellite, which produces higher resolution data than previous C-band sensors, has been successfully applied in PS-InSAR-based surface deformation analyses (Hasanlou et al., 2021). Using two satellites, Sentinel 1A and 1B, the temporal resolution can be reduced to 6 days. However, only Sentinel-1A satellite images will be used in this study since data cannot be obtained from the Sentinel-1 B satellite in the current situation. Therefore, the temporal resolution would be 12 days. It is also an important advantage that Sentinel images are provided free of charge by the European Space Agency.
Although the PS-InSAR technique has many advantages over other interferometric methods, PS points do not occur where the backscatter reaching agricultural areas, grasslands, and forests are constantly changing. (Vicari et al., 2019) In arid climatic regions, favorable conditions may occur for more PS points due to less humidity and sparse vegetation (Plank et al., 2013).
Stanford Method of Persistent Scatter (StaMPS) developed by A.Hooper et al. (2010) can measure a wide range of deformation rates in different seasons and terrains at high latitudes. In this method, more stable PS points can be determined by using both phase and amplitude instead of only amplitude in selecting PS points as in PS-InSAR. Thus, PS points representing soil properties in the StaMPS/MTI method are resistant to different climatic and atmospheric conditions and small changes in satellite angle. In addition, although PS points are usually formed in structures such as concrete and metal, they can also be created in other land types (Ansar et al., 2021). For example, this feature in the StaMPS method allows forming of some PS points in forest lands.
In the literature, Abdikan et al. (2013) have aimed to determine the deformations that occur due to subsidence from mining activities in the study area with the aid of the SAR interferometry. In this context, 21 European Space Agency satellite images were used between 1993 and 2001 for the study area. Analyses have been performed employing StaMPS software and the Persistent Scatterers (PS-InSAR) technique. In light of the obtained results, displacements of up to 7–8 mm per year have been calculated in the work area. The mass movement that happened at the end of 7.5 years has been calculated as 50–65 mm. Sara et al. (2014) have aimed to modernize a pre-available LIM. This study focuses on landslides inventory map updating, combining the explication of data acquired with the PS-InSAR technique with ground research. The consequences indicate the utility of the PSI technique to modernize LIM, owing to new satellite knowledge with millimeter precision. Aslan et al. (2020) have used a StaMPS method to run the groups of Sentinel images gained between 2015 and 2019 round orbits for the determined area in the French Alps. Results have been matched to pre-available LIM. More than 100 unreported landslides, their spatial pattern, deformation ratio, and speed status have been correctly categorized in a region of 4000 km2 located in the French Alps. Bianchini et al. (2014) have examined landslide actions in the study area, thanks to Persistent Scatterer Interferometry (PSI) data. This study has integrated InSAR data, which maps the effective landslide over 20 years, with detection of destruction monitored in the buildings. The results have allowed the development of the LIM of the area, warrantable providing a recent evaluation of the instability of the study area. Bianchini et al. (2017) have proposed a way at a particular scale for generating a map by using present satellite PSI data obtained from Sentinel-1. Movements have been obtained from PSI data of Sentinel-1 sensors in the C-band have been generated for the intensity map of motions, necessary for evaluations of the deficiency in the study area. The southwest of the city was identified as the highest risk area. Cigna et al. (2013) employed ENVISAT images to obtain the PSP technique for the study area. These images have been integrated with supplemental supporting data. Nine hundred eighty landslides have been improved through PSI data, and 64 current landslides have also been determined. In addition, the status of the action and the density of these landslides have also been interpreted, displaying that 22% and 2% of the current map contain effective (919 landslides) and reactivated (93 landslides) cases separately. Darvishi et al. (2018) have examined the combination of interferometric and SAR remote sensing for landslide hazard reduction in the Alps. Results have indicated that offset prediction precision can be significantly developed until less than 3% of the pixel dimension employing the integration of the various detectors and descriptors. Fárová et al. (2019) have employed traditional DInSAR and the PSI methods to observe the vertical movements of the current highway in this study. Sentinel-1 radar data has been used in this study. The traditional DInSAR technique has been analyzed using the SNAP program. The amount of maximum collapse has been determined when the highway body reaches 20 mm over the epoch. Greif and Vlcko (2011) have employed InSAR data to investigate the post-failure deformations in Slovakia in this study. The disastrous landslide of 1977 in the Lubietova has been investigated using the PSInSAR technique to analyze current and ancient deformations of the indecisive slopes. Results show that the landslide state in the Hutna River is in sustained heavy movement, and they need continuous monitoring. However, the recent hazard to people is not considered quite loud.
In conclusion, they have confirmed that the InSAR could be many advantageous technologies for better governance of landslide regions. Lu et al. (2012) have aimed to employ PSs obtained from 4 years of RADARSAT (The RADARSAT programme consists of a pair of remote sensing satellites from the Canadian Space Agency satellite images). The result has included two kinds of maps (separate orbit map and compound map). One hundred ten points have been detected from ascending PSs, and 155 points have been obtained from descending data. Accuracy evaluation has been performed by checking the points results with the current LIM, aerial photos, historical data, optical images, and topographic maps. Of the ascending and descending points, 79.1% and 63.2%, severally, have been related to available movements. Notti et al. (2014) have presented a workflow for the development of PSI data analysis for landslide actions in this paper. In order to analyze and confirm the results from this methodology, they have employed diverse PS datasets from varied satellites, which have been generated by employing three PSI techniques to map and observe landslides in different regions in Spain and Italy. Righini et al. (2011) have aimed to the LSI updating by combining the explication of information obtained from PSI with optical imagery and thematic maps. The outcomes have emphasized the features of multi-interferometry InSAR data, combined with traditional methods, to assist landslide investigation. Tofani et al. (2013) used PSI techniques and traditional monitoring techniques to observe the Santo Stefano d’Aveto landslide. The study has proposed that the PSI technique can have great potential in landslide monitoring, but employing only the method for this aim requires cautious interpretation. Bhattacharya and Mukherjee (2016) have aimed to analyze the subjects interested in SAR-based deformations works until now conducted in the Himalayan area and potential improved options in this study. This work has demonstrated the possibility of density monitoring techniques employing ERS-1/2 C-band SAR data to determine landslide-related deformations. The outcomes of SBAS and density monitoring techniques have been checked in many positions from steady and landslide-prone regions. Jennifer et al. (2020) have aimed to investigate surface deformations in Kerala, India. The PSI investigations on a group of 20 Sentinel-1 images have been processed to evaluate the preliminary movements of the landslides. According to the obtained outcomes, the work area has been calculated to occur at a deformation of -26 to + 19 mm/year, which shows the region's instability. Jiaxuan et al. (2020) have suggested a landslide susceptibility dynamic map to decrease LSM use and modernize the map in a period series. The PSInSAR method has been employed to forecast the visibility velocity. The DLSM has been generated by combining, employing total slope values, the landslide susceptibility estimated by logistic regression, and the PSInSAR technique. Meghanadh et al. (2019) investigated the arrangement of LIMs in India, employing the PSInSAR method. The LIM, arranged by the Geological Survey of India, is obtained by combining 15 Envisat ASAR and 10 Sentinel-1A images corresponding to ~ 11*24 sq. km extensive river located in Uttarakhand with the PSInSAR technique. In light of the results, nine landslide regions widespread to LIM were obtained by GSI as a section of the approval operation, and 22 recent landslide areas have been determined. Mohammed et al. (2013) have examined the improved interferometry SAR method with two significant ingredients. They have selected six available DInSAR methods employed for observing diverse displacement kinds. The result of this work has indicated that more robust techniques and satellite sensor improvement projects further decrease the technology's inherent limitations. Oliveira et al. (2015) have aimed to evaluate the possibility of the Persistent Scatterer deformation maps to the determination of recent landslides regions and in the description of landslide limits, to modernize the landslide efficiency, and to assess the capability of the Persistent Scatterer displacement maps in evaluation landslide susceptibility. Yazici and Gormus (2020) have examined the performance of PS-inSAR by suggesting various techniques for observing the harmony of the detected displacement velocities gained from a group of Sentinel-1A images. This study aims to resolve the compelling problem, display the effect of the PS-inSAR technique and define the residual clear in PS analysis of landslide practices in dam regions. Results have indicated that total displacement in the study area is calculated between − 25 and 28 mm/year by employing the PS-inSAR technique. In total, 1608 PS points have been produced by PS-inSAR.
The primary aim of this study is to detect new landslide areas that have not been reported with the help of StaMPS technique and thus to improve the landslide inventory map (LIM) of the study area. In this study, the correlation between current landslide areas and PS-inSAR points falling on these areas was examined. For this purpose, 25 Sentinel-1A images were used. The LIM has been made more up-to-date and realistic by determining the ground displacements in the study area. In this context, the workflows of the study are summarized Fig. 1. The workflow given in Fig. 1(a) is a chart for obtaining statistical results for PS-inSAR points falling within landslide areas. In this work flow chart, first of all, landslide susceptibility map and PS-inSAR points for the study area were produced. Then, these two data will be overlaid with each other and statistical results will be presented regarding the PS-INSAR points falling within the landslide areas. Thus, the correlation between the existing landslide areas and the PS-INSAR points falling on these areas will be determined. The workflow given in Fig. 1(b) is a chart for identifying new landslide areas within the boundaries of the study area and processing these new landslide areas into the landslide susceptibility map. In this context, first of all, spatial information about potential landslide areas determined by public institutions is transferred to the computer. In the second step, new potential landslide areas will be determined with the help of PS-inSAR points produced for the study area. In the next step, the potential landslide areas produced in different ways in the previous process steps will be overlapped and the new landslide areas will be determined. In the last step, the new landslide areas identified were overlapped with the land use capability maps (LUCM) and the slope and soil classes of the new landslide areas were examined.