Divya Sekhar Vaka

Land subsidence significantly threatens urban infrastructure, agricultural productivity, and environmental sustainability. This study develops a land subsidence susceptibility model by integrating Small Baseline Subset (SBAS) Interferometric Synthetic Aperture Radar (InSAR) data with key geospatial factors using machine learning approaches. The study focuses on the Attica prefecture, Greece, and utilizes SBAS InSAR data from 2015 to 2021 to extract ground deformation velocities by classifying them into four susceptibility levels: stable, low, moderate, and high. The susceptibility results indicate that stable zones constitute 58.2% of the study area, followed by low (27.2%), moderate (11.2%), and high susceptibility zones (3.4%), predominantly concentrated in areas undergoing hydrological stress and urbanization. Random Forest (RF) and XGBoost (XGB) models incorporate a comprehensive set of causal factors, including slope, aspect, land use, groundwater level, geology, and rainfall. The evaluation of the models includes accuracy metrics and confusion matrices. The XGB model achieved the highest performance, recording an accuracy of 94%, with well-balanced predictions across all susceptibility classes. Addressing class imbalance during model training improved the recall of minority classes, though with slight trade-offs in precision. Feature importance analysis identifies proximity to streams, land use, aspect, rainfall, and groundwater extraction as the most influential factors driving subsidence susceptibility. This methodology demonstrates high reliability and robustness in predicting land subsidence susceptibility, providing critical insights for land-use planning and mitigation strategies. These findings establish a scalable framework for regional and global applications, contributing to sustainable land management and risk reduction efforts.

Landslides pose significant threats to life and property, particularly in mountainous regions. To address this, this study develops a landslide susceptibility model integrating Earth Observation (EO) data, historical data, and Multi-Temporal Interferometric Synthetic Aperture Radar (MT-InSAR) ground movement results. The model categorizes areas into four susceptibility classes (from Class 1 to Class 4) using a multi-class classification approach. Results indicate that the Xtreme Gradient Boosting (XGB) model effectively predicts landslide susceptibility with area under the curve (AUC) values ranging from 0.93 to 0.97, with high accuracy of 0.89 and a balanced performance across different susceptibility classes. The integration of MT-InSAR data enhances the model’s ability to capture dynamic ground movement and improves landslide mapping. The landslide susceptibility map generated by the XGB model indicates high susceptibility along the Pacific coast. The optimal model was validated against 272 historical landslide occurrences, with predictions distributed as follows: 68 occurrences (25%) in Class 1, 142 occurrences (52%) in Class 2, 58 occurrences (21.5%) in Class 3, and 4 occurrences (1.5%) in Class 4. This study highlights the importance of considering temporal changes in environmental conditions such as precipitation, distance to streams, and changes in vegetation for accurate landslide susceptibility assessment.

The coseismic surface displacement of the 2019 M _w 5.4 Mirpur earthquake is derived using Differential Synthetic Aperture Radar Interferometry (DInSAR) technique. Two Sentinel-1 radar images before and after the earthquake acquired in interferometric wide swath mode are used for displacement map generation. Two definite lobes of deformation corresponding to subsidence and uplift are observed from the displacement map. The results indicate an uplift of 9.5 cm and subsidence of -6.2 cm in the earthquake epicentral region. Using a forward elastic dislocation model the causative source parameters of the earthquake are randomly searched using an iterative approach, which minimizes the error between the InSAR data and the modeled results. The inversion results indicate a rectangular fault of length ~10 km and width ~5 km is responsible for the earthquake. Other source parameters such as strike, dip, depth, and the slip of the earthquake are also calculated during the coseismic inversion.

The 2017 M_w 7.3 Iran-Iraq earthquake occurred in the seismically less active northern Zagros Mountains. In this study, DInSAR technique is used to derive the surface displacements of the earthquake using the Sentinel-1 and ALOS-2 images. One preseismic, three coseismic and a postseismic interferogram are generated. The coseismic interferograms did not show any signs of a surface rupture and indicate a maximum of 80 cm uplift and 50 cm subsidence. The Zenith Total Delay maps derived from the GACOS are used to correct tropospheric delays in the interferograms. After correction, an error of 0-2 cm is observed in the LOS measurements. The preseismic displacement reveals an indication of ground motion before the earthquake occurrence. The postseismic phase indicates afterslip after the earthquake. Localised displacements corresponding to landslides are observed on the mountainous slopes. The 3D displacements reveal approximately 40 cm of surface motion suggesting westward movement and reverse faulting.

Persistent Scatterer Interferometry (PSI) is an advanced technique to map ground surface displacements of an area over a period. The technique can measure deformation with a millimeter-level accuracy. It overcomes the limitations of Differential Synthetic Aperture Radar Interferometry (DInSAR) such as geometric, temporal decorrelation and atmospheric variations between master and slave images. In our study, Sentinel-1A Interferometric Wide Swath (IW) mode descending pass images from May 2016 to December 2017 (23 images) are used to identify the stable targets called persistent scatterers (PS) over Bengaluru city. Twenty-two differential interferograms are generated after topographic phase removal using the SRTM 30 m DEM. The main objective of this study is to analyze urban subsidence in Bengaluru city in India using the multi-temporal interferometric technique such as PSI. The pixels with Amplitude Stability Index ≥ 0.8 are selected as initial PS candidates (PSC). Later, the PSCs having temporal coherence > 0.5 are selected for the time series analysis. The number of PSCs that are identified after final selection are reduced from 59590 to 54474 for VV polarization data and 15611 to 15596 for VH polarization data. It is interesting to note that a very less number of PSC are identified in cross-polarized images (VH). A high number of PSC have identified in co-polarized (VV) images as the vertically oriented urban targets produce a double bounce, which results in a strong return towards the sensor. The velocity maps obtained using VV and VH polarizations show displacement in the range of ± 20 mm year⁻¹. The subsidence and the upliftment observed in the city shows a linear trend with time. It is observed that the eastern part of Bengaluru city shows more subsidence than the western part.