This session invites contributions across various disciplines and techniques to quantify, monitor and model the above-mentioned natural and human-induced processes, with particular emphasis on coastal vertical land motion and subsidence-related hazards. Interdisciplinary studies bridging tectonics, geodesy, volcanology, engineering geology, remote sensing, hydrology, ocean sciences, geomorphology and AI for enhanced risk assessment are strongly encouraged. We welcome contributions on a wide range of topics, including but not limited to: 1) Novel algorithms for mitigating SAR/InSAR errors, including deep learning approaches; 2) Advanced strategies for processing and analyzing SAR big data; 3) Integration of AI and machine learning with GNSS and InSAR observations to improve time series interpretation, identify deformation patterns, and disentangle driving processes, 4) multi-sensor and in-situ monitoring using geomorphologic, geodetic, field-based, and modeling approaches; 5) hazard assessments and disaster risk reduction, focusing on vulnerability, capacity, and resilience.
Orals: Thu, 7 May, 14:00–18:00 | Room 1.31/32
Water scarcity, land subsidence, and desertification constitute major environmental challenges in arid and semi-arid regions worldwide, with profound impacts on ecosystems, agricultural productivity, infrastructure, and long-term sustainable development. In many of these regions, intensive groundwater extraction has become a dominant driver of land subsidence, exacerbating water insecurity and environmental degradation.
Over the past decades, multi-decadal satellite observations from remote sensing and gravity missions have played a crucial role in estimating groundwater storage changes and quantifying the extent and rates of land subsidence at both local and regional scales. More recently, machine-learning (ML) approaches have been increasingly applied to map and assess land-subsidence hazards using diverse geospatial, hydrological, and satellite-derived datasets. While these models offer promising new capabilities, their results can vary substantially depending on model design, input data, and training strategies, sometimes leading to conflicting or uncertain outcomes.
In this contribution, we first focus on Iran, where land subsidence and water scarcity have emerged as widespread and critical issues, currently affecting more than 260 of the country’s 429 counties. We present results from a multi-decadal satellite-based analysis of land subsidence and groundwater dynamics and systematically compare these observations with outputs from several published machine-learning models. This comparison highlights both consistencies and discrepancies between observation-driven assessments and data-driven predictive approaches.
We then extend the analysis to selected regions in Central Asia, including Uzbekistan and Afghanistan, where similar hydrogeological and socio-environmental pressures are present but data availability and monitoring capacities are more limited. Finally, we discuss the key challenges and opportunities associated with integrating remote-sensing observations and machine-learning models for land-subsidence assessment, with particular emphasis on data quality, model transferability, uncertainty quantification, and implications for regional-scale hazard monitoring and water-resources management.
How to cite: Motagh, M. and Haghshenas Haghighi, M.: Observation-Driven Versus Machine-Learning Approaches for Land Subsidence Assessment in Arid Regions, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21423, https://doi.org/10.5194/egusphere-egu26-21423, 2026.
Abstract: Synthetic Aperture Radar Interferometry (InSAR) is a critical tool for monitoring geohazards. However, Phase Unwrapping (PU) remains a significant impediment. Conventional algorithms frequently encounter failure in regions characterised by low coherence or pronounced topographic gradients, resulting in substantial error propagation. In order to address these challenges, the present study proposes an advanced framework that integrates three optimised deep learning (DL) architectures: Attention-enhanced U-Net (UA), Generative Adversarial Network (GAN)-based restoration (GL), and Convolutional Neural Network (CNN) with channel attention (CUA).
The performance of these models was systematically evaluated using a large-scale, diverse InSAR benchmark dataset. Quantitative results demonstrate a substantial leap in accuracy compared to traditional methods. All three proposed DL models achieved a Peak Signal-to-Noise Ratio (PSNR) exceeding 28 dB and a Structural Similarity Index (SSIM) above 0.85. Specifically, the CUA model demonstrated the highest level of precision, achieving a PSNR of 38.24 dB and effectively suppressing noise in complex interferograms. In order to preserve structural integrity in areas of sharp terrain, the UA model (incorporating a 5-layer attention mechanism) achieved an edge SSIM of 0.8888, thereby demonstrating a significant improvement over the Minimum Cost Flow (MCF) algorithm, which frequently encounters difficulties with phase residues in high-gradient regions.
In order to validate the practical applicability of the models, they were tested on real TanDEM-X data from the Weinan region in China. The UA model exhibited a high average SSIM of 0.95, successfully recovering subtle terrain features where traditional MCF demonstrated a mean PSNR of only 18.08 dB. Moreover, a gradient accumulation strategy was introduced with a view to optimising the training process. A thorough efficiency analysis reveals that the GL model (at a 1:1 ratio) can reduce training time by approximately 92% compared to the high-complexity CUA-Accum2 configuration, offering a scalable solution for SAR big data processing.
In conclusion, this research provides a robust, automated, and high-precision methodology for InSAR PU. The present work offers novel insights into the generation of reliable geodetic products for disaster risk reduction in challenging environments by bridging the gap between advanced DL processing and real-world hazard monitoring.
Keywords: U-Net, generative adversarial network (GAN), convolutional neural network (CNN), phase unwrapping(PU), synthetic aperture radar interferometry (InSAR).
(Corresponding author: Ruiqing Niu)
How to cite: Wu, C. and Niu, R.: InSAR Phase Unwrapping via Integrated Multi-Model Deep Learning: Advancing Accuracy in Complex Topographic Hazards, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-3444, https://doi.org/10.5194/egusphere-egu26-3444, 2026.
Many low-lying coastal plains worldwide contain abundant organic facies, often peat, in the Holocene subsurface. These facies are very susceptible to soil deformation processes, some of which leading to irreversible land subsidence. Most important irreversible processes are oxidation of organic matter, shrinkage (in the unsaturated soil zone) and compaction (in the saturated soil zone). Reversible soil deformation processes are shrinkage and swell and poro-elastic deformation.
In cultivated areas, deformation processes leading to land subsidence are often driven, and accelerated, by human activities such as drainage for agriculture and loading of the subsurface. To reduce land subsidence, the first step is to quantify subsidence rates in space and time and to identify the relative contribution of soil deformation processes to total subsidence. Next, measures may be developed and applied.
At various locations in the Dutch coastal plain, extensometers specifically designed for soft organic facies are used to measure vertical movement of multiple levels in the Holocene subsurface at high temporal resolution. Resulting multiyear timeseries are used to quantify the amount of irreversible and reversible deformation over time for different soil intervals, which subsequently may be linked to soil deformation processes. Results demonstrate a large variability in the relative contribution of deformation processes to total subsidence, due to spatially variable geological and hydrological circumstances, indicating that site-specific measures are needed to reduce land subsidence. This spatial variability also requires spatially explicit mapping approaches, e.g. models, to predict deformation behaviour in soft soil sequences.
How to cite: van Asselen, S. and Erkens, G.: Unravelling shallow subsurface deformation processes leading to land subsidence in organic-rich coastal plains , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1131, https://doi.org/10.5194/egusphere-egu26-1131, 2026.
Ground deformation (GD) represents a significant geohazard, arising from both natural mechanisms such as tectonic movements and anthropogenic activities, including excessive groundwater extraction. Globally, GD threatens geological stability and civil infrastructure. Consequently, continuous monitoring of GD is vital for characterizing its spatial and temporal behaviour, enabling hazard assessment and improving regional safety. Time-Series Interferometric Synthetic Aperture Radar (TS-InSAR) has emerged as a robust remote sensing approach for long-term GD monitoring. Despite its effectiveness, many existing TS-InSAR processing platforms suffer from notable constraints, including limited geographic flexibility, commercial licensing, and the absence of a comprehensive end-to-end processing framework. Although GMTSAR, one of the most widely used TS-InSAR processing platforms, overcomes some of these shortcomings, it is highly user-driven, remains dependent on manual user input, requires command-line execution via C-shell, lacks a graphical user interface, and does not consider essential processing steps such as interferogram network pruning and unwrapped interferogram anchoring. These limitations reduce usability and may affect processing accuracy. To address these challenges, this study presents DefoEye (v1), an open-source, Python-based toolkit integrated with GMTSAR to enable a complete TS-InSAR processing workflow for Sentinel-1 data through an easy-to-use interface. DefoEye offers a unified end-to-end framework incorporating parallelized processing, interferogram network pruning, and multiple anchoring strategies. Its performance was tested across multiple regions characterized by diverse geological environments, GD drivers, atmospheric conditions, and climatic regimes over varying temporal scales. Validation results show strong agreement between DefoEye-derived GD measurements and independent GNSS observations. Additionally, the results closely match those obtained from other widely used TS-InSAR software packages. Unlike existing processing platforms, which require fragmented workflows across multiple tools, DefoEye streamlines the entire process within a single integrated platform. Overall, the findings demonstrate that DefoEye produces reliable TS-InSAR results applicable to a wide range of geological, hydrological, and environmental studies. The toolkit is publicly available at: https://github.com/ATDehkordi/DefoEye.
How to cite: Taheri Dehkordi, A., Hashemi, H., and Naghibi, A.: An End-to-End Python-Based Toolkit for Facilitated Time-Series Interferometric Synthetic Aperture Radar (InSAR) Analysis of Sentinel-1 Remote Sensing Data, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-3070, https://doi.org/10.5194/egusphere-egu26-3070, 2026.
Groundwater over-pumping from aquifer system is the primary driver of land subsidence in alluvial plains, causing serious impacts on engineering geology and/or environment, typically shallow aquifers within 300 m. However, in the Choushui River Alluvial Fan of Taiwan, subsidence rates reaching 20 mm per year have been recorded at depths greater than 300 m, indicating deep compression. This observation highlights the importance of evaluating not only the responses of shallow aquifers but also the contribution of deep compression to total subsidence. In this study, a stochastic heterogeneous hydrogeological model (HHM) is developed using 468 geological borehole data to assess and quantify the influence of shallow groundwater pumping on deep compression. The model simulates transient groundwater flow and compaction and is calibrated and validated using monitoring data from 2018 to 2021. Simulation outcomes indicate that shallow pumping accounts approximately 1.265 billion m3 annually. This contributes 6-35% of deep compression along the Taiwan High-Speed Rail corridor, with spatial variability governed by hydrogeological structure and pumping area. The HHM successfully captures depth-dependent groundwater flow and subsidence behavior. Future work will extend the model to scenario-based predictions to support high-speed rail safety and promote sustainable groundwater resource management.
How to cite: Tran, D.-H., Wang, S.-J., Wu, I.-Y., Wu, S.-C., and Lin, C.-W.: Modeling and Quantifying Deep Subsurface Compression Induced by Shallow Pumping Through a Stochastic Approach, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-2097, https://doi.org/10.5194/egusphere-egu26-2097, 2026.
Differential radar interferometry (dInSAR) enables to derive ground displacements maps all over
the globe. This method is particularly revelant on high active volcanoes such as the Piton de la Four-
naise (Reunion Island) for monitoring volcanic unrest. The dInSAR method consists of computing the
phase difference between two satellites radar images acquied at distinct overflight passes to produce
an interferogram showing cumulative displacements during the time period.
Nevertheless, atmospheric variability between the txo epochs, especially water vapor in the tro-
posphere, introduces a significant bias in the calculation of interferograms. To correct this, various
methods have been developed using empirical formulas based on the phase-elevation correlation or
global atmospheric models provided by the ECMWF. However, these methods are not well suited to
the context of Reunion Island, which features significant topography and highly variable atmospheric
conditions.
Here, we propose a GPS tomography algorithm specifically adapted to the island’s context in terms
of orography, spacial distribution of the GPS stations and inversion method, in order to reconstruct a
reliable 3-D water vapor field above Reunion Island, (See corresponding abstract in Session G5.2).
Hence, this water vapor field is then used to provide atmospheric corrections to individual Sentinel-
1 interferograms by tracing each delays through the 3D tomographic grid according to the specific
SAR geometry.
Finally, local delays maps and the associated corrected interferograms at Reunion Island will
be compared to those obtained from the common approaches (empirical, global models) in order to
quantify the benefits of our approach.
How to cite: Van Baelen, J., Gerville, H., Albino, F., Durand, F., and Morel, L.: Improving Tropospheric Corrections of InSAR Observations on Reunion Island using 3D GPS Water Vapor Tomography, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9701, https://doi.org/10.5194/egusphere-egu26-9701, 2026.
Shallow subsidence is a key problem in many coastal plains, such as those of the Mississippi (U.S.A), the Mekong (Vietnam), Po (Italy), and Rhine deltas (Netherlands). Managing it is as important as managing anticipated sea-level rise. Accurate prediction of shallow land subsidence into the coming century, requires robust parameterization of a series of complexly interacting physical and biochemical processes (leading to consolidation, oxidation, shrinkage), that operate across heterogeneous shallow subsurface conditions. Traditional physics-based models depend on parameters that are difficult to constrain spatially, while purely data-driven approaches might give physically inconsistent results and often lack physical interpretability. We present a hybrid modeling framework that balances this tradeoff by combining a fully differentiable version of the process-based subsidence model Atlantis with neural network components for learning spatial parameter heterogeneity, enabling gradient-based parameter optimization directly from observational data.
Our approach converts established isotach consolidation model and peat-oxidation calculation methods into a differentiable computational graph, allowing automatic differentiation to propagate observational constraints through the physics model. This enables joint inversion of spatially distributed InSAR-derived observations and vertical extensometer profiles to constrain process parameters at voxel scale. Crucially, we incorporate observation uncertainty (the full variance-covariance structure) of InSAR-derived measurements through a statistically rigorous loss function (Mahalanobis distance), properly accounting for spatial and temporal correlations that traditional calibration approaches neglect. First results confirm that the framework can recover peat oxidation parameters from synthetic subsidence observations; integration of InSAR-derived data with full uncertainty characterization is underway.
The differentiable architecture offers several advantages: 1) principled uncertainty quantification by accounting for the error structure of input observations, 2) efficient optimization through gradient descent rather than computationally expensive sampling methods, and 3) flexible integration of heterogeneous data sources within a unified modeling framework. We demonstrate the approach spatially, using various observations for a long-managed mainly agricultural peat meadow polder (Krimpenerwaard, The Netherlands; current average elevation 1.5 m below MSL and sinking). Our methodology bridges geodetic remote sensing with process-based geotechnical modeling, contributing to improved projections of coastal relative sea-level rise by constraining subsurface processes at operationally relevant scales. The approach is computationally efficient and can be scaled to larger areas and longer timeframes, depending on the availability of novel InSAR-derived observations and subsurface data.
How to cite: Kilic, D., Pomarol Moya, O., Erkens, G., Karssenberg, D., Nussbaum, M., Cohen, K. M., and Stouthamer, E.: Inverse Modelling of Shallow Land Subsidence using a Hybrid Differentiable Physics-Machine Learning Approach, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19716, https://doi.org/10.5194/egusphere-egu26-19716, 2026.
Over the last decades, synthetic aperture radar (SAR) images and SAR interferometry (InSAR) have revolutionized Earth observation, allowing for geophysical monitoring of Earth surface processes with centimeter-to-millimeter precision. Accurate measurement of ground displacement is essential for the understanding of natural hazards, such as earthquakes. In particular, the detection of small ground (transient) displacements is of utmost importance for better imaging the dynamics of active faults, especially in tectonic settings undergoing low deformation rates. However, detecting small deformation signals in InSAR data remains a significant challenge due to the high noise level in the data (e.g., speckle noise, tropospheric and ionospheric perturbations). Multiple and successful InSAR mass processing methods, including state-of-the-art noise correction methods, have been developed over the last decade, but all rely on intensive computing of massive databases, a tedious procedure that cannot yet be applied at a global scale. Furthermore, because of the low probability of finding earthquakes in intraplate continental settings, automatic detection of such signals with InSAR data is currently out of the question, mostly due to the low signal-to-noise ratio.
Here, we develop a deep-learning-based method to denoise InSAR time series. We design a spatiotemporal attentive convolutional U-Net to retrieve small-scale deformation in noisy interferometric SAR time series, trained in a hybrid supervised and self-supervised manner on synthetic data and evaluated first on synthetic and, finally, on real InSAR time series. When applied to a time series in the North Anatolian Fault, the method effectively extracts millimeter-scale deformation associated with fault creep. The extracted deformation is consistent with independent ground truth measurements, thereby validating our method and opening the possibility of its application to diverse tectonic settings globally, as well as targeting the method to the detection of dislocation-like signals in raw SAR data, possibly optimizing the SAR interferometry processing chain, reducing the need to process entire datasets, and significantly accelerating computation.
How to cite: Costantino, G. and Jolivet, R.: Denoising of interferometric SAR time series: towards global (slow) fault slip detection, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17951, https://doi.org/10.5194/egusphere-egu26-17951, 2026.
Marshlands in the Venice Lagoon (Italy) are among the most valuable morphological environments, yet they face the risk of disappearing by the end of the century. This risk is associated to relative sea-level rise due to climate change, land subsidence caused by various processes, and decrease in sedimentation rate above the marshland platform because of an increasing frequency of MoSE activation. In the micro-tidal conditions characterizing the Venice Lagoon, marshlands need to maintain an elevation between 20 and 40 cm above mean sea level to keep pace with relative sea level rise,. Recent research worldwide has clearly revealed how aggradation (i.e., net elevation gain) of transitional landforms can be significantly smaller than the accumulation rate of newly deposited sediments on their surface. The difference between aggradation and sedimentation rate is primarily related to the self-compaction of Holocene deposits induced by the progressive load applied by subsequently deposited (younger) sediments. To investigate the amount of sediments needed for saltmarshes to keep pace with the expected (relative) rise in lagoon water level, we applied the NATSUB3D simulator, that is based on finite element discretization and accounts for sediment deposition and consolidation over time in the context of large vertical deformations. NATSUB3D uses an adaptive mesh, with the hydro-geomechanical properties (porosity, hydraulic conductivity, compressibility) of the heterogeneous growing sedimentary body that vary in space and over time depending on the actual vertical effective stress. NATSUB3D is applied to three representative sections of the Holocene sequence in the Venice Lagoon, recostructed through borehole lithostratigraphy, facies analysis, C14 datings, in-situ and lab geomechanical tests. The model allows quantifying the sedimentation needs based on different climatic scenarios and characteristics of the future sedimentation. It also highlights the significant difference between marshlands located in the different sections depending on the Holocene thickness and composition
How to cite: Qayyum, U., Cosma, M., Baldan, S., Zoccarato, C., Ferronato, M., Tosi, L., and Teatini, P.: Holocene-dependent sediment requirement for supporting the resilience of the Venice Lagoon marshlands against expected (relative) sea-level rise, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20312, https://doi.org/10.5194/egusphere-egu26-20312, 2026.
Wide-area, high temporal resolution ground deformation measurements are crucial for geological hazard identification, continuous monitoring, and risk assessment. Interferometric Synthetic Aperture Radar (InSAR) has become an important technique for acquiring ground deformation information due to its all-weather, day-and-night observation capability and high measurement precision. LuTan-1 (LT-1) is China’s first interferometry-oriented SAR twin-satellite formation operating at L-band. Since its launch in 2022, LT-1 has collected a large volume of high-quality SAR data, demonstrating strong potential for deformation measurement. However, delivering stable, reliable, and operational wide-area ground motion services with LT-1 remains challenging because of the limited swath width in stripmap imaging, the complexity of multi-track acquisitions, and pronounced long-wavelength systematic errors that lead to inconsistencies across tracks.
To address these issues, we propose a provincial-scale ground motion service framework for geological hazard monitoring using LT-1 InSAR data. The framework enables automated multi-track InSAR processing, systematic errors correction, and routine generation of consistent wide-area deformation products. We first process LT-1 SAR data to generate multi-track InSAR deformation results. We then apply a large-look-based method to correct systematic errors in each InSAR deformation result. Exploiting the distinct spatial characteristics of the long-wavelength error component versus the true deformation signal, we select an appropriate large-look window and upsample to estimate and remove systematic errors, thereby reducing inter-track discrepancies. A unified reference frame is subsequently established, and the corrected multi-track results are resampled and integrated using weighted averaging to produce a seamless provincial ground motion result.
We used Shaanxi Province as the study area. The results showed that after correction, the mean absolute error (MAE) decreased by approximately 2 mm, and long-wavelength systematic errors were effectively suppressed. Comparison with contemporaneous Sentinel-1 (S-1) deformation results showed strong consistency in deformation trends. Approximately 150 deformation results covering the entire province were mosaicked in about 2 hours, demonstrating a good balance between accuracy and efficiency. The results were integrated into routine geological hazard monitoring workflows, and joint interpretation with optical imagery enabled the detection and delineation of potential hazard sites, providing data support for hazard surveillance and risk assessment. Similar to the European Ground Motion Service (EGMS), we have developed a provincial-scale ground motion monitoring service based on LT-1 data. The system can generate monthly updated deformation maps, providing a basis for near-real-time monitoring.
How to cite: Wang, K., Qin, Y., Zhu, H., Zhang, W., Yang, S., Yang, T., Zhang, J., Han, J., and Zhu, N.: A Framework of Provincial Ground Motion Service Using L-Band LuTan-1 InSAR Data for Geological Hazard Monitoring, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19156, https://doi.org/10.5194/egusphere-egu26-19156, 2026.
Land subsidence driven by intensive groundwater abstraction remains a major concern in Taiwan, particularly in Yunlin County. This study integrates multidisciplinary observations with artificial intelligence (AI)-driven modeling to support land-subsidence management. A hydrogeological conceptual model was developed to simulate groundwater-level dynamics and aquifer-system compaction, and an AI approach was used to capture the nonlinear relationship between groundwater fluctuations and soil-layer compression. Results indicate that subsidence is influenced by climate extremes and pumping intensity. The strong positive correlation and synchronized temporal variations between groundwater level and soil compression suggest a coupled hydro-mechanical response. To identify mitigation measures, five scenarios were evaluated, focusing on crop conversion and pumping regulation. Compared with current pumping conditions, both crop conversion and rotational pumping reduce groundwater drawdown and associated compression. Among the alternatives, conversion to sweet potato combined with rotational pumping yields the smallest drawdown, indicating a practical pathway for sustainable groundwater management and land-subsidence mitigation.
How to cite: Liu, C.-Y. and Ku, C.-Y.: Integrating Multidisciplinary Observations and AI-Driven Modeling for Land Subsidence Management in Taiwan, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-4790, https://doi.org/10.5194/egusphere-egu26-4790, 2026.
Suffusion-related subsidence and collapse processes represent a major geomorphological hazard in salt-bearing environments, particularly where natural dissolution is amplified by anthropogenic factors such as mining legacy and urban development. The town of Slănic (Prahova County, Romania) exemplifies a complex salt-karst landscape affected by ground deformation, impacts on the built environment, and localized collapses, including the major April 2024 event in the city center. The spatially heterogeneous evolution of suffusion features, combined with high societal exposure, requires quantitative monitoring strategies capable of resolving both slow trends and rapid deformation episodes.
The proposed monitoring system includes two complementary components: real-time and recurrent/event-based monitoring. Real-time monitoring relies on permanently operating instruments that continuously measure parameters such as displacement, acceleration, or inclination and transmit data to a centralized platform. Recurrent monitoring consists of measurements performed at predefined intervals or triggered by hazardous phenomena such as structural cracking, landslides, or ground subsidence.
The monitoring concept combines (i)permanent and temporal GNSS campaigns, (ii)terrestrial laser scanning (TLS), (iii)recurrent high-precision topographic measurements and geometric levelling, and (iv)real-time seismic monitoring designed to capture multi-scale deformation signals from neighbourhood scale down to structural detail.
GNSS data are processed using a PPP strategy (GipsyX) to obtain daily solutions in ITRF14 and derive horizontal and vertical deformation time series, with expected precisions of ~2 mm(H) and ~7 mm(V), enabling detection of subtle trends in the unstable urban setting. Campaign GNSS points on dedicated pillars densify the network where continuous deployment is not feasible. Repeated TLS surveys generate multitemporal point clouds for 3D change detection, capturing fractures, localized settlements, and infrastructure deformation linked to salt dissolution and suffosion processes. Topographic measurements using fixed pillars and prisms, together with precise digital levelling, provide independent constraints on vertical displacement and validate GNSS and point-cloud-derived signals.
The preliminary results demonstrate that an integrated multi-method monitoring concept provides a robust, reproducible, and scalable framework for geomorphological monitoring of suffosion-driven deformation in salt-karst terrain. In the monitored sector, the pillar closest to the active suffosion zone recorded a ~2.5 cm permanent displacement in a 3-month period, confirming measurable near-field ground instability. In parallel, prism-based tracking of instrumented buildings indicates systematic inclinations directed toward the suffossion center, consistent with progressive differential settlement and deformation gradients around the collapse-prone area. Data complementarity is essential: three-dimensional displacements measured by the total station on buildings can be validated through tilt measurements, while accelerometers provide short-term confirmation of structural stability.
Overall, the proposed multidisciplinary monitoring system demonstrates feasibility and significant added value, combining real-time detection capabilities with long-term observations. This integrated framework supports informed decision-making, risk mitigation, and targeted monitoring strategies, while offering strong potential for scalability, standardization, and replication at national and international levels through future prototype development, automated alert systems, and intelligent data integration platforms.
Keywords: Suffosion, Salt karst, Ground deformation monitoring, GNSS-(PPP), TLS
Acknowledgements: This work was carried out within the project No.28Sol(T28)⁄2025, funded by the Ministry of Education and Research, through UEFISCDI (Romanian Executive Agency for Higher Education, Research, Development and Innovation Funding).
How to cite: Nastase, I. E., Tiganescu, A., Muntean, A., Grecu, B., Poiata, N., Neagoe, C., and Tataru, D.: Integrated Monitoring Concept of Suffusion-Driven Ground Deformation and built environment in the Slănic Salt-Karst Area (Romania), EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13572, https://doi.org/10.5194/egusphere-egu26-13572, 2026.
Conventional InSAR time series methods constrain pixel selection to scatterers maintaining coherence across the entire observation period, preventing monitoring of newly constructed or demolished infrastructure. We present a method for estimating the displacement time series of Temporarily Coherent Scatterer (TCS) pixels implemented in the free and open-source research software SARvey that overcomes this limitation for changing infrastructure.
The TCS approach detects significant changes in the SAR signal time series with a coherent change detection method that exploits the phase noise level of a scatterer. The phase noise is estimated from the spatial neighbourhood which is also used to estimate each pixel's coherent lifetime from the period before and after the change. Displacement time series are then retrieved within each pixel's coherent lifetime from a small baseline interferogram network allowing to retrieve transient displacement signals.
Validation over Miami, USA (239 Sentinel-1 ascending images, track 48, April 2016–June 2025) demonstrates accurate detection of both, construction and demolition, of high-rise buildings along the coastline, validated against high-resolution optical satellite imagery. Post-construction settlement rates confirm previously reported infrastructure displacement patterns.
SARvey's TCS implementation combines automated change detection with robust time series inversion, delivering displacement maps for structural health monitoring and risk assessment. The modular, open-source framework supports multi-mission SAR data (Sentinel-1, TerraSAR-X). The methodology presented in this work contributes to overcoming the critical gap in operational InSAR services in case of changing environments.
How to cite: Motagh, M., Piter, A., and Haghshenas Haghighi, M.: Temporarily Coherent Scatterer Analysis for Monitoring Changing Infrastructure, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-14492, https://doi.org/10.5194/egusphere-egu26-14492, 2026.
Global mean sea level has accelerated to record-high rates over the past decade, with 2024 exhibiting an anomalous rise linked to exceptional ocean heat content. Relative sea-level rise (RSLR) in deltaic megacities, such as Bangkok, is further amplified by rapid land subsidence caused by groundwater extraction and urban development. This study develops a novel integrated framework to assess future coastal risks in Bangkok by combining bias-corrected CMIP6 sea-level projections with a hierarchy of five flood models ranging from simple static bathtub approaches to advanced shallow-water equations. The framework also incorporates subsidence-adjusted probabilistic retreat modeling and machine-learning-based downscaling of population data to approximately 200-meter resolution, allowing detailed spatial analysis of exposure. Our results indicate that by 2100, retreat probabilities exceed 90% in coastal districts such as Samut Prakan and Samut Sakhon under moderate emissions scenarios (SSP2-4.5), escalating to near-universal land loss (>99%) under high emissions (SSP5-8.5). Population exposure peaks at over 30 million people in these scenarios. Validation using satellite-derived NDWI data demonstrates the highest predictive skill for the shallow-water model (R² = 0.92). Policy analysis uncovers an “urban resilience paradox” where investments in protective infrastructure encourage expansion into vulnerable zones, increasing long-term risks to build equitable and resilient futures in subsiding deltaic megacities like Bangkok.
How to cite: Pramanik, M.: Probabilistic Retreat and Population Risk in Subsiding Bangkok: Multi-Scenario Sea-Level Rise and Flood Modeling, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21082, https://doi.org/10.5194/egusphere-egu26-21082, 2026.
This study investigates crustal deformation associated with moderate to large earthquakes along the Africa–Eurasia plate boundary in the western Mediterranean using Synthetic Aperture Radar interferometry (InSAR). We integrate multi-temporal Sentinel-1A/B SAR time series (MT-InSAR) with GPS measurements to obtain high-resolution deformation estimates in the central Tell Atlas of northern Algeria, a region characterized by significant seismicity driven by oblique plate convergence. A dataset of 120 Sentinel-1 C-band SAR images acquired between 2015 and 2023 from ascending and descending orbits was processed to capture deformation from multiple viewing geometries. Interferograms were selected using baseline thresholds to maximize coherence and detect subtle ground motion (< 5 mm/yr). Mean horizontal velocity profiles were modeled using a nonlinear least-squares inversion to estimate key fault parameters. The results indicate slip rates ranging from 3.5 to 6.0 mm/yr, with an average of ~5 mm/yr, and shallow fault locking depths (< 20 km). The deformation field reveals dominant E–W right-lateral motion and NNW–SSE contraction at rates of 2–3 mm/yr, consistent with transpressional tectonics associated with oblique convergence relative to the stable High Plateaus. These findings provide new constraints on strain accumulation and the long-term behavior of active faults, with important implications for seismic hazard assessment in northern Algeria.
Keywords: MT-InSAR, GPS, Tell-Atlas, oblique convergence, seismic hazard
How to cite: Miloudi, S., Çakir, Z., and Meghraoui, M.: Oblique Plate Convergence and Strain Accumulation in Northern Algeria from InSAR Analysis, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20345, https://doi.org/10.5194/egusphere-egu26-20345, 2026.
At a national scale, homogeneous data for sinkhole susceptibility mapping are scarce in Germany. The individual German geological surveys collect relevant data in their respective federal states, resulting in heterogeneous datasets. However, certain tasks, such as the search for a final nuclear waste repository, require homogeneous data coverage across the entire country.
To enable such an approach, we homogenised the karst feature inventories from multiple federal states and selected publications. This compilation provides point data on the spatial occurrence of generalised karst features, such as sinkholes and caves. We derived information on the presence of subrosion-prone rocks from both the General Geological Map of the Federal Republic of Germany and the Hydrogeological Map of Germany.
Due to differing subrosion rates, we distinguish between three types: carbonate karst, chloride karst and sulphate karst. We assigned one or more karst type to each feature in the merged karst inventory and generated a separate susceptibility map for each type.
Using average nearest neighbour analysis, we demonstrate that karst features are spatially clustered and derive a buffer distance to delineate areas of high susceptibility around these features. We classified areas underlain by known karst-prone rocks as having medium sinkhole susceptibility. The final sinkhole susceptibility map is generated by combining these two binary layers, thus depicting karst-prone areas in Germany susceptible to sinkhole formation at a scale of 1:250,000.
The results are validated using borehole data from the Borehole Map of Germany, including information on karst-prone horizons and geohazard maps from individual federal states.
Our results demonstrate a pathway for sinkhole susceptibility mapping in data-scarce regions.
How to cite: Schüßler, N., Fuchs, M., Torizin, J., Kuhn, D., Anschütz, H., and Mohr, C. H.: Sinkhole Susceptibility Mapping in Data-Scarce Areas at Small Scales, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-16409, https://doi.org/10.5194/egusphere-egu26-16409, 2026.
This study investigates surface deformation phenomena in the towns of Messolonghi and Aitolikon, located in Aitoloakarnania, western Greece, by applying Persistent Scatterer Interferometry (PSI) techniques. No systematic remote sensing–based investigation of land subsidence has been conducted in these areas, despite recurring reports of flooding, ground deformation, and structural damage.
Multitemporal Interferometric Synthetic Aperture Radar (MT-InSAR) analysis was performed using Sentinel-1A and Sentinel-1B satellite data covering the period 2015–2022. Persistent Scatterer (PS) time series were extracted to quantify Line-of-Sight (LOS) deformation rates and were cross-validated with results from the European Ground Motion Service (EGMS) of the Copernicus. Geological, geotechnical, and hydrogeological data were also acquired from the Hellenic Survey of Geology and Mineral Exploration, the Central Laboratory of Public Works archives, and private geotechnical consultants’ reports. These findings enabled a more robust interpretation of the observed deformation patterns and their controlling mechanisms.
The InSAR analysis results reveal ongoing subsidence in both towns, with spatially variable deformation rates. In Messolonghi, LOS deformation rates reach up to −5 mm/yr, particularly in the eastern and southern sectors of the town, while northern areas exhibit stability (0.3 to −1.3 mm/yr). Subsidence rates increase towards the coastline, reflecting the presence of younger, unconsolidated alluvial deposits. In Aitolikon, mean deformation rates reach −4.5 mm/yr, with pronounced subsidence observed in both the southern and northern parts of the island, where significant structural damage has been reported. These areas coincide with zones of artificial fillings.
Geological and geotechnical data indicate the presence of laterally continuous Quaternary deposits, consisting of clay, clayey silt to silt, and clayey sand to sand horizons, with a thickness of 80-100m. The observed deformations are primarily attributed to the natural compaction and consolidation of these sediments, further intensified by anthropogenic interventions such as river diversion in Messolonghi.
Beyond the subsidence processes identified, the long‑term impacts of climate change further intensify the vulnerability of Messolonghi and Aitolikon. Continuous sea level rise, documented at rates of 2–4 mm/yr in the wider Mediterranean, combined with coastal erosion and increasingly frequent flooding events, places additional stress on both cities. The combination of the formations’ compaction with rising water levels and extreme precipitation has led to recurrent inundations of the low‑lying areas. Both towns have already been declared in a state of emergency during major flood events in recent years, underscoring the severity of the hazard. Projections under high‑emission scenarios (SSP5–8.5) suggest that by 2100, sea level rise could exceed 0.8 m in the Messolonghi lagoon, significantly expanding the existing flood‑prone zones.
Overall, the study demonstrates that land subsidence in Messolonghi and Aitolikon is an ongoing process with steady deformation rates, posing a risk to the infrastructure and buildings in both areas. In addition, the continuous rise in sea level, combined with coastal erosion and increasingly frequent flooding events driven by climate change, further exacerbates the vulnerability of these towns.
How to cite: Antoniadis, N., Alatza, S., Loupasakis, C., and Kontoes, C. (.: Cascading Hazards of Land Subsidence and Relative Sea‑Level Rise: Flooding Risks in the Coastal Towns of Messolonghi and Aitolikon, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-2940, https://doi.org/10.5194/egusphere-egu26-2940, 2026.
Ground deformation resulting from groundwater extraction is a critical yet frequently under-monitored process in semi-arid agricultural regions. Although Interferometric Synthetic Aperture Radar (InSAR) is a key tool for detecting surface deformation, its integration with predictive modelling frameworks remains limited, especially in basins with limited in situ hydrogeological data. This research introduces a data-driven framework that integrates InSAR time series with machine learning techniques to examine the temporal dynamics of ground deformation and its sensitivity to hydroclimatic variability. The proposed framework is applied to the Ararat Valley (Armenia), a transboundary agricultural basin characterized by semi-arid climatic conditions, strong seasonal variability in precipitation and evapotranspiration, and intensive groundwater-dependent irrigation. These conditions make the valley particularly sensitive to groundwater stress and associated land deformation, while the limited availability of long-term groundwater observations poses challenges for conventional hydrogeological analyses. Surface deformation is extracted from Sentinel-1 imagery using a time-series InSAR approach and combined with satellite-derived hydroclimatic and environmental variables, such as precipitation, temperature, evapotranspiration, vegetation dynamics, and soil moisture. Long Short-Term Memory (LSTM) neural networks are utilized to model non-linear temporal relationships between deformation and environmental drivers, enabling the capture of delayed and cumulative responses to hydroclimatic forcing. For exploratory future assessments, additional machine learning and empirical models estimate potential trajectories of vegetation and soil moisture based on regional climate projections, which are then incorporated into the deformation modelling framework. The methodology is designed to be scalable and transferable, facilitating deformation analysis in regions with sparse or unevenly distributed groundwater observations. Instead of prioritizing site-specific calibration, the framework emphasizes process representation and scenario exploration. A dedicated InSAR validation strategy, involving the comparison of deformation signals from ascending and descending Sentinel-1 acquisition geometries (ASC versus DESC), is used to assess the internal consistency and robustness of the InSAR-derived time series. This work advances methodological development and highlights the potential of integrating satellite-based deformation monitoring with machine learning to enhance groundwater-related risk assessment under evolving hydroclimatic conditions in poor monitored regions.
How to cite: Romano, G. and Hashemi, H.: An InSAR–Machine Learning Framework for Ground Deformation Modeling and Scenario-Based Projections in the Ararat Valley (Armenia), EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-3505, https://doi.org/10.5194/egusphere-egu26-3505, 2026.
This study demonstrates the capability of the Small Baseline Subset (SBAS-InSAR) technique to resolve low-magnitude surface deformation in regions prone to high spatial and temporal decorrelation. By processing Sentinel-1 data (October 2020–June 2023) at dual resolutions of 80 m and 160 m for the Munich geothermal region, we resolved vertical and horizontal displacement rates of up to ±2 mm/year. The rates are low compared to many geophysical signals, but they are clearly resolvable with SBAS-InSAR. This demonstrates the technique's utility for monitoring complex environments such as urban infrastructures. The results are evaluated through comparison with external data obtained using the PS-InSAR technique.
In the Munich region, geothermal operations have previously triggered minor seismic events, such as in Unterhaching and Poing. We aim to study if non-linear stress changes that precede seismic failure can be identified in the time series and isolated from other processes, such as construction works (e.g., the new railway line), as well as other geophysical processes like hydrology. Subtle velocity gradients could serve as critical indicators of subsurface stress accumulation and provide an important empirical basis for validating geomechanical models and ensuring the long-term safety of geothermal energy expansion.
How to cite: Semenova, Y. and Seitz, F.: Application of SBAS-InSAR for monitoring geothermal-induced ground deformation in the Munich Region, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-3577, https://doi.org/10.5194/egusphere-egu26-3577, 2026.
Surface deformation monitoring is a critical component of natural hazard management. While Interferometric Synthetic Aperture Radar (InSAR) provides high-resolution observations, the traditional spatial phase unwrapping process remains a potential source of measurement bias and error. To circumvent the complexities of the process, this study presents an alternative workflow for estimating relative surface deformation by differencing the phase of a target area from a reference area on a pixel-by-pixel basis. Building on a preliminary experiment using a single corner reflector (CR) installed on top of a building, we designed an experiment site in an area characterized by low SAR backscatter to minimize background interference. A dual CR setup was deployed: one with manual height adjustments to simulate vertical displacement, while the other was maintained at a constant elevation to serve as a high-stability reference. By utilizing Sentinel-1 imagery with a 6-day revisit cycle, we evaluated the ability of the proposed workflow to detect relative movement between the two reflectors. Whereas the previous experiment using natural adjacent pixels as a reference yielded a standard deviation of approximately 0.6 cm, the current dual CR setup allows for more precise signal localization and stricter control over the stability of the reference point. Our findings suggest that this method is capable of detecting localized deformations and provides a pathway toward more automated and error-resistant disaster management tools.
How to cite: Tsai, T. Y. and Tseng, K. H.: Estimating Relative Surface Deformation via Pixel-Based InSAR Phase Differencing, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9752, https://doi.org/10.5194/egusphere-egu26-9752, 2026.
Natural gas remains a critical transitional energy source in the shift towards renewable systems, addressing seasonal demand fluctuations and supporting energy security. Underground Gas Storage (UGS) facilities, particularly salt caverns, play a key role in this framework by enabling rapid injection and withdrawal cycles. However, UGS activities can induce ground deformation, including subsidence and seasonal displacements. This raises questions about geomechanical stability, long-term sustainability, and environmental impacts. In this context, Interferometric Synthetic Aperture Radar (InSAR) represents a promising technology for large-scale, cost-effective, and high-precision monitoring of surface displacements associated with UGS operations. This study introduced a novel methodological framework for UGS monitoring by the use of Sentinel-1 SAR images and the advanced SqueeSAR algorithm. Two case studies from Lower Saxony (Germany), Jemgum and Nüttermoor, were selected as representative salt cavern UGS sites characterised by high injection/withdrawal rates and long operational histories. Multi-temporal InSAR analyses revealed subsidence velocities of up to 28 mm/year, resulting in distinct cone-shaped deformation that encompasses both UGS facilities. Decomposing the ascending and descending datasets allowed the vertical and east-west horizontal components of displacement to be quantified. The results confirm that salt cavern convergence induces a long-term subsidence trend, while operational cycles generate seasonal displacement patterns correlated to injection and withdrawal phases. To standardise the detection of UGS-affected areas, a semi-automatic thresholding procedure was implemented within a GIS environment, combining displacement velocity, cumulative deformation, and seasonal correlation criteria. This approach allowed the systematic identification of areas affected by UGS operations, including subsidence zones close to storage wells and seasonal deformation fields further away. Building on this, the interpretation of displacement time series in relation to UGS curves of gas in storage was refined using a cross-correlation technique. The RTK parameters, correlation (R), time delay (T) and proportionality (K), allowed the isolation of displacement signals directly attributable to UGS operations, filtering out unrelated processes. High R values (>0.8) and positive K indices close to the centres of the caverns highlighted the strong correlation between the volumes of gas injected/withdrawn and measured surface displacements. T values quantified the temporal lag in the surface response. The integrated methodology demonstrates the operational value of InSAR for UGS monitoring, offering insights into both the static subsidence regime and the dynamic seasonal behaviour of salt caverns. These results provide operators with a robust basis for optimising injection and withdrawal strategies, mitigating geomechanical risks, and extending the operational lifetime of storage assets. From a regulatory perspective, the proposed framework supports the adoption of standardised monitoring best practices, enabling proactive risk management and guaranteeing adherence to environmental safety standards. In addition, the proposed approach can be adapted to other geological contexts, including depleted reservoirs, aquifers, and emerging applications such as Carbon Capture and Storage (CCS) and Underground Hydrogen Storage (UHS). In conclusion, this research demonstrates the potential of InSAR as a primary monitoring tool for UGS activities. The study establishes a reproducible, scalable and cost-effective monitoring framework that integrates multi-temporal satellite data, automated threshold-based mapping and cross-correlation analyses.
How to cite: Fibbi, G., Montalti, R., Del Soldato, M., Cespa, S., Ferretti, A., and Fanti, R.: Satellite InSAR Data for Monitoring Ground Displacement in Salt Cavern Underground Gas Storage Sites, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-10078, https://doi.org/10.5194/egusphere-egu26-10078, 2026.
Interferometric Synthetic Aperture Radar (InSAR) enables surface deformation monitoring over large areas with high spatial and temporal resolution, in addition to traditional long-term in-situ monitoring methods such as inclinometers. Conventional Persistent Scatterer InSAR (PS-InSAR) techniques rely on phase-stable point targets and provide reliable results under low deformation rates and high coherence. However, the use of this method is constrained in environments characterized by rapid deformation or vegetation cover. In such environments, the Small Baseline Subset (SBAS) InSAR approach offers a more robust alternative by utilizing interferogram pairs with small spatial and temporal baselines, allowing distributed scatterers to be included in time-series deformation analysis. The main objective of this study is to evaluate surface deformation dynamics in active urban landslide areas through joint InSAR analysis by integrating both SBAS and InSAR results with in-situ inclinometer measurements and to investigate the temporal relationship between deformation behavior and rainfall conditions. The study area is located in the Büyükçekmece region, in the south-western part of Istanbul, Türkiye, where ongoing landslide activity affects a densely urbanized environment and is characterized by predominantly south-westward movement patterns. Historical inclinometer data acquired between 2014 and 2016 were used to characterize subsurface deformation. These measurements were analyzed with SBAS-InSAR LOS (line-of-sight) displacement time series derived from Sentinel-1 descending-orbit data acquired during the same period. Open-source MintPy software within the ASF OpenSARLab was used as a virtual computing environment to process InSAR data and time-series analyses. Low-coherence pixels were excluded, and standard atmospheric phase corrections were applied. Deformation velocities were analyzed by considering their non-linear temporal behavior, evaluating both average velocity patterns and temporal changes in deformation rates. Rainfall data from nearby meteorological stations were incorporated to assess the correlation between precipitation events and deformation acceleration. The novelty of this study lies in the preference for the SBAS-InSAR approach over PS-InSAR; this choice was driven by the rapid deformation characteristics of the landslide area, where the high phase stability required for PS-InSAR is often compromised. Although InSAR does not provide direct subsurface deformation data, the results demonstrate that when integrated with inclinometer measurements, SBAS-InSAR time-series analysis offers a reliable and efficient framework for continuous surface deformation assessment in active urban landslide environments.
How to cite: Kılıç, E., Doğu, M. M., Mutlu, B., Korkut, M., Zengin, E., and Ündül, Ö.: Integration of InSAR and Inclinometer Measurements for Evaluating Surface Deformation in an Active Landslide Area , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-14151, https://doi.org/10.5194/egusphere-egu26-14151, 2026.
Interferometric Synthetic Aperture Radar (InSAR) provides millimeter-scale measurements of line-of-sight (LOS) surface displacement, enabling detailed investigation of tectonic and volcanic processes [1]. However, interferograms are frequently contaminated by atmospheric delays and other noise sources whose amplitudes can be comparable to the deformation signal, especially over regions with complex topography [2]. Effective noise mitigation is therefore essential for extracting reliable geophysical information.
We developed a supervised deep-learning framework based on a modified Denoising Convolutional Neural Network (DnCNN) [3], with residual learning [4], designed to learn and remove atmospheric noise embedded in unwrapped interferograms automatically. The model was trained to estimate noise components directly and subtract them from the original interferograms, avoiding explicit physical modeling of atmospheric effects.
To evaluate performance, we applied the model to two ALOS-2 PALSAR-2 datasets: an ascending track (path/frame 126-710, 8 images) and a descending track (20-2890, 15 images) spanning 2014–2017. After baseline filtering (720 days, 150 m), 18 and 59 interferograms were generated. Linear correction [5], Generic Atmospheric Correction Online Service for InSAR (GACOS) [6], and the deep-learning (DL) method were applied, followed by conversion to 8-bit (uint8) format to standardize contrast for comparison. For ascending interferograms, the DL method produced the lowest mean standard deviation (SD = 11.59), outperforming GACOS (15.49), linear correction (16.78), and uncorrected results (15.89). Similar improvements were observed for descending interferograms (DL: 14.37; GACOS: 14.46; linear: 16.53; uncorrected: 16.45).
These results demonstrate that the proposed deep-learning approach effectively mitigates atmospheric noise in InSAR unwrapped phase maps and can outperform conventional correction methods.
References:
[1] Bürgmann, Roland, Paul A. Rosen, and Eric J. Fielding. "Synthetic aperture radar interferometry to measure Earth’s surface topography and its deformation." Annual review of earth and planetary sciences, 2000.
[2] Chaussard E, Wdowinski S, Cabral-Cano E, et al. Land subsidence in central Mexico detected by ALOS InSAR time-series. Remote sensing of environment, 2014.
[3] Zhang K, Zuo W, Chen Y, et al. Beyond a gaussian denoiser: Residual learning of deep cnn for image denoising. IEEE transactions on image processing, 2017.
[4] He, Kaiming, et al. "Deep residual learning for image recognition." Proceedings of the IEEE conference on computer vision and pattern recognition, 2016.
[5] Takada Y, Sagiya T, Nishimura T. Interseismic crustal deformation in and around the Atotsugawa fault system, central Japan, detected by InSAR and GNSS. Earth, planets and space, 2018.
[6] Yu C, Li Z, Penna N T. Interferometric synthetic aperture radar atmospheric correction using a GPS-based iterative tropospheric decomposition model. Remote Sensing of Environment, 2018.
How to cite: Li, Y. and Sagiya, T.: Deep Learning–Based Mitigation of Atmospheric Noise in InSAR Unwrapped Phase Maps, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-16679, https://doi.org/10.5194/egusphere-egu26-16679, 2026.
The Cimandiri Fault Zone (CFZ) is an active 100 km long fault system in western Java, Indonesia. Its location along the Indo-Australian–Eurasian plate boundary and proximity to densely populated areas make it a major seismic hazard. We present an integrated reassessment of the CFZ’s structure, seismicity, and crustal deformation to address unresolved questions about its geometry. Our study integrates a comprehensive review of past work, new geological field mapping, analysis of local seismicity, and geodetic observations from ongoing GNSS campaigns, complemented by an ongoing LiCSBAS InSAR time-series analysis, to better constrain the fault’s characteristics.
Preliminary results indicate the CFZ’s structural configuration is more complex than previously assumed. Although historically identified as predominantly sinistral (left-lateral) strike-slip, the fault actually comprises multiple segments with oblique and reverse-slip components. Recorded seismicity (e.g.1982 M5.5 and 2000 M5.4 earthquakes) confirms the CFZ’s activity and underscores its capacity to generate damaging earthquakes.
Previous GNSS-based studies have reported regional horizontal deformation on the order of 1–2 cm/year across western Java, with inferred slip rates of 4–5 mm/year along segments of the Cimandiri Fault Zone, indicating active strike-slip deformation and strain accumulation. We integrating geological, seismic, and geodetic insights and refining the CFZ’s segmented fault model and slip estimates that offer an improved basis for seismic hazard assessment and disaster risk reduction in West Java, ultimately enhancing regional resilience. Ongoing InSAR analysis will further give supporting results for the interseismic strain distribution along the CFZ and provide a forward look at evolving deformation patterns. The multidisciplinary approach yields new insights into the behavior of this active fault that will highlight how the combination of structural, seismological, and geodetic data enhances understanding of seismic hazards in complex tectonic settings.
How to cite: Shomim, A. F., Aribowo, S., Hanifa, N. R., Gunawan, E., Ratna, P. N., Ou, Q., Hidayat, E., Muttaqy, F., and Ramadhani, N.: The Cimandiri Fault Zone West Java (Indonesia) Revisited - An Integrated Structural, Seismic, Geodetic and InSAR-Derived Deformation Reassessment, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19563, https://doi.org/10.5194/egusphere-egu26-19563, 2026.
The Mekong Delta is highly exposed to land subsidence (i.e. negative Vertical Land Motion, VLM), which increases flood risk, groundwater and soil salinization, land degradation, infrastructure damage, and accelerates relative sea-level rise. InSAR time-series have been widely used to estimate VLM in the delta. However, most studies used short investigation periods of up to 4 years and reported VLM derived from a single viewing geometry by projecting line-of-sight velocities using the incidence angle, which assumes horizontal motion is negligible. Such results are useful, however they can be uncertain where horizontal motion is non-zero and where time-series are noisy or non-linear.
We overcome the limitations of previous InSAR studies in the Mekong Delta by processing a Sentinel-1 time series with SBAS InSAR for an extended observation period of 9 years. The descending dataset covers 27th February 2015 to 18th December 2023 (12-day sampling) and provides 942,978 coherent points, while the ascending dataset spans from 13th March 2017 to 31st December 2023 and provides 511,972 coherent points. We then apply an ascending-descending (dual-geometry) decomposition to separate VLM and an east-west horizontal component for 190,533 SBAS points, limited to locations coherent in both geometries.
Both tracks show widespread subsidence in the delta with strong spatial variability and local hot spots. In single-geometry results, maximum VLM reaches about -9.5 cm/yr, with mean rates of about -3.3 cm/yr (descending) and -3.6 cm/yr (ascending). The dual-geometry decomposition yields a consistent VLM field with maximum VLM of about -8.4 cm/yr and a mean of -3.2 cm/yr (linear trend), while the east-west component ranges from -3.3 cm/yr (westward) to +2.9 cm/yr (eastward). The lower maximum values in the decomposed solution are expected, because decomposition is only possible for the reduced set of points available in both tracks. VLM is not uniform in time: time series show local accelerations and slowdowns. VLM patterns also differ between land-use types (e.g., urban areas, rice fields, mangroves, aquaculture), suggesting that drivers vary across the delta.
These InSAR-derived land motion estimates should be interpreted with care, because InSAR captures the combined surface response to multiple mechanisms of both natural and human-induced origin. Still, long-term VLM and horizontal motion maps can help compare displacement patterns, support interpretation of potential drivers, and provide a check for subsidence scenarios used in planning, adaptation and mitigation. Robust use of the results requires integration with independent information, including aquifer-compaction modelling outputs and reliable, locally corrected elevation data for relative sea-level rise studies.
How to cite: Walczak, S., Guzy, A., Witkowski, W., Łucka, M., Teatini, P., Baldan, S., Seeger, K., and Minderhoud, P.: Vertical and horizontal land motion in the Mekong Delta: A long-term record from dual-geometry Sentinel-1 InSAR, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20136, https://doi.org/10.5194/egusphere-egu26-20136, 2026.
Local land subsidence (LLS) is caused by groundwater overextraction, long-term oil and gas extraction without pressure maintenance, and the natural compaction of sediments due to self-weight. Among these factors, the groundwater-induced LLS has intensified markedly in urbanized landscapes and groundwater-irrigated agricultural regions worldwide. Although LLS contributes to vertical land motion, its spatial extent, temporal persistence, and magnitude are generally smaller than those associated with regional tectonics or glacial isostatic adjustment. Consequently, global concern regarding groundwater-driven LLS has remained limited, particularly for non-coastal cities where subsidence does not directly exacerbate local relative sea-level rise. However, LLS increasingly threatens the long-term integrity of aquifer systems, urban infrastructure, and the sustainability of cities dependent on alluvial groundwater resources. In Indian metropolitan regions, LLS has been discussed over the past two decades, yet interpretations have largely been confined to InSAR-derived displacement velocities and their correlations with groundwater-level fluctuations. Here, we reevaluate these prevailing assumptions in Ahmedabad, the economically vibrant city of the western Indian state of Gujarat, by integrating stratigraphic, hydrogeologic, geodetic, geochemical, and demographic datasets. We combine eight years of InSAR observations with three years of continuous GPS measurements to characterize the spatial and temporal evolution of subsidence across the city. Our results show persistent subsidence footprints in the southwest sector of Ahmedabad, coinciding with a major industrial hub, and in the western outskirts, which have undergone rapid residential development since 2017. Subsidence initiated in the southwest, corresponding to the historic urban core, whereas the maximum subsidence rate, reaching 2.7 cm/year, occurs in the western peripheral zone of Bopal. Time-series analysis of InSAR-derived displacements reveals a superposition of inelastic, elastic, and uplift components, indicating that subsidence is nearly irreversible in some sectors while substantial recovery is observed elsewhere. Contrary to conventional interpretations, no direct relationship is identified between land displacement and groundwater-level fluctuations in Ahmedabad. Instead, a strong positive relationship emerges between subsidence-uplift patterns and the proportions of clay and sand in local lithofacies. In May 2004, which perhaps marks the pre-consolidation head, the groundwater levels show dominant recovery trends within the confined-1 and confined-2 aquifer systems, accompanied by seasonal variability. Recovery in shallow aquifers could be due to severe groundwater pollution associated with textile industries in the Vatva and Lambha localities, rendering these waters unsuitable for consumption. The present study develops a numerical model that calibrates delayed clay compaction relative to pre-consolidation head, skeletal storage coefficients, and the number of compacting aquitards. This framework is transferable to alluvial aquifer systems globally, enabling improved assessment of residual compaction and recharge dynamics beyond traditional interpretations in the Indian subcontinent.
How to cite: Karunakalage, A., Sharma, R., Jaboyedoff, M., Daqiq, M. T., Derron, M.-H., and Nayak, R.: Unveiling the Impact of Groundwater Extraction on Local Land Subsidence and Calibrating a Model for Understanding the Mechanics of Mixed Subsidence and Uplift Phenomena in Gujarat, India, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18988, https://doi.org/10.5194/egusphere-egu26-18988, 2026.
Posters virtual: Mon, 4 May, 14:00–18:00 | vPoster spot 3
EGU26-16451 | Posters virtual | VPS12
Near-Decadal Land Subsidence Susceptibility and Trends Using Physics-Informed LSTMMon, 04 May, 14:12–14:15 (CEST) vPoster spot 3
Land subsidence poses growing risks to urban infrastructure, water resources, and long-term resilience, requiring assessment frameworks that link present-day observations with planning-relevant forecasts. This study develops an integrated approach for land subsidence susceptibility mapping and trend forecasting over multi-year horizons. The analysis uses SBAS-InSAR deformation time series derived from Sentinel-1 observations from 2017 to 2025 to characterize subsidence patterns across East Baton Rouge Parish, Louisiana. Subsidence susceptibility is modeled using an ensemble machine-learning framework that combines Extra Trees and Random Forest regressors and incorporates geological, topographic, hydrological, land use, infrastructure, and climatic conditioning factors. The susceptibility results highlight the dominant influence of land use, elevation, proximity to faults and rivers, and terrain-hydrology interactions on subsidence patterns. To extend assessment beyond observation periods, a physics-informed long short-term memory (LSTM) ensemble is introduced for forecasting. The model integrates data-driven learning with physically motivated constraints to ensure stable and realistic deformation trajectories. The forecasts preserve observed spatial patterns while exhibiting physically consistent temporal evolution and quantified uncertainty. The results demonstrate that combining InSAR observations with physics-informed deep learning enables robust, planning-scale subsidence assessment and forecasting. The proposed framework is transferable to other urban settings where long-term subsidence poses increasing societal risk.
How to cite: Kangah, D. and Abdalla, A.: Near-Decadal Land Subsidence Susceptibility and Trends Using Physics-Informed LSTM, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-16451, https://doi.org/10.5194/egusphere-egu26-16451, 2026.
