The Copernicus European Ground Motion Service (EGMS) provides millimetre-accuracy line-of-sight displacement measurements for over five billion coherent scatterers across Europe, derived from Sentinel-1 SAR interferometry over the period 2015--2023. Despite the unprecedented spatial coverage and measurement density of this dataset, no open-source integrated toolchain exists for processing, classifying, clustering, and interactively visualising EGMS data at the regional to continental scale without resorting to subsampling or proprietary software. In this paper, we introduce EGMSpy, a fully open-source Python pipeline that ingests raw EGMS L2b CSV tiles, compresses them into a federated split-GeoParquet database (metadata and time series stored separately, linked by a unique point identifier), applies a physics-informed hybrid rule/GMM time series classifier to assign each point to a geophysical deformation class, groups spatially and dynamically coherent deforming areas using a velocity-weighted three-dimensional DBSCAN algorithm, and serves the entire dataset to an interactive Leaflet.js web viewer via a Flask/DuckDB REST bridge capable of sub-100\,ms viewport queries over hundreds of millions of points. The pipeline is demonstrated on the northern Italy EGMS L2b dataset (2019--2023), comprising 434 tile pairs (approx. 300 million coherent scatterers, 104GB on disk), processed on a single workstation. Internal quality control confirms that the toolkit's OLS-derived velocity features are consistent with the official EGMS pre-computed velocities (RMSE=0.096mm\yr^-1, bias=-0.001mmyr^-1, r$=0.9993, n$=99,751). All code is released under the MIT licence and is available at https://github.com/fcatani/EGMSpy (repository in preparation).

Deltas are particularly susceptible to extensive land subsidence and consequently face heightened flood risks. Interferometric synthetic aperture radar (InSAR) techniques allow for multidecadal detection of land surface displacements with high accuracy, demonstrating their power in revealing dynamics of land subsidence in delta areas. However, achieving a comprehensive understanding of the spatial and temporal patterns of the measured land displacements in these areas and assessing their potential impacts remain a challenge due to the superposition of various drivers of land movement and the high compressibility of the deposits forming these landforms. In this study, taking the Po Delta (Italy) as a case study, we analyze the spatiotemporal trend, seasonality, and abrupt changes in InSAR time series from the European Ground Motion Service, derived from Sentinel-1 images spanning 2018-2023. Particular attention has been given to the displacements affecting the extensive levee system that shields the territory from flooding. Temporally, a regression model is applied to the InSAR time series to track their trend and seasonality, while the Bayesian estimator of abrupt change, seasonality, and trend algorithm is used to detect their abrupt changes. Spatially, we propose a robust estimation method for modeling spatial distribution of land subsidence associated with Holocene sediment ages. Four models, i.e., power, logarithmic, hyperbolic, and exponential functions, have been tested. The results reveal that, in the temporal domain, the behavior of the time series can be classified into trend dominated, seasonal dominated, and irregular patterns. Seasonality mainly characterizes the deformation of bridges and transmission towers and shows either a positive or negative correlation with the land surface temperature. Abrupt changes in the time series can be effectively detected, thereby enhancing alerts of levee failures. In the spatial domain, anomalous subsidence, which typically results from recent anthropogenic activities, can be identified by comparing the observed and modeled subsidence. The proposed framework improves the organization and interpretation of InSAR time series and thus provides comprehensive spatiotemporal analysis and direct detection of anomalies, helping identify critical areas and plan appropriate mitigation measures.

Coastal alluvial plains underlain by unconsolidated deposits are prone to land subsidence, a geohazard that can damage infrastructure and alter drainage patterns. One such example is the Venetian–Friulian coastal plain (NE Italy), where natural sediment compaction and anthropogenic activities have led to ground deformation across multiple zones. From this perspective, this study presents a 30-year analysis of land subsidence across the Venetian–Friulian plain, particularly highlighting municipalities such as Portogruaro, Concordia Sagittaria, San Stino di Livenza, Eraclea, and Caorle. The dataset comprises multi-source SAR data from ERS, Envisat, COSMO-SkyMed (CSK), Sentinel-1, and the European Ground Motion Service (EGMS), covering the period from 1992 to 2021. The study integrates multi-platform SAR observations with ADAFinder-based extraction of Active Deformation Areas (ADAs), data quality evaluation using the Quality Index (QI), building-scale analysis based on LOS-derived vertical displacement time series, and orthophotos to confirm the building’s presence and evolution. By using the adopted extraction thresholds, a total of 57, 16, 83, 33, and 72 ADAs were identified from the ERS, ENVISAT, COSMO-SkyMed, Sentinel-1, and EGMS datasets, respectively. The result suggests that the strongest deformation occurred during the earlier observation periods in Zones 1 to 3, then progressively stabilized, whereas some parts of Zone 4 remained active and showed renewed deformation during the later periods. The research highlights the importance of conducting long-term analysis using multi-platform interferometric datasets to refine and personalize outcomes in geohazard monitoring. The findings from this research offer invaluable insights into the ongoing surveillance of geohazards, which are progressively related to urban development and planning.

Landslides are widespread geomorphic processes in mountainous regions, whose occurrence varies over time in response to both conditioning and triggering factors. Nevertheless, the majority of spatial prediction (susceptibility) models remain static and do not explicitly account for the spatiotemporal evolution of the factors governing slope stability. To overcome this limitation, we developed a multi-temporal landslide inventory based on very high-resolution (50 cm) orthophotos and used it to produce dynamic landslide susceptibility maps that capture temporal changes in terrain instability. The study focuses on the Belluno area (northeastern Italian Alps) and applies a spatiotemporal Generalized Additive Model (GAM) to analyze non-linear relationships between landslide occurrence and key environmental variables, including slope, aspect, curvature, elevation, and cumulative rainfall calculated over multiple temporal windows. Landslide occurrences derived from historical records and image interpretation were used as the dependent variable, while predictor variables were obtained from DEM-based terrain analysis and rainfall data. The GAM framework allows flexible modeling of non-linear effects and temporal dependencies, particularly those associated with rainfall-triggered landslides. Results indicate that slope steepness and short- to mid-term cumulative rainfall exert the strongest control on landslide susceptibility in the study area. Model performance evaluation using ROC–AUC metrics shows that the spatial–temporal GAM outperforms conventional cross-sectional susceptibility models. The proposed approach highlights the importance of incorporating temporal dynamics into landslide susceptibility assessment and provides a refined framework for spatially explicit evolution of evolving slope instability.

Precise assessment of landslide potential in tectonically active mountain areas like Darjeeling Sikkim Himalaya (DSH) is a scientific challenge due to the complexity of different landslide conditioning factors that control the slope stability. Despite several studies for landslide susceptibility mapping, most of the conventional methods struggle to capture the nonlinear relationships and spatial heterogeneity that characterize landslides. Besides, the current use of pixel-based methods is insufficient to depict geomorphological units and slope-scale processes, thus limiting their effectiveness in boundary demarcation of landslide-prone areas. These limitations highlight the need for more robust machine learning frameworks that integrate geomorphology-based terrain segmentation with advanced machine learning models, which would not only facilitate modeling the multifaceted interactions among environmental components but also improve the understanding of the landslide driving forces. In this study, we have used slope unit based landslide susceptibility mapping with 4380 slope units integrated with 17 conditioning factors, and 8373 total updated inventories using six models Random Forest (RF), Generalized Additive Model (GAM), Categorical Boosting (CatBoost), Tabular Neural Network (TabNet), Bayesian Additive Regression Trees (BART), and Convolutional Neural Network (CNN). The model hyperparameters were optimized using Bayesian optimization, except for the BART model. Among the six models, RF (AUC = 0.848) and CatBoost (AUC = 0.846) were the best two performing models. Furthermore, SHAP analysis reveals that elevation, aspect, slope, distance to faults, NDVI, and proximity to roads and drainage networks are the main landslide controlling factors in DSH. The interaction analysis using SHAP indicates that the occurrence of landslides is controlled by nonlinear and threshold-dependent relations, especially among slope-rainfall, rainfall-soil moisture, and slope-distance to roads and faults, which represents a complex interaction between the hydrological triggering factor, geomorphic processes, tectonic activity, and human interventions.

Classifying ground deformation processes, such as landslides, subsidence, deep-seated gravitational slope deformations (DSGSDs), and mining-induced deformations, is key for large-scale hazard assessment and national land-use management. Earth observation provides heterogeneous data over the same geographic region, including interferometric synthetic aperture radar (InSAR) time series (TS), multispectral imagery, and terrain products. However, effectively integrating such spatiotemporal information from multimodal datasets remains a major challenge. To fully use the rich information contained in the TS and to exploit the complementary strengths of spatial and temporal data, we propose a dual-branch deep learning (DL) approach that integrates InSAR ground deformation TS with geospatial information for classifying slow-moving ground deformation processes. To validate the approach, we construct a ground deformation dataset containing over 26 000 active deformation areas (ADAs), labeled into four deformation types: landslide, subsidence, DSGSD, and mining. Results demonstrate that our model achieves an overall classification accuracy exceeding 90% on both ascending and descending test datasets, though confusion remains between certain classes, such as landslides and DSGSD. Explainable AI (XAI) analysis indicates that spatial and morphological features contribute more significantly to classification performance than temporal deformation patterns, with clearer distinctions for subsidence and mining, but more overlap between landslides and DSGSDs. This work highlights the strength of multimodal data fusion method to classify ground deformation processes, while setting the stage for future research.

Ground deformations, such as landslides, subsidence, and mining-related deformations, pose significant risks to communities and infrastructure. Accurate classification of these deformations is crucial for hazard management and land use planning. Existing classification methods primarily rely on thresholding or traditional machine learning models, failing to fully capture the rich temporal and spatial information available from spaceborne remote sensing data.This study proposes a deep learning method that integrates both ground motion time series (European Ground Motion Service - EGMS) and geospatial data (spaceborne optical imagery, and morphological features) to classify ground motions. The method employs a dual-branch model, where 1D CNNs extract temporal features from ground motion time series, and 2D CNNs capture spatial characteristics from corresponding satellite imagery and topographic data. The features extracted by both branches are fused and fed to a multilayer perceptron to classify deformation processes, i.e., landslides, deep-seated gravitational slope deformations (DSGSD), subsidence, and mining-related deformations. To inform the model, we used a dataset over 26,000 Active Deformation Areas (ADAs), defined with the ADA finder tool(Navarro et al., 2020). We annotate each ADA by crossing it with existing inventories such as the Italian Landslide Inventory (IFFI) and CORINE Land Cover map. Corresponding time series data and imagery were subsequently extracted for each and fed to the model. Results, using cross-validation, show that the model achieves an overall accuracy of over 90%. This demonstrates its effectiveness and robustness in handling diverse deformation types. We finally deployed the validated model and classify all the ADAs generated for the entire Italy.This research provides a scalable and automated framework for ground motion classification, and the classification achieved can lead to better-targeted risk mitigation strategies, and improved ground motion forecasting and early warning systems.References: Navarro, J. A., Tomás, R., Barra, A., Pagán, J. I., Reyes-Carmona, C., Solari, L., Vinielles, J. L., Falco, S., & Crosetto, M. (2020). ADAtools: Automatic Detection and Classification of Active Deformation Areas from PSI Displacement Maps. ISPRS International Journal of Geo-Information, 9(10), 584. https://doi.org/10.3390/ijgi9100584

Within the framework of the RETURN (multi-Risk sciEnce for resilienT commUnities undeR a changiNg climate) project funded by the Italian National Plan for Recovery and Resilience (PNRR), the Spoke2 “Ground Instabilities” aims at restituting temporal and space distribution of effects induced by ground instabilities (scenarios), i.e. landslides, sinkholes, subsidence and liquefaction. Preliminary results obtained for sinkholes are here reported, as these processes represent the most diffused geological hazard in karst areas. These are widespread in Italy in all areas characterized by outcrops of soluble rocks. Sinkholes are being analyzed in the project in terms of the predisposing factors as well as of preparatory and triggering processes. The distinction among them is made on a temporal basis: predisposing factors are considered invariable at the observation scale, while the preparatory factors show trends or periodic variations in the same time interval. As a consequence, a trigger is considered as a process which acts over a very short and well-defined time. Sinkholes are distinguished also on the basis of their kinematics, following the main internationally accepted sinkhole classification. The activities carried out in the first year of RETURN project brought us to collect a number of learning cases dealing with sinkholes in Italy, and to deduce rationales of process to be considered as tools for operative chains in view of the implementation of a scenario producer to be designed through an IT-technology. The outputs expected in the simulated scenarios are quantitative values attributed to ground instability effects (among which the sinkholes) in a multi-hazard perspective, with the final goal to offer multi-risk mitigation strategies to national stakeholders.

Ground deformation, encompassing sudden and gradual shifts in the Earth's surface, poses significant global geohazard risks. These phenomena demand thorough investigation and monitoring and are influenced by a range of natural and anthropogenic factors such as mining, excessive groundwater extraction, seismic activities, structural loads, and subsurface geology. Our research is centered on the location in the Venetian-Friulian Plain (Veneto Region, NE Italy). This area is of interest because it represents a transitional zone where sedimentary deposits from both river systems (fluvial) and lagoon/coastal environments are found, marking the transition from the alluvial plain to the coastal plain. Ground displacement maps are generated using pre-event data from the Veneto Region Sentinel 1-PS data Service and the European Ground Motion Service (EGMS), allowing us to analyze the heightened susceptibility of areas undergoing deformation. Our approach integrates artificial intelligence techniques with InSAR-derived data to create comprehensive pre- and post-event multi-temporal deformation inventories and susceptibility maps. This fusion offers exceptional accuracy and timeliness in identifying, modeling, and predicting ground deformation events. Utilizing insights from InSAR data and AI techniques, we aim to project future trends and potential risks, contributing valuable insights to geohazard assessment and management within the study region.

In the past three decades, the city of Addis Ababa, a capital city of Africa, has grown significantly in population, facilities, and infrastructure. The area involved in the recent urbanization is prone to slow natural subsidence phenomena that can be accelerated due to anthropogenic factors such as groundwater overexploitation and loading of unconsolidated soils. The main aim of this study is to identify and monitor the areas most affected by subsidence in a context, such as that of many areas of emerging countries, characterized by the lack of geological and technical data. In these contexts, advanced remote sensing techniques can support the assessment of spatial and temporal patterns of ground instability phenomena, providing critical information on potential conditioning and triggering factors. In the case of subsidence, these factors may have a natural or anthropogenic origin or result from a combination of both. The increasing availability of SAR data acquired by the Sentinel-1 mission around the world and the refinement of processing techniques that have taken place in recent years allow one to identify and monitor the critical conditions deriving from the impressive recent expansion of megacities such as Addis Ababa. In this work, the Sentinel-1 SAR images from Oct 2014 to Jan 2021 were processed through the PS-InSAR technique, which allows us to estimate the deformations of the Earth’s surface with high precision, especially in urbanized areas. The obtained deformation velocity maps and displacement time series have been validated using accurate second-order geodetic control points and compared with the recent urbanization of the territory. The results demonstrate the presence of areas affected by a vertical rate of displacement of up to 21 mm/year and a maximum displacement of about 13.50 cm. These areas correspond to sectors that are most predisposed to subsidence phenomena due to the presence of recent alluvial deposits and have suffered greater anthropic pressure through the construction of new buildings and the exploitation of groundwater. Satellite interferometry techniques are confirmed to be a reliable tool for monitoring potentially dangerous geological processes, and in the case examined in this work, they represent the only way to verify the urbanized areas exposed to the risk of damage with great effectiveness and low cost, providing local authorities with crucial information on the priorities of intervention.

Mapping landslide-depleted source areas is pivotal for refining predictive models and volume estimations, yet these critical regions are often conflated with the landslide runouts, leading to sub-optimal assessments. The source areas are typically the regions where the actual failure occurs, providing crucial information on the initiation mechanisms and the nature of landslide propagation. Catering to this objective, we built a method based on a landslide’s topology and morphological information. We develop and test this method in geomorphologically distinct regions such as Dominica, Turkey, Italy, Nepal, and Japan (Niigata) to showcase the model’s robust adaptive capacity. The model can demarcate the source and runout zones from landslide planforms found in inventories with accuracy deviations under 15–20%. While distinguishing landslide source and runout areas, the model also considers triggering information and movement types. We also deploy the model in Chile, Japan (Hokkaido), Colombia, Papua New Guinea, and China. In those new regions, we found the mean area of the scarp to be consistently under 30% of the total landslide area. We additionally showcased the application of our model to the area-volume scaling of the coseismic landslides triggered by the 2018 Hokkaido Eastern Iburi Earthquake (MW 6.6) in Japan. Our analysis revealed that area-volume fitting using the landslide source areas instead of the total landslide planforms or polygons improves the linear fit from R2=0.49 to R2=0.81. The model could improve diverse landslide analysis, such as hazard models, and facilitate a deeper understanding of landslide behaviour.

Understanding landslide failure processes is pertinent to predict and minimize the effects of landslides. A variety of elements, such as geology, topography, and soil conditions, can lead to slope failures triggered via natural causes e.g., rainfall and earthquakes, setting off the failure movements. Proper geotechnical analysis requires knowledge of both the triggering event and the subsequent movement patterns of the landslide. This information is vital for accurately predicting when and where landslides might occur. To integrate this information into existing landslide inventories, we introduce Landsifier 2.0, a tool designed to meet the needs of the landslide research community. This Python-based library allows seamless usage of machine learning models to extract information regarding landslide triggers and failure movements solely based on inventories of landslides. Powered by topology, a high-dimensional feature extraction module encapsulated within our library, information accessed via a landslide's shapes and configurations allows the identification of triggers (e.g., earthquake-and rainfall-triggered landslides) and failure movements (e.g., rotational slides, translational slides, debris flows, rock falls) of undocumented landslide inventories through continuous remote sensing missions. We showcase the library’s application in diverse geomorphological and climatic settings e.g., South-western China, Denmark, Turkey, Japan, Italy and more. We anticipate that Landsifier 2.0 will be particularly useful in the predictive modelling domain (including susceptibility and hazard modelling) of landslide studies, where precise information about triggers and failure dynamics is essential for developing reliable predictive models. References:Rana, Kamal, Uğur Öztürk, and Nishant Malik. 2021. “Landslide Geometry Reveals Its Trigger.” Geophysical Research Letters 48(4). doi: 10.1029/2020gl090848.Rana, Kamal, Nishant Malik, and Uğur Öztürk. 2022. “Landsifier v1.0: A Python Library to Estimate Likely Triggers of Mapped Landslides.” Natural Hazards and Earth System Sciences 22(11):3751–64. doi: 10.5194/nhess-22-3751-2022.Rana, Kamal, Kushanav Bhuyan, Joaquin Vicente Ferrer, Fabrice Cotton, Uğur Öztürk, Filippo Catani, and Nishant Malik. 2023. “Landslide Topology Uncovers Failure Movements.” arXiv (Cornell University). doi: 10.48550/arxiv.2310.09631.

Ground-based global navigation satellite system (GNSS) and remote sensing interferometric synthetic aperture radar (InSAR) techniques have proven to be very useful for deformation monitoring. GNSS provides high-precision data but only at a limited number of points, whereas InSAR allows for a much denser distribution of measurement points, though only in areas with high and consistent signal backscattering. This study aims to integrate these two techniques to overcome their respective limitations and explore their potential for effective monitoring of critical infrastructure, ensuring the protection of people and the environment. The proposed approach was applied to monitor deformations of the shoulder structures of the MOSE (MOdulo Sperimentale Elettromeccanico) system, the civil infrastructure designed to protect Venice and its lagoon from high tides. GNSS data were collected from 36 continuous GNSS (CGNSS) stations located at the corners of the emerged shoulder structures in the Treporti, San Nicolò, Malamocco, and Chioggia barriers. Velocities from February 2021/November 2022 to June 2023 were obtained using daily RINEX data and Bernese software. Three different processing strategies were applied, utilizing networks composed of the 36 MOSE stations and eight other continuous GNSS stations from the surrounding area (Padova, Venezia, Treviso, San Donà, Rovigo, Taglio di Po, Porto Garibaldi, and Porec). InSAR data were sourced from the European ground motion service (EGMS) of the Copernicus program and the Veneto Region database. Both services provide open data related to the line of sight (LOS) velocities derived from Sentinel-1 satellite imagery using the persistent scatterers interferometric synthetic aperture radar (PS-InSAR) approach. InSAR velocities were calibrated using a reference CGNSS station (Venezia) and validated with the available CGNSS data from the external network. Subsequently, the velocities were compared along the LOS at the 36 CGNSS stations of the MOSE system. The results showed a strong agreement between the velocities, with approximately 70% of the comparisons displaying differences of less than 1.5 mm/year. These findings highlight the great potential of satellite-based monitoring and the effectiveness of combining GNSS and InSAR techniques for infrastructure deformation analysis.

Landslides are complex and dynamic natural hazards that require a comprehensive understanding of their temporal changes for effective assessment and management. Traditional landslide inventories often focus on static analysis, providing a snapshot of landslide occurrences at a specific point in time. However, to capture the dynamic nature of landslides and assess their evolution over time, multi-temporal inventories are essential. This study aims to go beyond static analysis by proposing the use of multi-temporal inventories for dynamic landslide assessment. The approach involves the integration of remote sensing data, advanced modeling techniques, and deep learning algorithms to analyze and map landslides over multiple time periods. By considering the temporal dimension, the proposed method enables the identification of changes in landslide patterns, movements, and susceptibility over time. We used orthophotos retrieved from WMS and WMTS services provided by the Italian national portal, covering the period from 1989 to 2021, for a study conducted in the Cordevole and Alpago areas (Belluno province, NE Italian Alps). These areas were impacted by two extreme meteorological events (return period > 100 years) in 2018 (October 27th–30th) and 2020 (December 4th–6th). The first, known as windstorm VAIA, has induced severe damage to the forest cover. The generated multi-temporal inventories provide valuable information for understanding the temporal dynamics of landslides, which is crucial for accurate landslide hazard assessment and risk management. The findings of this study highlight the importance of incorporating multi-temporal inventories into landslide assessment methodologies to enhance our understanding of landslide behavior and improve decision-making processes. Acknowledgement: This study was carried out within the PNRR research activities of the consortium iNEST (Interconnected North-Est Innovation Ecosystem) funded by the European Union Next-Generation EU (Piano Nazionale diRipresa e Resilienza (PNRR) – Missione 4 Componente 2, Investimento 1.5 – D.D. 1058 23/06/2022, ECS_00000043). This manuscript reflects only the Authors’ views and opinions; neither the European Union nor the European Commission can be considered responsible for them.

In alpine regions, Deep-seated Gravitational Slope Deformations (DsGSDs) pose significant risks due to their continuous deformation rates, potentially leading to sudden and accelerated transformations that can cause unpredictable damage to local communities and infrastructure. Monitoring DsGSDs is crucial for effective risk assessment and land-use planning. Advances in remote sensing technologies, particularly InSAR (Interferometric Synthetic Aperture Radar), offer substantial advantages in monitoring and studying these widespread and slow processes. The European Ground Motion Service (EGMS), which provides Europe-wide ground motion data, emerges as a viable tool for detecting, monitoring, and characterizing DsGSDs. This study aimed to develop and evaluate an automated workflow for identifying and analyzing trends in DsGSDs in alpine areas using deformation time series datasets. The approach involves utilizing advanced statistical methods to characterize DsGSD phenomena in alpine regions. Focusing on the Carnic Alps area in the northern part of the Veneto Region and Friuli-Venezia Giulia Region (NE, Italy), our objective is to explore supervised machine learning (ML) and deep learning (DL) algorithms to automatically identify DsGSD areas and analyze the spatiotemporal behavior of long time series of ground deformations. The findings will be compared with data from the Italian Landslide Inventory (IFFI), serving to not only validate the newly extracted information but also assess the potential of integrating multi-source datasets. This work sets the foundation for further analyses on how transient climatic factors could influence DsGSDs regimes.

Repeated temporal mapping of landslides is essential for investigating changes in landslide movements, legacy effects of the landslide triggering events, and susceptibility changes in the area. However, in order to perform such investigations, multi-temporal (MT) inventories of landslides are required. The traditional approach of visual interpretation from cloud-free optical remote sensing imageries is time consuming and expensive. Recent endeavors exploring Convolutional Neural Networks and deep learning models have made rapid and accurate mapping of landslides feasible but have not been applied for multi-temporal landslide mapping in the Himalayas, yet. Earlier models used a standard supervised learning approach, with a small landslide inventory over a limited area used for training, which is then utilized to predict landslides in nearby areas. We propose a new strategy, using geographically separate training samples to design a standard approach which can be utilized to create multi-temporal landslide inventories. RapidEye images of 5-meters spatial resolution are used to generate MT landslide inventories in the study area of Rasuwa district, Nepal. We test the effectiveness of the model by training with only 55 landslides and predicting for a different area. Then, using the weights attained from this first training phase, we use transfer learning to map landslides over a time period between 2013 and 2019 in the Rasuwa district attaining an average F1-score of 0.69 for the study area. We also perform a spatial comparison between the manual (observed) and predicted inventories to evaluate the differences between landslide densities and overall landslide statistics of landslide area distribution. The benefit of a transfer learning-based model training is that it circumvents the need for generating annual inventories for training a deep learning. A single event-based inventory is enough to generate landslide inventories over a number of years, at least until landslide conditioning factors do not change significantly. This application can enable automated workflows to generate MT landslide inventories of particular areas as the basis for landslide evolution and movement change analysis.

This study aims to define a preliminary geological model of the Passo Mauria tunnel, designed as an alternative underground route to the current S.S.52 to ensure a safer connection over the top way Belluno (A27 Highway) Tolmezzo (A23 Highway), on which various tourist and economic activities revolve. The survey stage has allowed us to describe and represent the cartographic elements that describe the geology of the study area. The study has mainly focused on the investigation of the presumed entrances of the tunnel, characterized by specific geological and geomorphological conditions. The aerial photogrammetric survey supported the reconstruction of the geometry of discontinuities in the non-accessible areas. Besides, this study has allowed the definition of the possible behaviour of the rock mass subjected to the tunnel excavation and of sections with homogeneous mechanical properties. The drafting of a survey plan supplements the study to determine a geological and geotechnical model, for the subsequent investigation and design phases.

Accurate early warning systems for landslides are a reliable risk-reduction strategy that may significantly reduce fatalities and economic losses. Several machine learning methods have been examined for this purpose, underlying deep learning (DL) models' remarkable prediction capabilities. The long short-term memory (LSTM) and gated recurrent unit (GRU) algorithms are the sole DL model studied in the extant comparisons. However, several other DL algorithms are suitable for time series forecasting tasks. In this paper, we assess, compare, and describe seven DL methods for forecasting future landslide displacement: multi-layer perception (MLP), LSTM, GRU, 1D convolutional neural network (1D CNN), 2xLSTM, bidirectional LSTM (bi-LSTM), and an architecture composed of 1D CNN and LSTM (Conv-LSTM). The investigation focuses on four landslides with different geographic locations, geological settings, time step dimensions, and measurement instruments. Two landslides are located in an artificial reservoir context, while the displacement of the other two is influenced just by rainfall. The results reveal that the MLP, GRU, and LSTM models can make reliable predictions in all four scenarios, while the Conv-LSTM model outperforms the others in the Baishuihe landslide, where the landslide is highly seasonal. No evident performance differences were found for landslides inside artificial reservoirs rather than outside. Furthermore, the research shows that MLP is better adapted to forecast the highest displacement peaks, while LSTM and GRU are better suited to model lower displacement peaks. We believe the findings of this research will serve as a precious aid when implementing a DL-based landslide early warning system (LEWS).

Structural Health Monitoring (SHM) represents a very powerful tool to assess the health condition of buildings. In recent years, the growing availability of high-resolution SAR satellite images has made possible the application of multi-temporal Interferometric Synthetic Aperture Radar (MT-InSAR) techniques for structural monitoring purposes, with high precision, low costs, timesaving, and the possibility to investigate wide areas. However, a comprehensive validation of the effectiveness of MT-InSAR in this application field has not been achieved yet. For this reason, in this paper a comparison between interferometric data and on-site measurement of displacements is proposed. The application case study is the Civic Tower of the city of L’Aquila (Abruzzo Region, Italy). After the seismic events that affected the area in 2009, an on-site monitoring system was installed on the tower to detect any changes in the damage pattern in the period 2010–2013. Furthermore, images acquired by COSMO-SkyMed constellation in Stripmap mode (~3 m resolution) during the same period were processed by the Permanent Scatterer-InSAR (PSI) technique to estimate the deformation of the structure and the surrounding area. The obtained results indicate that both methods are consistent in the measurement of displacement trends of the building and a slight rotation/displacement of the tower was detected. Such evidence highlights both the huge potential and the limitations of using InSAR techniques for SHM.

Multiple landslide events occur often across the world which have the potential to cause significant harm to both human life and property. Although a substantial amount of research has been conducted to address mapping of landslides using Earth observation (EO) data, several gaps and uncertainties remain with developing models to be operational at the global scale. The lack of a high-resolution globally distributed and event-diverse dataset for landslide segmentation poses a challenge in developing machine learning models that can accurately and robustly detect landslides in various regions, as the limited representation of landslide and background classes can result in poor generalization performance of the models. To address this issue, we present the High-Resolution Global landslide Detector Database (HR-GLDD), a high-resolution (HR) satellite dataset (PlanetScope, 3 m pixel resolution) for landslide mapping composed of landslide instances from 10 different physiographical regions globally in South and South-East Asia, East Asia, South America, and Central America. The dataset contains five rainfall-triggered and five earthquake-triggered multiple landslide events that occurred in varying geomorphological and topographical regions in the form of standardized image patches containing four PlanetScope image bands (red, green, blue, and NIR) and a binary mask for landslide detection. The HRGLDD can be accessed through this link: https://doi.org/10.5281/zenodo.7189381 (Meena et al., 2022a, c). HRGLDD is one of the first datasets for landslide detection generated by high-resolution satellite imagery which can be useful for applications in artificial intelligence for landslide segmentation and detection studies. Five stateof-the-art deep learning models were used to test the transferability and robustness of the HR-GLDD. Moreover, three recent landslide events were used for testing the performance and usability of the dataset to comment on the detection of newly occurring significant landslide events. The deep learning models showed similar results when testing the HR-GLDD at individual test sites, thereby indicating the robustness of the dataset for such purposes. The HR-GLDD is open access and it has the potential to calibrate and develop models to produce reliable inventories using high-resolution satellite imagery after the occurrence of new significant landslide events. The HR-GLDD will be updated regularly by integrating data from new landslide events.