The escalating frequency and severity of extreme weather events, induced by climate change, pose significant threats to global societies, economies, and ecosystems. Europe, and more specifically the Mediterranean area, has witnessed a surge in natural disasters resulting in substantial human casualties and economic losses. Despite advancements in disaster risk management, inadequate investment in early warning and detection systems have led to prolonged, costly, and frequent emergency responses, straining resources.
The UNICORN project, started in October 2024, focuses on the development of Copernicus emergency applications, using Earth Observation technologies and data to address the increasing frequency and intensity of extreme events (fires, floods) and geohazards (volcanoes) and their impact on society, the economy and the environment. UNICORN develops tools and applications for early warning, forecasting, and hazard monitoring that enable a resilient society, better-informed emergency services, and effective short-term recovery. It proposes innovative solutions for local authorities, policy makers, citizens, and industries which will increase their preparedness for extreme events and geohazards. UNICORN's approach involves creating state-of-the-art, scalable and transferable services tailored to user needs, pushing technological boundaries for precise, timely, and actionable results from data and knowledge. UNICORN is based on four use cases from different European regions, hazards, target stakeholders, and technologies, through an end user validation method to build a resilient European landscape.
UNICORN's foundation lays on the development of four strategically selected Copernicus emergency applications corresponding in 4 use cases which incorporate specific areas, regions, and countries from the Mediterranean area of Europe that has a long history of natural hazards and extreme events. These use cases through which the applications are implemented, monitored and validated in real world conditions are diverse due to the scale of operation (local, regional, sub-national), the hazards, the type of engaged stakeholders and the applied technologies:
- Flood forecasting integrating Copernicus data and weather forecast fusion - Attica region, Greece.
- Copernicus-based wildfire early detection, mapping and nowcasting - Corsica Island, France.
- High resolution fire danger forecasting - Northwestern Spain and Northern Portugal.
- Lava flow emergency management tool based on Copernicus data merged with numerical modelling - Sicily Island, Italy.
Acknowledgement: "This work has been supported by the European research project UNICORN. This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No. 101180172. This article reflects only the authors’ views and the EU Agency for the Space Programme (EUSPA) and the European Commission are not responsible for any use that may be made of the information it contains."
How to cite: Kontoes, C., Kaskara, M., and Pissaridi, K.: Copernicus emergency Applications for Resilience addressing businesses’ needs and policy making, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1293, https://doi.org/10.5194/egusphere-egu25-1293, 2025.
Our research institute has been promoting the national disaster monitoring and recovery support projects funded by Japanese government. Final goal of these projects is to establish the resilient social system in both before and after phase of national hazards. In this project, integrated system called "One stop system for disaster management" is under development, which enables us the optimum target observation and disaster situation assessment. With the rapid development of the small satellite industry, the use of remote sensing is dramatically changing in disaster monitoring. One of the key concepts is satellite constellation within or beyond single satellite series. Among this aspect, Japanese flagship satellite, ALOS-2 and -4, and small satellites from Japanese companies, are working together for effective observations under "One stop system". The effectiveness of this cooperative observation was evaluated in natural disasters caused by the Noto Peninsula earthquake (Jan. 2024), and other typical large-scale flooding in 2024. This paper summarizes the past research results and discusses future developments.
How to cite: Rokugawa, S., Taguchi, H., Sakai, N., and Boriigin, H.: Disaster monitoring initiative in Japan by the earth observation of multi satellite-series constellation., EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8691, https://doi.org/10.5194/egusphere-egu25-8691, 2025.
In Italy, "Telecommunications, Earth Observation, and Navigation" (TLC/EO/NAV) assets are currently exploited to realize services and applications (the so-called downstream) providing benefits to citizens and public institutions, translating space technology investments into significant social and economic gains. Among the downstream satellite-based services and products of most interest for the Italian institutions, there was the support to civil protection and environmental safeguard from natural and anthropogenic hazards.
In this context, the Italian Space Agency (ASI) promotes the development of downstream services through the "Innovation for Downstream Preparation" (I4DP) program. I4DP foresees an active engagement of the user community in the demonstration projects since the user requirement consolidation phase, until the testing of the developed technological solutions in real-world scenarios. The I4DP_SCIENCE stream, addressed to the Scientific User Community (i.e., Italian Universities and Public Research Bodies) and designed to develop joint projects with ASI, is aimed at demonstrating the usefulness of novel methods and algorithms in supporting applications of user's interest regarding topics of national relevance (e.g., defined by the National Copernicus User Forum), and/or international agendas (e.g., the UN Sustainable Development Goals). The I4DP_SCIENCE program has been developed through the issue of two Calls for Ideas, focused on the themes of "Sustainable Cities" and "Agriculture and Sustainable Use of Water Resources" [1]. Among the selected projects, three address the theme of natural and anthropogenic hazard assessment and mitigation: GEORES, SatellOmic, and GRAW. GEORES (Geospatial Application to Support the Improvement of Environmental Sustainability and Resilience to Climate Change in Urban Areas, Agreement n. 2023-42-HH.0) [2], led by the University of Bari and CNR-IREA, aims at developing a geospatial application to improve environmental sustainability in urban areas, through a multi-risk platform, with the synergistic use of EO data, machine learning techniques, and artificial intelligence. SatellOmic (Integration of Satellite and Metagenomic Systems for the Monitoring and Safeguarding of Water Basins, Agreement n. 2023-36-HH.0) [3], led by the Istituto Superiore di Sanità (ISS) and the Scuola di Ingegneria Aerospaziale (SIA) of Sapienza University of Rome, aims at combining EO-based products with metagenomics analyses, to evaluate and monitor the quality of inland and coastal waters, assessing the presence of oil spills or algal blooms. GRAW (Geomatics for Resilience Against Water Scarcity, Agreement n. 2023-52-HH.0) [4], led by Sapienza University of Rome, aims at developing specific approaches and algorithms to monitor and forecast hazards due to hydrological and agricultural drought. The paper outlines how the three I4DP_SCIENCE projects are addressing natural hazard and risk using EO and geospatial technologies and accounting for the specific user requirements and needs.
[1] D. Tapete et al. (2024) The Italian Space Agency’s programs of scientific downstream applications for water resources and hydraulic hazard management. 14° Workshop tematico di Telerilevamento “Telerilevamento applicato alla gestione delle risorse idriche”, ENEA, Bologna, Italy, pp. 5-9. https://www.eventi.enea.it/images/presentazioni2024/2024_06_06_telerilevamento/Abstract_AIT2024_def.pdf
[2] R. Lafortezza et al. (2024), doi: 10.1109/IGARSS53475.2024.10642728.
[3] E. D’Ugo et al. (2024), doi: 10.1109/IGARSS53475.2024.10642700.
[4] F. Bocchino et al. (2024), doi: 10.1109/IGARSS53475.2024.10641154
How to cite: Ursi, A., Tapete, D., Sacco, P., Virelli, M., Coletta, A., Guarini, R., Licciardi, G., Longo, F., Siciliani de Cumis, M., and Zoffoli, S.: Italian Space Agency's Program for the Development of Novel EO-based Scientific Products for Natural Hazards Applications, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10904, https://doi.org/10.5194/egusphere-egu25-10904, 2025.
Remote sensing data is a key tool in disaster response and preparedness. For fire detection and monitoring, however, public satellite missions have significant coverage gaps in the afternoon. As most fires start in the afternoon, however, many can burn potentially undetected for a long period of time.
This is why OroraTech is building the Forest Constellation to close this afternoon gap. With two successful launches completed and an additional 9 satellites scheduled to be in orbit by the time of EGU 25, the constellation will achieve a 12-hour revisit time for any location on Earth focusing on late afternoon orbits. FOREST-2, the current sensor generation in orbit covers a swath of 410km in a single scan at a resolution of 200m per pixel. Future launches will further enhance the system, eventually enabling a global revisit time of just 30 minutes. This increased temporal and spatial coverage will allow for significantly earlier fire detection.
The fire detection is run on board on a GPU to keep latencies minimal. This is because downlink bottlenecks are easier to circumvent with bites of fire location files then GB size satellite images. Otherwise, the file transfer of a satellite image to the ground would introduce a significant bottleneck. To cut down communication latencies further, we rely on a dedicated ground station network with OroraTech’s Fire Link technology. With the upcoming satellite launches, we therefore enable fire detection within minutes after the satellite overpass. At EGU, we aim to showcase current and future capabilities of our constellation to detect fires with first impressions from upcoming launches.
How to cite: Laux, D., Wahbe, J., Liesenhoff, L., Bereczky, M., Spichtinger, A., Würl, K., Gottfriedsen, J., and Langer, M.: Near-Real Time Active Fire Detection from Space with the Forest-Constellation, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18898, https://doi.org/10.5194/egusphere-egu25-18898, 2025.
Coastal erosion and accretion remain critical challenges for sustainable coastal zone management, particularly in regions facing intensified environmental changes and human interventions. With rising sea levels projected to exacerbate these challenges, this study investigates the shoreline dynamics of Ghana’s eastern coastline over a nearly four-decade period. Using automated shoreline extraction via CoastSat and change analysis with the Digital Shoreline Analysis System (DSAS), the research quantified erosion and accretion rates and revealed significant spatial variability across the study area.
To facilitate a focused analysis, the coastline was divided into three zones (Zone A, Zone B, and Zone C), categorized based on geomorphological characteristics and coastal management practices. Three statistical models—Linear Regression Rate (LRR), End Point Rate (EPR), and Net Shoreline Movement (NSM)—were employed to quantify shoreline changes. Zone A exhibited a balance of erosion and accretion, with rates ranging from −10.5 m/year to +10.8 m/year (EPR) and −10.4 m/year to +12.0 m/year (LRR). Zone B showed relatively stable dynamics, with EPR values from −3.1 m/year to +4.1 m/year and LRR values from −1.6 m/year to +5.0 m/year. Zone C displayed the most pronounced variability, with erosion rates peaking at −30.5 m/year (EPR) and −28.7 m/year (LRR), alongside accretion rates up to +9.7 m/year (LRR) and +8.8 m/year (EPR). Cumulative shoreline movements (NSM) averaged 15.2 m, 20.42 m, and 33.5 m for Zones A, B, and C, respectively.
This study underscores the value of remote sensing and GIS in monitoring shoreline dynamics. By automating shoreline extraction with CoastSat, human error is minimized, and reproducibility is enhanced. The findings provide actionable insights into shoreline management and highlight the potential of these techniques for broader applications in coastal monitoring. This approach can empower stakeholders to devise effective, data-driven strategies for resilience against coastal erosion and sustainable coastal management practices.
How to cite: Asare-Bediako, B. and Boateng, C.: Spatiotemporal Analysis of Coastal Erosion and Accretion Patterns in the Keta Region, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-376, https://doi.org/10.5194/egusphere-egu25-376, 2025.
Extreme meteorological events, such as heat and drought, can induce significant damage to vegetation and ecosystems. The frequency and intensity of extreme events are subject to change due to anthropogenic global warming. It is therefore crucial to quantify the impact of such events for better preparedness.
Here, we focus on forest damage in Europe, defined as negative anomalies of the normalized difference vegetation index (NDVI, a measure of vegetation greenness). Compound drought and heat wave events are known to trigger low NDVI events in summer. A dry summer combined with moist conditions during the previous autumn can also have a negative impact. Hence, the goal of our study is to find the most relevant predictors for forest damage in Europe at monthly to annual timescales. Using a Random Forest approach, we pinpoint hydro-meteorological conditions associated with low NDVI events. We train the model using remote sensing observations of NDVI (from the Advanced Very High-Resolution Radiometers, AVHRR) as the predictand, and a range of variables from the ERA5 and ERA5-Land reanalysis as hydro-meteorological predictors.
We provide an automated procedure with strong predictive performance for identifying low-greenness events during summer based on prior hydro-meteorological conditions. The most essential preceding periods and variables are location and forest-type dependent. Notably, warm and dry conditions in spring and early summer emerge as essential predictors. Additionally, we emphasize a longer-term relationship between hydro-meteorological conditions and forest damage. For instance, low dewpoint temperatures one year before the studied summer impact broad-leaved forests, while soil moisture during the preceding autumn influences low greenness events in coniferous forests, albeit with location-specific variations.
How to cite: Rivoire, P., Dupuis, S., Guisan, A., Vittoz, P., and Domeisen, D.: Hydro-Meteorological Drivers of Forest Damage over Europe, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16296, https://doi.org/10.5194/egusphere-egu25-16296, 2025.
This study focuses on the fifth eruption of the Sundhnúkur volcano on the Reykjanes Peninsula, Iceland, which occurred between May 29 and June 22, 2024. Multi-satellite imaging techniques were used to analyze the activity of this volcano, which erupted a total of seven times between 2023 and 2024. Multiple Landsat-9 satellite images were acquired before and after the eruption, and Support Vector Machine (SVM) techniques were applied to calculate changes in the volcanic plateau. In addition, a series of Sentinel-1 satellite images was used to detect coherence changes during the eruption period and compared with the expanded volcanic plateau derived from Landsat-9 to determine the area of change. The results show that the fifth eruption was characterized by a large number of lava flows and significant volcanic plateau expansion in the early stages, with relatively little lava eruption in the later stages. This study may contribute to the calculation of changes caused by volcanic eruptions in the Icelandic region and could potentially help determine the affected area using a combination of different satellite images. This approach might be useful for future volcanic activity monitoring and disaster management.
Acknowledgment: This work was supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIT) (No. NRF–2023R1A2C1007742).
How to cite: Kim, B. and Lee, C.-W.: Analysis of the May 2024 Iceland Volcano Eruption Using Coherence Change Detection Techniques, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14167, https://doi.org/10.5194/egusphere-egu25-14167, 2025.
Remotely sensed nighttime lights (NTL) are widely used as a proxy for human activity. A key application is assessing disaster impacts, but its potential has been limited by uncertainties in estimating baseline NTL intensity (the counterfactual without disasters) and challenges in isolating disaster impacts from other factors influencing NTL variation. To address these challenges, we used a synthetic control modeling framework with daily NTL images from NASA's Black Marble VIIRS product suite. We enhanced the traditional model by optimizing donor selection with the Dynamic Time Warping algorithm and incorporating random forest regression to better capture target-donor relationships. Testing on 20 severe disasters across diverse contexts, our model outperformed existing methods, achieving a correlation coefficient of 0.94 and a covariate difference of just 0.47%. It also excelled at detecting low-intensity and short-term disaster impacts often missed by other methods. The resulting metrics—impact duration, intensity, and severity—revealed significant regional variations in disaster resilience and coping capacity. This model provides valuable insights for disaster relief and supports broader climate resilience and sustainability efforts.
How to cite: Mu, T., Zheng, Q., and He, S. Y.: Robust disaster impact assessment with synthetic control modeling framework and daily nighttime light time series images, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2264, https://doi.org/10.5194/egusphere-egu25-2264, 2025.
Mapping of neotectonic faults is critical to the scientific study of earthquake processes and surface rupture hazard analysis. Geologists commonly map fault traces from remote sensing datasets by interpreting tectonic landforms formed from past earthquakes. However, the evidence for faulting is not always straightforward to observe and interpret. Even experts map faults differently. We seek to understand the variability in fault trace mapping by mappers with different experience, ranging from undergraduate students to professional geologists with decades of experience. An individual’s understanding of faulting is impacted by their experience, yet people with similar experiences can interpret areas differently. No matter their experience level, mappers all have gaps in knowledge, and faults can rupture in unexpected ways. We anticipate that the results will improve the development of standardized mapping practices.
To evaluate the effect of differing knowledge on mapped fault trace locations, 23 mappers of varying experience levels produced fault maps from pre-rupture topography and imagery acquired before the earthquake of interest. Mappers include four undergraduate students, eleven graduate students, two postdocs, and three mid- and three senior-level professional geologists. The mappers used a systematic approach to map faults based on geomorphology.
To assess map quality, we compared the pre-rupture fault maps to published coseismic rupture maps. We evaluated 1) the percentage of coseismic ruptures that were predicted by the mapped faults, 2) the percentage of the mapped faults that ruptured in the recent earthquake, 3) the distribution of mapped faults around indicative geomorphic landforms, and 4) the impact of data type and the use of a geologic map. We found slight improvement by experience level in the portion of ruptures predicted and mapped faults that ruptured. For ruptures near geomorphic landforms, professional geologists best predicted the rupture location, and undergraduate students mapped with the highest error. Less experienced mappers tended to misinterpret some geomorphology. Despite these differences in experience level, all participants mapped some faults that did not rupture. Mappers of all experience levels were most successful with the high-resolution topography (~1m/pix). Aerial imagery was less useful in areas with high vegetation and anthropogenic activity. Using a geologic map did not improve the maps.
While experience level has a small effect on fault trace mapping accuracy, with our results we can start defining the epistemic uncertainty for geomorphic fault mapping. Our results also suggest that mapping fault traces remains challenging regardless of expertise and highlight the need for improved and standardized mapping practices.
How to cite: Zuckerman, M., Scott, C., Arrowsmith, R., Madugo, C., Koehler, R., and Kottke, A.: Variability in the location of mapped fault traces based on geomorphic mapping of remote-sensing datasets from 23 mappers, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14423, https://doi.org/10.5194/egusphere-egu25-14423, 2025.
This study aims to enhance the precision of semantic segmentation in remote sensing by evaluating advanced deep learning models on high-resolution datasets, addressing a critical need in geoscience applications. Accurate spatial identification of flood-affected areas is vital for timely disaster response, yet traditional methods often fail to capture the intricate patterns and scales of flood events. Advanced architectures like Convolutional Neural Networks (CNNs) and transformer models have proven transformative in overcoming these limitations.Using high-resolution imagery from the ISPRS dataset, this research compares CNNs and transformers, including the Vision Transformer (ViT), to identify the most effective architecture. While CNNs excel in extracting localized features, they struggle with capturing long-range dependencies. Transformer models, leveraging self-attention mechanisms, address this gap by modeling complex spatial relationships and global contexts, crucial for segmenting large-scale flood scenarios. Additionally, a novel transformer-based framework will be introduced to further enhance segmentation accuracy to detect flooding.To test robustness, the best-performing model is applied to flood detection tasks using lower-resolution datasets, simulating real-world disaster scenarios where data quality varies. Flood detection through advanced deep learning is essential given the growing frequency of climate-driven disasters. These models enable precise and timely mapping of inundated areas, critical for effective resource allocation, evacuation planning, and post-disaster recovery. Transformers’ ability to process fine-grained and large-scale spatial features complements CNNs, delivering more reliable and detailed flood mapping.Focusing on coastal and urban flooding from Hurricane Milton, the study demonstrates the practical utility of these models in diverse scenarios. By optimizing model selection for flood detection, this research advances remote sensing methodologies, bridging the gap between theoretical advancements and real-world applications, and contributing to disaster preparedness and climate resilience efforts.
How to cite: Chakraborty, T. and Southworth, J.: Semantic Segmentation for Disaster Response: Evaluating CNNs and Transformers for Flood Mapping, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14424, https://doi.org/10.5194/egusphere-egu25-14424, 2025.
Taiwan's geological setting, characterized by rapid tectonic uplift and among the world's most intense precipitation patterns and recurring extreme rainfall events, offers a natural laboratory for studying sediment flux and erosion rates in mountain river basins. The availability of open-source satellite-derived digital elevation models (DEMs) provides an invaluable opportunity to evaluate their suitability for constraining sediment flux in these dynamic environments. The Laonong River, one of Taiwan's prominent and vulnerable watersheds, has been selected as a representative study area due to its history of past and ongoing landslides, making it ideal for understanding erosion processes and sediment transport dynamics. This study assesses erosion rates in the Laonong River Basin over the past two decades using satellite-derived DEMs from diverse optical and radar sources. By evaluating the suitability of underutilized global DEMs, including ASTER GDEM, NASADEM, SRTM, ALOS World 3D DEM (AW3D30), Copernicus DEM, FAB DEM, and TanDEM-X EDEM, and benchmarking them against a high-accuracy LiDAR DEM, we aim to enhance our understanding of the erosional processes. Accuracy assessments are conducted in stable areas through spatial domain analysis, utilizing comprehensive metrics, including RMSE, bias, and standard deviation, to quantify discrepancies and ensure rigorous error evaluation. Additionally, metadata analyses identify voids and artifacts filled from external sources, while Fourier analysis is applied to detect and mitigate vertical biases, enabling a robust examination of DEM suitability in this complex terrain.
Our findings revealed that while Copernicus DEM, FAB DEM, and TanDEM-X EDEM exhibited good vertical accuracy in the spatial domain, their reliance on external DEMs for void-filling rendered them unsuitable for multitemporal analysis. Similarly, ASTER GDEM was excluded due to its high standard deviation, significant negative bias, and prolonged acquisition period, averaging over 13 years. As confirmed through Fourier analysis and in the spatial domain against LiDAR DEM, AW3D30 demonstrated excellent vertical accuracy and minimal vertical bias. NASADEM, being the successor of SRTM, was preferred over its predecessor due to lower vertical bias and minimal external void-filling. Consequently, NASADEM and AW3D30 were identified as the most reliable DEMs for capturing topographic changes across different decades in the Laonong River Basin. Horizontal co-registration was refined to sub-pixel accuracy using the Nuth and Kääb method, while Fourier analysis was employed for vertical alignment, effectively minimizing biases across DEMs acquired at different time points. Spectral analysis identified long-wavelength topographic features crucial for correcting offsets and enhancing the accuracy of DEM differencing. Our results estimate that about 119 Mm3 of sediment volume has been transported out of the system over 20 years, as calculated from NASADEM and LiDAR DEM. We documented the spatial pattern of erosion and deposition across the whole Laonong River basin in the DEMs of Differences (DoD) maps, and the results were validated from the Google Earth imageries. These findings highlight the capability of underutilized satellite-derived DEMs in capturing sediment erosion rates over multiple decades, demonstrating their utility in environments where erosional signals are dominant over the inherent noise in the dataset.
How to cite: Kumar, G., Chan, Y.-C., Sun, C.-W., and Chen, C.-T.: Evaluating Erosion Rates Through Advanced DEM Differencing and Co-Registration Techniques Using Underutilized Satellite Data: A Case Study from Southern Taiwan, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3377, https://doi.org/10.5194/egusphere-egu25-3377, 2025.
Long-term monitoring of ecosystems is the only direct method to provide insights into the system dynamics on a range of timescales from the temporal resolution to the duration of the record. Time series of typical environmental variables reveal a striking diversity of trends, periodicities, and long-range correlations. Using several decades of observations of water chemistry in first-order streams of three adjacent catchments in the Harz mountains in Germany as example, we calculate metrics for these time series based on ordinal pattern statistics, e.g. permutation entropy and complexity, Fisher information, or q-complexity, and other indicators like Tarnopolski diagrams. The results are compared to those obtained for reference statistical processes, like fractional Brownian motion or ß noise. After detrending and removing significant periodicities from the time series, the distances of the residuals to the reference processes in this space of metrics serves as a classification of nonlinear dynamical behavior, and to judge whether inter-variable or rather inter-site differences are dominant. The classification can be combined with knowledge about the processes driving hydrochemistry, elucidating the connections between the variables. This can be the starting point for the next step, constructing causal networks from the multivariate dataset.
How to cite: Lange, H. and Hauhs, M.: Classification of environmental time series using ordinal pattern dynamics and complexity metrics, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3448, https://doi.org/10.5194/egusphere-egu25-3448, 2025.
Monitoring large-scale floods and tracking their evolution are essential for effective disaster response, particularly in regions where floods have widespread and dynamic impacts. Satellite-based flood detection using Synthetic Aperture Radar (SAR) and optical data faces challenges such as low spatial and temporal resolution, incomplete coverage, and cloud interference, which complicates the reliability of optical data. These issues hinder timely flood monitoring, which is critical for disaster management. This study introduces the Improved Knowledge-Driven Flood Intelligent Monitoring (KDFIMv2) method, which integrates SAR and optical data to improve flood monitoring by enhancing both spatial and temporal resolution.
The main challenge in large-scale flood monitoring is low spatiotemporal resolution, caused by limited SAR sensor coverage, low temporal observation frequency, and cloud interference affecting optical data. KDFIMv2 addresses these challenges through three key modules: 1) Surface Scattering Knowledge-Driven Flood Inundation Extraction, 2) Physical Knowledge-Driven Feature Fusion, and 3) Mathematical Knowledge-Driven Flood Information Extraction. The Surface Scattering Knowledge-Driven Flood Inundation Extraction module integrates SAR and optical data to extract flood information from satellite images. It tackles cloud cover and cloud shadows, which hinder water surface extraction in optical data, especially during floods. By combining SAR’s surface scattering capability with optical image spectral data, this module ensures accurate flood detection even under cloudy conditions. The Physical Knowledge-Driven Feature Fusion module improves adaptability by extending potential flood areas based on existing data. Using knowledge of flood dynamics, it infers the evolution of flood levels across a basin, filling gaps caused by cloud interference or incomplete satellite coverage, offering a more comprehensive flood monitoring solution. The Mathematical Knowledge-Driven Flood Information Extraction module uses mathematical models to calculate flood parameters such as depth, duration, and spread, providing a holistic assessment of the flood’s impact. This allows authorities to quantify flood disasters and track their evolution over time.
KDFIMv2 was applied to monitor floods in Bangladesh from June to December 2024. Results showed that KDFIMv2’s flood depth estimates had a mean error of only 0.1 meters, with 75% of the area within 0.2 meters and 95% within 0.5 meters. The method mitigated cloud cover and observational limitations, enabling flood tracking with a 30-meter resolution every two days. KDFIMv2 overcomes the limitations of current flood monitoring systems, offering high-accuracy flood evolution tracking. This study advances flood monitoring techniques and contributes to a better understanding of the impacts of floods on climate change adaptation and disaster resilience. By enhancing flood monitoring accuracy, KDFIMv2 plays a crucial role in reducing risks for vulnerable populations and contributes to achieving the United Nations Sustainable Development Goals (SDGs), particularly SDG 1 (No Poverty) and SDG 11 (Sustainable Cities and Communities).
How to cite: Jiao, Z., Zhang, Z., Chen, B., Miao, Z., and Wu, L.: Improved Knowledge-Driven Flood Intelligent Monitoring (KDFIMv2): A Case Study of the 2024 Bangladesh Flood, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6141, https://doi.org/10.5194/egusphere-egu25-6141, 2025.
Various Land surface anomalies have destroyed the stable and balanced state of human living, resulting in fatalities and serious destruction of property. Remote sensing technique has been proven useful in many studies with time-series and large-scale observation advantages. However, existing studies are limited in anomaly recognition of certain categories, lacking the important generalization ability to recognize rare or unseen anomalies. To tackle this problem, we have built a multi-modal land surface anomaly recognition foundation model, which connects the images and anomaly caption words in an open-world setting. A global scale multi-modal dataset is constructed to train the model, which refers to 1000 large-scale monitoring regions covering over 2000 km2 in total, with rich text caption collected from offical news report. After the self-supervised contrastive learning with image and text modalities, the foundation model can describe both the anomaly category and attributes directly given any monitoring image, without the need for further tuning. These open-world and tuning-free settings promote the ability of rapid anomaly monitoring.
How to cite: Li, J.: Recognizing any Land surface anomaly with multi-modal foundation model, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11003, https://doi.org/10.5194/egusphere-egu25-11003, 2025.
Earthquakes inflict extensive damage on urban settlements, posing significant risks to human life and infrastructure. This study investigates the destruction caused by the Mw=6.7 Sivrice earthquake, which struck on January 24, 2020. It focuses on its impact within the city center via remote sensing (RS) and geographic information system (GIS) methodologies for damage assessment. In the study, post-earthquake building damage data and building inventory, collected through field surveys and provided by the Ministry, were compared with the collapsed areas which were identified using Sentinel-2 optical imagery. The surface velocity rates of the study area over the past five years have been derived using advanced InSAR techniques, specifically Persistent Scatterer (PS) and Small Baseline Subset (SBAS) methods. Moreover, a comprehensive analysis of factors influencing building damage was conducted by integrating geological maps, active fault maps, Coulomb stress distribution, Vs30 velocities, and surface velocity rates in the city center with the building inventory. It was revealed that the damage estimation map produced by analyzing pre-earthquake data and the damage map produced by field studies carried out after the earthquake showed a similar pattern. The findings demonstrate that buildings constructed on alluvial soils having low Vs30 velocities with high surface velocity rates experienced the most severe damage. This analysis highlights critical geological and structural parameters that exacerbate earthquake-induced structural damage, offering valuable insights for urban planning and seismic risk mitigation.
How to cite: Gedik, Y., Özcan, O., Özmen, M., and Özcan, O.: Earthquake Damage Assessment Using SAR Imagery, Structural Data, and Geospatial Parameters, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11698, https://doi.org/10.5194/egusphere-egu25-11698, 2025.
Typhoon damage refers to the extensive destruction of natural and environmental resources caused by powerful tropical storms, which are characterized by high winds and heavy rainfall. The impacts of these natural disasters on mountainous land and forest resources are severe and far-reaching. From a forest management perspective, common issues resulting from typhoon strikes include uprooted or broken trees, significant soil erosion, and landslides of varying scales across sloped areas. These events lead to substantial changes in the structure of forest stands and the topography of forested regions, resulting in increased vulnerability of forest ecosystems and fragmented land. This fragmentation endangers wildlife habitats and undermines biodiversity conservation. This study analyzes changes in the growth competition index of trees resulting from damage caused by a typhoon in a subtropical forest. We use multitemporal airborne lidar scanning data to assess the impact of severe weather events on stand density and tree growth dynamics over time. By examining variations in the competition index over a decade, we aim to provide valuable insights into the resilience of forest ecosystems after such disturbances. Our research will deepen the understanding of forest recovery processes and inform management practices in areas frequently affected by typhoons.
How to cite: Lin, C., Ma, S.-E., and Liao, W.: Examining changes in the growth competition index of trees caused by typhoon damage using multitemporal airborne lidar scanning data, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11767, https://doi.org/10.5194/egusphere-egu25-11767, 2025.
The coronavirus disease 19 pandemic (COVID-19) has caused many deaths worldwide and has had a huge impact on society and the economy. The COVID-19 was caused by a new type of coronavirus (Severe Acute Respiratory Syndrome Coronavirus 2; SARS-CoV-2), which has been found that these viruses can be effectively inactivated by ultraviolet (UV) radiation of 290~315 nm. In this study, 90% inactivation time of the SARS-CoV-2 virus was analyzed using the UV Index data from Geostationary Environmental Monitoring Spectrometer (GEMS) satellite. The inactivation time of SARS-CoV-2 varies significantly with the seasonal, temporal and regional changes in the amount of radiation reaching the surface. In regions with lower latitudes, the higher amount of solar radiation was more effective in inactivating the virus, whereas in higher latitude regions, a longer duration was required for the same level of inactivation. Also in winter season, the natural prevention effect was meaningless because the intensity of UV radiation weakened, and the time required for virus inactivation increased. The spread of infectious diseases such as COVID-19 is related to the diverse and complex interactions of various variables. However, the natural inactivation of viruses by ultraviolet radiation presented in this study, particularly the seasonal, temporal and regional differences, needs to be considered as major variables.
How to cite: Park, S. and Lee, Y. G.: Estimation of the Inactivation time for SARS-CoV-2 using the GEMS UVI data, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14998, https://doi.org/10.5194/egusphere-egu25-14998, 2025.
Sinkholes represent a significant geohazard, especially in urban environments where their sudden formation damages infrastructure, property, and even loss of life. These features, often caused by natural processes such as the dissolution of soluble rocks (e.g., limestone, gypsum) or human-induced activities like water extraction and construction, pose unique challenges in urban areas.
In April 2024, Slănic Prahova, Romania, experienced a significant geological event. Near the local police headquarters, a portion of 23 August Street collapsed, creating a crater approximately 2 meters deep and over 60 square meters. Due to safety concerns, around 42 residents were evacuated from nearby buildings.
Following the event, several campaigns were started to monitor and assess the ground deformations of the sinkhole and its surroundings. This event underscores the importance of continuously monitoring and assessing natural hazards in urban areas, particularly in regions with known subsurface vulnerabilities.
How to cite: Necula, N. and Niculiță, M.: Investigation of ground instability and sinkhole monitoring in Slănic Prahova, Romania with InSAR and in-situ measurements, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16812, https://doi.org/10.5194/egusphere-egu25-16812, 2025.
Introduction
Monitoring and detecting rockfalls on cliffs is essential for reducing risks to infrastructure, human safety, and ecosystems in steep terrains. Traditional methods often rely on expensive equipment and labor-intensive surveys, restricting their use to high-risk areas. Single-camera monitoring offers a cost-effective and scalable alternative, using advancements in image processing and change detection algorithms to identify rockfall events. Challenges such as lighting variations and environmental noise, require adapting existing algorithms for effective deployment. This study introduces a low-cost, autonomous system using high-resolution images from a single camera combined with an unsupervised deep learning approach to efficiently detect rockfalls.
Methodology
Prototype Design and Installation
The proposed monitoring system integrates commercially available components for ease of deploymen. A Sony-Alpha-7RM4A camera paired with a 400mm lens captures high-resolution images of fine-scale changes on the rockface. A Gigapan pan/tilt mechanism provides precise control for acquiring mosaics covering large-areas. A Raspberry-Pi controller automates image capture, data transfer, power monitoring, and remote transmission. Solar panels mounted on an adjustable frame provide continuous power, while weatherproof housing protects the components.
Operating autonomously, the system captures hourly images between 6:00-AM and 6:00-PM daily. Images are stored locally and transmitted remotely when connectivity permits. The system also logs operational data to support maintenance. Installed at the St. Eynard cliff in Biviers, France, the prototype captures mosaics of 102 high-resolution images per acquisition, enabling daily monitoring.
Image Preprocessing and Change Detection
The system captures around 40 GB of images daily, with a resolution of about 1 cm, under varying conditions such as lighting changes, vegetation growth, weather effects, and camera vibrations. These factors pose challenges for detecting rockfall-related changes, necessitating a robust image processing chain.
First, images are organized into tiles corresponding to specific regions of the mosaic and renamed using timestamps. Blurry or poorly lit images are filtered out using methods like Laplacian variance and gradient analysis. Images are then coregistered within each tile using the Scale-Invariant-Feature-Transform method, ensuring consistent pixel-level correspondence across the time series. Preprocessed tiles are assembled into a coherent mosaic of the study area.
Traditional threshold-based change detection methods are ineffective in large study areas due to diverse changes and lack of ground truth. To overcome this, a Siamese-Variational-Autoencoder (SVAE) was developed. The SVAE uses a U-Net-like architecture to extract latent features, an attention mechanism to focus on critical features, and a change-detection branch to generate precise change maps. Loss functions, including Kullback-Leibler divergence, perceptual, and texture ensure robust latent representations and preserve image fidelity, enabling effective detection of subtle changes while minimizing false positives.
Finally, processed images are georeferenced, translating detected changes into geographic coordinates to extract attributes such as rockfall size and location.
Applications
This framework has been successfully implemented at the St. Eynard cliffs. Detection results were validated against complementary datasets, including lidar and seismic data, demonstrating the system's reliability and effectiveness in real-world applications. This research was funded, in whole or in part, by the French National Research Agency (ANR) under the project C2R-IA (https://anrc2ria.fr/, grant ANR-22-CE56-0005-06).
How to cite: Le, T. T. and Stelzenmuller, N.: Rockfall Monitoring Using Multitemporal Single-Camera Terrestrial Images and Deep Learning, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-20266, https://doi.org/10.5194/egusphere-egu25-20266, 2025.
Posters virtual: Wed, 30 Apr, 14:00–15:45 | vPoster spot 3
EGU25-8599 | Posters virtual | VPS13
The Italian Space Agency Contribution to CEOS WGDisasters for Disaster Monitoring and ResponseWed, 30 Apr, 14:00–15:45 (CEST) | vP3.13
The Working Group on Disasters (WGDisasters) has been established since 2013 by the Committee on Earth Observation Satellites (CEOS, https://ceos.org) to ensure the sustained coordination of disaster-related activities undertaken by the CEOS Agencies as well as to act as an interface between CEOS and the community of stakeholders / users involved in risk management and disaster reduction.
In this framework, CEOS WGDisasters has initiated, promoted and supported a series of concrete actions for Disaster Risk Management (DRM) and Disaster Risk reduction (DRR) oriented to disaster monitoring, preparedness and prevention. These actions have been translated in single-hazard Pilot and Demonstrator projects (currently focusing on fires, floods, landslide, volcanoes and seismic hazards) as well as multi-hazards projects as the Recovery Observatory (RO) and Supersites for Geohazard Supersites and Natural Laboratories (GSNL).
Since 2012 ASI participates and contributes to the above-mentioned initiatives in terms of project selection and evaluation (as part of Data Coordination Team); data provision of COSMO-SkyMed, SAOCOM (only within the ASI Zone of Exclusivity defined in agreement with CONAE within SIASGE program) and PRISMA images; scientific activities in DRM and RO projects.
In coordination with WG members and CEOS Agencies, ASI has delivered more than 20.000 EO products until now and is actively involved in demonstrating novel scientific products based on a tailored exploitation of COSMO-SkyMed radar images. Several showcases will be presented at the time of the conference dealing with volcano monitoring (e.g. Mount Agung in Indonesia, Sierra Negra at Galapagos, St. Vincent in Caribbean), seismic activities (e.g. 2023 Turkey-Syria earthquake), multi-hazards “Supersite” initiatives (e.g. Reykjanes Peninsula, Kilauea and Mauna Loa volcanoes in Hawaii, Nyamuragira and Nyiragongo volcanoes) and RO initiative (e.g. 2016 Hurricane Matthew and 2021 Hurricane Grace in Haiti).
How to cite: Montuori, A., Tapete, D., Frulla, L., Czaran, L., Eddy, A., Virelli, M., Pari, G., and Zoffoli, S.: The Italian Space Agency Contribution to CEOS WGDisasters for Disaster Monitoring and Response, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8599, https://doi.org/10.5194/egusphere-egu25-8599, 2025.
EGU25-7731 | ECS | Posters virtual | VPS13
Accuracy Analysis of Photogrammetry and LiDAR Point Clouds Using an iPhone 13 Pro MaxWed, 30 Apr, 14:00–15:45 (CEST) | vP3.14
Subsidence is a geological phenomenon that continuously affects Mexico City. Over time, the impact of this phenomenon has been extensively studied using various methodologies, primarily at a regional scale. In recent years, efforts have shifted toward mapping subsidence at a local scale using technologies such as photogrammetry and LiDAR. These studies aim to establish a reference database to validate or complement regional-scale initiatives.
Field-based studies on subsidence often involve identifying problematic areas and analyzing topographical changes and structural damage over time. However, it is crucial to quantify and understand the limitations and capabilities of these techniques to establish a reference framework and ensure the reliability of the obtained data. Currently, precision methodologies are within everyone's reach thanks to technologies like photogrammetry and LiDAR from smartphones.
To achieve this, two controlled experiments (one conducted in the field and one in a laboratory setting) were carried out, in which 3D reconstructions of a box with known dimensions were made. Ten photogrammetry and ten LiDAR surveys were performed to compare the measurements obtained from the digital model with those taken from the physical object.
In the laboratory experiments, the average percentage error using photogrammetry was 1.03% (0.20 cm). Specifically, the error for a 16-cm-tall box was 1.44% (0.27 cm), while for a 20-cm-tall box, it was 0.61% (0.12 cm). For LiDAR, the average percentage error was 1.51% (0.27 cm), with errors of 1.50% (0.26 cm) for the 16-cm box and 1.52% (0.27 cm) for the 20-cm box. In field experiments, photogrammetry yielded an average percentage error of 0.88% (0.3 cm), whereas LiDAR showed an average error percentage of 2.17% (0.62 cm).
These findings confirm LiDAR and photogrammetry's potential for high-precision subsidence monitoring, providing a robust and accessible validation method. Utilizing mobile devices such as the iPhone 13 Pro Max extends the reach of these methodologies, enabling more accessible and practical research in urban contexts where subsidence poses significant challenges to infrastructure and quality of life.
How to cite: Vidal, G., Ramírez, N. L., Jácome, M. P., López, N., Reyes, T. A., and Yépez, F. D.: Accuracy Analysis of Photogrammetry and LiDAR Point Clouds Using an iPhone 13 Pro Max, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7731, https://doi.org/10.5194/egusphere-egu25-7731, 2025.
EGU25-4913 | ECS | Posters virtual | VPS13
Quantifying Surface Mining Expansion and Reclamation Using Deep Learning-based ConvoLSTM Model and Satellite Images: A Case Study in Lapland Region of Finland.Wed, 30 Apr, 14:00–15:45 (CEST) | vP3.15
Mining conflicts sustainable environment and causes disturbances for the livelihoods of people. Given the adverse impact on environment, indigenous community including Sami people and domesticated reindeer, it is of critical importance to peruse mining expansion and reclamation in Lapland, Finland. For the first time, this study employs a spatial-temporal deep learning architecture called ConvoLSTM, which enables accurate predictions of mining activities by capturing spectral, spatial, and temporal dependencies. Our custom model integrates a 2-Dimensional Convolutional Neural Network (2D-CNN) with a Long Short-Term Memory (LSTM) component. Using 10-meter Sentinel-2 imagery, we generated time-series land use/land cover (LULC) maps from 2015 to 2024 to track changes in mining extent. The performance of the spatial-temporal model was carefully evaluated against a Random Forest (RF) and a standalone 2D-CNN model, where it achieved superior accuracy. In the post-analysis phase, the Change Vector Analysis (CVA) technique was applied to quantify the magnitude and direction of change in mining activities over the past decade. The unique contribution of this study lies in implementing a custom spatial-temporal deep learning model to map decade-long mining activities and detect changes using publicly available satellite data. The resulting time-series maps demonstrate significant conversion of forest land and bare soil into mining areas, highlighting the rapid expansion of mining activities in Lapland which indicates a growing environmental concern in the arctic-boreal forest region. These findings offer critical insights and a valuable resource for policymakers, researchers, and reindeer herders, facilitating informed decision-making for sustainable environmental management and natural resource conservation in Finland.
Keywords: Mining Mapping, Environmental Impact, Remote Sensing, Deep Learning, CVA.
How to cite: Hasan, I. and Liu, D.: Quantifying Surface Mining Expansion and Reclamation Using Deep Learning-based ConvoLSTM Model and Satellite Images: A Case Study in Lapland Region of Finland., EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4913, https://doi.org/10.5194/egusphere-egu25-4913, 2025.
