Conference Agenda

Interferometric Synthetic Aperture Radar (InSAR), a geodetic technique that measures ground deformation using phase differences of repeated radar acquisitions, provides a powerful tool for imaging geohazards and responding to crises. The 2024–2025 Fentale-Dofen magma intrusion in Ethiopia generated widespread and rapidly evolving surface deformation, triggered intense seismic swarms, raised the potential for an eruption, and resulted in large-scale evacuations of approximately 75,000 people. The intrusion propagated for ~13 days to a length of ~50 km, while the associated deformation (InSAR line-of-sight direction) reached to ~3 m over ~60 days and affected an area spanning ~10,500 km². Such events, with large spatial extent and rapid deformation, challenge any single SAR mission due to inherent trade-offs between spatial resolution, swath width, wavelength, and revisit time. Here we demonstrate the value of synergistic multi-mission SAR interferometry by integrating X-band COSMO-SkyMed (CSK and CSG), C-band Sentinel-1, and L-band SAOCOM observations to characterize deformation across spatial and temporal scales, using the 2024–2025 Fentale-Dofen magma intrusion as a case study.

X-band COSMO-SkyMed Stripmap data provide high spatial resolution (~3 m), allowing detailed observation of spatially localized deformation features including fault structures, crater deformation, and graben subsidence. The dataset was acquired in a tasked mode following detection of early unrest, resulting in short revisit intervals during the event (1, 8, and 16 days), which allow deformation to be tracked through the early intrusion stage when the magma propagated rapidly. Notably, the 1-day interferograms reveal short-timescale deformation transients associated with the intrusions and rapid earthquake-related deformation that longer-repeat satellites cannot capture. However, the relatively narrow swath (~40 km) limits regional spatial continuity, as a single track cannot cover the whole deformation field. Interferograms from adjacent tracks are acquired at different times and therefore cannot be merged during rapidly evolving deformation. In some cases, the large spatial baselines can also lead to coherence loss, even when temporal sampling is dense.

In addition, the short X-band wavelength results in many phase cycles for meter-scale displacement, generating large phase gradients and challenging the phase unwrapping. To overcome this limitation, we implement offset-supported phase unwrapping. In this approach, offsets derived from cross-correlation provide an unambiguous estimate of the long-wavelength deformation, which reduces phase gradients prior to unwrapping, while the interferometric phase preserves high spatial resolution and precision. The results demonstrate that offset-supported unwrapping is practical and effective for X-band data, stabilizing unwrapping in areas characterized by dense fringes and locally discontinuous deformation.

In contrast, C-band Sentinel-1 Interferometric Wide (IW) mode provides a large swath width (~250 km), enabling the entire deformation field to be captured in a single acquisition. While the original resolution is ~5 × 20 m, the interferograms used in this study are multilooked to ~30 m, which is sufficient for regional monitoring but cannot resolve localized features. With a consistent 12-day revisit time, Sentinel-1 cannot track the rapid propagation of the intrusion away from the central volcano, but offers a reliable and stable monitoring for long-term evolution, particularly when X-band acquisitions become less frequent after the deformation rates significantly decreased. The precise orbit control of Sentinel-1 ensures small spatial baselines and helps maintain coherence, which is a capability that not always guaranteed in newer missions. With the spatially broad and temporally consistent observations, Sentinel-1 supplements the detailed but localized X-band data, providing a continuous framework for the regional deformation assessment.

L-band SAOCOM observations further strengthen the analysis by providing an independent long-wavelength constraint on cumulative deformation. With ~10 m resolution, ~50 km swath width, and a revisit time of approximately 3–7 months in this region, the interferograms capture the cumulative displacement rather than the evolution of the event. The longer L-band wavelength produces fewer phase cycles for the meter-scale displacement, resulting in reduced phase gradients and more stable unwrapping. Although the long revisit time limits time-series analysis, L-band data provide an independent cross-validation on deformation magnitude and spatial pattern, improving the confidence in the multi-mission integration.

Together, these datasets illustrate how multi-frequency SAR integration mitigates individual mission limitations and allows for imaging the full spatio-temporal complexity of rapid magmatic deformation: X-band reveals localized details and temporal evolution, C-band provides regional continuity and consistent monitoring, and L-band maintains coherence over long timescales to constrain total displacement of the event. Looking forward, the Sentinel-1 Next Generation mission will further improve C-band observations, while the forthcoming ROSE-L mission is expected to deliver L-band data with revisit time of 3–6 days. These missions will further enhance the capability to establish a reliable framework for multi-mission geohazard observation in the future.

Lava dome–building eruptions pose significant hazards to communities due to dome collapse events, pyroclastic density currents, and sudden transitions to explosive activity. Magma ascent rate is a key control on eruptive behaviour and changes in ascent rate may precede transitions between effusive dome growth and explosive phases. Therefore, the timely detection and interpretation of ground deformation proximal to growing lava domes is critical for hazard assessment.

Ground deformation associated with magma flux during dome growth is typically small in magnitude and occurs over short timescales. Previous observations have therefore relied on proximally deployed ground-based instruments, such as tiltmeters. However, installing and maintaining instrumentation near active vents is hazardous, logistically challenging, and requires a dense sensor network to adequately resolve the spatial distribution of deformation. Typically, lava dome eruptions have been challenging targets for InSAR due to poor maintenance of coherence close to the vent, limited temporal and special resolution, and steep topography.

Here, we use high-resolution X-band TerraSAR-X and TanDEM-X InSAR observations to measure near-field ground deformation during three recent dome-building eruptions: prior to the July 2015 dome collapse at Volcán de Colima; during dome growth at Sinabung (2020–2021); and during the effusive phase of the 2021 eruption of La Soufrière, Saint-Vincent.

Understanding the distribution of excess pressure and shear stress along the conduit walls provides a direct link between magma ascent and ground deformation. However, multiple mechanisms may contribute to ground displacement during dome growth, such as the loading of newly extruded material. Distinguishing between these source processes is essential for interpreting the state of magma ascent.

To investigate the physics underlying the observed deformation, we use numerical modelling to model magma ascent in the conduit and resulting surface deformation in COMSOL Multiphysics. These models provide a framework for interpreting both ground-based and InSAR-derived deformation signals in terms of subsurface magmatic processes and improving assessments of dynamics during lava dome growth.

The high spatial resolution and small perpendicular baselines of the TerraSAR-X and TanDEM-X SAR mission allows detection of small magnitude and short-wavelength deformation signals that are unlikely to be captured by lower-resolution systems or sparse ground networks. Our results demonstrate the spatial distribution of possible deformation signals occurring proximally to growing domes. These observations provide a framework for validating magma ascent models and inform strategic deployment of future ground-based monitoring infrastructure. These findings support the acquisition of frequent X-band SAR images during volcanic crisis events, particularly at dome-building events where deformation is subtle and transient.

The U.S. Geological Survey (USGS) Volcano Hazards Program operates five volcano observatories across the United States under the umbrella of the Volcano Science Center (VSC). These observatories integrate in situ and satellite-based observations to evaluate volcanic activity, monitor hazards, and issue timely warnings to enhance public safety and reduce social and economic disruption. In addition to seismic, gas, and thermal measurements, geodetic observations play a central role in detecting pressurization and depressurization of magmatic systems before, during, and after episodes of volcanic unrest.

Almost half of the active volcanoes in the U.S. are not currently equipped with geodetic ground-based instrumentation. For these volcanoes, Interferometric Synthetic Aperture Radar (InSAR) is often the only geodetic data source available to monitor the evolution of surface displacement and evaluate potential unrest and eruption hazards. However, operational InSAR processing at volcano observatories is commonly performed manually or semi-manually. Such approaches can limit scalability, require significant computational resources, and present challenges for near-real-time hazard assessment, particularly during periods of escalating unrest when rapid situational awareness is critical.

Here we present VolcSARvatory, a cloud-based InSAR time series analysis service designed to support operational volcano monitoring using data from Sentinel-1 and NISAR. VolcSARvatory is a collaborative effort between the University of Alaska Fairbanks, the Alaska Satellite Facility (ASF), and the volcano observatories within the USGS VCS. The service and its workflows are automatically triggered when a new SAR acquisition intersects a user-defined Area of Interest (AOI). For each monitored site, VolcSARvatory retrieves Sentinel-1 and NISAR data from ASF’s cloud-based archives and constructs an optimized interferogram network based on expected coherence between reference and secondary acquisitions. Selected InSAR pairs are submitted to ASF’s HyP3 service for low-latency interferogram generation.

Generated interferograms are subsequently ingested into a small baseline subset (SBAS) InSAR framework to derive displacement time series. To improve computational efficiency and enable continuous updates, the time series analysis is divided into two-year-long segments with a one-year overlap. These segments are processed separately and merged during post-processing to maintain temporal continuity. Final displacement products are automatically transmitted to volcano observatory servers for integration into decision-support portals, where InSAR-derived deformation is combined with seismic, gas, and thermal datasets to inform hazard assessments.

We will present the conceptual framework, automated workflows, and cloud implementation of VolcSARvatory. Its anticipated operational impact is demonstrated through retrospective analyses of recent volcanic unrest episodes in Alaska, highlighting how automated, scalable InSAR time series generation can enhance situational awareness and strengthen volcano hazard assessment capabilities.

Observations of ground deformation provide insights on the triggering mechanisms of eruptions. The vast majority of the observations of ground deformation in volcanoes are made with C-band data, usually from missions like ENVISAT and Sentinel-1. These missions have provided data of excellent coherence for many targets across the Earth, but there are certain types of ground deformation that can only be studied by means of L-band data. This includes large deformation in regions that do not sustain coherence. For example, the onset of large eruptions due to large strain (e.g., Sierra Negra, 2018) or heavily vegetated and snowy terrain (e.g., Cordon Caulle 2011). One of these examples occurred at Aniakchak Crater in the Aleutians. Unrest between 2022 and 2023 resulted in 72 cm of uplift in less than 7 months recorded by ALOS-2 and SAOCOM-1 L-band data. Sentinel-1 data were decorrelated during the boreal winter and due to the large strain. The uplift was likely produced by the inflation of a point pressure source at a depth of 3.1 km. The infill of weak material in the caldera plays a minor role in amplifying the ground deformation signal, so magma injection is the most likely mechanism responsible for the uplift, with a time-averaged magma flux of 1.7 m³/s. A global compilation of magma injection rates in subduction zones derived from satellite geodetic data indicates that rates such as those of Aniakchak are not sustained over periods of time longer than one year. Therefore, pulses like that on their own are usually not likely to reach the conditions that promote eruptions, unless the reservoir is very close to failure.

Episodes like that of Aniakchak can only be observed with L-band data due to the high strain and the winter conditions of the volcano. Experiments with winter ALOS-4 images acquired during the end of the 2024-2025 winter and SAOCOM-1 during the 2022-2023 winter show that even 14 day-long pairs cannot sustain coherence during the winter, except for specific pairs when the dielectric properties of the snow remain constant. This is the shortest repeat period currently available for L-band data and it implies that several NISAR pairs acquired every 12 days should not be able to sustain coherence in the winter.

The interaction between tectonic earthquakes and volcanic systems represents a fundamental feedback loop in crustal dynamics. While it is generally accepted that large earthquakes can influence volcanic activity and vice versa, the underlying mechanisms remain a subject of debate. The 2018 earthquake sequence adjacent to the Rinjani-Samalas volcanic complex on Lombok Island, Indonesia, provides a unique opportunity to investigate these interactions in a near-field setting. This highly active system, which features one of the world's largest crater lakes, experienced a significant sequence of four earthquakes (M_w 6.2–6.9) on the island’s north coast. The InSAR-derived rupture models indicate a maximum slip of 2.5 m at a depth of 22.2 km directly beneath the Rinjani-Samalas edifice, raising critical questions regarding the mechanical response of the adjacent volcanic and hydrothermal system. In this study, we present the first long-term, systematic satellite-based characterization of the Rinjani system’s response, utilizing a multi-parametric time-series analysis spanning over a decade of multi-SAR and optical satellite datasets. Our methodology integrates InSAR ground deformation, subaerial and subaqueous gas emissions, Land Surface Temperature (LST), and SAR backscatter on a unified temporal scale to provide a holistic view of the system’s evolution. We utilized the 12 m resolution WorldDEM product based on TanDEM-X as the foundational digital elevation model and processed 1232 descending and 1553 ascending Sentinel-1 interferograms to generate a high-resolution 10-year InSAR time series. Modelling the geometry and volume changes of volcanic and hydrothermal sources from InSAR cumulative displacements quantifies the unrest of the volcanic complex in different periods; additionally, calculating the co- and post-seismic volumetric strain changes allows us to investigate the possible near-field impact of the earthquake sequence on the magmatic-hydrothermal system. Our analysis of multi-platform remote sensing data, including subaerial SO₂ emissions based on TROPOMI COBRA datasets and crater lake colour/clarity changes driven by subaqueous degassing illustrates a high consistency with the time-dependent displacement of the Rinjani caldera. The observed volcanic unrest, characterized by simultaneous degassing and temperature variations, challenges simple post-seismic tectonic triggering mechanisms. Instead, our results suggest that the influence of significant earthquakes on volcanic systems can be traced back to the early pre-seismic phase, highlighting the indispensable role of long-term geodetic monitoring in understanding complex regional stress interactions and enhancing volcanic hazard assessments.