Search

SUMMARYOne of the main constraints on the use of satellite radar data for monitoring natural hazards is the existence of atmospheric signals. In particular, volcanic deformation can be difficult to identify because atmospheric phase delays can mask or even mimic ground deformation signals. Eliminating atmospheric signals is particularly crucial for high-relief volcanoes such as Ağrı, Tendürek, Acigöl, Göllüdağ and Hasandağ in the Eastern and Central Anatolia. To overcome the atmospheric effects, we use high-resolution ECMWF weather models coupled with an empirical phase-elevation approach for correcting Sentinel-1 interferograms. We apply these methods to two areas of Turkey, the first of which covers three volcanoes in Central Anatolia (Acigöl, Göllüdağ, Hasandağ) between January 2016 and December 2018 and the second covers two volcanoes in Eastern Anatolia (Ağrı, Tendürek) between September 2016 and December 2018. The reduction in standard deviation (quality factor) is calculated for both ascending and descending tracks and the atmospheric corrections are found to perform better on descending interferograms in both cases. Then, we use a least-squares approach to produce a time-series. For Central Anatolia, we used 416 ascending and 415 descending interferograms to create 144 and 145 cumulative displacement maps, respectively, and for Eastern Anatolia, we used 390 ascending and 380 descending interferograms to produce 137 and 130 cumulative displacement maps, respectively. We find that the temporal standard deviation before atmospheric corrections ranges between 0.9 and 3.7 cm for the five volcanoes in the region and is consistently higher on ascending track data, which is acquired at the end of the day when solar heating is greatest. Atmospheric correction reduces the standard deviation to 0.5–2.5 cm. Residual signals might be due to the ice-cap at Ağrı and agriculture near Acigöl. We conclude that these volcanoes did not experience significant magmatic deformation during this time period, despite the apparent signals visible in individual uncorrected interferograms. We demonstrate that atmospheric corrections are vital when using InSAR for monitoring the deformation of high-relief volcanoes in arid continental climates such as Turkey.

Merapi is a strato-volcano rising at 2900 m a.s.l, located on the South coast of Java island, Indonesia. Only 30 km north to the city of Yogyakarta (2 millions inhabitants), it is considered one of the most dangerous dome building stratovolcanoes, as summit domes almost continuously grow and destruct. Merapi is therefore closely and routinely monitored by InSAR (Interferometric Synthetic Aperture Radar) to track ground deformation. To retrieve ground deformation from the full wave path, the delay due to the radar wave crossing the atmosphere needs to be corrected. In the case of Sentinel-1, interferograms are mostly biased by the tropospheric variations. Tropospheric variations are expected to be stronger in tropical regions and where topographic gradient is high, which is the case at Merapi. They can be estimated thanks to various methods, including global weather models (ERA-5 and GACOS), a linear model regarding topography, and GNSS networks.In this work, we compare the performance of atmospheric corrections derived from two weather-based models, ERA-5 and GACOS, and those derived from the empirical method based on a linear phase-elevation correlation. The aim is to evaluate the efficiency of each model in correcting this tropospheric bias. To this end, we choose to study a period between 2016 and 2018 during which no deformation occurred on the Merapi, so that most of the phase delays corresponds to tropospheric signals.We use three criteria to evaluate the performance: i) the reduction of the standard deviation, ii) the reduction of the sill of the semi-variogram, iii) the slope reduction of the phase-elevation correlation. We show that corrections with ERA and GACOS are efficient on only half of the interferograms. Finally, we also use the local network of 5 GNSS stations to rely on an independent dataset. We show there is a linear relation between the GNSS tropospheric delays and the global weather models delays. However, the GNSS network at Merapi is too small to provide an efficient correction on the whole volcanic edifice. For this reason, a similar workflow has been carried on the Piton de la Fournaise, Réunion island, using a wider GNSS network. The final aim of this study would be to implement a strategy on which the most suitable tropospheric model is chosen routinely based on the evaluation of the performance criteria to obtain atmospheric-free interferograms during volcanic unrest.

Countries with low to lower-middle income have limited resources to deploy and maintain ground monitoring networks. In this context, satellite-based techniques such as Radar interferometry (InSAR) is a great solution for detecting volcanic ground deformation at regional-scale. With the launch in 2014 of Sentinel-1 mission, regional monitoring of volcanic unrest becomes easier as SAR data are freely available with a revisit time of 6-12 days. Here, we develop a tuned processing workflow to produce Sentinel-1 InSAR time series and to automatically detect volcanic unrest over 80 volcanic systems located along the East African Rift System (EARS). First, we show that the correction of atmospheric signals for the arid and low-elevation EARS volcanoes is less important than for other volcanic environments. For a 5-year times series (between Jan. 2015 and Dec. 2019), we show that statistically uncertainties in InSAR velocities are around 0.1 cm/yr, whereas uncertainties associated with the choice of reference pixel are typically 0.3–0.6 cm/yr. For the automatic detection, we found that volcanic unrest can be detected with high confidence in the case the cumulative displacements exceed three times the temporal noise (threshold of 3σ). Based on this criterion, our survey reveals ground unrest at 16 volcanic centres among the 38 volcanic centres showing historical evidence of eruptive or unrest activity. A large variety of processes causing deformation occurs in the EARS: (1) subsidence due to contraction of magma bodies at Alu-Dalafilla, Dallol, Paka and Silali; (2) subsidence due to lava flows compaction at Kone and Nabro; (3) subsidence due to fluid migration at Olkaria and Aluto or fault-fluids interactions at Haludebi and Gada Ale; (4) rapid inflation due to magma intrusions at Erta Ale and Fentale; (5) short-lived inflation of shallow reservoirs at Nabro and Suswa; (6) long-lived inflation of large magmatic systems at Corbetti, Tullu Moje and Dabbahu. Except Olkaria and Kone, all these volcanoes were identified as deforming by previous satellites missions (between late 90’s and early 2000), which is an indication of the persistence of activity over long-time scales (>10 years). Finally, we fit the time series using simple functional forms and classify seven of the volcano time series as linear, six as sigmoidal and three as hybrid, enabling us to discriminate between steady deformation and short-term pulses of deformation. We found that the characteristics of the unrest signals are independent of the expected processes, which means that additional information (structural geology, seismicity, eruptive history and source modelling) will be necessary to characterize the processes causing the unrest. Our final objective will be to improve the transfer of this information to local scientists in Africa, which can be achieved by integrating our tools to an existing monitoring system and by developing web-platform where the InSAR products can be freely available.

With the evolution of InSAR into a tool for active hazard monitoring, new methods are sought to quickly and automatically interpret the large number of interferograms that are created. We present a convolutional neural network (CNN) that is able to both classify the type of deformation, and to locate the deformation within an interferogram in a single step. We achieve this through building a “two headed model", which returns both outputs after one forward pass of an interferogram though the network. We train our model by first creating a dataset of synthetic interferograms, but find that our model’s performance is improved through the inclusion of real Sentinel-1 data. When building models of this type, it is common for some of the weights within the model to be transferred from other models designed for different problems. Consequently, we also investigate how to best organise interferograms such that the filters learned in other domains are sensitive to the signals in interferograms, but find that using different data in each of the three input channels degrades performance when compared to the simple case of repeating wrapped or unwrapped phase across each channel. We also release our labelled Sentinel-1 interferograms as a database named VolcNet, which consists of ∼500,000 labelled interferograms. VolcNet comprises of time series of unwrapped phase and labels of the magnitude, location, and duration of deformation, which allows for the automatic creation of interferograms between any two acquisitions, and greatly increases the amount of data available compared to other labelling strategies.

Satellite-based earth observation sensors are increasingly able to monitor geophysical signals related to natural hazards, and many groups are working on rapid data acquisition, processing, and dissemination to data users with a wide range of expertise and goals. A particular challenge in the meaningful dissemination of Interferometric Synthetic Aperture Radar (InSAR) data to non-expert users is its unique differential data structure and sometimes low signal to noise ratio. In this study, we evaluate the online dissemination of ground deformation measurements from InSAR through Twitter, alongside the provision of open access InSAR data from the Centre for Observation and Modelling of Earthquakes, Volcanoes and Tectonics (COMET) Looking Into Continents from Space with Synthetic Aperture Radar (LiCSAR) processing system. Our aim is to evaluate (1) who interacts with disseminated InSAR data, (2) how the data are used and (3) to discuss strategies for meaningful communication and dissemination of open InSAR data. We found that InSAR Twitter activity was primarily associated with natural hazard response, specifically following earthquakes and volcanic activity, where InSAR measurements of ground deformation were disseminated, often using wrapped and unwrapped interferograms. For earthquake events, Sentinel-1 data were acquired, processed, and tweeted within 4.7±2.8 days (shortest was one day). Open access Sentinel-1 data dominated the InSAR tweets and were applied to volcanic and earthquake events in the most engaged with (retweeted) content. Open access InSAR data provided by LiCSAR was widely accessed, including automatically processed and tweeted interferograms and interactive event pages revealing ground deformation following earthquake events. The further work required to integrate dissemination of InSAR data into longer-term disaster risk reduction strategies is highly specific, both to hazard-type, international community of practice, and local political setting and civil protection mandates. Notably, communication of uncertainties and processing methodologies are still lacking. We conclude by outlining the future direction of COMET LiCSAR products to maximise their useability.

East Africa hosts significant reserves of untapped geothermal energy. Most exploration has focused on geologically young (

Radar (SAR) satellites systematically acquire imagery that can be used for volcano monitoring, characterising magmatic systems and potentially forecasting eruptions on a global scale. However, exploiting the large dataset is limited by the need for manual inspection, meaning timely dissemination of information is challenging. Here we automatically process ~ 600,000 images of 1000 volcanoes acquired by the Sentinel-1 satellite in a 5-year period (2015-2020) and use the dataset to demonstrate the applicability and limitations of machine learning for detecting deformation signals. Of the 16 volcanoes flagged most often, 5 experienced eruptions, 6 showed slow deformation, 2 had non-volcanic deformation and 3 had atmospheric artefacts. The detection threshold for the whole dataset is 5.9 cm, equivalent to a rate of 1.2 cm/year over the 5-year study period. We then use the large testing dataset to explore the effects of atmospheric conditions, land cover and signal characteristics on detectability and find that the performance of the machine learning algorithm is primarily limited by the quality of the available data, with poor coherence and slow signals being particularly challenging. The expanding dataset of systematically acquired, processed and flagged images will enable the quantitative analysis of volcanic monitoring signals on an unprecedented scale, but tailored processing will be needed for routine monitoring applications.The online version contains supplementary material available at 10.1007/s00445-022-01608-x.

Satellite radar backscatter has the potential to provide useful information about the progression of volcanic eruptions when optical, ground-based, or radar phase-based measurements are limited. However, backscatter changes are complex and challenging to interpret: explosive deposits produce different signals depending on pre-existing ground cover, radar parameters and eruption characteristics. We use high temporal- and spatial-resolution backscatter imagery to examine the emplacement and alteration of pyroclastic density currents (PDCs), lahar and ash deposits from the June 2018 eruption of Volcan de Fuego, Guatemala, using observatory reports and rainfall gauge data to ground truth our observations. We use a temporally dense time series of backscatter data to reduce noise and extract deposit areas. We observe backscatter changes in six drainages, the largest deposit was 11.9-km-long that altered an area of 6.3 urn:x-wiley:21699313:media:jgrb55183:jgrb55183-math-0001 and had a thickness of 10.5 urn:x-wiley:21699313:media:jgrb55183:jgrb55183-math-00022 m in the lower sections as estimated from radar shadows. The 3 June eruption also produced backscatter signal over an area of 40 urn:x-wiley:21699313:media:jgrb55183:jgrb55183-math-0003, consistent with reported ashfall. We use transient patterns in backscatter time series to identify nine periods of high lahar activity in a single drainage system between June and October 2018. We find that the characterisation of subtle backscatter signals associated with explosive eruptions are best observed with (1) radiometric terrain calibration, (2) speckle correction, and (3) consideration of pre-existing scattering properties. Our observations demonstrate that SAR backscatter can capture the emplacement and subsequent alteration of a range of explosive deposits, allowing the progression of an explosive eruption to be monitored.

This article contains brief overview on current key improvements in the interferograms generator part of LiCSAR, a system to generate open access moderate resolution Sentinel-1 interferograms aiming towards monitoring tectonic and volcanic deformation. We were focusing on improvements in performance and quality of unwrapping, and quality control of LiCSAR data.

There are ~ 1400 subaerial volcanoes with the potential to erupt, but less than 10% are instrumentally monitored. Routine acquisition by the European Sentinel-1 radar mission now offers the potential to monitor most of them with at least two acquisitions every twelve days. We have developed a system to routinely apply radar interferometry (InSAR) whenever a new image is acquired by Sentinel-1 over a volcano. Displacement of the ground between images shows up in the resulting interferograms, but there are too many of these to inspect individually. We have therefore developed an automated approach to identify signs of both new deformation patterns and changes in rate for existing deformation patterns, using independent component analysis. We first use a set of training interfero-grams to identify components associated with background deformation and common atmospheric patterns. We then analyse new interferograms in the context of those components, and flag changes in rate above the background variation, and significant unexplained new signals. We have found this approach to be successful even without having to first explicitly estimate the atmospheric signal in individual interferograms. We have also developed an alternative algorithm using a deep-learning approach, with the aim of extending this approach to forecasting of volcanic activity. This is implemented with a “double-headed” convolutional neural network, which is able to both locate deformation and classify the type of deformation, in single interferograms. We achieve this through creating a large dataset of synthetic interfero-grams, which features labels of both the type and location of any deformation, and is used for initial training of the network. We then fine-tune the network's performance through the inclusion of a smaller amount of real data.

A major challenge in using satellite interferometry (InSAR) for volcanic monitoring in the tropics is distinguishing volcano deformation from atmospheric noise. We reanalysed InSAR time series from 2007-2009 from Agung volcano, Indonesia, which had previously been used as evidence for unrest. Using uncorrected data, we find an apparent velocity of 5.0 ± 2.7 cm/yr consistent with previous reports, but we show that this signal is consistent with predictions of atmospheric contributions derived from weather models (ECMWF-HRES). Following the correction, we find a velocity of 1.4 ± 4.2 cm/yr and conclude that there was no significant deformation related to the inflation of a shallow magma source from 2007-2009. We discuss the implications for the inferred magma storage system at Agung and consider which other reported signals might have been wrongly attributed to deformation and should be reanalyzed.

Volcanoes have dormancy periods that may last decades to centuries meaning that eruptions at volcanoes with no historical records of eruptions are common. Baseline monitoring to detect the early stages of reawakening is therefore important even in regions with little recent volcanic activity. Satellite techniques, such as InSAR, are ideally suited for routinely surveying large and inaccessible regions, but the large datasets typically require expert interpretation. Here we focus on Turkey where there are 10 Holocene volcanic systems, but no eruptions since 1855 and consequently little ground-based monitoring. We analyse data from the first five years of the European Space Agency Sentinel-1 mission which collects data over Turkey every 6 days on both ascending and descending passes. The high relief edifices of Turkey’s volcanoes cause two challenges: 1) snow cover during the winter months causes a loss of coherence and 2) topographically-correlated atmospheric artefacts could be misinterpreted as deformation. We propose mitigation strategies for both. The raw time series at Hasan Dag volcano shows uplift of ~ 10 cm between September 2017 and July 2018, but atmospheric corrections based on global weather models demonstrate that this is an artefact and reduce the scatter in the data to

Space-borne Synthetic Aperture Radar (SAR) Interferometry (InSAR) is now a key geophysical tool for surface deformation studies. The European Commission’s Sentinel-1 Constellation began acquiring data systematically in late 2014. The data, which are free and open access, have global coverage at moderate resolution with a 6 or 12-day revisit, enabling researchers to investigate large- scale surface deformation systematically through time. However, full exploitation of the potential of Sentinel-1 requires specific processing approaches as well as the efficient use of modern computing and data storage facilities. Here we present LiCSAR, an operational system built for large-scale interferometric processing of Sentinel-1 data. LiCSAR is designed to automatically produce geocoded wrapped and unwrapped interferograms and coherence estimates, for large regions, at 0.001° resolution (WGS-84 system). The products are continuously updated in a frequency depending on prioritised regions (monthly, weekly or live update strategy). The products are open and freely accessible and downloadable through an online portal. We describe the algorithms, processing, and storage solutions implemented in LiCSAR, and show several case studies that use LiCSAR products to measure tectonic and volcanic deformation. We aim to accelerate the uptake of InSAR data by researchers as well as non-expert users by mass producing interferograms and derived products., This work was partially undertaken on ARC4, part of the High Performance Computing facilities at the University of Leeds, UK.

Satellite interferometric synthetic aperture radar (InSAR) can be used for measuring surface deformation for a variety of applications. Recent satellite missions, such as Sentinel-1, produce a large amount of data, meaning that visual inspection is impractical. Here we use deep learning, which has proved successful at object detection, to overcome this problem. Initially we present the use of convolutional neural networks (CNNs) for detecting rapid deformation events, which we test on a global dataset of over 30,000 wrapped interferograms at 900 volcanoes. We compare two potential training datasets: data augmentation applied to archive examples and synthetic models. Both are able to detect true positive results, but the data augmentation approach has a false positive rate of 0.205% and the synthetic approach has a false positive rate of 0.036%. Then, I will present an enhanced technique for measuring slow, sustained deformation over a range of scales from volcanic unrest to urban sources of deformation such as coalfields. By rewrapping cumulative time series, the detection performance is improved when the deformation rate is slow, as more fringes are generated without altering the signal to noise ratio. We adapt the method to use persistent scatterer InSAR data, which is sparse in nature, by using spatial interpolation methods such as modified matrix completion Finally, future perspectives for machine learning applications on InSAR data will be discussed.

Radar satellites, such as Sentinel-1, are now able to produce time-series of ground deformation at any volcano around the world, but atmospheric effects still limit the real time detection of unrest at tropical volcanoes. Here, we test two approaches to correct atmospheric errors - phase-elevation correlations and global weather models - and assess the ability of InSAR time-series to detect deformation anomalies using either a fixed threshold or a cumulative sum control chart. We use the 2017-2018 crisis at Agung volcano as a case example because strong atmospheric signals were originally misidentified as true deformation, and obscured the subtle deformation pattern associated with magmatic activity. We assess the Receiver Operating Characteristics (ROC) of each method and found the average area under the ROC curve to be about 0.5 for the uncorrected data (corresponding to no discrimination capability), around 0.8 after combined atmospheric corrections (weather model and phase-elevation approaches), and more than 0.95 using a cumulative sum control chart (where 1 corresponds to ideal separation between classes). Our results retrospectively show that uplift could have been detected to a 95% level of confidence for both ascending and descending time series by October 2017, 15 days after the start of the seismic swarm and one month prior to the eruption. Thus our approach successfully flags anomalous behavior without relying on visual inspection or selection of an arbitrary threshold, and hence shows potential as a monitoring tool for volcano observatories globally.

The accuracy of elastic analytical solutions and numerical models, widely used in volcanology to interpret surface ground deformation, depends heavily on the Young’s modulus chosen to represent the medium. The paucity of laboratory studies that provide Young’s moduli for volcanic rocks, and studies that tackle the topic of upscaling these values to the relevant lengthscale, has left volcano modellers ill-equipped to select appropriate Young’s moduli for their models. Here we present a wealth of laboratory data and suggest tools, widely used in geotechnics but adapted here to better suit volcanic rocks, to upscale these values to the scale of a volcanic rock mass. We provide the means to estimate upscaled values of Young’s modulus, Poisson’s ratio, shear modulus, and bulk modulus for a volcanic rock mass that can be improved with laboratory measurements and/or structural assessments of the studied area, but do not rely on them. In the absence of information, we estimate upscaled values of Young’s modulus, Poisson’s ratio, shear modulus, and bulk modulus for volcanic rock with an average porosity and an average fracture density/quality to be 5.4 GPa, 0.3, 2.1 GPa, and 4.5 GPa, respectively. The proposed Young’s modulus for a typical volcanic rock mass of 5.4 GPa is much lower than the values typically used in volcano modelling. We also offer two methods to estimate depth-dependent rock mass Young’s moduli, and provide two examples, using published data from boreholes within Kīlauea volcano (USA) and Mt. Unzen (Japan), to demonstrate how to apply our approach to real datasets. It is our hope that our data and analysis will assist in the selection of elastic moduli for volcano modelling. To this end, our new publication (Heap et al., 2019), which outlines our approach in detail, also provides a Microsoft Excel© spreadsheet containing the data and necessary equations to calculate rock mass elastic moduli that can be updated when new data become available. The selection of the most appropriate elastic moduli will provide the most accurate model predictions and therefore the most reliable information regarding the unrest of a particular volcano or volcanic terrain.Heap, M.J., Villeneuve, M., Albino, F., Farquharson, J.I., Brothelande, E., Amelung, F., Got, J.L. and Baud, P., 2019. Towards more realistic values of elastic moduli for volcano modelling. Journal of Volcanology and Geothermal Research, https://doi.org/10.1016/j.jvolgeores.2019.106684.

Satellites enable widespread, regional or global surveillance of volcanoes and can provide the first indication of volcanic unrest or eruption. Here we consider Interferometric Synthetic Aperture Radar (InSAR), which can be employed to detect surface deformation with a strong statistical link to eruption. Recent developments in technology as well as improved computational power have resulted in unprecedented quantities of monitoring data, which can no longer be inspected manually. The ability of machine learning to automatically identify signals of interest in these large InSAR datasets has already been demonstrated, but data-driven techniques, such as convolutional neutral networks (CNN) require balanced training datasets of positive and negative signals to effectively differentiate between real deformation and noise. As only a small proportion of volcanoes are deforming and atmospheric noise is ubiquitous, the use of machine learning for detecting volcanic unrest is more challenging than many other applications. In this paper, we address this problem using synthetic interferograms to train the AlexNet CNN. The synthetic interferograms are composed of 3 parts: 1) deformation patterns based on a Monte Carlo selection of parameters for analytic forward models, 2) stratified atmospheric effects derived from weather models and 3) turbulent atmospheric effects based on statistical simulations of correlated noise. The AlexNet architecture trained with synthetic data outperforms that trained using real interferograms alone, based on classification accuracy and positive predictive value (PPV). However, the models used to generate the synthetic signals are a simplification of the natural processes, so we retrain the CNN with a combined dataset consisting of synthetic models and selected real examples, achieving a final PPV of 82%. Although applying atmospheric corrections to the entire dataset is computationally expensive, it is relatively simple to apply them to the small subset of positive results. This further improves the detection performance without a significant increase in computational burden (PPV of 100%). Thus, we demonstrate that training with synthetic examples can improve the ability of CNNs to detect volcano deformation in satellite images, and propose an efficient workflow for the development of automated systems.

Forecasting explosive eruptions relies on using monitoring data to interpret the patterns and timescales of magma transport and mixing. In September 2017, a distal seismic swarm triggered the evacuation of around 140,000 people from Agung volcano, Bali. From satellite imagery and 3D numerical models, we show that seismicity was associated with a deep, sub-vertical magma intrusion between Agung and its neighbour Batur. This, combined with observations of the 1963 eruption which caused more than thousand fatalities, suggests a vertically and laterally interconnected system experiencing recurring magma mixing. The geometry of the 2017 dyke is consistent with transport from a deep mafic source to a shallow andesitic reservoir controlled by stresses induced by the topographic load, but not the regional tectonics. The ongoing interactions between Agung and Batur have important implications for interpretation of distal seismicity, the links between closely spaced arc volcanoes, and the potential for cascading hazards., Using seismic data and numerical modelling, here, the authors characterize the three-month period of unrest occurring prior to the 2017 Agung eruption (Bali, Indonesia). They observe a large uplift signal located at ~5 km from Agung summit corresponding to the emplacement of a 10 km deep magma intrusion between Agung edifice and Batur caldera, suggesting a potential magmatic connection between the two volcanic systems.

Automated systems for detecting deformation in satellite InSAR imagery could be used to develop a global monitoring system for volcanic and urban environments. Here we explore the limits of a CNN for detecting slow, sustained deformations in wrapped interferograms. Using synthetic data, we estimate a detection threshold of 3.9cm for deformation signals alone, and 6.3cm when atmospheric artefacts are considered. Over-wrapping reduces this to 1.8cm and 5.0cm respectively as more fringes are generated without altering SNR. We test the approach on timeseries of cumulative deformation from Campi Flegrei and Dallol, where over-wrapping improves classication performance by up to 15%. We propose a mean-filtering method for combining results of different wrap parameters to flag deformation. At Campi Flegrei, deformation of 8.5cm/yr was detected after 60days and at Dallol, deformation of 3.5cm/yr was detected after 310 days. This corresponds to cumulative displacements of 3 cm and 4 cm consistent with estimates based on synthetic data.

The estimation of the volume of volcanic flows during an ongoing eruption is challenging but this information is crucial for improving risk assessment and for forecasting future events. Although previous studies have shown the ability of TanDEM-X satellite data to derive the thickness and the volume of lava flow fields during effusive eruptions, the method has not been explored yet for pyroclastic flows. Using bi-static interferometry, we produce TanDEM-X DEM on Fuego volcano (Guatemala) to measure the significant topographic changes caused by the 3rd June 2018 eruption, which destroyed the town of San Miguel Los Lotes. We estimate the volume of the Pyroclastic Density Currents (PDCs) to be 15.1 ± 4.2 × 106 m3. The deposits are likely to be the source of lahars during future rainy seasons. We identify the main channel of deposition (positive elevation changes) and the source region of pyroclastic material, areas of significant substrate erosion, and vegetation destruction (negative elevation changes). Our results show that the June 3rd 2018 pyroclastic flow was predominantly composed of material which had gravitationally collapsed from a location close to the vent. The eroded material increased the volume of the flow (bulking) and likely caused the run-out distance of the 2018 PDC to be larger than previous eruptions (1999–2017). This study highlights the potential of remote sensing techniques for actively monitoring topography changes in inaccessible locations and to rapidly derive deposit volumes.

The strong tidal triggering of mid-ocean ridge earthquakes has remained unexplained because the earthquakes occur preferentially during low tide, when normal faulting earthquakes should be inhibited. Using Axial Volcano on the Juan de Fuca ridge as an example, we show that the axial magma chamber inflates/deflates in response to tidal stresses, producing Coulomb stresses on the faults that are opposite in sign to those produced by the tides. When the magma chamber’s bulk modulus is sufficiently low, the phase of tidal triggering is inverted. We find that the stress dependence of seismicity rate conforms to triggering theory over the entire tidal stress range. There is no triggering stress threshold and stress shadowing is just a continuous function of stress decrease. We find the viscous friction parameter A to be an order of magnitude smaller than laboratory measurements. The high tidal sensitivity at Axial Volcano results from the shallow earthquake depths., Tidal triggering of earthquakes at Axial Volcano on the Juan de Fuca ridge is shown to be driven by tidally induced magma chamber inflation. Fitting the data to theory requires that the frictional parameter A be much smaller than laboratory measurements indicate.

Recent improvements in the frequency, type and availability of satellite images mean it is now feasible to routinely study volcanoes in remote and inaccessible regions, including those with no ground-based monitoring. In particular, Interferometric Synthetic Aperture Radar (InSAR) data can detect surface deformation, which has a strong statistical link to eruption. However, the dataset produced by the recently-launched Sentinel-1 satellite is too large to be manually analysed on a global basis. In this study, we systematically process >30,000 short-term interferograms at over 900 volcanoes and apply machine learning algorithms to automatically detect volcanic ground deformation. We use a convolutional neutral network (CNN) to classify interferometric fringes in wrapped interferograms with no atmospheric corrections. We employ a transfer learning strategy, and test a range of pretrained networks, finding that AlexNet is best suited to this task. The positive results are checked by an expert and fed back for model updating. Following training with a combination of both positive and negative examples, this method reduced the number of interferograms to ~100 which required further inspection, of which at least 39 are considered 'true positives'. We demonstrate that machine learning can efficiently detect large, rapid deformation signals in wrapped interferograms, but further development is required to detect slow or small deformation patterns which do not generate multiple fringes in short duration interferograms. This study is the first to use machine learning approaches for detecting volcanic deformation in large datasets, and demonstrates the potential of such techniques for developing alert systems based on satellite imagery.

Sentinel-1 represents a major step forward in enabling us to monitor the Earth's hazardous tectonic and volcanic zones. Here, we present the latest progress from the Centre for the Observation and Modelling of Earthquakes, Volcanoes and Tectonics (COMET), where we provide deformation results to the community for volcanoes and the tectonic belts. For the estimation of seismic hazard, we require relative accuracy on the order of 1 mm/yr between points 100 km apart. This requires mass processing of long time series of radar acquisitions. As of January 2018, we are producing in-terferograms systematically for the entire Alpine-Himalayan belt (~9000 × 2000 km) and the majority of subaerial volcanoes. Currently we make interferograms and coherence products available to the community, but we plan to also provide average deformation rates and displacement time series, in the future. The results are made available through a dedicated COMET portal, and we are in the process of linking them to the ESA G-TEP and EPOS. COMET also responds routinely to significant continental earthquakes, larger than ~Mw 6.0. The short repeat interval of Sentinel-1, together with the rapid availability of the data, allows us to do this within a few days for most earthquakes. For example, after the Mw 7.8 Kaikoura earthquake we supplied a processed interferogram to the community just 5 hours and 37 minutes after the Sentinel-1 acquisition. In this paper we provide an overview of some of the latest results for tectonics and volcanism and discuss how the accuracy of these products will improve as the number of data products acquired by Sentinel-1 increases.

We use numerical models to study the mechanical stability of magma reservoirs embedded in elastic host rock. We quantify the overpressure required to open tensile fractures (the failure overpressure), as a function of the depth and the size of the reservoir, the loading by the volcanic edifice, and the pore fluid pressure in the crust. We show that the pore fluid pressure is the most important parameter controlling the magnitude of the failure overpressure rather than the reservoir depth and the edifice load. Under lithostatic pore fluid pressure conditions, the failure overpressure is on the order of the rock tensile strength (a few tens of megapascals). Under zero pore fluid pressure conditions, the failure overpressure increases linearly with depth (a few hundreds of megapascals at 5 km depth). We use our models to forecast the failure displacement (the cumulative surface displacement just before an eruption) on volcanoes showing unrest: Sinabung and Agung (Indonesia) and Okmok and Westdahl (Aleutian). By comparison between our forecast and the observation, we provide valuable constraint on the pore fluid pressure conditions on the volcanic system. At Okmok, the occurrence of the 2008 eruption can be explained with a 1,000 m reservoir embedded in high pore fluid pressure, whereas the absence of eruption at Westdahl better suggests that the pore fluid pressure is much lower than lithostatic. Our finding suggests that the pore fluid pressure conditions around the reservoir may play an important role in the triggering of an eruption by encouraging or discouraging the failure of the reservoir.

Stress transfer between separate magma bodies is evaluated by considering how pressure changes related to magma accumulation/propagation influence the stability of a separate nearby magma body. Three-dimensional numerical models are used to evaluate the stability evolution of a magma body through the calculation of two variables: (i) the variation of the threshold pressure needed to cause failure around the magma body and (ii) the magma pressure change. A parametric study indicates that stress interactions are strongly dependent on the distance between magma bodies as well as the body's shape. Such models are then applied to evaluate stress influence of intrusive activity in 1994, 1999, and 2010 at Eyjafjallajokull volcano, which preceded two eruptions there in 2010. Two cases are considered: influence of these intrusions on (i) a magma reservoir at 20 km distance under the Katla volcano and (ii) a silicic magma body under Eyjafjallajokull. The distance between the Eyjafjallajokull intrusions and the Katla reservoir is sufficiently long to reduce the stress interaction to insignificant levels, with an amplitude of the same order as Earth tides (a few kilopascals). However, cumulative stress transfer due to the intrusions to a remnant silicic shallow body situated below the Eyjafjallajokull is much larger (0.5 2.5 MPa). This mechanical transfer could have contributed to the failure of the silicic body and promoted the chemical mixing/mingling between different magma types, which is commonly interpreted as the main cause of the 2010 explosive eruption of Eyjafjallajokull.

There are some 1,500 volcanoes with the potential to erupt, but most are not instrumentally monitored. However, routine acquisition by the Sentinel‐1 satellites now fulfils the requirements needed for interferometric synthetic aperture radar (InSAR) to progress from a retrospective analysis tool to one used for near‐real‐time monitoring globally. However, global monitoring produces vast quantities of data, and consequently, an automatic detection algorithm is therefore required that is able to identify signs of new deformation, or changes in rate, in a time series of interferograms. On the basis that much of the signal contained in a time series of interferograms can be considered as a linear mixture of several latent sources, we explore the use of blind source separation methods to address this issue. We consider principal component analysis and independent component analysis (ICA) which have previously been applied to InSAR data and nonnegative matrix factorization which has not. Our systematic analysis of the three methods shows ICA to be best suited for most applications with InSAR data. However, care must be taken in the dimension reduction step of ICA not to remove important smaller magnitude signals. We apply ICA to the 2015 Wolf Volcano eruption (Galapagos Archipelago, Ecuador) and automatically isolate three signals, which are broadly similar to those manually identified in other studies. Finally, we develop a prototype detection algorithm based on ICA to identify the onset of the eruption.

International audience; The behaviour of a magma plumbing system during a cycle of volcanic edifice growth is investigated with a simple physical model. Loading by an edifice at Earth's surface changes stresses in the upper crust and pressures in a magma reservoir. In turn, these changes affect magma ascent from a deep source to the reservoir and from reservoir to Earth's surface. The model plumbing system is such that a hydraulic connection is maintained at all times between the reservoir and a deep magma source at constant pressure. Consequently the input rate of magma into the reservoir is predicted by the model rather than imposed as an input parameter. The open hydraulic connection model is consistent with short-term measurements of deformation and seismicity at several active volcanoes. Threshold values for the reservoir pressure at the beginning and end of eruption evolve as the edifice grows and lead to long-term changes of eruption rate. Depending on the dimensions and depth of the reservoir, the eruption rate follows different trends as a function of time. For small reservoirs, the eruption rate initially increases as the edifice builds up and peaks at some value before going down. The edifice size at the peak eruption rate provides a constraint on the reservoir shape and depth. Edifice decay or destruction leads to resumption of eruptive activity and a new eruption cycle. A simple elastic model for country rock deformation is valid over a whole eruptive cycle extending to the cessation of eruptive activity. For large reservoirs, an elastic model is only valid over part of an eruptive cycle. Long-term stress changes eventually lead to reservoir instability in the form of either roof collapse and caldera formation or reservoir enlargement in the horizontal direction.

Most recent SAR sensors use wide band signals to achieve metric range resolution. One can also take advantage of wide band to split it into sub-bands and generate several lower-resolution images, centered on slightly different frequencies, from a single acquisition. This process, named Multi Chromatic Analysis (MCA) corresponds to performing a spectral analysis of SAR images. Split-Band SAR interferometry (SBInSAR) is based on spectral analysis performed on each image of an InSAR pair, yielding a stack of sub-band interferograms. Scatterers keeping a coherent behaviour in each subband interferogram show a phase that varies linearly with the carrier frequency, the slope being proportional to the absolute optical path difference. This potentially solves the problems of phase unwrapping on a pixelperpixel basis. In this paper, we present an SBInSAR processor and its application using TanDEM-X data over the Nyiragongo volcano.

International audience; We reanalyzed the surface displacements observed at the Lazufre volcanic complex in the Southern Andean Central Volcanic Zone using GPS measurements made between 2006 and 2008 and a large InSAR data set. We performed a detailed spatiotemporal analysis of the displacements using a principal component analysis inversion method (PCAIM). The PCAIM reveals a source with no significant changes in shape and dimension and with a remarkably linear strength increase over the whole period of observation (i.e., 2003-2010). Then we used a three-dimensional mixed boundary element method (MBEM) to invert the first component of surface displacement as obtained from PCAIM. We explored a continuum of geometries from a shallow elliptic crack to a deep massive truncated elliptical cone that could represent a sill or a large magma chamber, respectively. The best models indicate a large flat-topped source with a roof area between 40 and 670 km 2 and a depth of between 2 and 14 km below ground surface. Lastly, on the basis of the limited data available for the thermomechanical structure of the crust in the Southern Andean Central Volcanic Zone, we consider some possible scenarios to explain the spatial and temporal pattern of displacements at Lazufre.

Arenal Volcano is a small (~1750 m above sea level, ~10 km3) stratovolcano that continuously erupted between July 1968 and October 2010. During this longlasting eruption (over 42 yr), a large volume of material--~5.6 × 108 m3 of dense rock equivalent--has been extruded and has produced a thick and extended lava fi eld, mainly on the western fl ank of the edifi ce. Measurements of ground deformation obtained using a network of dry-tilt stations are presented for the period 1986-2000. They show a continuous subsidence of the volcano with maximal amplitude on the western side. The load effect of the lava fi eld is calculated and explains the largest part of the observed tilts. Once the data are corrected by this load effect, pressure source models are not supported by the observations and by quality criteria on the models. Although the dry-tilt data from Arenal Volcano give limited constraints on the deformation models, they are representative of a long period of activity that cannot be recovered by other means. Moreover, the corresponding interpretative model is consistent with results obtained by geotechnical studies and modern ground deformation methods like interferometric synthetic aperture radar (InSAR).

Ice retreat on volcanoes reduces pressure at the surface of the Earth and induces stress changes in magmatic systems. The consequences can include increased generation of magma at depth, increased magma capture in the crust, and modification of failure conditions of magma chambers. We review the methodology to evaluate each of these effects, and consider the influence of ongoing ice retreat on volcanoes at the Mid-Atlantic divergent plate boundary in Iceland. Evaluation of each of these effects requires a series of assumptions regarding the rheology of the crust and mantle, and the nature of magmatic systems, contributing to relatively large uncertainty in response of a magmatic system to climate warming and associated ice retreat. Pressure release melting due to ice cap retreat in Iceland may at present times generate a similar amount of magma as plate tectonic processes; larger than realized previously. However, new modelling shows that part of this magma may be captured in the crust, rather than being erupted. Gradual retreat of ice caps do steadily modify failure conditions at magma chambers, which is highly dependent on their geometry and depth, as well as the details of ice load variations. A model is presented where long-term ice retreat at Katla volcano decreases the likelihood of eruption, as more magma is needed in the magma chamber to cause failure than in the absence of the ice retreat.

International audience; Major volcano flank collapses strongly affect the underlying magmatic plumbing system. Here, we consider the magma storage zone as a liquid pocket embedded in an elastic medium, and we perform numerical simulations in two-dimensional axisymmetric geometry as well as in three dimensions in order to evaluate the consequences of a major collapse event. We quantify the pressure decrease induced within and around a magma reservoir by a volcano flank collapse. This pressure reduction is expected to favor replenishment with less evolved magma from deeper sources. We also estimate the impact of the magma pressure decrease, together with the stress field variations around the reservoir, on the eruptive event associated with the edifice failure. We show that, for a given magma reservoir geometry, the collapse of a large strato-volcano tends to reduce the volume of the simultaneous eruption; destabilization of large edifices may even suppress magma emission, resulting in phreatic eruptions instead. This effect is greater for shallow reservoirs, and is more pronounced for spherical reservoirs than for vertically-elongated ones. It is reduced for compressible magmas containing a large amount of volatiles. Over a longer time scale, the modification of pressure conditions for dyke initiation at the chamber wall may also explain an increase in eruption rate as well as an apparent change of magma storage location.