Search

Automated systems for detecting deformation in satellite interferometric synthetic aperture radar (InSAR) imagery could be used to develop a global monitoring system for volcanic and urban environments. Here, we explore the limits of a convolutional neural networks for detecting slow, sustained deformations in wrapped interferograms. Using synthetic data, we estimate a detection threshold of 3.9 cm for deformation signals alone and 6.3 cm when atmospheric artifacts are considered. Overwrapping reduces this to 1.8 and 5.2 cm, respectively, as more fringes are generated without altering signal to noise ratio. We test the approach on time series of cumulative deformation from Campi Flegrei and Dallol, where overwrapping improves classification 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 8.5 cm/year was detected after 60 days and at Dallol, deformation of 3.5 cm/year was detected after 310 days. This corresponds to cumulative displacements of 3 and 4 cm consistent with estimates based on synthetic data. We adapt a pretrained convolutional neural networks for detecting slow, steady deformation in InSAR imagesSynthetic tests show the detection threshold is 6.3 cm in volcanic environments but is reduced to 5.0 cm by decreasing wrapping intervalWe demonstrate the method on cumulative time series of interferograms from Campi Flegrei, Italy (8.5 cm/yearr) and Dallol, Ethopia (3.5 cm/year)

We use ALOS‐1 interferometric synthetic aperture radar data spanning the period of 2007–2011 to obtain time‐dependent ground deformation data over Taal caldera related to the 2010–2011 volcanic unrest, which did not lead to an eruption. We employ finite element models that account for subsurface thermomechanical properties to test for viscoelastic deformation. We also examine the variability of crustal rheological parameters of the temperature‐dependent Arrhenius formulation for viscosity, to investigate the effects on surface deformation. Deformation at Taal is observed to be time dependent and explained by a linear rate of pressure increase in a shallow magma reservoir at 5‐km depth within a hot and viscoelastic upper crust. The best‐fitting rheological properties of the local setting are consistent with either a felsic or hydrothermally altered uppermost crust. We show the significantly different effects that the variation of rheological parameters has on surface deformation, highlighting the importance of incorporation into future studies of deformation at different volcanic settings. Deformation during the 2010–2011 unrest at Taal volcano is explained by inflation of a shallow magmatic sourceNumerical models of viscoelastic deformation revealed rheological properties that fit the local geologic settingSurface deformation effects from the variability of Arrhenius parameters were demonstrated for different volcanic settings

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 data can detect surface deformation, which has a strong statistical link to eruption. However, the data set produced by the recently launched Sentinel‐1 satellite is too large to be manually analyzed 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 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 data sets and demonstrates the potential of such techniques for developing alert systems based on satellite imagery. We present a machine learning framework to detect volcanic ground deformation in wrapped interferograms using convolutional neural networksThe classification model is initialized with Envisat data set and then tested and retrained with Sentinel‐1 data set covering over 900 volcanoesThis framework can reduce the number of interferograms for manual inspection from more than 30,000 to approximately 100

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. Numerical calculation of the failure overpressure required to cause magma intrusion is dependent on pore fluid pressureHigh pore fluid pressure encourages eruptions by reducing the failure overpressureDifferent pore fluid pressure conditions can explain the difference of eruptive behavior between volcanoes

Nyamulagira and Nyiragongo are two of the most active volcanoes in Africa, but their eruptive histories are poorly known. Assessing lava flow volumes in the region remains difficult, as field surveys are often impossible and available Digital Elevation Models (DEMs) do not have adequate spatial or temporal resolutions. We therefore use TerraSAR‐X add‐on for Digital Elevation Measurement (TanDEM‐X) interferometry to produce a series of 0.15 arc sec (∼5 m) DEMs from between 2011 and 2012 over these volcanoes. TanDEM‐X DEMs have an absolute vertical accuracy of 1.6 m, resulting from the comparison of elevation with GPS measurements acquired around Nyiragongo. The difference between TanDEM‐X‐derived DEMs from before and after the 2011–2012 eruption of Nyamulagira provides an accurate thickness map of the lava flow emplaced during that activity. Values range from 3 m along the margins to 35 m in the middle, with a mean of 12.7 m. The erupted volume is 305.2 ± 36.0 × 106m3. Height errors on thickness depend on the land covered by the flow and range from 0.4 m in old lavas to 5.5 m in dense vegetation. We also reevaluate the volume of historical eruptions at Nyamulagira since 2001 from the difference between TanDEM‐X and SRTM 1 arc sec DEMs and compare them to previous work. Planimetric methods used in literature are consistent with our results for short‐duration eruptions but largely underestimate the volume of the long‐lived 2011–2012 eruption. Our new estimates of erupted volumes suggest that the mean eruption rate and the magma supply rate were relatively constant at Nyamulagira during 2001–2012, respectively, 23.1 m3s−1and 0.9 m3s−1. TanDEM‐X DEMs were used to quantify 2011–2012 lava flow volume at NyamulagiraFor long‐lived eruptions, lava thickness is much larger than initially supposedMagma supply rate at Nyamulagira is low and constant during the 2001–2012 period

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 Eyjafjallajökull 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 Eyjafjallajökull. The distance between the Eyjafjallajökull 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 Eyjafjallajökull 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 Eyjafjallajökull. Stress transfers between magma bodies are simulated through 3‐D mechanical modelsMagma movements during an eruption promote failure of magma reservoirs nearbyThis process may explain the two successive eruptions at Eyjafjallajokull in 2010

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 Volcán 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 km2and had a thickness of 10.5 ±2 m in the lower sections as estimated from radar shadows. The 3 June eruption also produced backscatter signal over an area of 40 km2, 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 characterization 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. Volcanic eruptions cause changes to the Earth's surface that can be observed using satellite‐based radar instruments. Changes to the radar scattered back from the surface can be caused by new volcanic deposits or changes to the ground caused by an eruption. However, such signals are also affected by what was there before the eruption and variations in the satellite positions. Deposits from explosive eruptions are particularly hard to identify because they can affect radar signals in different ways. We use a high spatial and temporal resolution satellite radar dataset to identify different volcanic deposits from the June 2018 eruption of Volcán de Fuego, Guatemala. We were able to identify three types of volcanic deposit from radar backscatter: hot mixture of solid particle and gas (pyroclastic density currents), water saturated flow with solid particles (lahars), and ash. We observe backscatter changes over six drainage systems that were affected by pyroclastic density currents in June 2018. Combining our radar dataset with rainfall data, we identify nine periods of lahars in one drainage between June and October 2018. We describe what information and corrections are helpful to identify volcanic changes in backscatter and especially to monitor the progression of an explosive eruption. Radar backscatter observed 3 pyroclastic density currents and 9 periods of lahar activity in a single drainage system between January and October 2018Backscatter noise is reduced by up to 42% by using dense time series, which aids in the extraction of subtle signals in explosive depositsBackscatter corrections and understanding of pre‐eruption scattering properties are necessary for meaningful analysis of explosive deposits Radar backscatter observed 3 pyroclastic density currents and 9 periods of lahar activity in a single drainage system between January and October 2018 Backscatter noise is reduced by up to 42% by using dense time series, which aids in the extraction of subtle signals in explosive deposits Backscatter corrections and understanding of pre‐eruption scattering properties are necessary for meaningful analysis of explosive deposits

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 Interferometric Synthetic Aperture Radar (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 1 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 use of high‐resolution weather models is crucial for correcting atmospheric signals in Sentinel‐1 interferograms at tropical volcanoesSequential anomaly methods on corrected time series flag uplift at Agung with 95% confidence 1 month before the eruptionOur study contributes to the development of tools suitable for real‐time monitoring of volcanic unrest in remote areas

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. Through monitoring volcanoes, we can detect signs that they may be likely to erupt in the future, yet many of the world's volcanoes remain unmonitored. However, the latest generation of RADAR satellites produces regular measurements of ground displacement around the majority of the world's volcanoes, and these measurements can be used to detect signs of unrest at many previously unmonitored volcanoes. As the volume of data produced in this undertaking is vast, an automatic detection algorithm is required to detect these signals of interest. In this work, we present methods to separate signals indicative of volcanic unrest from other nuisance signals and to then monitor the geophysically important signals to detect if a volcano has entered a period of unrest. Independent component analysis (ICA) is the best of a suite of blind signal separation methods for use with InSAR dataIndependent component analysis can be used to automatically isolate signals of geophysical interest during the 2015 Wolf Volcano eruptionThrough monitoring the changes in the signals recovered by ICA, signs of unrest at a volcano can be automatically detected