Tectonics/Volcanoes 4

Pico do Fogo, on Fogo Island, is the most active volcano in the Cape Verde archipelago and in the Central Atlantic region, with at least 26 eruptions in the past 500 years. The last two eruptions, in 1995 and 2014-2015, were both characterized by the opening of linear eruptive fissures on the southwester flank of Pico do Fogo cone.

In our work we use interferometric synthetic aperture radar (InSAR) data from European Space Agency’s (ESA) satellites to constrain the geometry of the intrusions that fed the most recent eruptions and to shed light on the plumbing system of the volcano. We first use ERS data to study the 1995 event and reinterpret results from previous studies. We then analyse data from the Sentinel-1A satellite, which had been operative for only a few weeks at the time of the onset of the eruption in November 2014, to study the most recent event. The Sentinel-1 InSAR data was acquired in the TOPS (Terrain Observation by Progressive Scans) mode and the eruption at Fogo represents the first volcanic eruption imaged by Sentinel-1 in its standard acquisition mode. To accurately model TOPS data, variations in both incidence and squint angle of the satellite line-of-sight (LoS) vector should be taken into account when projecting 3D displacements into the LoS direction. Following this approach, we perform a Bayesian inversion of the InSAR data and infer that the measured deformation is best explained by the intrusion of a sub-vertical dike beneath the south-western flank of Pico do Fogo cone. This intrusion seems to have first propagated upwards beneath the cone and subsequently laterally towards the south-western flank of Pico do Fogo, where it reached the surface. Very similar results are also obtained when analysing surface displacements associated with the 1995 eruption, implying that very similar processes must have controlled both eruptions. These findings set the basis for further research aimed at understanding the pattern of eruptive fissures and underlying intrusions that have characterised the past 300 years of activity at Fogo volcano, as well as assessing the stability of the volcanic edifice.

Finally, we use pre- and post-eruptive digital elevation models (DEMs) of the volcano to estimate the volume of the most recent lava flows. The DEMs are formed using synthetic aperture radar imagery from the German Space Agency’s TanDEM-X (TerraSAR add-on for Digital Elevation Measurements) satellite mission and tri-stereo optical imagery from the French National Centre for Space Studies’ Pleiades satellite constellation. Preliminary results show that during the 2014-2015 eruption almost 50 million cubic meters of lava were emplaced at the surface of the volcano over an area of ~5 square kilometres.

For much of the first 20 years of Kīlauea’s 1983–present ERZ (East Rift Zone) eruption, deformation was characterized by subsidence at the summit and along both rift zones. Using a 3D-Mixed-Boundary Element model, we speculate that deep rift opening and basal fault slip caused subsidence of the rift zones. At the end of 2003, however, Kīlauea’s summit began a 4-year-long period of inflation that culminated in an ERZ dike intrusion and small eruption during 17–19 June 2007—the “Father’s Day” (FD) event.The FD event was interpreted as the result of forcible intrusion of magma driven by high pressure within the summit magma storage area, as opposed to a passive response to deep rift zone opening. To investigate the sequence of events that led to this drastic change in deformation behavior, define the active deformation sources and their interactions with faulting; we integrate contemporary geodetic data from InSAR and GPS with double-couple fault-plane solutions for volcano-tectonic earthquakes and Coulomb stress modeling.

Volcán de Colima is a steep sloped explosive stratovolcano located in southern Central Mexico, and one of the most active volcanoes in North America. Since the beginning of the most recent eruptive period in 1998, the type of activity has been varying between predominantly explosive, dome building and dome collapse, interrupted by longer periods of no eruptions. Since 2013, a sequence of domes and lava flows were extruded from the summit crater, until a large explosion in July 2015 produced an amphitheater-shaped crater, partly open to the South.

Since the beginning of 2012, we have acquired a combination of radar image datasets, including TerraSAR-X spotlight, Cosmo SkyMed spotlight and stripmap, as well as Sentinel-1 TOPS data. The presented acquisition modes have a wide range of spatial resolution and temporal coverage, ranging from 1-day revisit pairs of Cosmo SkyMed, to 11-day intervals, to monthly acquisitions from background acquisition programs. Here we analyze the recent eruption episodes of Volcán de Colima by using a variety of data analysis methods, including Persistent Scatterers (PS), radargrammetric DEM and pixel offset (PO) time series.

Volcan de Colima has also been intensely monitored by ground and near-ground observations, such as optical and infrared camera installations and overflights, that were recording time lapse imaging data at high resolution since 2012. This data we analyzed using digital image correlation and photogrammetric approaches, such as Structure from Motion, allowing the calculation of elevation models and mapping of displacements occurring on the lava flows and the summit dome.

We combine and compare results derived from satellite radar datasets and from the camera monitoring system, and investigate three main stages at this highly active volcano: (i) the preparation phase before eruptions as observed by inflation of the summit region and associated surface displacements, (ii) the eruption phase as shown by dome extrusion and lava flow formation quantified by pixel velocity mapping and topographic change monitoring, and (iii) the posteruptive cooling phase as observed by opening of cooling fractures and thermal contraction of the summit. Techniques and processes investigated can be similarly applied to other volcanoes and show the strong value of remote sensing datasets at high resolution.

On April 3, 2014, the new Senitnel-1 mission was successfully launched and it started its operational life. The onboard SAR instruments is capable to acquire images of the Earth surface with e new mode: the Terrain Observation with Progressive Scan (TOPS) SAR. TOPSAR can take wide swath image of large areas (250 km of swath) with medium-high (about 12 m per pixel) spatial resolution.

The first monitored natural events were the Mw 6.0 Napa (California, USA) earthquake and the Pico de Fogo (Cape Verde) volcano eruption. This latter is the object of our study.

On 23rd of November, after almost 20 years from the previous eruption, the volcano began its activity, discharging a big amount of lava in the caldera. The village of Chã das Calderias was evacuated, saving the population (about 1000 inhabitants) but the houses and the unique connection road were damaged by the lava flow.

We used two pair of Sentinel-1 images in TOPSAR mode, along ascending and descending orbit.  The first pair is dated November 3 and November 27, and the second pair was acquired on November 8 and December 2. The Sentinel-1 images were exploited to estimate the emplacement of the lava filed between the beginning of the eruption and 2 December 2014. It is worth to point-out that the lava field was partially covering the caldera and that the amount of changed pixels was a very small fraction of the entire image. This latter is a frequent issue with images with a huge swath width such as Sentinel-1 mission, preventing the application of statistical based approaches because it makes impossible to model the distribution of changed pixels.

Here we propose a novel technique, previously applied for flood mapping purposes, based on a multiscale tiling of the change image to identify bimodal regions, meaning that changed and unchanged pixels are well represented in the area to fit their distribution. Later, based on the statistic of previously selected areas, the extraction of the lava extent in the entire scene is performed by using an automatic algorithm which is a combination of distribution fitting, backscatter thresholding and region growing.

The resulting lava-emplacement maps were also compared with external information for validation purposes.

The main advantage of the approach is that it renders the applicability of statistical-based change detection approaches independent from the different technical characteristics of the image scene such as spatial resolution or percentage of the changes extension. 

Volcanic eruptions are the surface expressions of magma movements at depth. Analyzing ground deformations associated with volcanic eruptions allows to understand the mechanisms of magma emplacement and individual magma plumbing systems.

InSAR  is particularly well suited tool for studying ground displacement in areas that are difficult to access for geographical, economical or political reasons, and little or not monitored with complementary ground-based instruments.

This is the case of Nyamulagira, an active shield volcano with a central caldera, located in the eastern part of the Democratic Republic of Congo, in the western branch of the East African Rift System. From 1948 to 2012, Nyamulagira showed a particular eruptive cycle with a short-lived flank eruption, sometimes accompanied with intracrateral activity, every 1 - 4 years on average. Less frequent long-lived eruptions usually emitting larger volumes of lava may also take place at larger distance (>8 km) from the central caldera. The 2011-2012 Nyamulagira eruption is of that last type. This eruption began on November 4 2011 and ended sometimes in April 2012; the exact date couldn’t be determined precisely by visual or ground-based observations, mainly because of the political unrest affecting the region at that time.

In the framework of the REmote Sensing and In Situ detection and Tracking of geohazards (RESIST) project we used InSAR data from different satellite (Envisat, Cosmo SkyMed, Terrasar-X and Radarsat) to measure pre-, co and post-eruptive ground displacement associated with the Nyamulagira 2011-2012 eruption. By inverting the measured ground deformations using analytical models, we evaluated the source parameters and the mechanism of magma emplacement for this eruption. We also compared InSAR measurements with available seismic and GPS data in order to constrain the temporal evolution of the eruption. The three datasets suggest the presence of a deep magmatic source that possibly fed the shallower magmatic system. This mechanism involving a deep source contrasts with the usual shallow plumbing system identified during the recent flank eruptions.