Earth and Space Science Open Archive This work has been accepted for publication in Journal of Geophysical Research - Solid Earth. Version of RecordESSOAr is a venue for early communication or feedback before peer review. Data may be preliminary. Learn more about preprints. preprintOpen AccessYou are viewing an older version [v1]Go to new versionWhat triggers caldera ring-fault subsidence at Ambrym volcano? Insights from the 2015 dike intrusion and eruptionAuthorsTaraShreveiDRaphaëlGrandiniDDelphineSmittarelloValérieCayolVirginiePineliDMarieBoichuiDYuMorishitaSee all authors Tara ShreveiDCorresponding Author• Submitting AuthorInstitut de Physique du Globe de ParisCarnegie Institution for ScienceiDhttps://orcid.org/0000-0003-2103-2434view email addressThe email was not providedcopy email addressRaphaël GrandiniDInstitut de Physique du Globe de ParisiDhttps://orcid.org/0000-0002-1837-011Xview email addressThe email was not providedcopy email addressDelphine SmittarelloUniversity Grenoble Alpes, University Savoie Mont BlancEuropean Center for Geodynamics and Seismologyview email addressThe email was not providedcopy email addressValérie CayolLaboratoire Magmas et Volcansview email addressThe email was not providedcopy email addressVirginie PineliDInstitut de Recherche pour le DéveloppementiDhttps://orcid.org/0000-0002-4928-9584view email addressThe email was not providedcopy email addressMarie BoichuiDUniversité de LilleiDhttps://orcid.org/0000-0003-3163-8325view email addressThe email was not providedcopy email addressYu MorishitaGeospatial Information Authority of Japan,Hokkaido Universityview email addressThe email was not providedcopy email address
Les systemes volcaniques rift-caldera basaltiques fournissent les conditions propices a l’etude de plusieurs processus volcaniques, comme le transport de magma, les effondrements de caldera et le remplissage magmatique. Certaines des plus grandes calderas dans le monde, cependant, sont situees dans des regions isolees dont l’acces peut etre dangereux ou logistiquement complexe. La teledetection des deformations du sol, du degazage et des anomalies thermiques, offre une alternative pour y suivre l’activite volcanique. Ambrym, une ile volcanique du Vanuatu isolee mais tres active, a subi de nombreux episodes de deformation du sol au cours des 20 dernieres annees. Depuis 2015, deux eruptions ont eu lieu a l’interieur de sa caldera de 12 km de diametre. La premiere eruption a eu lieu apres 15 ans de degazage passif et d’activite des lacs de lave. L’eruption la plus recente, en decembre 2018, a vidange les lacs de lave des crateres sommitaux, provoquant l’intrusion d’un volume >0.4 km3 dans la zone de rift sud-est. Le dike engendre a parcouru une distance de plus de 20 km, et s’est ouvert de plus de 4 metres en profondeur. La vidange du magma a produit une subsidence de la caldera a grande echelle, associee a une activation des failles bordant la caldera, et a alimente une eruption sous-marine de ponces basaltiques. Une eruption plus modeste a eu lieu en fevrier 2015, activant egalement une portion de la caldera, et extrayant du magma depuis une chambre situee a une profondeur de ∼4.1 km. Ces deux evenements confirment que les failles bordieres de la caldera d’Ambrym sont des structures actives. L’activite de ces failles contribue a la topographie de la caldera d’Ambrym, dont le mecanisme de formation est discute (eruption Plinienne initiale a 2ka, suivie d’eruptions phreatomagmatiques). La detection d’une activation des failles de caldera par la geodesie spatiale nous permet de formuler l’hypothese que des centaines d’intrusions de tailles moderees a grandes peuvent contribuer a un approfondissement de la caldera, en drainant le magma stocke temporairement sous la caldera d’Ambrym. Outre ces deux evenements eruptifs majeurs, nous mettons en evidence deux episodes (2004–2007, 2015–2017) de subsidence rapide (∼10 cm an-1), mesures par InSAR. Aucune de ces deux periodes n’est associee a une eruption repertoriee. A partir des informations glanees au cours des eruptions de 2015 et 2018 (e.g., la profondeur des zones de stockage, le volume des intrusions, la relation entre l’activite eruptive et les lacs de lave), nous explorons les mecanismes physiques pouvant expliquer cette subsidence inter-eruptive. L’episode de 2004–2007 est probablement associe a une intrusion de dike (en l’absence d’eruption), engendrant la depressurisation d’un sill superficiel, hydrauliquement connecte aux lacs de lave. A partir d’un modele theorique propose par Girona et al, 2014, en couplant le degazage passif (mesure par spectroscopie satellitaire) et la depressurisation du reservoir magmatique (deduite de la geodesie spatiale), nous proposons que l’episode de 2015–2017 ait pour origine la depressurisation d’un reservoir magmatique de grande taille (>10 km3). Par contraste, de courtes periodes de soulevement pourraient etre limite aux periodes de temps pendant lesquelles le systeme est ferme, par exemple en 2019–2020 apres l’episode de vidange des lacs de lave en 2018, et potentiellement en 2007–2010 a la suite d’un evenement d’intrusion non repertorie en 2005. En comparant les phases de regain d’activite d’Ambrym avec celles observees dans les autres systemes rift-caldera basaltiques (Kīlauea, Barðarbunga, Sierra Negra, etc.), les resultats obtenus dans cette dissertation permettent de mieux comprendre l’activite des lacs de lave, le developpement de la caldera, les processus de deformation induits par le degazage, le remplissage magmatique, et la geometrie et le fonctionnement des systemes rift-caldera basaltiques.
The following files were used in the analysis from "Trapdoor fault activation: a step towards caldera collapse at Sierra Negra, Galápagos, Ecuador", Journal of Geophysical Research: Solid Earth: alos2_csk/alos2_sm1_dsc_20180504_20180713/: Includes DEM used in processing of the ALOS-2 SM1 descending interferogram spanning 4 May 2018–13 July 2018 (dem.2alks_2rlks.crop.*); geocoded SAR offsets in pixels (range resolution=1.43 m/pixel; azimuth resolution=2.01 m/pixel; denseOffsets.bil.2alks_2rlks.geo*); geocoded SNR of SAR offsets (denseOffsets_snr.bil.2alks_2rlks.geo.*); geocoded, unwrapped interferometric phase (filt_topophase.unw.2alks_2rlks.geo.*); geocoded incidence and heading angle for the interferogram (los.rdr.2alks_2rlks.geo.*); all in ISCE format. alos2_csk/alos2_sm3_asc_20180114_20180701/: Includes DEM used in processing of the ALOS-2 SM3 ascending interferogram spanning 14 January 2018–1 July 2018 (dem.crop.*); geocoded, unwrapped interferometric phase (filt_topophase.unw.geo.*); geocoded incidence and heading angle for the interferogram (los.rdr.geo.*); all in ISCE format. alos2_csk/alos2_wd1_dsc_147_180518_180629/: Includes DEM used in processing of the ALOS-2 WD1 descending interferogram spanning 18 May 2018–29 June 2018 (crop.dem.*); geocoded, unwrapped interferometric phase (filt_180629-180518_2rlks_14alks.unw.geo.*) ; geocoded incidence and heading angle for the interferogram (180629-180518_2rlks_14alks.los.geo.*); geocoded coherence for the interferogram (180629-180518_2rlks_14alks.cor.geo.*); geocoded mask for the interferogram (filt_topophase.unw.masked.geo.*); all in ISCE format. alos2_csk/csk_asc_20180617_20180703/: Includes DEM used in processing of the COSMO-SkyMed ascending interferogram spanning 17 June 2018–3 July 2018 (dem.crop.*); geocoded SAR offsets in pixels (range resolution=1.54 m/pixel; azimuth resolution=2.48 m/pixel; denseOffsets.bil.geo.*); geocoded SNR of SAR offsets (denseOffsets_snr.bil.geo); geocoded incidence and heading angle for the interferogram (los.rdr.geo.*); all in ISCE format. alos2_csk/csk_asc_20180703_20180719/: Includes DEM used in processing of the COSMO-SkyMed ascending interferogram spanning 3 July 2018–19 July 2018 (dem.crop.*); geocoded SAR offsets in pixels (range resolution=1.54 m/pixel; azimuth resolution=2.48 m/pixel; denseOffsets.bil.geo.*); geocoded SNR of SAR offsets (denseOffsets_snr.bil.geo); geocoded incidence and heading angle for the interferogram (los.rdr.geo.*); all in ISCE format. alos2_csk/csk_dsc_20180618_20180704/: Includes DEM used in processing of the COSMO-SkyMed descending interferogram spanning 18 June 2018–4 July 2018 (dem.crop.*); geocoded SAR offsets in pixels (range resolution=1.70 m/pixel; azimuth resolution=2.45 m/pixel; denseOffsets.bil.geo.*); geocoded SNR of SAR offsets (denseOffsets_snr.bil.geo); geocoded incidence and heading angle for the interferogram (los.rdr.geo.*); all in ISCE format. alos2_csk/csk_dsc_20180704_20180720/: Includes DEM used in processing of the COSMO-SkyMed descending interferogram spanning 4 July 2018–20 July 2018 (dem.crop.*); geocoded SAR offsets in pixels (range resolution=1.70 m/pixel; azimuth resolution=2.45 m/pixel; denseOffsets.bil.geo.*); geocoded SNR of SAR offsets (denseOffsets_snr.bil.geo); geocoded incidence and heading angle for the interferogram (los.rdr.geo.*); all in ISCE format. S1.zip: Unwrapped, geocoded interferometric phase in meters for Sentinel-1 ascending and descending interferograms, spanning time periods of interest. S1_20180630_20180706_asc_mask_nan_ref.grd: Unwrapped, geocoded, and masked interferometric phase for Sentinel-1 ascending interferogram spanning 30 June 2018–6 July 2018. S1_20180701_20180707_desc_mask_nan_ref.grd: Unwrapped, geocoded, and masked interferometric phase for Sentinel-1 descending interferogram spanning 1 July 2018–7 July 2018. tandemx12m_crop.grd: TanDEM-X 12 meter DEM in meters. pleaides_tandemx12m_diff.grd: Difference between the TanDEM-X 12 meter DEM and the Pléiades-derived DEM, computed from images on 29 October 2018 and 6 December 2019. trapdoorFaultSlip.zip: Discretized trapdoor fault patch dip-slip modeled to fit deformation from Sentinel-1 ascending interferograms, estimated using the Classic Slip Inversion software. trapdoorFaultTraces.zip: Caldera and trapdoor fault traces, derived from Bell et al. 2021. SN14_tilt_10s_2018-19.txt: Text filt containing date-time (sampled at 10 s, in matplotlib date-time number format), N-S tilt and E-W tilt. Tilt values can be converted to microradians by multiplying by a factor of 0.00129. Tilt data obtained from authors of Bell et al. 2021. For use of this dataset, please cite https://doi.org/10.1038/s41467-021-21596-4.
<p><span>Located about 30 km North of the city of Yogyakarta on Java island, Merapi is considered one of the most dangerous dome building stratovolcanoes, as about 2 million people live less than 30 km away from the crater. Its recent eruptive activity consists in cyclic effusive growth of viscous lava domes, followed by partial or total destruction of domes. Dome destruction favors gravitational collapses (VEI 2) every 4-5 years, or bigger explosive eruptions (VEI 3-4) every 50-100 years resulting in pyroclastic density currents (PDCs) driven downhill at high velocities that are a major risk for surrounding population. Therefore, it is crucial to assess precisely the location, the shape, the thickness, and the volume of emplaced lava in order to prevent populations from sudden PDCs. </span></p><p><span>The last major explosive eruption (VEI 3-4) occurred in November 2010, resulting in a horseshoe-shaped crater of 500 m wide and 250 m depth hosting a lava dome shaped like a plateau. Within the crater, a new dome appeared on 11 August 2018 and was partially destroyed as of late 2019. In this study, we take advantage of 2 high resolution remote-sensing datasets, Pl&#233;iades (optical acquisitions in tri-stereo mode, 1 m resolution) and TanDEM-X (radar acquisitions in StripMap mode, 2 m resolution), to produce 19 Digital Elevation Models (DEMs) between July 2018 and December 2019. We calculate the difference in elevation between each DEM and a reference DEM derived from Pl&#233;iades images acquired in 2013 in order to track the evolution of the dome in the crater between 2018 and 2019. Uncertainties are quantified for each dataset. We show that the DEMs derived from Pl&#233;iades (optical) and TanDEM-X (radar) data are consistent with each other and provide good spatio-temporal constraints on the evolution of the dome. Furthermore, the remote-sensing estimate of lava volume is consistent with local drone measurements carried on by BPPTKG at the time of dome growth. </span></p><p><span>The time period covered by the TanDEM-X data is larger than that covered by the Pl&#233;iades acquisitions, allowing coverage of the growth and destruction of the dome. However, the Pl&#233;iades data allow us to evidence an accumulation zone below the crater that is not well imaged by TanDEM-X. We show the dome reached 40 meters (+-5 m) high and 0.5 Mm</span><sup><span>3 </span></sup><span>(+- 0.1Mm</span><sup><span>3 </span></sup><span>) between August 2018 and February 2019, corresponding to an effusion rate of 3000 m</span><sup><span>3</span></sup><span>/day. Its shape was initially radial</span><span>,</span><span>then extended asymmetrically to the northwest and southeast from October 2018. From February 2019 onwards, the dome elevation remained constant, but lava was continuously emitted, as evidenced by TanDEM-X amplitude maps. Lava supply was balanced by destabilization southwards downhill in an accumulation zone of 400 meters long and 15 meters (+-5m) high maximum. In late 2019, several minor explosions partially destroyed the center of the dome. This study highlights the strong potential of the combination of TanDEM-X and Pl&#233;iades DEMs to quantitatively monitor domes at andesitic stratovolcanoes.</span></p>
<p>Satellite-based UV spectrometers can constrain sulphur dioxide (SO<sub>2</sub>) fluxes at passively degassing volcanoes over decadal time scales. From 2005 to 2015, more than 15 volcanoes had mean passive SO<sub>2 </sub>fluxes greater than 1 kiloton per day. Although the processes responsible for such high emission rates are not clearly established, this study aims to investigate the impact of strong degassing on the pressurization state of volcanic systems and the resulting ground deformation. One possible result of high degassing rates is the depressurization of the region where the melt releasing gas is stored, which may result in subsidence at the Earth&#8217;s surface. Passive degassing may depressurize pathways between deep and shallow magma storage regions, resulting in magma ascent and possibly eruption.</p><p>A lumped-parameter model developed by Girona et al., 2014 couples the mass loss by passive degassing with reservoir depressurization in an open volcanic system. However, this model has yet to be tested using real measurements of gas emissions and ground deformation. In our study, we focus on Ambrym volcano, the past decade&#8217;s top passive emitter of volcanic SO<sub>2</sub>, which exhibits intriguing long-term subsidence patterns and no obvious pressurization preceding eruptive periods. We compare subsidence rates measured by InSAR to the system&#8217;s average daily SO<sub>2</sub> flux, focusing on a subsidence episode spanning 2015 to 2017 that is not clearly linked to magma removal from the system. Using realistic input parameters for Ambrym&#8217;s system constrained by petrology and gas geochemistry, a range of reservoir volumes and conduit radii are explored. Large reservoir volumes (greater than 30 km<sup>3</sup>) and large conduit radii (greater than 300 m) are consistent with depressurization rates obtained from geodetic modelling of InSAR measurements using the Boundary Element method. By comparing these values of reservoir volume and conduit radius with those estimated from geodesy, gas geochemistry, and seismology, we test the applicability and discuss uncertainties of the aforementioned lumped-parameter physical model to interpret the long-term subsidence at Ambrym volcano as a result of sustained passive degassing.</p>
