On 19 September 2021, the island of La Palma, one of the Canary Islands off the northwestern coast of Africa, saw its first volcanic activity after about 50 years of dormancy.
The eruption lasted about 3 months and created a new volcanic cone, named Tajogaite, amid the island’s Cumbre Vieja volcanic ridge. It released large plumes of ash and gas, as well as roughly 200 million cubic meters of lava that spread over 12 square kilometers of the island, burying or severely damaging urban settlements and roads. The event forced the evacuation of about 7,000 people and caused more than 842 million euros in damages.
Data reported by gas geochemists and geophysicists had foretold a possible eruption (although its likely timing was uncertain) since 2017. Research identified an anomalous gas composition in a spring used as a monitoring site, as well as dispersed seismic activity caused by the magmatic reactivation of the volcano (Figure 1). However, actual volcanic precursors, such as well-localized ground deformation, seismicity, and gas emissions, were detected only 8 days before the eruption [D’Auria et al., 2022].
Real-time seismic readings, satellite-based analyses, and other data collected on site during the eruption of Tajogaite provided vital information not only for keeping surrounding communities safe but also for guiding and managing associated scientific efforts. By comparison, the first petrographic and geochemical analyses of Tajogaite lavas, done using rocks collected only in the first week of the eruption, were published in early 2022 for the purpose of describing the erupted material. Months later, other petrological and geochemical studies of the eruption were published, describing the conditions of origin of the magmas, their rheological properties, and their dynamics during the eruption.
Petrologists’ work with respect to volcanoes can be likened to conducting autopsies.
Petrologists’ work with respect to volcanoes can be likened to conducting autopsies, providing insights into the causes and behavior of eruptions after the fact. This information is highly valuable, but the typically lengthy time lags before it is available reduce the usefulness of petrological monitoring as an effective scientific tool during volcanic eruptions.
In recent years, however, some scientists—including us—have been investigating potential applications of petrological approaches in near-real-time volcanic monitoring to help understand what’s happening underground and manage eruption responses. Questions we’ve considered include, for example, whether petrology can be used to study variations in magma ascent paths during eruptions and what additional information can be gleaned by combining frequent petrological analyses of erupted material with monitoring of syneruptive seismicity (seismic activity occurring during an eruption).
The Tajogaite eruption was an excellent opportunity to test our approach [Zanon et al., 2024]. To do so, we collected samples daily from the eruptive area and sent them to a laboratory for rapid analysis of fluid inclusions within the samples (Figure 2).
Visualizing a Volcano’s Interior
Petrological studies, with help from geochemistry, petrography, mineralogy, and thermodynamics, allow scientists to gain deep understanding of the conditions in which rocks form and transform. For example, they can decipher the conditions of magma formation and develop complex, accurate genetic models for volcanoes. And they can reconstruct magma ascent paths and precisely visualize volcanic anatomy.
Although these sorts of studies can be applied only after materials have been erupted, they provide a level of detail not achievable with other methods. They can also complement the work of geophysicists and geochemists, who rely upon networks of high-quality sensors distributed on volcanoes for signs of impending eruptions. These sensors can, for instance, observe the propagation of volcanic dykes by tracking the migration of seismic swarm hypocenters, observe ground deformation caused by volcanic activity, and detect other sorts of anomalies.
During eruptions, real-time acquisition and interpretation of petrological data can provide missing information.
Once an eruption begins, however, analyzing some geophysical and geochemical signals becomes more challenging. When magma starts flowing rapidly through the magmatic system, these signals may no longer fully capture internal processes with sufficient rapidity. As an example, the seismic “noise” caused by the movement of gas bubbles through the magmatic system and by the gas released from it hides the signals of low-intensity seismicity. In addition, sensors can be damaged or destroyed by explosions or buried by lava flows and volcanic ash, further limiting the availability of these signals. During eruptions, real-time acquisition and interpretation of petrological data can provide missing information.
Scientific personnel in volcanological observatories worldwide understand the importance of efficient and rapid petrological monitoring. This monitoring currently combines multiple analytical techniques, including basic petrographic observations using optical microscopy and other analyses to track variations in magma chemical composition and mineralogical constitution during an eruption. This approach has been applied to study several eruptions in the past 25 years, including the Etna flank eruptions in 2001 and 2002–2003, the 2011–2012 submarine eruption at El Hierro, and violent periodic explosions of Stromboli [Corsaro et al., 2007; Andronico et al., 2009; Martí et al., 2013].
However, several obstacles to real-time petrological monitoring are challenging to resolve. Sample collection, preparation, and analysis, as well as data processing, are time-consuming. Some observatories do not possess the necessary instrumentation to perform analyses themselves and therefore must rely on preexisting scientific collaborations or commercial analytical laboratories, which may be far from the site, booked for other research, or expensive to utilize. In addition, analyzing single samples on a daily (or frequent) basis over an extended period—as needed for petrological monitoring during an eruption—is never cost-effective.
Furthermore, the architecture of magmatic systems can be highly complex [Cashman et al., 2017; Sparks et al., 2019], and the information gathered from monitoring approaches may not be useful immediately without an existing thorough understanding of the magma’s path to the surface.
Fluid Inclusions Provide a Way Forward
Microthermometric analysis of many crystals allows us to constrain the depths where magma ponded and reveal the underground architecture of volcanoes.
The 2021 eruption on La Palma occurred along a small eruptive fissure of the Cumbre Vieja volcanic ridge. This ridge consists of a roughly north–south-oriented series of monogenetic cones (each generated during a single eruptive episode) along the dominant volcano-tectonic system on the island.
Olivine and pyroxene crystals in the lavas of this volcano often contain fluid inclusions—small amounts of fluids (mainly carbon dioxide and water) trapped in minerals as they form or as a result of recrystallization (Figure 3).
These fluid inclusions offer valuable information that can be revealed using a technique called microthermometry. A microthermometric study involves heating and cooling mineral crystals under a microscope to determine the temperatures of phase changes in fluid inclusions. These temperatures provide insights into the barometric conditions of crystallization and/or reequilibration specific to the host crystal—that is, the pressures under which these processes occurred, which relate to the depth below the surface. Microthermometric analysis of many crystals thus allows us to constrain the depths where magma ponded and reveal the underground architecture of volcanoes.
Fluid inclusions have been used successfully to reconstruct the magma storage systems of volcanoes in various locations, including the Aeolian Islands, the Azores, the Canary Islands, Cape Verde, and Piton de La Fournaise [e.g., Boudoire et al., 2019; Hildner et al., 2012; Klügel et al., 2000; Zanon et al., 2003, 2020]. These studies used rock samples from different eruptions that occurred within narrow time intervals, and the resulting snapshots clearly delineated magma ascent paths, including dykes and ponding zones, during those intervals.
In our study [Zanon et al., 2024], the results of petrological analyses of fluid inclusions in Tajogaite samples were promising: Estimates of the depths at which magma had ponded below the surface before erupting were consistent with the depths of the two seismically active zones that had been defined through geophysical monitoring.
Our reaction to the method’s potential after the successful test was much the same as Gene Wilder’s Dr. Frederick Frankenstein in Young Frankenstein: “It could work!”
Microscopy + Seismology = Success
Combining highly accurate microthermometric data from fresh lavas and pyroclastic fallouts sampled frequently during the eruption with concurrent data on the locations of seismic hypocenters was the secret to success.
Seismological imaging techniques allow us to identify lithological discontinuities and the presence of partially melted rocks beneath a volcano. Microthermometric analyses of many fluid inclusions reveal the pressures (i.e., the depths) at which the magmas have ponded. In addition, tracking features of seismicity during an eruption, including hypocenter locations, energy release, and seismic wave frequency, is important for characterizing the lithology of magma ponding zones (both temporary and long-lasting accumulation zones, i.e., magma chambers) and the modality of magma extraction. Indeed, seismic wave analyses can even reveal episodes of compression of porous lithologies, such as mush layers composed of crystal aggregates, which frequently host magma.
Collecting and cross-referencing information about depths of magma ponding obtained from seismic data on subsurface discontinuities and fluid inclusion data from materials erupted by Tajogaite allowed us to define the architecture of the magmatic system with greater accuracy than we could have using either technique individually. The approach also allowed us to show the path of magma ascent in near-real time and to use changes in mineral content and in composition of trapped fluids to discriminate different batches of magma and their distinct ascent rates (Figure 4).
Using this approach, we identified five magma batches that emerged from a main accumulation zone 27–31 kilometers deep, spanning the boundary into the lithospheric mantle. This deep location is established by fluid inclusions hosting nitrogen and carbon monoxide, markers of mantle outgassing.
Magma also accumulated over different durations at depths of 22–27 and 4–16 kilometers. The deeper of those ranges comprises layers of porous rocks made of clinopyroxenes and olivines in different proportions. The shallower depth range contains a series of amphibole-bearing mush layers and crystallized magma bodies from much older intrusions.
This study demonstrates how integrating these two fundamentally different methodologies can set a new standard for monitoring volcanic activity.
Time-integrated magma ascent velocities (including ponding times) were estimated by calculating the time between peaks of deep and shallow seismicity in clusters of earthquakes. These data showed that velocities were between 0.01 and 0.1 meter per second. Differences in the velocities determined the partial mingling of the magma batches and, at the surface, the occurrence of changes in magma flow rate related to the speed, mobility, and size of erupted volumes. Differences in magma ascent velocities also affected the formation of new eruptive fractures and the occurrence of cyclical explosivity during the eruption. New eruptive fractures signaled changes in eruption sites in the areas threatened, and changes in explosivity meant potential hazards for air traffic, human health (e.g., breathing problems), and infrastructure (e.g., possible roof collapses).
Considering these consequences, early recognition of the rise of new magma batches would have been very important for civil defense authorities.
This study represents the first time such a result blending seismological and petrological data has been obtained, and it demonstrates how integrating these two fundamentally different methodologies can set a new standard for monitoring volcanic activity.
Improving and Integrating Petrological Monitoring
Unfortunately, this new methodology is not applicable for monitoring all volcanoes because not all magmas contain crystals with trapped fluid inclusions. In large active magma reservoirs below some volcanoes, fractionation processes (e.g., gravitational settling of minerals contained in magmas) filter out heavier crystals that form first and that mostly trap fluid inclusions.
For example, it is unlikely to work for silicic volcanoes, in which viscous magmas are rich in silica and gas—making them prone to dangerously explosive eruptions—but devoid of early-formed crystals likely to host fluid inclusions. However, this approach might apply to any volcano that erupts mafic (olivine- and pyroxene-rich) magmas, which typically rise rapidly through the crust without pausing for long in ponding zones, thus limiting how many inclusion-bearing crystals settle out.
Improving on our method will involve accelerating the workflow, from sample collection and preparation to interpreting the data. If a volcano erupts at a location without research infrastructure, logistics may become problematic, and days may be lost transporting samples. The best solution in such cases could be to develop a mobile laboratory equipped with the necessary instruments for microthermometric fluid inclusion analysis and interpretation.
Furthermore, as more people collaborate, more time is saved during sample preparation and analysis. At best, it will always take at least a day to perform the necessary petrological analyses on a single sample. However, this time lag is short enough to allow the approach to be applied during the scientific management of a volcanic eruption.
Challenges aside, we are confident that the developments in petrological analysis techniques demonstrated recently add petrology to the disciplines that can be used for near-real-time eruption monitoring, greatly enhancing our ability to understand internal volcanic dynamics while they are still in action.
References
Andronico, D., et al. (2009), Shifting styles of basaltic explosive activity during the 2002–03 eruption of Mt. Etna, Italy, J. Volcanol. Geotherm. Res., 180, 110–122, https://doi.org/10.1016/j.jvolgeores.2008.07.026.
Boudoire, G., et al. (2019), Eruptive activity on the western flank of Piton de la Fournaise (La Réunion Island, Indian Ocean): Insights on magma transfer, storage and evolution at an oceanic volcanic island, J. Petrol., 60(9), 1,717–1,752, https://doi.org/10.1093/petrology/egz045.
Cashman, K. V., R. S. J. Sparks, and J. D. Blundy (2017), Vertically extensive and unstable magmatic systems: A unified view of igneous processes, Science, 355(6331), eaag3055, https://doi.org/10.1126/science.aag3055.
Corsaro, R. A., L. Miraglia, and M. Pompilio (2007), Petrologic evidence of a complex plumbing system feeding the July–August 2001 eruption of Mt. Etna, Sicily, Italy, Bull. Volcanol., 69, 401, https://doi.org/10.1007/s00445-006-0083-4.
D’Auria, L., et al. (2022), Rapid magma ascent beneath La Palma revealed by seismic tomography, Sci. Rep., 12(1), 17654, https://doi.org/10.1038/s41598-022-21818-9.
Hildner, E., A. Klügel, and T. H. Hansteen (2012), Barometry of lavas from the 1951 eruption of Fogo, Cape Verde Islands: Implications for historic and prehistoric magma plumbing systems, J. Volcanol. Geotherm. Res., 217, 73–90, https://doi.org/10.1016/j.jvolgeores.2011.12.014.
Klügel, A., et al. (2000), The chemically zoned 1949 eruption on La Palma (Canary Islands): Petrologic evolution and magma supply dynamics of a rift zone eruption, J. Geophys. Res., 105(B3), 5,997–6,016, https://doi.org/10.1029/1999JB900334.
Martí, J., et al. (2013), Correlation of magma evolution and geophysical monitoring during the 2011-2012 El Hierro (Canary Islands) submarine eruption, J. Petrol., 54, 1,349–1,373, https://doi.org/10.1093/petrology/egt014.
Sparks, R. S. J., et al. (2019), Formation and dynamics of magma reservoirs, Philos. Trans. R. Soc. A, 377, 20180019, https://doi.org/10.1098/rsta.2018.0019.
Zanon, V., M. L. Frezzotti, and A. Peccerillo (2003), Magmatic feeding system and crustal magma accumulation beneath Vulcano Island (Italy): Evidence from CO2 fluid inclusions in quartz xenoliths, J. Geophys. Res., 108(B6), 2298, https://doi.org/10.1029/2002JB002140.
Zanon, V., et al. (2020), Unravelling the magma feeding system of a young basaltic oceanic volcano, Lithos, 352–353, 105325, https://doi.org/10.1016/j.lithos.2019.105325.
Zanon, V., et al. (2024), Toward a near real-time magma ascent monitoring by combined fluid inclusion barometry and ongoing seismicity, Sci. Adv., 10(6), eadi4300, https://doi.org/10.1126/sciadv.adi4300.
Author Information
Vittorio Zanon (Vittorio.VZ.Zanon@azores.gov.pt), Instituto de Investigação em Vulcanologia e Avaliação de Riscos, Universidade dos Açores, Ponta Delgada, Portugal; and Luca D’Auria, Instituto Tecnológico y de Energías Renovables, Tenerife, Canary Islands, Spain; also at Instituto Volcanológico de Canarias, Tenerife, Canary Islands, Spain
