Lava dome cycles reveal rise and fall of magma column at Popocatépetl volcano

The unpredictable behavior of lava domes presents a significant challenge in understanding their formation and associated hazards. Lava dome morphologies vary widely, ranging from tall and steep peléan domes to flat and circular domes, influenced by magma rheology, extrusion rate, and substrate topography. Growth mechanisms include exogenous (extrusive) and endogenous (intrusive) processes, each leading to distinct morphologies and hazard profiles. Instabilities can trigger rockfalls and collapses, generating pyroclastic density currents, while rapid permeability changes can cause explosive events. While modeling and remote sensing (photogrammetry, optical, and infrared) have been used to study dome morphology and dynamics, limitations exist due to visibility issues. Synthetic Aperture Radar (SAR) offers a unique advantage, penetrating clouds and allowing imaging regardless of visibility obstructions. Interferometric processing (InSAR) has limitations for studying lava domes due to their small spatial extent and surface incoherence; therefore, reflected SAR intensity is used to track dome and crater morphology. However, visual interpretation is hindered by radar geometry and speckle noise. This study focuses on Popocatépetl volcano, known for its successive lava dome construction and destruction episodes since 1994, characterized by low lava domes susceptible to subsidence and explosive processes. The repetitive dome growth episodes are thought to result from varying buoyancy of the magma column, influenced by volatile content. Popocatépetl also exhibits strong excess degassing, suggesting magma degassing from deeper reservoirs or a shallower convecting magma column. This research investigates the shallow magma dynamics governing these cycles and excess degassing, utilizing multiparametric satellite datasets to quantify inner-crater depth variations and analyze dome morphological evolution with deep learning-enhanced SAR images, complemented by SO2 gas emission and thermal radiation observations. This integrated approach aims to provide a comprehensive understanding of the eruptive dynamics and magma conduit processes at Popocatépetl.

Previous research on lava domes has largely focused on understanding their varied morphologies and the processes driving their growth and destruction. Studies using laboratory models (Fink & Griffiths, 1998) and numerical simulations (Husain et al., 2019; Tsepelev et al., 2020) have provided valuable insights into the influence of magma rheology and extrusion rates. The role of gas pressure in dome stability and explosive events has also been investigated (Sparks, 1997; Voight & Elsworth, 2000). Remote sensing techniques, including photogrammetry (Diefenbach et al., 2012; Zorn et al., 2020) and thermal infrared imaging (Wright et al., 2002; Coppola et al., 2022), have been employed to monitor dome growth and activity. However, the use of SAR imagery for detailed analysis of lava dome evolution, particularly at high temporal and spatial resolution, has been less extensive. While InSAR has been widely used to study volcano deformation (Lu, 2013; Mania et al., 2021), its application to lava domes is limited due to the challenges posed by their small scale and rapidly changing surfaces. Previous work on Popocatépetl has documented the cyclic nature of dome growth and destruction (Gómez-Vazquez et al., 2016; Mendoza-Rosas et al., 2017), highlighting the importance of volatile content in driving these cycles (González-Mellado & De la Cruz-Reyna, 2008). Studies have also emphasized the volcano's strong excess degassing (Campion et al., 2018; Roberge et al., 2009), indicating the potential role of deep magma reservoirs or shallow magma convection in supplying volatiles.

This study utilizes a multi-parametric approach combining high-resolution satellite data with advanced image processing techniques. Speckle filtering of SAR images was achieved using a specifically designed convolutional neural network (CNN) trained on data from Colima volcano. The CNN removed granular noise from TerraSAR-X (TSX) and Sentinel-1 (S1) SAR images, improving the visibility of fine morphological details. A total of 158 descending track TSX acquisitions (2012-2019) were analyzed, along with Sentinel-1 data. Crater depth and diameter were estimated from the SAR images by analyzing radar shadows cast by the crater walls, accounting for the radar viewing geometry. A fixed intensity threshold was used for TSX images, while a graph cuts segmentation was applied to S1 images. Crater volume was estimated assuming a truncated cone geometry. Volcanic Radiative Power (VRP) was obtained from the MIROVA system using MODIS data (2005-2020), providing measurements of the heat flux. SO2 gas emissions were estimated from OMI satellite data (2005-2020) using the traverse method, correcting for seasonal variations. Magma volumes were estimated from SO2 emissions, thermal VRP, and SAR-derived extruded magma volumes. The volume of tephra was estimated semi-empirically. Magma fluxes were calculated from these smoothed monthly averaged volumes. To model the short-term crater depth variations associated with dome cycles, the study considered changes in exsolved gas fraction in the upper magma column, assuming a simple cylindrical model. The long-term crater deepening was modeled using the theoretical framework of Girona et al. (2014), considering magma reservoir depressurization due to steady gas loss via magma convection in the upper conduit. The model accounts for magma convection, degassing rate, conduit radius, reservoir volume, and magma/host rock properties. The study used parameters consistent with previous work on Popocatépetl to simulate pressure changes and the resulting magma column height variations.

The analysis of high-resolution SAR images revealed numerous cycles of dome construction and destruction at Popocatépetl. Dome construction was a rapid process (hours to days), resulting in pancake-shaped domes (45-270 m diameter, 3-6 m thickness) often accompanied by strombolian activity. Dome destruction involved two mechanisms: progressive subsidence (days to months) characterized by a central pit, peripheral bulges, ring fractures, and piston-collapse structures, and vulcanian explosions. In addition to these short-term cycles, a progressive deepening and widening of the inner crater was observed over the past decade (2012-2019), accompanied by a decrease in thermal VRP and monthly SO2 flux. The inner-crater volume loss during this period was estimated at ~8 Mm³. Analysis of VRP and SO2 fluxes over 15 years showed a strong linear correlation, with sporadic high VRP values associated with explosive events. Magma budget estimation revealed an imbalance between magma input (inferred from SO2 emissions) and output (extruded lava and tephra), indicating excess degassing and thermal radiation, suggesting that unerupted magma is degassing and cooling at shallow levels. The short-term dome construction-subsidence cycles were interpreted as gas-driven rise and fall of the upper magma column, where gas retention and escape control magma extrusion and withdrawal. The long-term crater deepening, along with decreasing heat/gas fluxes, was attributed to progressive depressurization of the magma conduit and reservoir due to gas depletion. A simple model based on gas depletion in the upper magma column successfully explained the short-term dome subsidence rates. A more complex model, based on the work of Girona et al. (2014), was used to simulate long-term crater deepening due to passive degassing, showing reasonable agreement with observations.

The findings of this study provide compelling evidence for a model of lava dome formation and destruction at Popocatépetl driven by variations in the gas content of the upper magma column. The short-term cycles of dome growth and subsidence are consistent with a buoyancy-driven process, where gas-rich magma rises and extrudes, followed by degassing, cooling, and magma drain-back. The long-term crater deepening is linked to the gradual depressurization of the magma system due to continuous degassing. The excess degassing and thermal radiation indicate that a significant portion of the magma's volatile content is released before eruption, highlighting the importance of considering the entire magma plumbing system when assessing volcanic hazards. These findings are consistent with observations at other low-viscosity lava dome volcanoes and suggest similarities with magmatic processes in more basaltic open-system volcanoes. This integrated approach improves our understanding of open-vent volcanic activity and provides valuable insights for volcano monitoring and hazard assessment.

This study provides a novel integrated analysis of high-resolution satellite data and advanced modeling to understand the dynamics of lava dome cycles at Popocatépetl. The results highlight the critical role of gas exsolution and escape in driving both short-term dome cycles and long-term crater evolution. The findings suggest a coupled system where gas-driven magma ascent and descent are responsible for short-term dome growth and collapse, while long-term degassing and depressurization lead to crater deepening and widening. Future research could focus on refining the models to incorporate additional factors, such as the influence of magma rheology and the complex interactions between gas bubbles and the surrounding melt. Further development of deep learning techniques for SAR image processing will also improve the accuracy and efficiency of volcano monitoring.

While this study offers valuable insights into the dynamics of Popocatépetl, certain limitations should be acknowledged. The semi-empirical estimation of tephra volume introduces uncertainty into the magma budget calculations. The simplistic model for short-term subsidence neglects factors such as gravitational load and magma drainage. The long-term model of crater deepening relies on assumptions about the magma reservoir and conduit geometry. The accuracy of the SAR-derived crater depth measurements depends on the chosen intensity threshold and the complex crater morphology. Despite these limitations, the integrated approach employed in this study provides a significant advancement in understanding lava dome evolution.