More Like this

1. Forecasting Eruptions at Steamboat Geyser: Time Scales, Differentiability, and Detectability of Seismic Precursors Through Data‐Driven Methods

   https://doi.org/10.1029/2024JH000569  

   Ardid, Alberto; Barth, Anna; Dempsey, David; Manga, Michael; Cronin, Shane J (June 2025, Journal of Geophysical Research: Machine Learning and Computation)

   Abstract Geyser eruptions provide a test bed for using geophysical data to forecast eruptions and to understand heat and mass transport in hydrothermal systems. We used time series analyses of seismic data at Steamboat Geyser, Yellowstone National Park, to identify short‐term precursors that are recurrent, detectable in real time, distinctly identifiable, as well as being rare during non‐eruptive periods. We analyzed seismic data from March to December 2018 to identify patterns that occurred before 31 eruptions. Four seismic amplitude measures and 700 time‐series features were computed from the seismic data. A template matching analysis identified an optimal 18‐hr window for detecting precursors. We applied a random forest to classify pre‐eruptive and non‐eruptive data for out‐of‐sample eruptions (eruptions that were not included in the model's training data), showing ability to distinguish between the two. This model performed better than a simpler amplitude‐based approach. Seismic features with the most predictive power include autocorrelations, longest strike above the mean, and change quantiles, particularly within the 4.5–16 Hz frequency range. We applied isotonic regression, a method that converts raw model outputs into calibrated probabilities, to improve the interpretability of eruption forecasting outputs. The likelihood of an eruption reaches 12.6% within 18 hr prior to the event, representing a marginal increase over the static 8% probability derived solely from eruption intervals. Unlike the interval‐based approach, our model does not rely on the time since the last eruption, instead using real‐time seismic features to detect precursory signals. Our study advances Machine Learning methodologies in eruption forecasting by integrating calibrated probability estimation through isotonic regression, which has advantages over traditional approaches for geysers with highly irregular eruption intervals. 

   more » « less
2. A shake and a surge: Assessing the possibility of an earthquake-triggered eruption at Steamboat Geyser

   https://doi.org/10.30909/vol.07.02.733748  

   Reed, Mara; Barth, Anna; Taira, Taka'aki; Farrell, Jamie; Manga, Michael (July 2024, Volcanica)

   When and why earthquakes trigger volcano and geyser eruptions remains unclear. In September 2022, Steamboat Geyser in Yellowstone, USA erupted 8.25 hours after a local M3.9 earthquake—an improbable coincidence based on the geyser’s eruption intervals. We leverage monitoring data from the surrounding geyser basin to determine if the earthquake triggered this eruption. We calculate a peak ground velocity of 1.2 cm s−1, which is the largest ground motion in the area since Steamboat reactivated in March 2018 and exceeds a threshold associated with past earthquake-triggered geyser eruptions in Yellowstone. Despite no changes in other surface hydrothermal activity, we found abrupt, short-lived shifts in ambient seismic noise amplitude and relative seismic velocity in narrow frequency bands related to the subsurface hydrothermal system. Our analysis indicates that Steamboat’s eruption was likely earthquake-triggered. The hours-long delay suggests that dynamic strains from seismic waves altered subsurface permeability and flow which enabled eruption. 

   more » « less
3. The 2018 reawakening and eruption dynamics of Steamboat Geyser, the world’s tallest active geyser

   https://doi.org/10.1073/pnas.2020943118  

   Reed, Mara H.; Munoz-Saez, Carolina; Hajimirza, Sahand; Wu, Sin-Mei; Barth, Anna; Girona, Társilo; Rasht-Behesht, Majid; White, Erin B.; Karplus, Marianne S.; Hurwitz, Shaul; et al (January 2021, Proceedings of the National Academy of Sciences)

   null (Ed.)

   Steamboat Geyser in Yellowstone National Park’s Norris Geyser Basin began a prolific sequence of eruptions in March 2018 after 34 y of sporadic activity. We analyze a wide range of datasets to explore triggering mechanisms for Steamboat’s reactivation and controls on eruption intervals and height. Prior to Steamboat’s renewed activity, Norris Geyser Basin experienced uplift, a slight increase in radiant temperature, and increased regional seismicity, which may indicate that magmatic processes promoted reactivation. However, because the geothermal reservoir temperature did not change, no other dormant geysers became active, and previous periods with greater seismic moment release did not reawaken Steamboat, the reason for reactivation remains ambiguous. Eruption intervals since 2018 (3.16 to 35.45 d) modulate seasonally, with shorter intervals in the summer. Abnormally long intervals coincide with weakening of a shallow seismic source in the geyser basin’s hydrothermal system. We find no relation between interval and erupted volume, implying unsteady heat and mass discharge. Finally, using data from geysers worldwide, we find a correlation between eruption height and inferred depth to the shallow reservoir supplying water to eruptions. Steamboat is taller because water is stored deeper there than at other geysers, and, hence, more energy is available to power the eruptions. 

   more » « less
4. Snow Suppresses Seismic Signals From Steamboat Geyser

   https://doi.org/10.1029/2023GL103904  

   Reed, Mara H.; Manga, Michael (June 2023, Geophysical Research Letters)

   Abstract Geyser and volcano monitoring suffer from temporal, geographic, and instrumental biases. We present a recording bias identified through multiyear monitoring of Steamboat Geyser in Yellowstone National Park, USA. Eruptions of Steamboat are the tallest of any geyser in the world and they produce broadband signals at two nearby stations in the Yellowstone National Park Seismograph Network. In winter, we observe lower eruption signal amplitudes at these seismometers. Instead of a source effect, we find that environmental conditions affect the recorded signals. Lower amplitudes for 23–45 Hz frequencies are correlated with greater snow depths at the station 340 m away from Steamboat, and we calculate an energy attenuation coefficient of 0.21 ± 0.01 dB per cm of snow. More long‐term monitoring is needed at geysers to track changes over time and identify recording biases that may be missed during short, sporadic studies. 

   more » « less
5. On the Hundred-Meter Sput and Its Temperamental Home: Geophysical Investigations of Steamboat Geyser and Norris Geyser Basin

   Reed, Mara H (December 2025, UC Berkeley)

   Geysers, hot springs that have periodic steam-driven eruptions, are a natural window into fluid flow within the crust and provide insight into common processes at volcanoes. Their eruptions also capture public interest, and millions of tourists visit geyser fields each year. Geysers are ephemeral features and even those with reputations for being regular performers are susceptible to perturbations from the hydrologic cycle, climate change, and earthquakes. Study of geysers and their associated hydrothermal systems is hampered by a lack of long-term monitoring, making it difficult to assess the factors that control their activity on timescales of days to decades. I take advantage of data available from an unusually well-monitored geyser basin in Yellowstone, United States to understand if, why, and how certain factors affect geysers. Steamboat Geyser is currently the tallest active geyser on Earth with a water jet that can reach a height of 137 m. It is located in the dynamic Norris Geyser Basin which is characterized by periodic events called disturbances during which the activity, temperature, chemistry, and turbidity of thermal features across the basin change suddenly. In 2018, Steamboat initiated an active phase that peaked in 2019-2020 with 48 eruptions in each year and then gradually relaxed to just 6 in 2024 and 2 in 2025 through November. I do not find evidence that Steamboat's active phases are triggered by changes in aquifer recharge or dynamic stresses related to earthquakes. The most recent active phase began after a period of uplift and slightly increased thermal radiance in Norris Geyser Basin, which possibly suggests a link between Steamboat's activity and magmatic processes. However, given that the activity of other geysers at Norris did not change, nor did the temperature of the deep reservoir feeding Steamboat, a causal link cannot be established. Before mid-2021, Steamboat's eruption intervals varied seasonally, indicating that the hydrologic cycle influences eruption timing. Abnormally long intervals during this period followed Norris Geyser Basin disturbances. At Steamboat, there is no relationship between eruption interval and the volume of water discharged during an eruption's initial water phase. The geyser's eruption jets are so tall because the depth at which water is stored in the shallow plumbing is deeper there than at other geysers, meaning more energy is available to drive eruptions. Powerful geysers like Steamboat produce air-ground coupled acoustic waves that can be detected at seismometers hundreds to thousands of meters away. Care must be taken in interpreting these signals. I found that Steamboat eruption signals had lower amplitudes in winter not because eruptions were weaker but rather because the signal was damped by 0.21 dB per cm of snow. Seismometers can also track local hydrothermal tremor, which provides insight into the state of a hydrothermal system. In September 2022, Steamboat erupted 8.25 hours after a local M3.9 earthquake produced a peak ground velocity of 1.2 cm/s at Norris Geyser Basin. The amplitude and frequency of hydrothermal tremor shifted immediately following shaking from the earthquake which indicates a change in fluid flow and thus heat transport. The earthquake likely triggered Steamboat to erupt. Finally, disturbances at Norris Geyser Basin since 2013 are periodic but not seasonal, dispelling a previous hypothesis that their timing is controlled by the hydrologic cycle. All disturbances are accompanied by changes to hydrothermal tremor recorded at two locations 600 m apart and surface features change within hours in both the Porcelain and Back Basins. Most but not all disturbances are followed by a drop in total liquid discharge from the hydrothermal system. Land managers and the scientific community should prioritize the creation and maintenance of robust monitoring networks to collect data needed to understand how geysers and subsurface hydrothermal systems change over time and how they are connected to surface hazards such as hydrothermal explosions. 

   more » « less