Formation and evolution of newly formed glaciovolcanic caves in the crater of Mount St. Helens, Washington, USA

A new and extensive system of glaciovolcanic caves has developed around the 2004-2008 lava dome in the crater of Mount St. Helens, Washington, USA. These systems offer a rare view into a subglacial environment and lead to a better 10 understanding of how glaciers and active volcanoes interact. Here, we present first results from geodetic and optical surveys done between 2014 and 2019 as well as climatologic studies performed between 2017 and 2019. Our data show that volcanic activity has altered subglacial morphology in numerous ways and formed new cave systems that are strongly affected by heat flux from several subglacial fumaroles. More than 2.3 km of cave passages now form a circumferential pattern around the dome, some several hundred meters long. Air and fumarole temperature measurements were conducted in two specific caves. 15 Whereas air temperatures reveal a strong seasonal dependency, fumarole temperatures are affected to a minor extent and are primarily regulated by changes in volcanic heat flux or the contribution of glacial melt. Related studies from Mount Hood, Oregon, and Mount Rainier, Washington, are used as comparison between glaciovolcanic cave systems. Fumarolic heat and resulting microclimates enable further genesis of this dynamic system. Already one of the largest worldwide, it is very likely that the system will continue to expand. As Mount St. Helens is the Cascade Volcano most likely to erupt again in the near 20 future, these caves represent a unique laboratory to understand glaciovolcanic interactions, monitor indicators of recurring volcanic activity and to predict related hazards.

combination thereof (Badino et al., 2007;Benn and Evans, 2010). Glaciovolcanic caves, their importance to understand the volcano's hydrothermal and magmatic system, or their significance to serve as important analogues for extraterrestrial phenomena (Curtis and Kyle, 2011;Zimbelman et al., 2000) have been identified and discussed in literature. First studies comprise research on the summit craters of Mount Rainier (Kiver and Mumma, 1971), and investigations of the geothermal 35 ice caves on Mount Erebus (Giggenbach, 1976). However, research in these harsh environments hold numerous hazards like volcanic gases, the potential beginning of an eruption, or any instability of the volcanic edifice. A problem unique to the Crater Glacier is the abundance of loose rock with an average content of 15 %, deriving from the surrounding amphitheater walls (Walder et al., 2008). Schilling et al., (2004) even suggested one-third of the glacier's volume to be rock debris from abounding rock avalanches. 40

60
We present the first overview of glaciovolcanic cave systems circumnavigating the 2004-2008 lava dome. We discuss the formation and evolution of cave systems recorded over six years through tacheometric survey methods and discuss the results of air and fumarole temperature measurements from 2017 to 2019 in two distinct caves. We also present essential cave characteristics identified by direct observations. The caves experience seasonal changes and although they share this feature 65 with other cave systems, e.g. on Mount Rainier (Florea et al., 2020, in review), several features unique to the cave systems on Mount St. Helens have been discovered. Our data reveal that these caves show incredible dynamic growth compared to other https://doi.org/10.5194/tc-2020-279 Preprint. Discussion started: 23 October 2020 c Author(s) 2020. CC BY 4.0 License.
glaciovolcanic cave systems and are trending to continue expansion. This paper provides a general summary to introduce main areas of research and only represents selected parts of our work.
2 Geological setting 70 Mount St. Helens, an active andesite-dacite volcano (Anderson and Vining, 1999), is located in Washington State in the Cascade Range, a magmatic arc resulting from the subduction of the oceanic Juan de Fuca Plate beneath the continental North American Plate (Miller and Cowan, 2017). Coming into existence between 40,000 and 50,000 years ago, Mount St. Helens remains the youngest of five major stratovolcanoes which developed in Washington State (Harris, 2005). In 1980 activity at the volcano was rekindled in terms of a catastrophic explosive eruption which dramatically modified the surrounding landscape 75 and the volcano itself (Voight et al., 1983). The eruption caused a loss of approximately 70 % of the volcano's ice cover (Brugman and Post, 1981) and resulted in a 1.6 x 3.5 km horseshoe-shaped and north-facing crater (Anderson et al., 1998).
Despite a relatively low elevation of the crater floor, this area represents an important accumulation zone for snow, rime, and avalanche deposits throughout the year due to deep shade and insulating characteristics of dust layers (Schilling et al., 2004). From 1980From until 1986, the unrest of Mount St. Helens has generated a new lava dome (Fig. 3), episodically growing on the 80 new crater floor (Swanson and Holcomb, 1990). Simultaneously, a progressive accumulation of snow heralded the beginning of a new glacier to form. This process was accompanied by a series of 16 dome-building eruptions (Anderson et al., 1998) and at least six smaller ash-producing explosions with minor fallout between 1989-1991 (Vallance et al., 2010). Another unrest started in September 2004, when a second lava-dome-building eruption initiated in the crater. This renewed activity greatly influenced the morphology of the glacier, which had already grown to a thickness of 150 m. Ongoing activity through 2008 85 bisected, fractured, and compressed the glacier, doubling the surviving east and west arms in thickness and increasing the flow rate (Scott et al., 2008;Vallance et al., 2010). The formation of the new lava dome, the whaleback spine, also involved small explosions as well as plumes of volcanic ash and gases (Iverson et al., 2006). Even though the volcano is quiescent at the moment, it is supposed that it will erupt again this century (Vallance et al., 2010). During the past 4,000 years it has been the most active volcano in the Cascade Volcanic Arc (Crandell and Mullineaux, 1978), and according to the Cascade Volcano 90 Observatory currently considered as the Cascade Volcano most likely to erupt again in the near future.

Cave survey
Travel on the Crater Glacier and cave exploration requires mountaineering on glacier ice, ice climbing techniques, and the capability to ascend and descend cave entrances and passages. Crevasse hazards ranged from minimal to extensive on the east 100 and west side the 2004-2008 lava dome, respectively. To mitigate some of the risks of severe weather, rockslides, and firn collapse, expeditions were confined to May and June. Potential cave entrances were located by reconnaissance via satellite imagery and by ground observations. Exploration to locate caves was not undertaken in the area forward of the 1980-1986 lava dome as there the glacier is heavily fractured.

Data collection in the crater 105
Tacheometric surveys were used to record the extent of 10 discovered glaciovolcanic caves (Fig. 2). Furthermore, surveys were used to produce fixed, recoverable stations and record the location of fumarole and temperature loggers inside Mothra Cave (Fig. 4) and Crevasse Cave (Fig. 5). GPS georeferenced stations were recorded at each cave entrance. Cave survey data https://doi.org/10.5194/tc-2020-279 Preprint. Discussion started: 23 October 2020 c Author(s) 2020. CC BY 4.0 License.
were collected using DistoX and DistoX2 to generate distance, azimuth, and inclination and to compute passage volumes at each station. The device is a Leica Laser Disto customized with a 3-axis electronic compass, clinometer, and a wireless 110 Bluetooth connection (Heeb, 2009) to communicate with PDA or PC based applications to manage and visualize the data.  The survey methods evolved during the project. Data in 2014 generated with the DistoX were manually recorded, with passage 120 cross sections hand drawn at key stations. Passage dimensions were recorded in four directions at each permanent station.
Survey data from 2017 onward were generated using the splay method with DistoX2 which includes more passage dimension measurements. The DistoX2 communicated with a Dell Axim X51 PDA and PocketTopo cave survey software or Samsung Galaxy Note 4 or similar Android OS devices and TopoDroid software. Challenges were given by interference of the DistoX laser from water spray, fog, mist, or other obstructions, occasionally arising in the caves. In periods of limited visibility, splay 125 measurements and detail were not possible. In those cases, basic passage measurements were estimated in four directions from the fixed stations. Other challenges included glacier movement and falling rocks which made it impossible to relocate some marked stations from previous surveys. https://doi.org/10.5194/tc-2020-279 Preprint. Discussion started: 23 October 2020 c Author(s) 2020. CC BY 4.0 License.

Data processing and cartography
Post processing of survey data was conducted in COMPASS cave survey project management software. Data was corrected 130 for annual magnetic declination variance. Georeferenced stations input within COMPASS provided references for cartography over multiple survey years. The software generates statistics including length, depth, and volume. Due to magnetic interference from volcanic rock it was expected that some survey measurements generated by DistoX2 could be erroneous. These shots were mitigated by relying on loop closure capability of the software and GPS referenced stations where possible. COMPASS makes this correction via the least square method (Schmidt and Schelleng, 1970). Finally, .kml files for ArcGis were generated, 135 3D visualizations in CaveXO software were produced, and corrected line plots and sketch data from software or manual recordings were used in Adobe Illustrator to complete final cartographic plan diagrams of each cave. With the use of historical imagery of the crater to identify the previous dimension of the crater and the location of the rock-ice-interface, it was also possible to calculate growth rates and the evolution of certain caves over the last years.

Set up of climatologic measurement devices and data processing 140
Two caves were subjected to climatologic studies, most inside Mothra Cave, and to a minor extent in Crevasse Cave. Between 2017 and 2019 GeoPrecision wireless data sensors (M-Log5W-CABLE) were placed in the caves to measure air and fumarole temperatures, adjusted to measurement intervals of five minutes. Locations were chosen to cover a large area inside each cave system, or to investigate interesting areas from a climatologic point of view. The installation of sensors and upload of data cover the period from May to June as this is the safest time for fieldwork. In an analogous manner, a GeoPrecision sensor 145 chain (digital thermistor string) of 20 to 30 sensors which are attached to the chain at intervals of one meter was installed to generate air and fumarole temperature profiles. In contrast to single data loggers, the chain was left in the caves during each expedition only, comprising three to five days. The chain was placed in an area with high fumarolic and individual loggers were positioned at fumaroles where possible. To detect hotspots inside the caves or outside on snow-free areas of the lava dome a thermal camera (InfraTec VarioCAM HiRes) view was usually observed before placing any sensors. The direction and 150 velocity of air flow was visualized by using orange smoke torches. Each implementation was documented by photography.

Cave survey and general cave morphology
We surveyed 10 caves with a combined length of 2,335.7 m in detail, the three most significant ones reaching more than 400 m each. Surveyed lengths ranged from 33.5 to 539.3 m. Depths varied between 7.7 and 65.2 m (Table 1) exhibit two or more entrances, exceptions are two of the smaller caves, The Cloaca and Gabara Cave. Entrance elevations 160 range from 2,098 (The Igloo) to 2,256 m a.s.l. (Mothra Cave). Main passages tend to be horizontal and circumferentially parallel to the ice-rock interface, large enough to be traversable by humans, with vertical sides and convex ceilings. All caves exhibited large scalloping in the ice walls and ceilings. Some featured scalloped ice floors mirroring those features. Occasional cryospeleothems were observed, including ice stalactites and stalagmites, and were associated with infiltrations of water.
However, the drainage of superficial water streams was not observed in the caves. Ambient air CO2 measured < 0.3 %. Rocks 165 embedded in the glacier reveal themselves in the cave ceilings and walls through ablation. This creates a hazard present in many of the caves. Various cave walls also contain diverse tephra layers. Cave characteristics are illustrated in figures 7a-f. Included Length: This is the included slope length of all the surveys processed. Slope length is the sum of all the tape lengths in the cave. It is the distance that you move through the cave, both horizontally and vertically.
Horizontal Length: This is the included horizontal length of all the surveys processed. Horizontal length is the length of the cave when it is flattened into a horizontal plane. Horizontal length includes no vertical component.

Cave Depth:
This is the absolute vertical distance between the highest and lowest points in the survey. It includes no horizontal movement.
Cave Volume: This statistic gives the volume of the cave surveys processed. It is based on the passage Left, Right, Up and Down dimensions. Surveys that are missing LRUD's for part or all of the data will give inaccurate volume calculations.

Features of remaining caves
Centrally located in between the crater rim and the lava dome, the Godzilla Hole was the first cave investigated in 2014. This 225 cave has not been entered since initial exploration in 2014 since the entrance pit, a 10 x 20 m opening, was no longer present.   (Fig. 10c). We assume that the data logger was frozen in ice and reveals stages of thawing and refreezing cycles. Amplitude scaling and offset translation indicate that temperature profiles at all three locations follow the same pattern during the year (Fig. 10d), and also 260 mirror diurnal trends with temperatures increasing at noon and afternoon and decreasing in the evening (Fig. 11). Over the course of one day temperatures in the entrance area reveal most distinct amplitudes (3.6°C), followed by the connection passage (2.1°C) and the tunnel area (0.9°C).  A correlation between monthly mean temperatures inside Mothra Cave and the average snow depth per month at June Lake exists. The correlation coefficient after Pearson was calculated as 0.77 for the entrance area and 0.91 for the connection 275 passage. Coefficients after Spearman analogue range between 0.73 and 0.87 (Fig. 12). Snow depths represent a periodicity with snow accumulation between November and June and maxima in March and April, conform with maxima of temperatures.
One exception is represented by spring 2019 where temperature maxima arise the month before and/or after the snow maximum. Equally to the entrance area and the connection passage, the tunnel area mirrors a correlation during the spring and summer season in 2017 but data were not considered for further calculation due to interference of the statistical series. 280 https://doi.org/10.5194/tc-2020-279 Preprint. Discussion started: 23 October 2020 c Author(s) 2020. CC BY 4.0 License. Figure 12: Monthly mean temperatures versus monthly mean snow depth at June Lake, Skamania County, south side of Mount St. Helens at an elevation of nearly 1,050 m. Correlations after Pearson and Spearman illustrate the relation between monthly mean air temperatures in the entrance area and the connection passage versus monthly mean snow depth at June Lake. A correlation for the tunnel area was not 285 calculated since the data logger probably got frozen into glacial ice after the first year. Snow depth data was provided by the NRCS National Water and Climate Center website, online accessible (https://wcc.sc.egov.usda.gov/nwcc/site?sitenum=553&state=wa).

Observation of ventilation effects and air flow movement
We were able to notice ventilation effects and air flow movement. Strongest ventilation occurred near entrances, minor effects 290 were experienced at higher elevations inside the caves and areas further away from the entrances. The release of orange smoke inside Mothra Cave (Fig. 7f) and Crevasse Cave in June 2018 similarly visualized the movement of air flow. A downward directed movement, following the cave morphology and the steep slopes, was observed inside Mothra Cave. Moreover, the air was flowing deeper into the cave system directing to the western part of the cave. The smoke was set free in visual range of the main entrance (0 m) in between the data loggers of the entrance area and the connection passage. In Crevasse Cave the 295 smoke equally was set free in one of the entrance areas. Similar to the situation inside Mothra Cave the smoke moved into the cave. At both cave systems a subsequent release of smoke through other entrances or openings was observed. Orange smoke at the surface was visible a few minutes after its release inside the caves.    Fumarole temperatures do not exceed 60.1°C. This is enough heat for cave systems to form. A wide range of temperatures is revealed by single fumaroles not only annually but also in the timeframe of a few days. They primarily fluctuate slightly abovezero up to 40°C. Higher temperatures are locally and temporarily limited. Some fumaroles even revealed temperatures subzero before continuing their heat output once more. We assume that ablation of cave ceilings and walls as well as rainfall are Furthermore, the crater floor is characterized by steep slopes. We were able to enter into the Crater Glacier to depths of more 350 than 65 m and did not detect streams or pooling in any of the caves. We conclude that the continuous presence of rainfall and glacial ablation in combination with the features of the crater floor essentially affect fumarole temperatures and cause this strong fluctuation. Cracks and fissures are constantly influenced by infiltrating water. We observed different situations in other glaciovolcanic cave systems. This once more indicates the significance of the hydrologic situation.

355
Mount Hood's cave systems feature several streams and cascades. The caves contain fog and reveal a strong humidity.
Fumaroles as observed on Mount St. Helens usually do not exist and rather arise as hot springs (Pflitsch et al., 2017). Intensified superficial water runoff has been observed. Altered volcanic material and loamy gravel exist inside the caves. Typical situations experienced on Mount Hood are illustrated in figures 15a-f. The crater floor on the summit of Mount Rainier is characterized by various clays which result from strong hydrothermal alteration (Zimbelmann, 1996). The morphology is featured by natural 360 depressions and is mainly controlled by the location inside the craters and the abundance of glacial ice plugs. Unique subglacial lakes are able to form (Figs. 16a-f). Since the summit craters are located at high elevations, snow precipitation dominates.
These contrasts to Mount St. Helens are revealed by fumarole temperatures. Although we equally observed variations over time, this apparently occurs at a much smaller scale. Fumaroles also remain at constantly high temperatures and fluctuate roughly between 40°C and 60°C. Some exceptions with abrupt temperature decreases may be related to sudden and strong 365 rainfall events, increased ablation, or temporary changes and disturbances of the hydrothermal system. Fumaroles furthermore depict a kind of dependency and show similar trends over a period of two years although a distinct seasonality is missing (Florea et al., 2020, in review). Fumaroles on Mount St. Helens seem to follow independent pathways in so far, that they do not reveal concordant trends. Although a distinct seasonality cannot be confirmed for the fumaroles in general, at least MSH 1 in the entrance area mirrors seasonal variations. We therefore conclude that fumaroles are affected by seasonality but this 370 applies more to the areas aside from the caves. Explicit reactions of fumaroles inside seem to be absent or hardly noticeable.
https://doi.org/10.5194/tc-2020-279 Preprint. Discussion started: 23 October 2020 c Author(s) 2020. CC BY 4.0 License.  390 Walder et al., (2008) suggested that much of the rubble underneath Crater Glacier is likely to be ice free because of the geothermal heat flow. Our observations turned out that this applies to most parts of the caves. An exception to this are fumarole-free areas in the vicinity of entrances where occasional cold air can lead to ice sheets that are several centimeters thick as well as icefall from ceiling and rime ice coatings. He also suggested that water that reaches the glacier bed probably flows out of the crater through the rubble layer or downward into the volcano's groundwater system rather than moving along 395 the glacier bed. His suggestions additionally support our assumption that fumarolic activity is significantly controlled by cave hydrology and the nature of the crater floor. Volcanic heat and fumaroles as the driving force initiate the onset of cave formation before further processes start to arise.

Ventilation effects and air temperature
Strong ventilation effects inside Mothra Cave and Crevasse Cave have been observed and visualized with smoke. This situation 400 results from a complex cave morphology with several cave openings in different elevations causing strong chimney effects.
Continuous transitions from ascending warm air heated by fumaroles and gravitational cold air exist as long as cave entrances are open and an exchange with external air is possible (Fig. 17). We assume this situation applies to most of the caves on  (Stenner et al., 2020, in preparation;Zimbelman et al., 2000). Compared to Mount St. Helens, ventilation effects here are likely to be less distinctive due to the different morphology, primarily minor elevation changes, and chimney effects arising to a lesser extent. Moreover, the crater floor is less steep and impermeable, thus providing more possibilities for CO2 410 to accumulate and to get trapped.  (Florea et al., 2020, in review;Pflitsch et al., 2017;Zimbelman et al., 2000).
In months without snow strong venting effects on Mount St. Helens influence the cave climate and reduce the concentration of fumarolic heat. These effects were equally observed inside the former cave systems (Anderson and Vining, 1999). We 425 assume that the absence of ventilation effects during winter and spring and resulting melting processes from the inside contribute much more to the transformation of the glacier than ablation processes at the surface in summer and autumn.
Temperature profiles at different locations inside Mothra Cave reveal a peculiarity and are subjected to the same seasonal cycles though they are fluctuating within a different amplitude. As expected, the area nearest the entrance as well as the 430 connection passage are most influenced by seasonal changes. Those comprise relative elevations of -16 and -24 m where the entrance represents a reference point with an elevation of 0 m. Another data logger placed inside the tunnel area, the highest of all three sensor elevations at -11 m, recorded the smallest temperature amplitude. The morphology and different elevations inside a cave system have distinct influence on the development of small-scale climatologic characteristics. During the year but also during one single day temperatures in the connection passage are usually the warmest. This is the location closest to 435 concentrated fumarolic activity. Occasional warmer temperatures near the entrance may be related to minor and temporary changes of fumarole degassing. The tunnel area also reacts to external factors, but to a much smaller extent. The absence of concentrated heat in this part of the cave is visible through the appearance of an icefall and is moreover revealed by lower monthly mean temperatures and the assumption that the data logger got frozen into glacial ice. Whereas the connection passage has the highest temperatures for most of the year, the entrance area is much more sensitive to external changes as long there is 440 a direct exchange with the outside air and entrances are open. This means cave air temperatures develop in an analogue pattern to outside air temperatures.

Cave morphology and further evolution
Ventilation effects, the distribution of air temperatures, the morphology of the crater floor, and volcanic heat interact in several ways. This causes the formation of caves around the dome, internal morphology changes, and circumferential evolution like 445 the cave system on Mount Rainier. Fumaroles as the driving force primarily influence the evolution of caves and their morphology. During our surveys distinct changes within cave systems were related to the appearance or disappearance of fumaroles. Although individual fumaroles changed, we conclude that major fluid pathways must have remained. Minor changes may be related to the growth or decay of cracks even though this could not be directly observed on the crater floor.
Distinct shifts related to changes of fumarolic heat have been observed for Crevasse Cave with passages now extending further 450 northwest (Fig. 8) and the disappearance of The Waterfall Room inside Mothra Cave (A1). Nevertheless, main cave passages remained, indicating that major fluid pathways did not change. This can also be confirmed for The Igloo.
Observations and resurveys of the caves revealed significant dynamics and the trend to expand in the near future. Historical imagery and the reconstruction of former cave dimensions (Mothra Cave) emphasize the immense growth during the last 455 decade. Arial photos from June and July 2020 moreover indicate further expansion. Crater glacier is growing and as the https://doi.org/10.5194/tc-2020-279 Preprint. Discussion started: 23 October 2020 c Author(s) 2020. CC BY 4.0 License. accumulation of snow continues, the boundary of the rock-ice interface is subjected to an ongoing shift towards higher elevations circumnavigating the 2004-2008 lava dome. It is very likely for the cave systems to follow these patterns and to continue extension towards the dome (Fig 17). This will probably happen within the next few decades, mostly controlled by fumarolic activity. To make more precise calculations for the evolution of glaciovolcanic caves on Mount St. Helens longer 460 time series data and additional resurveys are required. In comparison to our studies of other glaciovolcanic cave systems, the system on Mount St. Helens is likely the fastest growing of its type.
We assume that the rock-ice interface is most significant for the further evolution of cave systems and for generating circular patterns around the dome. Most caves appear to follow the same characteristic horizontal master passage network orientation 465 with dendritic entrance passages following the rock bed upwards to the rock-ice interface (Fig. 6). As a contact point for increased melting, due to volcanic heat or warming by sunlight, melting of lateral cave passages is expected to continue until caves have connected and a master passage has formed. The same situation probably applies to the crater rim of Mount Rainier and the East Crater Caves, the difference being that a master passage already exists. For Mount St. Helens a variety of circumstances may be responsible that caves are not interconnected so far. First, the cave system is too juvenile to have formed 470 a master passage connection. Second, the morphology of the lava dome floor could cause obstructions to required ventilation leading to formation of void spaces in the ice. Lastly, fumarole temperatures may be too high in some areas of the lava dome to support cave formation. Evidence is provided by Hedorah Cave and fumaroles of around 90°C which appear to stop ice accumulation in a crucial junction area that would otherwise connect three of the caves (Hedorah, Rodan, and Ghidorah).
Rodan Cave, where a laterally oriented master passage connects a section with upward trending morphology (Fig. 9), is 475 presumably the best example that cave systems are growing and that they will certainly interconnect in one or the other way.

Conclusions
Cave surveys and climatologic investigations of glaciovolcanic caves on Mount St. Helens reveal that the systems are extremely dynamic, trending to expand in the near future. Fumarolic activity is the driving force and is responsible for the formation and evolution of cave systems. Heat flux from fumaroles appeared to remain constant during the years of survey 480 with minor chances being present. Although individual fumaroles arose and disappeared during our survey, concentrated areas of activity were observed to be stationary. Fumarolic activity is closely related to other features which include internal as well as external factors. The permeable crater floor with steep slopes and a complex morphology with changing elevations is most significant for the drainage of water to the hydrothermal system. Annual snowfall with maximum heights in spring and the resulting sealing of entrances have the greatest influence on air temperatures inside the caves and induce a self-energizing 485 process due to the accelerated melting of glacial ice. Significant seasonal cycles have been observed. Similarities but equally several differences to other glaciovolcanic systemsbasically Mount Hood and Mount Rainiercould be identified. Those include differences in fumarole temperatures, morphology, the nature of the crater floor, and the hydrologic situation.
We expect that caves will interconnect during their further evolution similar to the East Crater Caves on Mount Rainier and 490 finally circumnavigate the new lava dome. Apart from lateral extension, passages will increase with annual snow accumulation.
The rock-ice interface significantly influences this process. The caves on Mount St. Helens are the most dynamic caves we have seen in the Pacific Northwest, holding hemispherical rooms and continuously changing passages. We therefore assume that a connection of systems to a master passage will probably happen within the next few decadesthe absence of major volcanic activity required. Mount St. Helens and its cave systems represents a unique natural laboratory and provides several 495 possibilities to expand future research. As many volcanoes not only in the Cascade Volcanic Arc but worldwide host glaciovolcanic cave systems, their contribution to changes of the cryosphere from underneath and the possibilities to study a subglacial environment, investigate related hazards, or to detect evidence for renewed volcanic activity can be essential.
Data availability. The datasets are available from the corresponding author on request. 500 Author contributions. LS and CS conceptualized and visualized this work and led writing of the manuscript. AP supervised the work. EC took part in project administration and led expeditions. CH and TB supported fieldwork and data acquisition.
CH helped with data processing and writing of the manuscript. All authors provided comments and suggested edits to the manuscript. 505 Competing interests. The authors declare that they have no conflict of interest.