Albert’s article Volcanoes and rain explores the possible links between volcanic eruptions and subsequent periods of wet weather. But there is another side to this story: rain can, in turn, have an impact on volcanic activity, and maybe even trigger eruptions!
The possibility of rainfalls causing eruptions has actually been speculated long ago. Georg Forster, who acted as a naturalist during Cook’s second voyage to the Pacific, wrote this observation about the Yasur volcano (Vanuatu) in his travel account A Voyage Round the World, published in 1777:
“We remarked that this was the second time the explosions of the volcano had recommenced after showers of rain; and were therefore led to suspect, that the rain in some measure excites these explosions, by promoting or encreasing the fermentation of various mineral substances in the mountain.”
The Landing at Tana one of the New Hebrides, by William Hodges, painter of Cook’s second voyage. It represents the arrival of the expedition on Tanna Island (Vanuatu) and shows the dark silhouette of the Yasur volcano blowing black smoke in the sky (behind the ship’s rigging). National Maritime Museum.
Of course, two observations of a phenomenon is not considered to be enough data to presume correlation, even less causation, by modern scientific standards. And the proposed mechanism appears somewhat amusing in regard of today’s knowledge. Nevertheless, Forster’s intuition may have held some truth, because rain can indeed influence volcanic activity, even if some of these purported links remain controversial. Let’s start with the more obvious ones.
Some watery hazards
Some volcanic hazards directly involve the presence of water, and I mean atmospheric water, not the water dissolved in magma as a gas phase, nor the marine or land water of phreatomagmatic eruptions. The two mains hazards that require rain are lahars and phreatic eruptions. It sounds like stating the obvious, but a study has shown that lahars are more frequent after heavy rainfalls (Farquharson & Amelung, 2022). More rain means more water can remobilise the ashes down the slopes of a volcano and generate the deadly volcanic mudflows. The same research article demonstrates a similar relationship between heavy rain and phreatic eruptions. In this type of eruption, water heated by a magmatic source suddenly flashes to steam, triggering an explosion of vapour and pulverising the rocks above—no “fresh” (juvenile) magma being emitted. Because this kind of eruption does not involve any movement of magma, it is very unpredictable, and can be deadly even in countries where volcanoes are well-monitored, such as Japan, where the phreatic eruption of Mount Ontake caused 63 victims in 2014. More rain means more water enter the porous, volcanic rocks above a magmatic source, enhancing the probability of a phreatic eruption.
Perhaps less obvious is the link between rainfall and dome collapse. Lava domes being extruded are unstable edifices, prone to crumble. A good example of this instability happened during the long-lived eruption of Soufrière Hills volcano (Montserrat), which produced dome after dome for more than 15 years (1995–2013). A study (Matthews et al., 2009) has shown a statically robust relationship between intense rain episodes and the collapse of these domes. And Soufrière Hills is not the only volcano where rain has been accused of causing dome collapse: it has also been proposed at Japanese Mount Unzen (Yamasato et al., 1998). A possible mechanism is “quench fragmentation”, where the rain provokes rapid cooling of the lava dome surface and its subsequent explosion. Another possible explantation is that water, penetrating the permeable rock of the dome, inhibits its degassing, resulting in overpressure. However, Matthews et al. recognise that “there was no rainfall-related change in deeper, volcano–tectonic activity”, meaning that this interaction remains surficial. As for the case of lahars and phreatic eruptions, rain-triggered dome collapses do not involve the magmatic system. Can rain spark a “true”, magmatic eruption?
Above: Records of rain (blue) and seismicity (blue) on July 12–13, 2003 (Matthews et al., 2009). An episode of intense rainfall is followed by an increase in seismicity corresponding to a dome collapse at Soufrière Hills. Below: The dome before and after the collapse of July 13, which involved 210 millions cubic meters of rocks. Montserrat Volcano Observatory/British Geological Survey.
When water forces magma out
This has been postulated by a study following the East Rift Zone eruption of Kīlauea in May 2018, which curtailed in dramatic fashion the 35-years-long eruptive cycle at Puʻu ʻŌʻō. According to this research (Farquharson & Amelung, 2020), the 2018 events where preceded by an intense rainy episode, with 2.25 meters of rain poured on the volcano during the first quarter of the year, compared to 0.90 meters on average. (These numbers always amaze me, as my city receives less than 0.6 meters annually. Clermont-Ferrand is located in the rain shadow of the Chaîne des Puys, a range of monogenetic cones, or when volcanoes prevent rain!) By means of modelling, Farquharson and Amelung argue that this colossal quantity of water has increased the pore-pressure of Kīlauea and, ultimately, triggered a dyke intrusion without a supply of magma from below. Other authors disagree though (Poland et al., 2022), and this result remains controversial.
If truth be told, this was not the first time that this hypothesis was proposed at a basaltic shield volcano. In the early 1990’s, a French team had already devised a similar idea for the Piton de la Fournaise volcano, on La Réunion island (Carbonnel et al., 1990). And their case did not rest upon one eruption, but upon dozens, giving it a more solid support. Their simple but elegant analysis consisted in splitting the eruptive record in two six-months periods: eruptions happening in November–April, when 70–75 % of the rain occur, and eruptions happening in May–October. For the record spanning from 1800 to 1987, the data yields a ratio of 1.7 times more eruptions during the wet season. For the 1956–1987 subset, for which the record is exhaustive, this ratio is even higher, at 1.85. (It would be interesting to verify if this relation holds up for the most recent period, as the volcano has erupted numerous times since 1987, as recently as February 2026—during the wet season.) This does not mean that rain is always a factor: eruptions do happen during the dry season. But they are less frequent. Again, the invoked mechanism is that the infiltration of meteoric water in the porous and fractured rocks of the edifice puts more pressure on the surficial magmatic system, forcing the magma up. A process that seems pretty intuitive once you think about it. Water adds weight to the volcano, increasing the lithostatic pressure on the reservoir—or should we rather say “litho-hydrostatic” pressure? This effect is not limited to water in its liquid form: a recent study showed a relationship between the snow-load at Ruapehu (New-Zealand) and its activity, which seems to follow a seasonal pattern (Yates et al., 2024). But let us focus on rain here.
An erupting Piton de la Fournaise in September 2016. Some eruptions of the volcano could be linked to rainy episodes. CC BY Martin Mergili, via imaggeo.
The long-term effect of rain on volcanic systems
Contrary to basaltic volcanoes, rain might inhibit the eruption of large silicic systems at a longer timescale (Glazner, 2020). Again, the mechanism is quite intuitive. Water circulating in porous rocks transfers heat to the surface by convection, a process that is way more effective at removing heat than simple thermal conduction through a dry, rocky crust. Over long timescales, the large silicic magmatic systems located in wet regions will tend to cool off and crystallise, yielding granitic plutons, while similar systems in dry conditions will be more prone to erupt. To illustrate this, one can look at a map of the Andes presenting the location of major late Cenozoic calderas (below): most of them are located in the arid parts of the mountain range.
Map showing mean annual rainfall (in mm, from worldclim.org) and the locations of major late Cenozoic calderas. They are mostly found in arid parts of the Andes. CC BY Glazner (2020).
Rain is so powerful that it could even displace an entire volcanic range! How? Orography has an impact on climate, creating a wet windward side and a dry leeward side. On the windward side, precipitation enhance erosion. The eroded crust, more heavily fractured, offers an easier path for magma to reach the surface. New volcanoes then tend to grow on the windward side, where the process will eventually repeat itself. Over millions of years, the whole range migrate. This effect of rain has been proposed to explain the westward migration of volcanism in the Cascades and in the Southern Andes during the late Cenozoic (Muller et al., 2022). In both cases, rain from the Pacific Ocean would have promoted erosion on the West side of the mountains, in turn promoting the ascent of magma on this side and ultimately the migration of the range.
According to this model of the Andes, the western part of the crust, more eroded by precipitations, favours an asymmetric rise of magma to the windward side of the continent. Over long timescales, this process makes volcanism migrate toward the West. CC BY Muller et al. (2022), modified.
This article shows that the volcano–climate relationship in not one-sided. Yes, volcanic eruptions can exert a hold over the climate. But in turn, there can be a climatic forcing on magmatic and volcanic activity. This influence actually extends beyond that if rain: glaciation cycles also play a major role in magma production, but that is another story. Anyway, it should be no surprise that water is a powerful agent in shaping a small marble called “the Blue Planet”, including its volcanoes.
Jean-Marie Prival, February 2026
Parts of this post have previously been published in Kipuka, a French magazine which ‘revue de vulgarisation scientifique dédiée aux volcans’. (If you feel unsure about vulgarising volcanoes, it means popularising, i.e. making it accessible to the general public. VC pleads guilty: we also share this goal.)
References
Farquharson, J. I.; Amelung, F. Volcanic Hazard Exacerbated by Future Global Warming-Driven Increase in Heavy Rainfall. R. Soc. open sci. 2022, 9 (7), 220275. https://doi.org/10.1098/rsos.220275
Matthews, A. J.; Barclay, J.; Johnstone, J. E. The Fast Response of Volcano-Seismic Activity to Intense Precipitation: Triggering of Primary Volcanic Activity by Rainfall at Soufrière Hills Volcano, Montserrat. Journal of Volcanology and Geothermal Research 2009, 184 (3–4), 405–415. https://doi.org/10.1016/j.jvolgeores.2009.05.010
Farquharson, J. I.; Amelung, F. Extreme Rainfall Triggered the 2018 Rift Eruption at Kīlauea Volcano. Nature 2020, 580 (7804), 491–495. https://doi.org/10.1038/s41586-020-2172-5
Poland, M. P.; Hurwitz, S.; Kauahikaua, J. P.; Montgomery-Brown, E. K.; Anderson, K. R.; Johanson, I. A.; Patrick, M. R.; Neal, C. A. Rainfall an Unlikely Factor in Kīlauea’s 2018 Rift Eruption. Nature 2022, 602 (7895), E7–E10. https://doi.org/10.1038/s41586-021-04163-1
Carbonnel, J.-P.; Soler, E.; Hubert, P. Hypothèse sur la genèse de certaines éruptions volcaniques du Piton de la Fournaise (île de la Réunion). C. R. Acad. Sci. Paris 1990, 310 (II), 75–80
Yates, A. S.; Caudron, C.; Mordret, A.; Lesage, P.; Pinel, V.; Lecocq, T.; Miller, C. A.; Lamb, O. D.; Fournier, N. Seasonal Snow Cycles and Their Possible Influence on Seismic Velocity Changes and Eruptive Activity at Ruapehu Volcano, New Zealand. Journal of Geophysical Research: Solid Earth 2024, 129 (12), e2024JB029568. https://doi.org/10.1029/2024JB029568
Glazner, A. F. Climate and the Development of Magma Chambers. Geosciences 2020, 10 (3), 93. https://doi.org/10.3390/geosciences10030093
Muller, V. A. P.; Sternai, P.; Sue, C.; Simon-Labric, T.; Valla, P. G. Climatic Control on the Location of Continental Volcanic Arcs. Sci Rep 2022, 12 (1), 22167. https://doi.org/10.1038/s41598-022-26158-2
Good JOb.
The influence of rain on volcanism reminds a bit to the extinct Neptunist geology worldview which regarded geological processes as predominantly water driven. Contrary to this, the Plutonist worldview proved to be predominantly correct.
The most important geological process caused by rain is the interaction with limestone: Karst geology. Croatia and South China have examples for the karst geology.
As far as I remember right, I saw an effect of rain on Kilauea after the last episode. There was an extensive field of steaming on the new lava field. It was an example for hydrothermal effects of rain. Hydrothermal effects are the smallest but most frequent effect of rain on volcanoes. Phreatic eruptions are a larger scale, but they need more than just rain; they need the abundance of groundwater or significant surface water (f.e. a lake, a swamp, a river …).
Water can also turn any dry magmatic eruption into a Phreato-magmatic eruption: a dry Hawaiian, Strombolian, Vulcanian, Plinian eruption looks very different, if the same magma meets water somehow. Grimsvötn 2011 is an example for a Phreato-Hawaiian eruption. Monte Nuovo 1538 was an example for a Phreato-Strombolian eruption, also the smaller Grimsvötn eruptions like 1983 are Phreato-Strombolian. Krakatau and HTHH were examples for Phreato-Plinian eruptions. Usually they happen in a wet climate. The water must be there with more mass than just a rain shower.
Water does have an effect on rock. It is of course important to magma formation in the first place: water lowers the melting temperature of rock, so magma can form where for dry rock it wouldn’t. For earthquakes, water reduces the effective friction coefficient, so putting water in a fault can trigger an earthquake. Adding water to a pre-eruptive volcano could also have an impact in different ways. where a volcano needs to break through a rock, water lubricating an existing fault can provide a pathway. Where there is already a lot of heat, adding water can cause a phreatic explosion, and this may also help providing pathways for magma. And water can destabilise a slope, and this could play a role in a flank collapse – which can also lead to an eruption. Whether these things happen is a matter for discussion. But I would not reject them based on plutonic priority. Just some thoughts. The question is perhaps whether it is rain or ground water.
It’s also a question whether it’s weather or climate. Weather can change from week to week. There can be showers, thunderstorms, floods, but they have their random time. But Climate is a rule that applies to many years, it supplies a regular weather. In the western US volcanic systems you have some volcanoes with dry desert climate (Arizona, New Mexico) and others with wet westwind conditions (Washington, Alaska).
The rain discussed 2018 about Kilauea was a weather phenomen, as far I imagine. But we can discuss the meaning of the wet subtropical side of Big Island (east rift zone) compared to the dry subtropical side (Ka’u desert and SWRZ).
The Andes volcanoes discussed in the article show the impact of different climate zones on the volcanoes. There is a good quasi-experimental situation to study the realtion that otherwise is often rare in volcanology.
Since the article has been published (thanks again Albert!), I took the time the look at the recent numbers for Piton de la Fournaise. Since March 1998, there has been 60 eruptions. 30 started in November–April, 30 in May–October, so it’s a perfect tie!
This article focuses on rain, but it should be noted that a major impact of water on volcanism is in the form of ice. It has long been shown that, at the beginning of an interglacial period, the removal of ice sheets lowers the pressure, thus increases magma production by decompression melting, and ultimately eruption rates. Except at oceanic island settings, where the effect is opposite (rising sea levels increase pressure and inhibit magma production).
Anyone here have an idea what’s up with the continuous sequence of M4-6 earthquakes NW of Miyakojima? They are occurring over a wide range of depths. There is an active fault on the island itself, but these seem too distant and/or deep for that, and I can’t find good information on a source for this current series.
The island is not volcanic, so this is tectonic.
UWD shows the unusual wave-shaped deformation on Kilauea:
HVO says that this up and down of inflation is unique compared to previous episodes. It increases uncertainty in the prediction of the next episode.
In hindsight this started during the previous inflation episode. I was wondering whether it is the weight of the new cone and surrounding lava, pushing the magma below sideways
It may also be a mixture between non-volcanic influences (weather …) and a slow shift towards DI deformation cycles.
They could also be an early indicator for a future continuous eruption style contrary to the current episodic eruption style. As Mauna Ulu and Pu’u O’o showed, the child stage of longterm eruptions often is episodic, while the adult stage is continuous.
If we look at Kilauea Iki, then the episodic eruption continued until the end. F.e. the five last episodes (13 to 17) December 16th to December 20th 1959 were all very similar to the current episodes with tall spectacular lava fountains and no quiet effusive activity. The 16th episode still produced 500m tall lava fountains. Like the 2024-2026 eruption the first initial episodes of Kilauea Iki were relatively calm and effusive. Kilauea Iki was in short time, what the current eruption is in long time.
E1-E3 in December 2024 to January 2025 had an average height of 113 meters.
E40-E42 in winter 2026 had an average height of 367 meters.
New map for Kīlauea here:
(Post Ep. 42)
(Post Ep. 38)
Vents seemed to grown 20 more meters above the floor from Ep. 38 and the lava level has risen by 17 meters, all in just 4 episodes!
They also mentioned something about a cone, too. I don’t think it’s that mound on the caldera rim but rather the one close to the North vent.
A cold lahar on Merapi struck today and wiped out some houses and cars. The power of rain.
On Iceland earthquakes on half way between Ljosufjell (Snaesfjellsnes) and Reykjanes Peninsula look like an evidence that the systems have a positive correlation and relation. I have the impression that during the “Reykjanes Fires” the whole west of Iceland functions like Vatnajökull, EVZ and NVZ during normal times. So during the next 500 years we can expect eruptions in the west of Iceland on average as frequent as in the east.
IMO has a new hazard assessment out. I haven’t looked at it but here’s a write up in RÚV:
Magma level approaching highest seen since start of eruptions (RÚV, 3 Mar)
Not many tremors on the Reykjanes Peninsula lately, so I don’t know where all this magma is going. Maybe after the last dike there’s more room for it to fit in along that line of weakness.
Imo announced that the length of refilling and the magma volume asks for a reevaluation of possible future activity, which theyll share in the coming weeks. Every real eruption since march 2024 ended up being between 2.5-4x larger than their initial injection volumes. April 2025 couldve become an event of 75-120million m³ based on that factor (initial injection measured 31million m³). Since svarts appears to be a layered sill stack inside a domain, how accurate can a point source based model even be especially over extended periods of time?
The speed of recharge is slowly decreasing. This spring looks like the last moment for an eruption, but then it will likely be exhausted for a while.
Maybe the next episode will be the last episode, and then after years a third (Fagra first, Svartsengi episodes second) discrete eruption will happen: “One episode of eruptive activity (‘Fires’) has occurred in Reykjanes-Svartsengi volcanic systems in historical times, 1210-1240 CE. During these Fires, at least six discrete eruptions occurred at 2 to 12 year intervals” https://icelandicvolcanos.is/#
1210-1240 eruptions sometimes happened on Svartsengi and sometimes on Reykjanes, so the location can shift. We’ve entered a possible 30 years (war) eruption 2020 that could last until 2050 with varying activity.
I expect that those midway earthquakes are wrongly located and combine independent tremors in the two locations.
How else can we observe if the west of Iceland (RVZ, SVZ) becomes more active together by a causal relation? There must be something, that drives both systems at the same time to increase activity. Just like the “big volcanoes” of the east are linked together by EVZ, NVZ and SISZ, there must be a link, if the western systems usually erupt in the same periods.
Not the best thing to report about Kilauea.
From Fox news
“Man dies after sneaking into closed section of popular national park
The 33-year-old’s death at Kilauea caldera remains under investigation by National Park Service officials”
No details released.
Mac
Surprisingly after the 79 Pompei eruption Vesuvius changed its behaviour the way that after long dormant periods the next eruption wasn’t a Plinian one. The 472 and 1631 Plinian eruptions were preceded by a minor eruption 100 resp. 60 years before. The long gap was before the minor eruptions. First 235 to 379, second between 1150 and 1500/1570.
The 79 eruption had a major rain component. Herculaneum was hit by Pyroclastic flows first, but second by Lahars caused by depressions running from the west into the Bay of Naples. Vesuvius had this in common with Pinatubo 1991, when a Taifun hit the Philippines during the Plinian eruption.
At https://pubs.usgs.gov/publication/pp1890C/full “Forecasting Volcanic Activity in Germany—A Multi-Criteria Approach” about the volcanism in Germany.
The study has a longterm approach of 1 million years. It is related to the public search for safe places for radioactive waste. They looked for all Cenozoic eruptions (66 million years) to determine the probability for eruptions in the next one million years.
New post is up! It is raining earthquakes. Will it or won’t it? Did it?
https://www.volcanocafe.org/the-recent-eruptions-of-teide/