I take it where I left it in my last post. Using volcanic signals (mainly sulfate spikes) in the Greenland and Antarctica ice cores, I listed the most powerful explosive eruptions between the years 1500 and 1900:

Many of these eruptions have written records, so it was easy to know the source volcanoes: Krakatau (1883), Cosigüina (1835), Tambora (1815), Parker/Mélébingóy (1641), Huaynaputina (1600), and Raung (1593). The 1831 spike has been linked to Zavaritskii based on tephra found in Greenland. While the 1694 and 1621 ones seem to match Long Island and Billy Mitchell, respectively, based on timing and the characteristics of the spike. The 1808 eruption is still a bit of a mystery, though I ventured Agrigan as a possibility.

However, as I reach back in time, it gets harder. Other than Samalas, none of the pre-1500 AD eruptions have written records, so it’s a matter of linking a particular volcanic signal to a source eruption based on timing and other characteristics of the given signal. To make identification possible, I need to extract as much information as possible from the ice-core signals. So, before I continue with the eruption list, I will talk about some of the clues I’m looking for in the ice core signals and the things I’ve improved and learned since the last article.

Size of sulfate injections

The amount of sulfur injected into the stratosphere by volcanic eruptions can be estimated from the amount of aerosol deposited in the ice. Here on, I’ll be using the estimates from Michael Sigl and Matthew Toohey (2024), which cover eruptions from 500 BC to 1900 AD, and are given in teragrams (Tg) of sulfur. The size of the injection more or less ramps up together with the overall volume of the eruption, though how gassy the magma is also seems to matter, which is something I’m not yet sure if it can be predicted from the chemistry or the type of volcano.

Sulfate dispersal between the poles

Eruptions at mid to high latitudes will evidently be deposited preferentially or solely in the pole closest to the eruption. For example, Greenland ice cores have a large number of volcanic signals from Icelandic volcanoes that are not visible in Antarctica, ranging from fissure eruptions to shield volcanoes and to the explosive events of Hekla, Öræfajökull, or Katla. One thing I was not sure of is whether signals from powerful mid-latitude northern hemisphere eruptions would show in Antarctica. The answer is that they do. The timing of the spikes of the Aniakchak II and Okmok II eruptions (Alaska) is known from tephra found in Greenland, and these two spikes show clearly in the WAIS and SPC14 ice cores in Antarctica, as does the Minoan eruption of Santorini, which is reasonably well constrained from high precision dating and the size of the spike. Zavaritskii (1831, Kuril Islands) shows in the WAIS, and perhaps in the SPC14, though somewhat unclear, while Mount Churchill’s White River Ash (Alaska) signal does show in the SPC14 ice core. So it seems that if an explosive eruption is powerful enough (ultraplinian?), it will show in both poles, even if the location is extratropical.

Because l was not too sure about the exact way sulfate is distributed, I’ve calculated the ratio between the estimated Greenland sulfate deposition and the total deposition using the SPC14 data to gauge Antartica’s, which I’m relying a lot on, due to the high temporal resolution and little noise from other sulfate sources (unlike the WAIS and maybe other Antarctic ice cores that have a lot of noise and seasonality).

After plotting the data, there was one surprise, and it’s that there’s a difference between tropical eruptions south and north of the Equator (which until now I had been treating the same). In particular, it’s possible that only the eruptions south of the Equator can have a Greenland/total ratio of less than ~0.54. This could have an explanation: the eruptions north of the Intertropical Convergence Zone would be dispersed more to the north, while the southern hemisphere tropical eruptions are generally dispersed preferentially to the south, but with perhaps the strong SE Asian monsoon? sometimes deflecting them mostly into the north.

Though it’s hard to tell, given the few extratropical events used since 800 AD (probably just two), I expect the northern hemisphere extratropical events to have Greenland sulfate/total >0.7, and southern hemisphere extratropical events <0.3. Almost all major volcanic signals seem to come, however, from tropical regions.

Sulfate duration and shape

This is one of the most interesting aspects to me, and where the SPC14 (ice core in the location of South Pole, 0º south) data from Dominic A. Winski (2021) shines the most. This data has very high temporal resolution and very little noise from non-volcanic sulfate sources. Two things became evident when first plotting the data: that the more powerful eruptions almost always have durations of 2.5-3.5 years, regardless of overall volume, and that spikes come in different “shapes”. Volcanic aerosol clouds are made of tiny sulfate particles suspended in the atmosphere, so it makes sense that the amount of time it takes for the cloud to disappear is related to how long it takes for these particles to fall out, which should be longer the farther up the stratosphere they are injected. It seems the duration of the volcanic signal may be related to this, with potent ultraplinian eruptions generally producing similarly long sulfate spikes in the poles, 2.5-3.5 years, even across an order of magnitude variation in volume and sulfur yield.

It turns out the shape may also be related to this. The timing of the peak in sulfate deposition varies: usually between 0.65 and 1.75 years after the onset of the spike. Some signals are more stretched out, and others are more focused. At first, I assumed it’d have to do with how far the source eruption was from Antarctica, taking longer to peak the further away from it. This was, however, “crushed” by the signal of the Okmok II eruption (Alaska), which has an early peak in Antarctica, and a shape that resembles similar eruption style tropical eruptions. Okmok is a basaltic-andesite caldera-forming ignimbrite, and the shape is very reminiscent to that of the volcanic signals that I’ve attributed to Long Island, based on timing, which is also a basaltic-andesite CFE ignimbrite; it also resembles Tambora, which is a trachyandesite CFE ignimbrite (so relatively close composition and style). That’s why I think the shape of the sulfate spike in the SPC14 core may be related to the eruption style, or more specifically, to the altitude distribution of the sulfur injection.

Differences are slight, and it can be a problem if the start of the signal onset is misidentified or if there’s some weather disturbance that distorts it, particularly in the weaker signals, but I do think the shape of the sulfate spike can be useful in many cases to recognize the eruption. I’ve also been dealing almost exclusively with tropical eruptions too, so I have not yet had much of a chance to see how extratropical eruptions may behave. There are few confidently identified eruptions for me to establish how different eruption types may be reflected in the ice, but some potential patterns are visible.

Andesitic/basaltic-andesitic ignimbrite-dominated eruptions seem to have early peaks followed by a steady decrease, so most sulfate is probably injected into relatively low altitudes by the dense scoriaceous coignimbrite clouds. Submarine eruptions may span a range of shapes, and I think they are more likely to be recognized by their large, disproportionate drops in nitrate, but one sulfate type that seems dominant among the large nitrate drop eruptions is that of a relatively short-lived signal, with a concentrated peak of deposition 1 year after the onset followed by a very fast decline and then a faint “tail” to 2.5 years. Other eruptions can be complicated and relatively unique looking, for example, Huaynaputina’s 1600 eruption probably reflects a lower altitude injection (1 year into the signal), possibly by PDCs and vulcanian activity, and a higher altitude injection (around 2 years), possibly by the more intense sustained plinian phases. The higher silica eruptions, particularly rhyolite CFEs, and maybe mafic plinians too, probably tend to peak towards later times. Lastly, all sorts of mafic (basaltic andesite) eruptions may have unusually long total durations of 3.5-4.5 years (Raung, Okmok, and the likely signals of Tofua and Long Island).

Nitrate drops

I’ve already mentioned this aspect in the previous installment. All or almost all major volcanic spikes (>2.4 yr duration) in the South Pole core come with drops in the nitrate concentration (another of the atmospheric aerosols usually measured in ice). The shape is also variable, but I think mostly related to the overall size of the drop. The bigger drops usually have a U shape with a smooth bottom that is unaffected by background nitrate variations. Smaller nitrate drops instead have more often V shapes, or U shapes with irregular floors, that let the background nitrate variations through.

I believe the origin of these drops may be a complex cascade: first, injection of chlorine into the stratosphere by a volcanic eruption; second, chlorine reacting with and destroying stratospheric ozone; third, reduced ozone letting in increased UV radiation that photolyzes stratospheric nitrate, breaking it into nitrogen dioxide and other compounds; and finally, reduced stratospheric nitrate downwelling over central Antarctica lowering the concentrations in the south pole ice. The mechanism is still speculative, but the usefulness in identifying eruptions would be that it may help pinpoint submarine eruptions.

In the scenario of chlorine-depleting ozone, it’d make sense for submarine eruptions to have a disproportionate impact on nitrate, because the interaction between lava and seawater is known to produce a haze that is rich in HCl gas, thus providing additional chlorine to the volcanic plume (to the magmatic chlorine already present in magma). Some evidence supports this notion. Krakatau (mostly submarine eruption), despite having only 1/3 of the estimated sulfur yield of Tambora, has twice as much nitrate drop. Another supporting evidence is that major sulfate spikes between 1100 and 1300 AD greatly outnumber the candidate eruptions, but with there being multiple major nitrate drops among these, some of the eruptions would therefore be submarine (and thus unstudied), mostly solving the mysterious spikes. And even then, two subaerial events of this interval remain enigmatic.

Subaerial eruptions can also produce major nitrate drops, and the CFE of Samalas/Rinjani is an example of such a case, but this is not necessarily in conflict with the idea of the large drops being characteristic of submarine eruptions. The Samalas eruption and gas content were massive, 59 Tg estimated sulfur yield, the largest of the last 1000 years, so it’s not unlikely that the proportionate amount of magmatic chlorine of Samalas managed an effect comparable to that of the submarine eruptions’ hydrochloric gas.

With this explained, I will now resume the list of the largest volcanic explosions our world has seen:

 

Largest explosive eruptions 1500-1000 AD:

1457, Kuwae (Vanuatu), 33 Tg of sulfur

The 1457 eruption is a massive sulfate spike, estimated at 33 Tg of stratospheric sulfate, which would be only second to Samalas in the last 1000 years, and similar to Tambora. It had long been suspected that the 1457 eruption was sourced by a collapse of the large Kuwae caldera in Vanuatu, dated around this time. Recently, the identification of tephra with chemistry similar to Kuwae has confirmed its identity (Petit, J.-R., et al.; 2025). This event also carries the largest nitrate drop in the SPC14 core (though only barely larger than Samalas), probably a combination of vast amounts of magmatic Cl and additional HCl flashed from seawater during the underwater eruption.

Local traditions suggest that the islands of Epi and Tongoa were once joined in a larger landmass, called Kuwae, separated in a large volcanic eruption, preceded by smaller explosions and accompanied by tsunamis. A complex, ~60 m thick sequence of pyroclastic deposits on Tongoa Island is thought to have been deposited in this eruption. The eruption started with minor maar activity of basaltic and andesitic composition, followed by a climactic ignimbrite eruption of crystal-poor (4-7 wt% phenocrysts) dacite.  Using an apparent average depth of 700 m (assuming no later resurgence and ~100 m of syn-eruption ignimbrite fill) the caldera-forming event of Kuwae had a volume of 40 km3 DRE, which makes it roughly equivalent to Tambora.

Most of the Kuwae eruption probably happened underwater, so it makes a point that maybe submarine calderas can still have much of their sulfur reaching the stratosphere. That said, Kuwae is a relatively shallow one (it started above water, and its submarine rim is generally no deeper than 100 m bsl.

 

1453, Cosigüina (Nicaragua)?, 10 Tg of sulfur

This eruption’s signal is very short-lived, small, and without a nitrate drop in the SPC14, but it is 2.4 years in duration in Greenland, so I decided to include it. It’s also a signal that has been very debated because it occurs around the start of a major cold period that continued through Kuwae’s eruption and greatly affected tree ring growth in the northern hemisphere. But I think it can be argued whether this cold period was volcanic in nature, given the modest size of the 1453 sulfur signal.

My preferred candidate for 1453 is Cosigüina. The Nicaraguan volcano has ~4 explosive deposits thought to date from 1500 AD and later, and to be from larger magnitude eruptions than the 1835 eruption of the volcano. However, the Spanish had already founded towns close to Cosigüina by 1522 AD, so I don’t think any similar or larger than the 1835 Cosigüina eruptions can date from after 1522. The eruption reports in 1609 and 1709 don’t match sulfate spikes, while the reported March 1809 eruption is unlikely to be responsible for the 1808 spike because it postdates sighting of the aerosol cloud. So I think some mistake has been made with the stratigraphy that has led to the multiplication of young major eruptions.

Nonetheless, there is at least one major, dated deposit >12 m thick (Sequence I), which contains broken animal bones and carbonized tree trunks when PDCs devastated the Cosigüina peninsula. It’s basaltic-andesite (possibly crystal-rich with more evolved melt composition). Sequence I carbon yields two ranges when calibrated, 1314 to 1362 and from 1387 to 1475, the second being the one with the highest probability, so I think the 1453 spike is from the Cosigüina Sequence I eruption. It could also be that Sequence I corresponds to the 1344 signal, while 1453 belongs to one of the later Cosigüina eruptions.

 

1344, Pinatubo (Philippines), 15 Tg of sulfur

The characteristics of the 1344 signal suggest a subaerial eruption somewhere in the tropics. The time distribution of the sulfate spike is a bit unique, but carries a resemblance to Huaynaputina, and, oddly enough, Long Island 1694 (which is almost as far as of Huaynaputina’s eruption style as possible). But the closest analogue is another mystery event in 1109 AD. This sulfate distribution is a bit of a wild card in this case, since the only two known similar peaks come from widely different eruption styles/composition. I haven’t looked to compare with the 1991 eruption because anthropogenic sulfate noise had become very strong by the time it happened.

So, I have chosen Pinatubo. The previous activity before Pinatubo’s recent collapse in 1991, is known as the Buag eruption cycle. It’s known that this cycle involved a major plinian eruption that carried ash to SE Asia, where it’s found in the Huguangyan Maar (Hainan Volcanic Field) as cryptotephra and estimated as roughly around 1000-1200 AD. It’s likely that the start of the eruption cycle commenced explosively, resulting in a caldera event similar to the 1991 eruption. The tephra fall of the eruption has been stripped away by erosion, and there don’t seem to have been major pyroclastic flows either. The only PDCs known are from dome collapses during effusion towards the end of the Buag cycle. However, there are numerous lahars that were likely produced when torrential rains reworked the tephra fall. Many lahars have been dated with radiocarbon; these have mostly similar ages and give an average that calibrates to 1350 AD  (1276-1415 AD). This matches perfectly the timing of the sulfate spike, so I think it makes a strong argument for the Buag plinian eruption happening in 1344. Another, weaker, argument in favour could be the slight spike resemblance to Huaynaputina, which is also a small crystal-rich dacite caldera system. Volume is likely to have been similar to 1991, ~5 km3 DRE (dense rock equivalent), a ~1 km deep collapse of its caldera.

 

1290, Okataina (New Zealand), 1.3 Tg of sulfur

The 1290 eruption is one of the faintest >2.4-year-long volcanic signals in the South Pole, somewhat below the 1621 spike that I’ve attributed to Billy Mitchell, and the two early 11th-century spikes. It’s only found in Antarctica, so it’s likely a southern hemisphere extratropical eruption, subaerial too. I imagine it’s visible in both poles, but since it’s already faint in Antarctica, then in Greenland it can’t even be recognized.  The sulfate time distribution is relatively unique, though with some similarities to Huaynaputina (1600) and the 1262 event. Something unusual is that it has a very abrupt start, rising to a peak 0.6 yr after the signal start, maybe due to the closeness to Antarctica, due to tropospheric sulfate arriving swiftly at the south pole.

I think this is the only major volcanic signal that can fit the Kaharoa eruption of Okataina volcano. This VEI 6 eruption has been previously dated to around 1305 or 1315 AD, very close to this southern hemisphere spike. So I think the identification is relatively evident in this case. The Kaharoa event involved a plinian fissure eruption, 8 km3 DRE and 16 km3 bulk, plus 1 km3 of lava that formed three lava domes and block and ash deposits. The composition was high silica rhyolite, crystal poor at first (5-25% phenocrysts) and crystal rich in its later phases (20-40 wt% phenocrysts). Unlike major eruptions 1000-1900 AD, it did not form a caldera because it was supplied by a very large rhyolitic reservoir of the Okataina volcano.

 

1285, Quilotoa (Ecuador), 15.06 Tg of sulfur

The characteristics of the 1285 spike suggest a subaerial eruption on the southern side of the tropics. The sulfate time distribution has certain similarities to many spikes: Mélébingóy/Parker (1641), Billy Mitchell? (1621), Kuwae (1457), Samalas/Rinjani, and the 1109 event. These spikes are dacitic eruptions ranging from crystal-rich (Parker, Billy Mitchell) to crystal-poor (Kuwae, Samalas?), and there was major pyroclastic flow activity in at least three of them.

My candidate is the Q-I eruption of Quilotoa (Ecuador). Q-I was the last of several caldera events of Quilotoa, happening at intervals of 10,000-15,000 years. The eruption has an estimated volume of 21 km3, and the caldera would suggest around 6 km3 DRE if ~1 km deep, but, in these small dacite calderas, it’s not really clear if the mechanism of volume ejection is really chamber collapse. The magmas of Q-I and older Quilotoa eruptions are always very similar, crystal-rich (10–50 vol.% phenocrysts) and with dacite composition (65-67 wt% silica). The eruption is estimated to have produced a 35 km tall plume and exceptional amounts of fine ash due to the violence of fragmentation. Minor phreatic explosions preceded the main event. The eruption then ejected a lake that existed inside the caldera, an estimated flood of 250 million m3 of water gushed out of the crater, carrying large boulders and excavating ridges and gullies. Then a sustained plinian column ensued, gradually transitioning into an ignimbrite phase dominated by pyroclastic density currents that travelled 10 km from the caldera.

The eruption has been previously suggested to occur roughly around 1280 AD, which is the average yielded by the radiocarbon ages when calibrated. The geochronology is in perfect agreement with the timing of the 1285 spike, and the sulfate time-distribution matches with eruptions that have very similar characteristics to the Q-I eruption (dacite composition with major PDC activity). With the estimated latitude in agreement too, I think it’s almost certain that Quilotoa is the 1285 volcano.

 

1276, Kermadec Arc submarine?, 11.53 Tg of sulfur

The 1276 signal comes with a large U-shaped drop in nitrate concentration, similar to Krakatau (1883), in the South Pole Core. This may indicate a submarine eruption with a large plume of HCl gas. The sulfate time distribution is very similar to most of the other eruptions I suspect to be submarine, also to the older presumed Tofua spike (but without the long duration), and also somewhat to the 1808 signal. This event dispersed the vast majority of the sulfur into the southern hemisphere: Greenland/(Greenland+Antarctica) is 0.14, meaning a southern hemisphere extratropical eruption. I think this was probably a caldera collapse in the Kermadec Arc (north of New Zealand) that concentrates a large number of submarine calderas, and most of the southern hemisphere extratropical ones.

Sometime around 1000 years ago, a large amount of sea-rafted dacite and rhyolite pumice washed onshore New Zealand, it’s known as the Loisels Pumice, and the chemical composition implies an origin in the Tonga-Kermadec Arc. I haven’t tried looking for a chemical match among the Kermadec calderas because most of them don’t have available samples, and even fewer have the glass samples that would be necessary to solidly match a distal tephra. The 1276 eruption was probably one of the main contributors to the Loisels Pumice, though the 1269 eruption probably contributed too. I wouldn’t fully rule out a South Sandwich Islands eruption instead, though.

 

1269, Kermadec Arc submarine?, 3.17 Tg of sulfur

About seven years earlier, there was another volcanic signal that was nearly identical in all aspects to the one in 1276. It too has a large U-shaped nitrate drop, the same sulfate time distribution, and a dispersal ratio that indicates a southern hemisphere extratropical eruption, Greenland/(Greenland+Antarctica) is 0.21. The only difference is that the 1276 has a larger estimated sulfur yield. Though in the SPC14 core, the 1269 sulfate signal is only slightly smaller, so the difference in estimated yield may not be as large as thought. Likely, this was another caldera event in the Kermadec Arc and another contributor to the Loisels Pumice.

 

1262, unknown, 1.05 Tg of sulfur

This is by far the faintest of the sulfate signals >2.5 years long in the South Pole, but since it has an apparent nitrate reduction and matches with the WAIS and Greenland ice cores, it was likely a powerful, albeit small, explosive eruption. The Greenland/total ratio of about 0.66 is around the expected limit between tropical and northern extratropical eruptions. I doubt the time distribution of the surface can be reliable in this weak signal. There’s much uncertainty, and I don’t have any good candidates for this one.

 

1257, Samalas/Rinjani (Indonesia), 59 Tg of sulfur

The 1257 eruption of Samalas is a massive event, with by far the largest estimated sulfur yield among the post-1000 AD eruptions. It is also the largest sulfate signal in the SPC14 core, albeit only slightly larger than Tambora, but this is because the 1257 eruption was one of the rare southern tropical events that are preferentially dispersed towards the north; its sulfur yield seemingly considerably higher than Tambora and Kuwae. It’s also the second-largest nitrate drop after Kuwae’s, likely accomplished through the vast amounts of volcanic gases injected into the stratosphere (including magmatic chlorine). The tephra in the ice cores has been chemically linked to the eruption that formed the spectacular Samalas caldera in the Rinjani volcano complex.

I find a caldera collapse volume of at least ~30 km3 DRE, which is the same amount as estimated from fieldwork, so maybe slightly smaller/comparable to Tambora (1815) and Kuwae (1457), and also probably around the size of Long Island (1694). But Samalas was likely a more gassy eruption than the other three, particularly Long Island. The eruption started with 8 km3 DRE of plinian fall pumice. The eruption then entered an ignimbrite phase that erupted 16 km3 DRE of pyroclastic flows and 8-9 km3 of coignimbrite ash (the hot plume that develops above an ignimbrite). This coignimbrite ash layer is visible in exposures as far as 660 km away from Samalas, in central Java. The eruptive column is estimated to have reached 43 km high, and pyroclastic flows reached at least 25 km away. Erupted magma is trachydacite (64 wt% silica).

1229, Tonga/Mariana submarine?, 24 Tg of sulfur

The next volcanic signal in 1229 is another major event. This is the 4th largest sulfate spike in the South Pole, as well as the 4th largest estimated stratospheric sulfur injection since 1000 AD (after Samalas, Kuwae, and Tambora). Nitrate characteristics of this event indicate that it was probably a tropical submarine eruption, most likely in the Tonga Arc, which has by far the greatest concentration of tropical submarine calderas, or less likely in the Marianas. The shape of the sulfate spike is distinctive, with only close similarities to Huaynaputina (1600) and 1109 events, and very different from the other suspected submarine eruptions. Because of this, I’d speculate that (maybe) it started as a subaerial (dacitic?) eruption before becoming submarine, or that it happened through very shallow water.

A cryptotephra has been found in the South Pole ice, but it occurs well after the onset of the signal, and the various glass shards have disparate chemistries, so I don’t think it’s tephra from the eruption responsible (tephra from a major explosive eruption falls at the start of the sulfate signal).

 

1191, Southern tropical mafic unknown?, 8.5 Tg of sulfur

There is an odd volcanic spike in 1191. The characteristics of the spike would suggest a subaerial eruption in the southern side of the tropics, and with a shape that only seems (somewhat) similar to the 1593 basaltic-trachyandesite plinian CFE of Raung volcano. There is an Antarctic cryptotephra around this time, but I don’t know where exactly it falls in relation to the south pole core, so it may not come from the source eruption of the 1191 sulfate signal. The tephra is rhyolitic (with mid-alkalinity within subalkaline compositions and high potassium to sodium ratio). Either I’ve displaced the onset of the 1191 signal or this is perhaps some mafic plinian undocumented eruption, maybe an earlier outburst of Raung itself, or some powerful eruption of Ambae or Ambrym.

 

1171, Tonga/Mariana submarine?, 18.05 Tg of sulfur

The characteristics of the 1171 signal suggest a tropical submarine eruption, most likely in the Tonga Arc, or less likely in the Marianas. The spike shape is similar to most of the other eruptions I suspect to be submarine.

There is a cryptotephra in the SPC14 that matches perfectly the onset of the signal and contains several shards of identical composition: rhyolite with very low alkalinity and high potassium-to-sodium ratio. The cryptotephra is probably from the same volcano that deposited the sulfate, but I haven’t found any matching submarine calderas, this should not be surprising, though, since it’s already few submarine calderas that have chemistry samples of their lavas and even fewer that have glass samples that could be linked to distal shards in Antarctica (only the glass, the molten portion of the magma upon eruption travels long distances through the air).

 

1126, Puyehue (Chile)?, 3.68 Tg of sulfur.

The 1126 event is, according to the signal characteristics, a subaerial, southern tropics eruption. The shape is very rare, seemingly peaking 2.5 years after the onset, which is the latest peak of any signal I’ve seen yet. It doesn’t really match the distribution of any other cases; it has some weak similarities to Raung 1593 and 1191, but the full signal doesn’t last as long. The 1290 event that I’ve attributed to Okataina has a substantial proportion of sulfate deposition 2.5 years into the event, so maybe this can be taken as a sign of both 1290 and 1126 being energetic, highly silicic eruptions, maybe extratropical locations also helping with the slow removal of the sulfate. So one far-fetched guess is Puyehue, which had a rhyodacite plinian eruption very poorly dated to around 1050 AD (Mil Hojas), though it could be anywhere 650-1300 AD. Puyehue doesn’t seem a bad match in terms of timing, but it still remains a wild guess for an event that is hard to identify.

The Mil Hojas eruption is thought to have formed the summit crater/caldera of Puyehue, and if a 1 km deep collapse is correlated with the eruption volume, that will be around 4 km3 DRE. The latitude suggested by the dispersal between the poles would have to be off if it’s Puyehue, though it could match Billy Mitchell, which was my first option for this event before changing it to the next older spike.

 

1109, Billy Mitchell (Papua New Guinea)?, ~19.16 Tg of sulfur*

*Hallmundarhraun probably goes into the sulfur yield.

The dispersal ratio of the 1109 eruption would indicate northern extratropical, but I think this is probably an error because the 1109 event of Antarctica is coincident with a strange plateau-shaped 5-year-long signal in Greenland that I think is the 9 km3 lava flow of Hallmundarhraun in Iceland. So I think it’s actually a tropical eruption+Iceland shield. 1109 is maybe the spike that has the most similar analogues in terms of sulfate deposition vs time in the South Pole. It resembles 1229, Quilotoa (1285), Pinatubo? (1344), Kuwae (1457), Huaynaputina (1600), Billy Mitchell? (1621), and Parker (1641). It’s perhaps a typical example of a dacitic (tropical?) explosive eruption. I had initially thought this was maybe a submarine eruption signal because the nitrate drop size is about halfway between the larger drop sizes of subaerial eruptions and that of Krakatau and other suspected submarine events, but I think it’s probably Billy Mitchell’s because it ticks a lot of its boxes.

In particular, I think this event is the second-last CFE of Billy Mitchell, the PM1 eruption. It was a large plinian dacite eruption with an estimated 7 km3 volume and did not involve ignimbrites. The present-day caldera would suggest 4-5 km3 DRE, but this is for the PM2 eruption, and it’s hard to tell if PM1 was comparably big or not.

I haven’t found the original source for the PM1 eruption date, but the radiocarbon date I’ve seen quoted calibrates to the 1038-1160 AD range, so the 1109 spike would fall in the middle, a very good match. A lot of the shape resemblances I’ve seen with other sulfate spikes seem to point to a dacite system for 1109 AD, the most common type of major explosive eruption source volcano. The PM1 spike sounds a lot like Pinatubo’s Buag eruption in that it was sourced by a “small” dacite system and did not involve PDCs, and the 1109 spike is the one that seemed to best match the 1344 spike shape that I’ve attributed to Pinatubo’s Buag eruption… So I think Billy Mitchell is a solid match here.

 

1039, Hunga Tonga–Hunga Ha’apai (Tonga)?, 2.1 Tg of sulfur.

In 2022, Hunga Tonga shook the world when it showed the power of a good caldera-forming blast. As it turns out, this must have been almost exactly 1000 years after its previous CFE. There are three prehistoric pyroclastic successions exposed on the Tonga and Ha’apai islands, which include PDCs that are thought to originate from earlier CFEs, andesite ignimbrite eruptions that occurred mostly underwater. The last of these is dated at 1040-1180 AD. Around this interval, there are two eruptions that I think are submarine (based on the large nitrate drops). One is the 1171 eruption, which, however, seems to be of rhyolite composition, so not Hunga Tonga. Thus, the 1039 signal seems to be the best option for Hunga Tonga’s second-last caldera-forming eruption. Like the 2022 eruption, this event must have been somewhere around 6 km3 DRE.

 

1029, Tofua (Tonga), 7.8 Tg of sulfur.

10 years earlier, and two volcanoes to the north, we find another CFE. This seems to be a tropical subaerial volcanic signal with the very distinctive, I think even unmistakable, sulfate deposition distribution of a basaltic-andesite ignimbrite. It’s a lot like Okmok’s and the presumed Long Island (1694) signal, which are both basaltic andesite ignimbrite eruptions. Like the 1694 spike, the 1029 signal is unusually long, 4.5 years, where I think a small portion of the sulfate was carried to extreme heights by plumes, which then took very long to fall back, maybe fueled directly by the hot basaltic andesite “fountains”.

The Tofua ignimbrite consisted of a highly welded ignimbrite of basaltic andesite magma, 52-57 wt% SiO2, and nearly crystal-free (<5 % phenocrysts). The island has two calderas nested within each other. The eruption must have collapsed the smaller caldera, which would give 9 km3 DRE, assuming 500 m collapse depth.

The calibrated age of a trunk in the Tofua ignimbrite would give 990-1200 AD as the potential range, which seems to fit well. Tofua has already been suggested to match the 1029 AD signal, and taken together with the distinctive shape, I think it’s the only event that ticks all the boxes.

 

A journey 1000 years back

Together, my last two articles give an overview of the explosive history of our planet since 1000 AD, I’m afraid I’ll have to leave any conclusions to you, our dear readers, given how long this second installment has grown. I’ll just include a final graph to sum up these 900 years of volcanic activity in one picture:

 

Ice core data references

Sigl, Michael; Toohey, Matthew (2024): Volcanic stratospheric sulfur injections from 500 BCE to 1900 CE, eVolv2k_version4 [dataset]. PANGAEA, https://doi.org/10.1594/PANGAEA.971968

Sigl, Michael, et al. “Volcanic Stratospheric Sulfur Injections and Aerosol Optical Depth during the Holocene (Past 11 500 Years) from a Bipolar Ice-Core Array.” Earth System Science Data, vol. 14, no. 7, 12 July 2022, pp. 3167–3196, https://doi.org/10.5194/essd-14-3167-2022

Winski, D. A. (2021) “South Pole Ice Core Sea Salt and Major Ions” U.S. Antarctic Program (USAP) Data Center. doi: https://doi.org/10.15784/601851

McConnell, J. (2017) “WAIS Divide Ice-Core Aerosol Records from 1.5 to 577 m” U.S. Antarctic Program (USAP) Data Center. doi: https://doi.org/10.15784/601009

Joseph R. McConnell. (2013). Elemental and chemical measurements of the North Greenland Eemian Ice Drilling core, NEEM-2011-S1, collected in summer 2011 near the NEEM deep drilling site in northwest Greenland. Arctic Data Center. doi:10.18739/A2ST7DX9N.

Joseph McConnell, & Nathan Chellman. (2024). Chemical, elemental, and isotopic measurements of the Tunu2022a and Tunu2022b ice core collected from northeastern Greenland; 2022-2024. Arctic Data Center. doi:10.18739/A22F7JT09.

Siggaard-Andersen, Marie-Louise; Hansson, Margareta E; Steffensen, Jørgen Peder; Jonsell, Ulf; Rasmussen, Sune Olander (2022): NorthGRIP ice-core record of major ions measured using ion chromatography covering the last two millennia and additional short Holocene sections [dataset]. PANGAEA, https://doi.org/10.1594/PANGAEA.944172

 

Eruption references

Naranjo, J., Singer, B., Jicha, B., Moreno, H., & Lara, L. (2017). Holocene tephra succession of Puyehue-Cordón Caulle and Antillanca/Casablanca volcanic complexes, southern Andes (40–41°S). Journal of Volcanology and Geothermal Research, 332, 109–128. https://doi.org/10.1016/j.jvolgeores.2016.11.017 (40-41 S

Caulfield, J. T., Cronin, S. J., Turner, S. P., & Cooper, L. B. (2011). Mafic Plinian volcanism and ignimbrite emplacement at Tofua volcano, Tonga. Bulletin of Volcanology, 73(9), 1259–1277. https://doi.org/10.1007/s00445-011-0477-9

Vidal, C. M., Komorowski, J., Métrich, N., Pratomo, I., Kartadinata, N., Prambada, O., Michel, A., Carazzo, G., Lavigne, F., Rodysill, J., Fontijn, K., & Surono, N. (2015). Dynamics of the major plinian eruption of Samalas in 1257 A.D. (Lombok, Indonesia). Bulletin of Volcanology, 77(9). https://doi.org/10.1007/s00445-015-0960-9

Lavigne, F., Degeai, J., Komorowski, J., Guillet, S., Robert, V., Lahitte, P., Oppenheimer, C., Stoffel, M., Vidal, C. M., Surono, N., Pratomo, I., Wassmer, P., Hajdas, I., Hadmoko, D. S., & De Belizal, E. (2013). Source of the great A.D. 1257 mystery eruption unveiled, Samalas volcano, Rinjani Volcanic Complex, Indonesia. Proceedings of the National Academy of Sciences, 110(42), 16742–16747. https://doi.org/10.1073/pnas.1307520110

Hall, M. L., & Mothes, P. A. (2008). Quilotoa volcano — Ecuador: An overview of young dacitic volcanism in a lake-filled caldera. Journal of Volcanology and Geothermal Research, 176(1), 44–55. https://doi.org/10.1016/j.jvolgeores.2008.01.025

Hradecký, P., & Rapprich, V. (2007). Historical Tephra-stratigraphy of the Cosigüina volcano (Western Nicaragua). Revista Geológica De América Central, 38. https://doi.org/10.15517/rgac.v0i38.4217

Sun, C., Wang, L., Plunkett, G., Zhang, E., & Liu, J. (2021). An Integrated Late Pleistocene to Holocene Tephrostratigraphic Framework for South‐East and East Asia. Geophysical Research Letters, 48(5). https://doi.org/10.1029/2020gl090582

Petit, J.-R., Savarino, J., Delmonte, B., Gautier, E., Ginot, P., and Batanova, V.: Cryptotephra fingerprinting of 1458 CE and 426 BCE volcanic events in East Antarctic ice cores

Robin, C., Monzier, M., & Eissen, J. (1994). Formation of the mid-fifteenth century Kuwae caldera (Vanuatu) by an initial hydroclastic and subsequent ignimbritic eruption. Bulletin of Volcanology, 56(3), 170–183. https://doi.org/10.1007/bf00279602

Ballard, C., Bedford, S., Cronin, S. J., & Stern, S. (2023). Evidence at source for the mid-fifteenth century eruption of Kuwae, Vanuatu. Journal of Applied Volcanology, 12(1). https://doi.org/10.1186/s13617-023-00138-1