Before climate change denial, there was volcano denial. This is the story about a forbidden eruption, the one that was deleted from history. This is the suppressed eruption, the one that got away.

Blame Iceland. This post had been up for barely one day when events in Iceland rapidly moved towards another -and potentially very damaging- Reykjanes eruption. Iceland demanded a new posting and with that, this post disappeared into the archives. Now that Iceland has decided that an eruption would be too much work, left the magma underground and instead moved parts of Grindavik a meter closer to centre of the Earth, we can give Guatemala another chance. Iceland was an eruption that almost happened. This post is about an eruption that did happen – but which was denied.

Guatemala is a volcanic country. It is also known for some of the best coffee in the world – the two are not entirely unrelated. Wikipedia, that infallible resource, lists 29 volcanoes in the nation of which 23 have been active during the holocene, and 8 in historial times. Fuego and Pacaya are perhaps the best known. Others, such as Tacana, are rarely mentioned. The fourth highest volcano of Guatemala is Santa Maria. Although frequently active at its subsidiary, Santiaguito volcano, Santa Maria itself only erupted once. Guatemala covered it up. Blame coffee, with sugar, no milk.

2023

The story began in October 2023. This month was extraordinary wet in the UK. East Scotland was hit hard with two exceptional rainfall events within two weeks. (It is remarkable how often disasters come in pairs.) There was flooding in the towns. England and Wales too had their share of flooding when swollen rivers broke their banks. I wrote this on a long distance train, two days after the rain event. From the train there were houses visible next to flooded fields, almost submerged. The houses looked new. Flood plains apparently are irresistible to housing developers, being flat, green and empty – and no questions asked.

But this October did not break records. Those records were set in October 1903, at the end of a very wet and exceptionally cool UK summer. One wonders why.

1902

It was a bad year for volcanoes. On 23 April, Mount Pelee erupted in a minor event that became catastrophic. The eruption culminated on 8 May, when a pyroclastic flow came down the side of the mountain. Such flows have happened many times in many different places. But this time it was in the wrong place. A city was directly in its path. 28,000 people perished.

But the main volcanic action that year was further west. It had started (in hindsight) shortly before the Pelee eruption. In the evening of 18 April, an earthquake shook much of Guatemala and Mexico. The damage was large: the towns of Quezaltenago and San Marcos were largely destroyed. The earthquake has an estimated magnitude of 7.5 and the epicentre was near Quezaltenago. There had been a number of foreshocks over the previous three months. The major volcanic eruption of the year would be in the same area.

Faults and volcanoes of Guatemala

Guatemala has many earthquakes and events this size are not unusual. The country lies on the intersection of three plates. The low-lying plains in the north of the country belong to the North American craton The rest of the country is part of the Caribbean plate, and out in the Pacific lies the Cocos plate. The major faults of the country lie between the North American and Caribbean plate: the Motague fault and 50 km north and parallel, the Polochic fault. Parallel to these are the Ixcan fault in Mexico (very active) and the Jocotan fault (presumed inactive). The Motague fault carries most of the major earthquakes. As it runs only some 60km from Quezaltenago, this was a plausible cause for the 1902 earthquake. However, at Huehuetenango, which is located directly on this fault, there was much less damage reported. That suggests another fault was responsible.

The actual cause may be some fault that runs along the volcanic arc, between the mountains and the Pacific coastal plain of Guatemala. There is a similar fault further along the coast, called the Jalpatague fault. There was also a large earthquake in southern Mexico in September 1902: this may have been the northern extension of the same fault that failed in April. The coastal plain is where the Cocos plate subducts underneath the Caribbean plate. The 1902 earthquake may have been triggered by this subduction, although it was too shallow to be located in the subducting plate. In a nation as fractured as this, there may also be more faults hiding, only revealing themselves when they fail.

The Pacific coastal plain is around 50 km wide. It ends quite abruptedly at the mountainous chain, where the volcanoes of the Central American Volcanic Arc are located. The volcanoes receive their magma from the subduction of the Cocos plate below. Behind the line of volcanoes, and partly mixed in with them, lies a series of calderas of all sizes, remnants of an older phase of volcanism. The plains contain a series of low ridges, fan shapes extending from the mountains. These are ancient lava flows and land slides. They suddenly end some 20 km from the sea. When they were emplaced, the shore was where the ridges end. The calderas and lava flows show that at times, this region was a mover and shaker of the world. Perhaps it still is!

The Central American Volcanic Arc is an impressive sight. The following is from Gustav Eisen, 1903, describing the volcanoes one by one.

Santa Maria is a tall stratovolcano, reaching over 3700 meters, high, located near the northwestern end of the Central American Volcanic Arc. The local name was Excanul or Gagxanul (‘the naked mountain’). The volcanoes here, near the end of the arc, are not among the most active in the country. The nearby systems that are quite similar to Santa Maria include Almolonga and Atacana. Atacana has occasional minor phreatic eruptions, whilst Almolonga (also known as Cerro Quemado) had an effusive eruption in 1818. But both can also do dangerous stuff. Atacana had a large explosive eruption 2000 years ago. Almolonga suffered a major flank collapse 1200 years ago.

There are more volcanoes here but the others are considered inactive. That included Santa Maria. There is a Mayan legend of spectacular activity at Gagxanul. This has been assumed to refer to the eruption of Almolonga but this is far from certain. Santa Maria is considered to have been inactive for 25,000 years. However, Eisen’s describes it as ‘beautiful’ and ‘perfect’. Volcanoes tend to age poorly: they quickly deteriorate from erosion, becoming rugged and impressive but not pretty. The pretty ones are young and active, which make them resurface themselves frequently with volcanic make-up. It makes me wonder whether Santa Maria has been more active than it is credited with. But there had been no eruption reported since the Spanish occupation.

That inactivity changed after the earthquake. There had already been earthquake swarms in the months before the April 1902 earthquake. It is not known whether this was due to re-activation of Santa Maria or from precursor activity on the fault. After the earthquake, the aftershocks focussed on the vicinity of Santa Maria and the activity seemed to have shifted over the months from tectonic to volcanic.

Eruption

The long silence came to a sudden end on 24 October. Earthquakes were felt and in the afternoon steam was seen on the southwestern flank of this supposedly extinct Santa Maria. Things build up very quickly from here. At 5pm came the first report of ‘retumbos’ near the volcano, meaning the underground rumbling which you can hear and feel during earthquakes. By 6:30pm, these were heard (or felt) all over Guatemala and El Salvador. For a volcano that had not done anything for 25,000 years (presumedly), this was already off the scale. By 7pm ash began to fall at a farming estate (‘finca’) 14 km west, quickly followed by pumice. A cloud was seen to rise from the volcano, triggering lightning. The cloud was measured at 10 km tall. Santa Maria had woken up from its long slumber.

This was only the beginning. The explosions were heard from further and further away. At 11pm, distant cannons were heard in central Guatemala, thought to come from celebrations in the capital for the fiesta of the following day. During the night and morning the eruption intensified. The finca that had previously been showered in ash and pumice, by 6 am reported hot rocks falling from the sky. The wind was blowing the ash northwest. The explosions and retumbos were now heard as far away as Costa Rica, Mexico City and the Yucatan. Earthquakes peaked at 3am, 7:30am and 11am. A sulfur smell pervaded the air at Coban, 160 km northwest, and in Guatemala city. The ash column could no longer be seen from nearby as everything had gone dark. Already in the morning of the 25^th there was no dawn, and candlelight was needed throughout the day up to 25 miles away. To the west, the near-total darkness would last for 3 days. However, the cloud was visible from a ship in the Pacific: they measured the column at 28 km tall.

Loud explosions were again reported in the evening of the 25^th, and on the hills in central Guatemala the ground was shaking by the force of the explosion. This shaking was caused by the sound waves, as the shaking was felt much less in the valleys. From the evening onward, these explosions grew less and they could not be heard from as far away. The extinct volcano was now winding down.

During the afternoon of the 25^th, the area south and west was hit by a hurricane, accompanied by tremendous lightning. Trees were destroyed up to 15 miles away and forests had become impassable. This storm which lasted four hours did more damage than the eruption itself. People also reported St Elmo’s fire within the falling ash. This rain storm is confirmed by studies of the ash, which show in one area, the upper layer of the ash has been partly washed away

In the morning of the 26^th there were further explosions, now producing a muddy ash which fell up to 12 miles away. Steam columns showed that these were phreatic events, triggered by water. Ash fall continued and destructive lahars came down to the coastal plains. But the ash became lighter and more local. The Sun was first seen near the volcano on the 30^th.

The frequent, watery explosions would continue for months, until April 1903. During November they were at times near continuous, still causing retombos. One of these later eruptions in January 1903 coincided with ashfall in San Fransisco, 4000 km away. However, an association with Santa Maria must be considered as doubtful. During the main eruption, the most distant ash was reported from Acapulco, 900 km away. But although there were hot rocks and hot water, there are no reports of lava. That would come 20 years later.

The land around Santa Maria seemed destroyed. All vegetation was covered by a blanket of white ash, appearing like a mid-winter landscape. Trees were stripped of their branches. The coffee and sugarcane plantations were gone, showing only as openings in the forest. Close to the crater, the ash was reported to be 20 meters thick. The ubiquitous rain quickly formed deep erosion channels in the ash. The mudflows destroyed all bridges along the rivers. Their replacements were not for the faint-hearted.

The crater

The area was difficult to access for some time after the eruption. The crater was first described only in December. It was located on the southwestern side of Santa Maria, and measure 1 km by 800 meter, with a depth of 250 meters. At the bottom of the crater were 6 holes, up to 30 meters in diameter, which were the location of vigorous fumaroles.

But the crater size is inconsistent with the severity of the eruption. It seems like an afterthought, caused only by the later phases of the eruption. Is there a much bigger crater buried under the ash? Williams and Self (1982) propose that the magma chamber was much deeper than usual, and that this meant that the area of subsidence is much larger. The magma was highly evolved, hence it’s white colour, which is something that is common among Guatemalan volcanoes that rarely erupt. And several volcanos in the chain have had these oversized eruptions without overly changing their topography. Following Williams and Self, this is often attributed to them having deep chambers that are fairly oversized.
As long as the eruption is not too large, the bedrock just flexes downwards and avoids caldera formation. The absence of a sufficiently large crater in flank eruptions is also seen in Huaynaputina.

But why a flank eruption? This is the flank towards the coastal plain. Was it caused by the April earthquake? It seems unlikely that that earthquake was triggered by magma supply to the volcano: it was too large and too extended for that. But it may have caused the eruption by destabilizing the flank of Santa Maria. We can speculate that this created a new magma pathway, and eventually led to a flank collapse.

This suggests another aspect of the lack of a crater. After the main event, the flank would have collapsed over the eruption site. This filled in much of the crater which had formed. Perhaps this was the cause of the later, phreatic explosions?

The eruption clearly did not involve the summit or any shallow magma chamber underneath it. The height of Santa Maria reported before the eruption by the International Railway of Guatemala (really) was 3777 meters. The current height is 3772 meters. There has been no real change here. The eruption came from the flank, and involved only the flank. The summit was unaffected.

Fatalities

The number of fatalities is very poorly known. Most were among the local Indian population, either in their villages or as seasonal workers (slave workers in all but name) on the coffee plantations, and they were never counted. The most distant casualty occurred on board the ship Acapulco, where an American refugee reportedly died from breathing! (The captain blamed the sulfur and the ash in the air.) Local reports mention deaths among labourers on one plantation (115 out of 122) and 350 fatalities in one Indian village. People died in collapsed or buried houses, and within 10 km of the volcano from rocks falling through the roofs of houses. The total number of documented fatalities is 650. But refugees streaming into Mexico and the US reported many more deaths than this.

After the eruption, local society broke down. There were deaths from starvation, disease, but also murder and alcohol. Malaria caused further casualties: one reporter attributes this to the lack of birds eating the mosquitos! We will probably never know how many died from these later consequences. Eisen estimates 1500 died, from burial (nearby houses were buried up to their roofs), asphyxiation and drunkenness. The commonly stated numbers of 5000-7000 are estimates from US newspapers in the months after the eruption. They all agree that the large majority of the casualties were among the Indian population, and were uncounted.

Ash

The image shows the distribution of the ash, ranging from 1 meter near Santa Maria itself (actually 10 meter or more near the crater), to 0.1 cm at the edge of the blue region. The two directions come from a change in the wind during the eruption. The numbers next to the lines indicate how many hours into the eruption the ashfall began. The ash extends a smaller distances than the sound did: the most distant reports of heard explosions are from Costa Rica and Mexico City.

In principle, adding up all the ash within the contours should provide the total amount that was erupted. The historical reports help considerably and make this more accurate than is normally the case. Berry et al find a total volume of 11.4km³ – a VEI 6. The dense rock equivalent, or the size of the hole, comes out as 6.4 km³.

The column height also suggest such a large eruption. The measurement of 28 km may have missed the peak height. There is in fact another measurement of 48 km, but it is not clear how reliable this is. The grain sizes suggest a peak height of 33 km. All of these rank among the highest of the 20th century!

There have been four VEI-6 events in the 20^th century: Santa Maria in 1902, Katmai in 1912, Cerro Azul in 1932, and Pinatubo in 1991. In terms of DRE, their volumes are 12 km³, 6.4 km³, 4 km³ and 5 km³. Assuming these values are accurate (and they are not that easy to measure!) the order is: 1. Katmai; 2. Santa Maria; 3. Pinatubo; 4. Cerro Azul. (Cerro Azul is borderline – it may have been just below VEI-6.)

But Santa Maria is often left out from this series, or relegated to a lower place. Why is that?

Suppression

Within hours of the start, the sound of explosions was heard across Central America . It made it immediately clear that a volcano had erupted. But which one? Already early on October 25, Gustav Eisen send a telegram to Guatemala City to inquire. The response was negative: the government had blocked any news of the event. Later that day, a telegram came from the government saying that a volcano in Mexico had erupted. That was clearly wrong as the sounds did not come from that direction. Within days it was clear that the eruption had come from the vicinity of Santa Maria. But Guatemala remained silent. In the UK, the first report of Santa Maria erupting came on 19 October, sourced from a journalist in Guatemala City who stated that the town and villages in the region had been abandoned. US newspapers began to report on thousands of casualties. Guatemala remained silent.

The government finally acknowledged the existence of the eruption on 3 November, with a statement that there had been a small eruption, there were no casualties and only some light ashfalls with a few damaged estates and that other reports were greatly exaggerated. This message was distributed as a cable by the department of Foreign Affairs. This was at the same time that newspapers abroad came out with detailed descriptions of coffee plantations buried 7 feet deep with immense loss of life (and of cattle). One reported ‘as far as the eye could see everything was blue and grey and dead, like a mammoth cemetery’. To this, Guatemala said that ‘in view of the facts, we can only censor such scaremongering’ and that the English and North American press were only interested in throwing shades over the towns of Latin America. Refugees mentioned that there was complete suppression of news from the region and that all cable stations had a strong censorship. The situation was very different from that after the April earthquake, when there had been extensive reporting in Guatemala of the damage done. In contrast, the Santa Maria eruption was first denied and afterwards was claimed to be only a minor one.

Why this denial? We don’t really know. The first censorship was immediate, when the impact of the eruption was not known. That day (25 October) there was a fiesta, and perhaps the president did not want that canceled as it was partly in his honour. The government also had a significant financial problem, and lacked the money to even pay for the damage from the April earthquake. Coffee provided one of the main income streams of Guatemala. Perhaps half of all plantations had been destroyed: if this news came out, it could destroy the confidence in the financial health of the nation. Add to this that the previous president had been assassinated and one can imagine some reluctance to announce another major disaster. In the end, the coffee harvest in the rest of the country that year was better than usual so there was some compensation.

Coffee was a major source of income at the time. Although not the only one, the government had focussed on it, to the degree of even re-instating forced seasonal labour for the Indian population. The sugarcane industry was also growing in importance, and this was cultivated on the Pacific coast and so was also affected, although not to the same degree. Bananas were becoming important. This was the time when the United Fruit Company began to almost take over much of Central America: in Guatemala they were exempt from paying tax, and by the time of the eruption, even ran the national postal service. The company focussed on growing bananas: the United Fruit Company is the cause of the later expression (1930’s) of ‘banana republic’. The company is known nowadays as Chiquita. But for export, coffee was the major crop. Indeed, the economic disruption after the eruption became a major cause of the famine. People could not afford to buy the food that was still available.

In the end, the dominant motivation for the suppression of the eruption must have been the president.

Denial has consequences. There was no help for the recovery of the local area: this is perhaps one reason that the number of fatalities is so poorly known. But it has also left us a legacy of a lack of knowledge. Santa Maria is still the least known of the major recent eruptions. And that lack of knowledge brings risks. The city of Quezaltenago is only 10 km from Santa Maria. It is at risk if the eruption would recur. How prepared is the region, after a history of denial? It is tempting to downplay risks if that means no action needs to be taken. It is a very human attitude. But it can lead to disaster.

Afterwards

One of the effects of the eruption was that a small Swedish spice company started to name their consumer line of spices as Santa Maria. It was a show of respect for the dead people at the spice farm from which they used to buy their cardamom. Something to ponder while eating texmex food..

The mountain silently recovered for 2 decades. But in 1922, the crater which had formed in 1902 became the site of a new eruption, this time with real lava. In 1927 a larger eruption here again caused several hundred casualties. Over the years, a new mountain grew consisting of four different peaks. Only the original one from 1922 is currently active. Santiaguito, the new mountain, is among the youngest and fastest growing volcanoes in the world. It is an undeniable volcano.

Weather

1903 is not in the list of years affected by volcanic eruptions. The climate of 1903 isn’t a major outlier. But it was a year that was colder than average. In the NASA annual averages, 1903 is -0.37C below the average for 1951-1980. And 1904 was even colder: it was -0.47C below that average, making it the second coldest year of the 20^th century. But since random year-on-year fluctuations can also be this large, it is hard to prove that the cold was volcanically induced. El Chicon, for instance, had its effect on the climate masked by a strong El Nino.

Looking at the four VEI-6 eruptions in the 20^th century, the years 1903, 1933 and 1992, one year after the eruption, were all about 0.1C cooler than the year before. Katmai (1912) had no such effect though. Agung in 1963 was also followed by a cold year. Is a VEI-6 really large enough to change our weather? Pinatubo did manage this, though, even affecting sea levels.

And how about local climate? When looking at a small region, the year-on-year random fluctuations become much larger, and it is therefore harder to detect small effects. But for instance a change in the location of the jet stream may affect a particular area while leaving others unchanged, Local changes are not impossible.

Looking at the British Isles, the weather of 1903 was notable. On 26 and 27 February, the British Isles (including Ireland) was hit by storm Ulysses. This happened long before storms started to be named, so having a name was rather exceptional. It is mentioned in Joyce’s book Ulysses, hence the name. Re-analysis of the records indicate that this was one of the strongest storms every recorded in the British Isles, on par with the English hurricane of 1987.

June 1903 was singled out in reports of the time for its unusual rain distribution. While parts of the UK had little rain, the south was drowning in it. From 13 to 15 June it rained continuously in London and the Thames Valley for 58 hours, at a near constant rate. It was caused by a low-pressure system getting stuck of the Isle of Wight. But this was just one weather event. The year overall, over the entire British Isles, was one of the wettest ever recorded. This was caused not so much by those three days in June, but by the summer and especially by October when it rained every day. Places where this is normal (such as the Lake District) had their usual amount. But this time the rain was far more widespread across the UK. The summer had also been unusually cool: it was for instance the coldest June for almost 50 years, 3C below the norm. All three summer months were much too cold.

But was this volcanic? It is impossible to be sure. But it was what drew me to look at the events of 1903. And to find out about more an attempt to suppress a VEI-6 eruption.

Albert & Carl, November 2023

References

Hannah C. Berry, Katharine V. Cashman, Caroline A. Williams,

The 1902 Plinian eruption of Santa María volcano, Guatemala: A new assessment of magnitude and impact using historical sources. Journal of Volcanology and Geothermal Research, Volume 414, 2021, 107167

Conway, F.M., Vallance, J.W., Rose, W.I., Johns, G.W., Paniagua, S., 1992. Cerro Quemado, Guatemala: the volcanic history and hazards of an exogenous volcanic dome complex. Journal of Volcanology and Geothermal Research, 52 (4), 303–323.

Rhodes, E., et al. Textural Insights Into the Evolving Lava Dome Cycles at Santiaguito Lava Dome, Guatemala. Frontiers of Earth Science, Section Volcanology, Volume 6 (2018).

Williams, S., Self, S., The October 1902 eruption of Santa Maria Volcano, Guatemala. Journal of Volcanology and Geothermal Research, 16 (1983), 33.

Eisen, G, The earthquake and volcanic eruption in Guatemala in 1902, Bulleting of the American Geographical Society, 35 (1903), 325-352.

Guzman-Speziale M, Molina E., Seismicity and seismically active faulting of Guatemala: A review. Journal of South American Earth Sciences, 115 (2022), 103740.