Explosive eruptions have killed perhaps 250,000 people in the past 400 years. They are often unpredictable in their timing, although not in their location. We need to ask whether the human response to such eruptions has been the most sensible.

Explosive eruptions occur when gas pressure under the eventual vent is released catastrophically. Often, but not always, there are warning signs. Magma rising below the eventual eruption site may meet ground water, and steam-generated explosions, often quite small, begin to crack open the existing surface rocks. A dome or plug in the crater may begin to crack, or expand. Part of the summit may begin to bulge. But as soon as the gas pressure is released, the magma itself will then expand explosively and on a much bigger scale, producing a major eruption almost instantaneously. Discharges of 100,000 tonnes/second have been calculated at this stage, at velocities up to several hundred meters/second. At the surface, observers see a cloud formed of fragmented surface rock, flash-frozen and exploded magma, steam, and gas. The blast may be directed sideways, as at Mount St. Helens, or vertically.

Vertical Eruptions.‹In vertical eruptions the explosion usually forms an immense mushroom-shaped cloud blasted vertically into the stratosphere. A height of 20 km is typical, but it can be as much as 30 km.

An eruption cloud 20+ km high can only be held up by the continued eruption of magma and gas below it. Some of the height is maintained simply by the momentum of the eruption, but most of the cloud is supported by the hot gas underneath it, which is buoyant enough to keep it moving upward. As long as the main part of the cloud is supported, the major effect is the fall-out of the denser or larger components. Ashes fall out of the cloud over the surrounding region. They vary in size from small dust-sized particles to fist-sized blocks. The ashfall darkens the sky but is not generally lethal as long as people have shelter. As it accumulates, however, it smothers plants and crops, overloads roofs, and stops most transport by choking streets, streams, people, horses, and engines.

Critical danger occurs as the explosive eruption begins to lose its first momentum, and cannot support the cloud any more. The cloud will collapse for any of a number of reasons. Perhaps the eruptive vent enlarges as the force of the erupting magma erodes its sides away; and as the vent enlarges, the velocity of the blast decreases. Perhaps the magma supply to the vent drops. In any case, once the eruption cloud cannot be supported, it will start to collapse, even if magma continues to erupt at a high rate. The collapse turns very quickly into an unstoppable runaway reaction, as 20+ km of erupted gas and ash cloud collapses downward, gaining heat energy as it falls. The cloud blasts out sideways as it hits the ground, and it is only at that point that it can be seen by observers and potential victims as it rushes at low level out of the masses of ash, gas, smoke, and steam that surround and hide the eruption.

The ground-level blast clouds are called nuées ardentes, or "burning clouds." The lighter part of the cloud forms a pyroclastic surge or base surge, a cloud of ash, gas, and dust travelling at perhaps 200 km/hr (over 100 mph), at over 100° C. The surge does not follow terrain accurately, but blasts directly across ridges and uneven topography. The denser part of the nuée ardente forms a pyroclastic flow or base flow, which is very hot and very dense, and moves very close to the ground at somewhere between 50 and 100 km/hr (30­60 mph), at up to 400° C, largely following any valleys downhill. The base surge quickly travels ahead of the base flow, so that a collapsing eruptive cloud produces two components that arrive successively at a locality and have different effects. Although it is devastating, a base surge may not carry much ash, and so may not leave much of an ash layer. The base flow that follows it typically leaves a much greater mass of volcanic debris.

During the course of an eruption, successive large vertical blasts form successive eruption clouds that are followed by successive base surges and nuées ardentes.

Lateral Eruptions.‹Pyroclastic flows and surges may be generated more directly than by collapse of a vertical eruption column, and they are described here under "lateral" eruptions, for want of a better official term.

A lateral eruption may occur if an eruption is released sideways: for example, by a vent that is not in the center of the crater; by a landslide from one side of the crater (as at Mt. St. Helens); or by a smaller explosion that is not powerful enough to blow vertically into the stratosphere but is still powerful enough to generate lethal pyroclastic flows and surges down the slopes of the volcano.

Obviously, lateral eruptions tend to be destructive mainly in one direction, whereas the collapse of a vertical eruption may be destructive all round the base of the volcano.


The eruption of Mt. St. Helens on 18 May 1980 had been predicted, but it still killed dozens of people and cause millions of dollars' worth of damage. Could any of that have been prevented?

Mt. St. Helens is one of the volcanoes in the active Cascade chain that runs from California to British Columbia, with Mt. Lassen at its southern end and Mt. Garibaldi at its northern. Only Lassen had had a major eruption in the 20th century, climaxing in 1915, but it was common knowledge that several Cascade volcanoes were anything but extinct. The first modern survey of Mt. St. Helens in the 1930s stressed that the mountain was dormant rather than extinct.

By the mid-1970s it was clear that Mt. St. Helens had had a long history of spasmodic violent eruptive activity, and that it was not only overdue for eruption, but was the most likely of all the Cascade volcanoes to erupt next. In 1975 three US Geological Survey (USGS) scientists warned that an eruption of Mt. St. Helens was likely within 100 years, and possible within 25 years. In 1978 the USGS issued a "Notice of Potential Hazard" for the volcano, and produced maps and reports describing the likely events of an eruption, and the hazards that would endanger people and property. Local authorities and emergency services understood them.

The volcano was prone to explosive eruption that would most likely be accompanied by nuée ardentes. These would be deadly in themselves, but they would also be likely to cause catastrophic melting of snow on the slopes and summit of the volcano, with especially severe effects in winter and spring. One would expect large floods of water choked with ash and mud (lahars) sweeping down the major valleys surrounding the volcano. If lahars swept into the reservoirs near the mountain, flash flooding might be expected if the water of the reservoir were to rise above the crest of the dam. Ash falls would be expected, and they might interfere with transportation at greater distances from the volcano, and might cause crop damage.

In retrospect, it is a tribute to the geologists and planners that the scenarios that followed the 1980 eruption were so accurately predicted. Some unexpected features were not predicted, but we understand how they happened, and they can be incorporated into future emergency planning. Altogether, this eruption can be used as a test case for dealing with a volcanic emergency of moderate size.

Mt. St. Helens was not provided with much instrumentation for detecting an imminent eruption. But there was one seismometer on the mountain, and it recorded small earthquakes beginning on March 20, 1980. The right conclusion was drawn: that molten magma was moving actively under the volcano, indicating that a really major eruption (probably explosive) was likely to happen soon. Many more instruments were put in place, aircraft began photographic surveys of the summit, and teams of geologists began daily observations. The general public, however, was warned to stay away from the volcano on March 24. On March 25, with earthquake activity increasing, the US Forest Service officially closed Mt. St. Helens to the public above timberline; many national forest roads were closed; and State highway 504 was closed up to 5 miles (8 km) from the summit.

Mount St. Helens began erupting on March 27. The first eruptions were small explosions of ash and dust, but they allowed local authorities time and valid reason to set up smooth procedures for observation, information, and action to minimize the hazards posed by the volcano. The fact that the eruption precursors had been so accurately interpreted gave authorities and public alike confidence in their scientific advisers.

Hundreds of people‹permanent residents, forest workers in logging camps, scientists, and news reporters‹were evacuated from a zone that had been forecast as dangerous (within 15 miles [24 km] of the summit). Visitors, whether they were casual tourists, volcano-watchers, or fishermen and hikers seeking recreation in the area, were either warned to keep out of certain areas or were actively prevented from reaching them.

On March 28, light ash was falling over a wide area, it was clear that the eruption was significant, and sightseers were jamming possible evacuation roads and trying to evade road blocks. At the weekend of March 29­30, as many as 70 aircraft were trying to fly over the volcano at the same time, and strict air traffic control was instituted. Tourists were flagrantly flouting the warnings: a helicopter landed on the crater rim, and another group climbed to the summit to make a TV commercial. On April 1, county officials said that they were unable to cope with people ignoring warnings, and asked for help in manning road blocks. Governor Ray sent in the National Guard on April 5, and announced that only property owners and scientists should be allowed within the restricted zone. Meanwhile, attorneys for logging companies were arguing that loggers should be allowed to make their own choice about entering the restricted zone, and the Weyerhaeuser company re-opened three logging camps on the lower slopes of the volcano, and moved 300 workers back into them.

The volcano showed a slow increase in the intensity of earthquakes, and its summit showed increasing cracking and swelling. The crater continued to erupt ash and dust, and grew in size. But it did not do anything spectacular, or overtly dangerous. The county authorities began to wonder aloud how long they could afford to (or should) keep people away from the mountain. The weekend of April 12­13 was clear, and record numbers of sightseers crowded the area.

The following week, the volcanic activity reached a plateau, and even declined a little. News updates issued by the USGS dropped to two a week, rather than daily. Naturally, this was seen as reducing the actual hazard, and social pressures for greater access to the area began to mount. But by this time the USGS had realized that the mountain had bulged 80 m (250 feet) on its northern flank, and they stressed the potential danger of avalanches of snow, ice, and rock down the northern slopes, especially toward Spirit Lake and down the north fork of the Toutle River. On April 30, authorities defined a "Red Zone" of extreme danger on the northern side of the summit, extending as much as 8 miles (13 km): only essential personnel were to be allowed in that zone. In a "Blue Zone," logging operations could continue, and residents could visit their property; but no-one was to stay overnight. The only persons staying overnight in the "Red Zone" were USGS geologists at a station called Coldwater II.

By May 8 the bulge on the northern side was so great that geologists felt it was only a matter of time before a major landslide. However, by mid-May, crews were being allowed into the "Red Zone" to remove equipment from the big institutional summer camps round Spirit Lake, and exiled residents were demanding the same kind of access to their properties. On May 17, 50 carloads of property owners were escorted in and out of the Red Zone to retrieve property, and another similar convoy was scheduled for the 18th. It never set off.

The major eruption, when it came on the morning of the 18th, could not have been accurately predicted because of unusual factors. Mt. St. Helens was already erupting from its summit crater, therefore it was clear that conduits for pressure release were open to the summit. The massive build-up of magma inside a volcano in this state would normally be released in an explosion, but one would expect a vertical explosion through the weak points that already existed, that is, through the crater.

However, Mount St. Helens had been "bulging" on its north side, so that there was always more danger that it would erupt on that side: the danger zone here was broader. But the ferocity of the eruption was not predictable. On the morning of May 18, a small earthquake, no bigger than several previous ones, triggered a major landslide on the bulging north side of the volcano at 8:32 am local time. The volume of rock sliced off the volcano on its northern side reached deep enough into the interior that it triggered the expected eruption in an unexpected direction. Within 20 seconds, the unexpected release of pressure on the northern side set off an enormous sideways blast that did most of the damage.

The landslide itself consisted of three cubic kilometers of broken rock and ice. It avalanched down over the geologist David Johnston, who was radioing a warning from Coldwater II, into and across Spirit Lake and into the valley of the North Fork of the Toutle River. Reaching speeds of 250 km/hr, it left a blanket of debris 1­2 km wide and up to 200 m thick, which completely filled the valley floor for 28 km downstream.

The lateral blast that followed the landslide blew northward at something like 200 mph. It utterly destroyed more than 500 sq km of forest. Everything within 10 km was completely stripped and covered with 2 m of rock debris. 15 km away, trees were uprooted and carried with the blast, or were snapped off instantly. Two people who stopped to watch the eruption were killed in a car 24 km from the summit. Some fishermen 26 km from the blast survived only because they jumped into the river, but they were burned when they came up for air. Plastic parts melted in a truck that was blown over, 21 km from the crater, probably reaching 360° C. This was a classic nuée ardente.

The predicted lahars (volcanic mud flows) occurred as the snow and ice on the mountain melted. Because the blast was to the north, the lahars that flowed down the Toutle drainage were larger than anyone had predicted. The giant landslide had contained billions of liters of water and ice, and a few hours after the slide, that water began to flow out of the slide area as an enormous rapid viscous flow. The flows were slow enough that prompt warnings allowed people to evacuate logging camps and homes, but the damage was immense. Flows reached 16 m (50 feet) above normal river level on the North fork of the Toutle, and floating trees destroyed nearly all the bridges over the river. The mud was still warm as it reached and clogged the Cowlitz River downstream, and killed all the fish. The flows continued unchecked down into the Columbia, and a freighter went aground on a newly formed mud bank in the navigation channel on the Columbia as the depth was cut from 40 feet to 15 feet (13 m to less than 5 m) over a distance of 2 miles (3 km). Dozens of ships were held up on each side of the obstruction for two weeks, until emergency dredging cleared away the blockage.

The ash that was blown upward fell over a wide tract of country. Visibility in Spokane, 430 km from the volcano, was reduced to 3 m (10 feet) in the middle of the afternoon. The Federal Aviation Authority closed all airports in the region, and issued a warning that aircraft should not fly into the ash cloud. The ash caused extensive damage to automobile engines of emergency vehicles that were forced to operate close to the volcano: 200 vehicles were disabled in two days.

In all, 57 people were killed and the damage estimate was $1 billion, mostly in timber that was destroyed. Lumber companies who tried to salvage wood from the thousands of trees that had been blown down found that their saws could hardly cut through some of the trees, because the wood had been impregnated millimeters deep with tiny particles of volcanic ash.

Overall, however, loss of life and property in this eruption were minimized. In the end, the people who were killed or injured in the Mt. St. Helens eruption were professionals doing their jobs in hazardous conditions (geologists and forestry workers), residents who refused to evacuate, and ignorant or reckless individuals who ignored the warnings of scientists and government officials at various levels. The death toll was cut down a lot by the predictions and warnings, and by prompt emergency action: about 100 people were rescued from life-threatening situations. The death toll would have been much lower if people had listened to warnings.

Most of the property damage was unavoidable: it involved structures that could not be moved or protected, such as standing trees and crops, homes and other buildings, bridges and roads. Some incidental damage to airplanes and automobiles occurred because individuals and companies ignored warnings of the damage that fine volcanic dust can cause inside an internal combustion engine.

There were unexpected benefits too. In the end, as much good as harm was done to crops: by chance, the volcanic ash that fell over the wheat fields of the northwest helped to retain moisture in the soil. The Federal money that poured into the area for reconstruction helped to relieve short-term hardship, and the burgeoning tourist industry that has grown in the region has brought more long-term dollars into the area, and provided more long-term jobs, than were lost by the destruction of timber in the eruption.

Only three weeks after the eruption of May 18, the volcano was relatively quiet, and new Red Zone/Blue Zone areas had been defined. The new Red Zone was much enlarged, and extended 17 to 27 miles (27 to 40 km) from the volcano. Logging corporations were already operating under Mt. St. Helens within the Red Zone. It's not clear whether this latter action is best interpreted as faith in the accuracy and efficiency of the volcano monitoring system, or persistent greed in the face of manifest danger. Since many people refused to evacuate Red Zone areas after a warning on June 12, one simply has to conclude that people simply will not listen to warnings, and do not learn from bitter experience. After this particular incident, the State Department of Emergency Services made the only response it could short of ignoring the problem: it temporarily withdrew entry permits to Red Zone areas.

Mt. St. Helens lies in a sparsely populated region, and long-term evacuation of most endangered residents was possible, rapid, and cheap, and caused little dislocation to the regional economy. It is not clear whether the timely and effective response to the predictions could as easily have been made in a more densely populated region where evacuation could mean as large an economic disaster as most possible eruptions.

The After Effects.‹The lahars had filled the choked the Toutle and Cowlitz river channels with 150­200 million cu m of sediment. The Cowlitz could hold only a flow of 13,000 cfs within its banks without flooding, instead of a pre-eruption 76,000 cfs. There was every prospect of devastating flooding in the winter and spring of 1980­81, even from a normal rainfall season. Any further lahars coming down the Toutle would also be much more damaging than they would have been previously.

Cowlitz County immediately banned any new construction on the Cowlitz floodplain, and the Army Corps of Engineers instituted an emergency program to dredge the river channels wider and deeper, dumping the dredge material on the banks to form higher levees and a higher flood plain. The bed was cleared enough to hold 50,000 cfs by December 1980, after close to 40 million cu m of sediment had been excavated at a cost of $100 million.

There was still an enormous amount of loose debris in the Toutle Valley, and this began to wash downstream, threatening to re-clog the Cowlitz. The Corps reacted by raising ten miles of the Cowlitz levees in 1982­83. The Corps then submitted proposals for comprehensive solutions to the flood hazard along the Toutle and Cowlitz, and with the help of local politicians, the plan was funded in 1985. A very large rock-fill dam was constructed high in the Toutle Valley to try to contain the major sediment content of any future lahars.

By 1986 Cowlitz County was being pressured to relax the moratorium on building in the Cowlitz floodplain, and did so‹a step that may have been politically expedient, but laid the foundation for further damage and potential disaster in years to come.

Upstream, the giant landslide of May 18, 1980 had blocked off the natural drainage behind a large dam made of fine, unconsolidated volcanic debris, about 170 m high. Water began to build up behind the dam, essentially re-forming Spirit Lake 60 m above its previous level. Smaller lakes formed in side valleys. There was great danger that the new lakes would build up, overflow their soft, unstable debris dams, and break out catastrophically. Once the flood water had churned up the volcanic ash in its path, it would have become a giant volcanic mudflow (lahar) that would have been much larger and more dangerous than the lahars that had immediately followed the eruption.

The small side lakes were allowed to drain slowly by constructing emergency spillways for them, but the new Spirit Lake was too big: it would have eroded any unreinforced spillway in a catastrophic break-out. Fortunately, there was time to realize the danger and act on it. Spirit Lake refilled quite slowly at first, and first estimates were several years for it to reach a dangerous height behind the enormous debris dam.

However, the debris dam began to erode quickly, and in the end emergency action had to be taken before the winter of 1981­82. The Corps installed floating barges on Spirit Lake carrying pumps that drained water over the debris dam through a 5-foot steel pipe. This kept the lake level fairly steady for the next three years, while a better solution was worked out. A 12-foot tunnel was driven a mile and a half through the debris dam to provide a permanent, safe, engineered outlet.

Once the danger of lahars had been recognized as one of the major hazards for people down the Toutle and Cowlitz drainages, geologists searched for, and found, evidence of previous lahar catastrophes of this sort from Mount St. Helens. There is evidence that more than 25 giant lahars from the volcano reached at least 50 km downstream. There had been good reason to fear a break-out from Spirit Lake.

In particular, about 2500­3000 years ago, four closely-spaced lahars devastated the Toutle valley and swept on down the Cowlitz River into the Columbia, each one covering the flood plains tens of meters deep in muddy water. In all, they dumped up to 12 m of new sediment on the valley floors. Judging by the way they affected trees, all four happened within a few years. The two largest flows had about the same peak flow as the Amazon in flood, except that they were moving much faster‹somewhere around 20 m/sec. The only likely source for lahars of this size and discharge is the failure of successive new volcanic dams across an ancestral Spirit Lake in an active period of eruption of Mount St. Helens.

It's clear that the events of 1980 could have been a great deal more destructive than they were. It's also clear that the potential dangers from a new eruption are much greater than were foreseen before the 1980 eruption, but it's not clear that any engineering will be able to prevent them. The rock-fill dam will help to contain an eruptive lahar (if the next one goes down the Toutle rather than in any of the other possible directions of eruption). But it could not contain a catastrophic break-out of Spirit Lake. For emergencies of that magnitude, the inhabitants downstream will have to rely on timely warning and rapid evacuation.

It's an interesting fact that the cost of cleaning up after the immediate eruption‹the river dredging, the flood prevention schemes, the draining of the lahar lakes, the building of the sediment retention dam, and so on‹was as great as the original damage from the eruption. The Federal Government alone will have spent $600 million by the time the Corps of Engineers has completed its projects. A good deal of this money, however, has gone toward mitigation of the next eruption, as well as cleaning up after the previous one.


The explosion of Krakatau in 1883 was smaller than Tambora (see below), but it is still the largest that has been studied scientifically. Krakatau was an island in the Sunda Strait, and the effects of the eruption were strongly influenced by its marine setting. Strong activity culminated on August 27, 1883, in a massive explosive eruption that was heard 3600 km away. A vertical blast resulted produced a dense cloud of ash and gas (possibly as much as 80 km high) that collapsed in an enormous base surge into the sea round about: in turn, this set off large tsunami as much as 30 m high, that devastated hundreds of miles of densely populated coastlines. Three port cities on Java were totally devastated. Only two people survived at the town of Merak when the tsunami swept right over a hill 40 m high on which some of the inhabitants had gathered for safety. The tsunami were recorded as little surges on tide gauges as far away as Europe, after travelling across the entire world ocean. The caldera that was left was 7 km across and 300 m deep, and estimates of the volume of rock that was blown into dust range up to several cubic km.

The smaller fraction of ash remained in the stratosphere for years, and produced the spectacular sunsets like those that had characterized the skies after the 1815 eruption of Tambora. People in Poughkeepsie, New York, called the fire brigade one evening, convinced that the brilliance of the sunset must be a nearby fire.


Tambora is a large volcanic pile that forms the peninsula of Sanggar, on the island of Sumbawa, east of Java, in Indonesia. In 1814 it was the highest mountain in the region, at 4300 m. But on the night of 10­11 April 1815 Tambora exploded in several phases that culminated in a cataclysmic blast. 150 cu km and 1500 m of the summit of the volcano blew away in an explosion that spread ash 1300 km away and was heard 2000 km away.

Pyroclastic flows devastated the peninsula, and reached the sea 20 km down the mountain with such volume and momentum that they generated a tsunami. The eruption was the largest explosive eruption in modern times. Although the local effects were catastrophic‹92,000 people were killed in eastern Indonesia alone‹the results were global. We know that because we can detect excess sulfuric acid in ice cores from the Greenland ice cap for the years 1815­1818, but there is a rich historical record of the aftereffects from several continents.

Some effects were immediate and direct: for example, the temperature dropped below freezing at Madras, India, one night at the end of April. But the most dramatic effects were those of the following summer. The summer of 1816 was a disaster for crops as far away as India, New England and Western and Central Europe. It was "The Year Without a Summer" in New England, "Eighteen Hundred and Froze to Death." Six inches of snow fell in New England on June 6, and Savannah, Georgia, had a high temperature of 46° on the Fourth of July. Dust in the stratosphere was almost certainly the link between these events, and it also caused spectacular sunsets that are recorded in the paintings of J. W. Turner. A devastating Hungarian blizzard of January 1816 had brown or flesh-colored snow. Snow fell at Taranto, in southern Italy, a rare event in any case, but the snow there was recorded as red and yellow. Throughout the Western world there were records of dust in the air, or "dry fog," or vapor. Many contemporary observers saw the weather of 1816 as related to a sunspot maximum. Benjamin Franklin had already suspected that volcanic effects were responsible for the bad weather of 1783-1784 after the eruption of Laki, in Iceland.

The 1816 crop failures led to unrest and violence all over Europe. The miserable weather even led Mary Shelley to such gloom that she wrote the classic Gothic novel Frankenstein while she was held indoors all summer in Switzerland by the cold wet weather.

This was not the first time that volcanic effects on climate had been felt over wide areas. It's likely that the eruptions of three volcanoes in 1707‹Vesuvius, Santorini, and Fujiyama‹were responsible for a cold winter in 1708­1709, one of the two worst in the century, and one that meant famine and high food prices in Western Europe at least. The terrible winter and famine of 1740 may be associated with eruptions in Kamchatka, but the precise timing and size of those eruptions remains to be determined. The eruption of Laki in 1783 had devastated Iceland and affected much of Western Europe.

The New England weather of 1816 was not the coldest of the century, though it was 0.9° C below normal: the two coldest years were 1.7°C below normal, in 1836 and 1837, after the eruption of Coseguina. It was the pattern of weather that was so devastating in 1816. Temperatures were a little warmer than normal before March and after September, but during the critical spring and summer months the temperatures were critically low, dropping nearly 3° C below normal throughout June and July at Salem, Massachusetts.

Judging from weather records throughout New England from Williamstown, Massachusetts to Bangor, Maine, a succession of cold fronts brought temperatures that were occasionally below freezing to the whole region. A major cold front apparently went through from June 5­11th. It snowed several inches on June 6th, and surface water and the ground were frozen on June 7th. Cucumbers and other vegetable crops were nearly destroyed, and there was another severe frost on June 10th, which frosted leaves on most of the trees. Frost struck again on June 29­30, and yet again on July 9th, August 22nd, and August 29th. Whole fields of corn were killed on valley floors, and hardly any corn ripened anywhere. The hay crop was devastated, though wheat and rye harvests were fair.

Weather records are not as good for the other regions of eastern North America, but it is clear that they had a severe summer too. Canadian crops were hard hit, but with a much sparser population the effects on society were not as extreme. The grain harvest in North Carolina was about one-third of normal. Even the sugar harvest in the British West Indies was affected.

1816 was marked on the Colorado Plateau by the greatest latewood (summer) growth in conifers in 500 years‹since 1487, to be precise (another probable giant eruption, by the way). This probably reflects lack of heat and water stress during an abnormally cool and wet growing season. In fact, there was abnormally large latewood growth in these conifers from 1815 to 1817, the largest three-year growth since the 1490s (for which records are not as reliable).

In Western Europe, low temperatures were accompanied by increased rainfall. Overall in Western Europe, 1816 was one of the coldest summers on record, especially in Western and Central Europe, where annual temperature averaged as much as 3° C below normal, especially in the Alps. Even so, for 1816 as a whole, temperatures were not as low as 1740, as we have seen for New England also. Once again, it was the pattern of weather that was devastating. In 1740 the extraordinarily cold months had been January­May and October­December, when most crops were dormant: in 1816 the cold months were July­September. July 1816 was the coldest July ever recorded in Midland England in nearly three hundred years of observations. The harvest date was six weeks late in southern England, and similar records are found throughout the British Isles.

In Belgium, the summer was cold and wet, so that potatoes rotted in the ground. In Holland, violent storms spoiled many crops. In France, where records in the wine industry go back centuries, the 1816 grape harvest began on the latest date in 325 years of record. (The latest harvest date in the 18th century was in 1740.) The grapes never ripened in the Champagne district round Verdun. Even comparatively small enterprises were hit hard: the "failure of the silk-worms" meant that silk production was slowed down in France that year, so that import duties on foreign silk had to be relaxed in 1817.

The wet weather prevented bees from foraging, creating an "almost unprecedented" shortage of honey. The chestnut crop failed, creating great hardship in some areas of France and Italy. The wet weather even prevented firewood from drying, and by November 1816 the price of firewood had doubled in the German Rhineland. Switzerland suffered greatly in 1816. Its economy had become dependent on importing grain, and exporting beef and cheese. Grain, hay, grape, and potato harvests failed. In central Europe the harvest was a month late, and in many places was greatly reduced by being wet or frosted, and was lost altogether in some places.

The Austro-Hungarian Empire was struck by a blizzard early in January 1816, which was not white but "brown or flesh-colored," probably because it contained volcanic dust. Harvests were small throughout the Empire, and although the Empire harvested enough grain to feed itself, there was no surplus for the normal export. With harvests generally failing in the Alpine regions, and poor communications, there were local famines. Harvests were very poor in most of Italy; the grape harvest failed in Portugal, and olive oil prices sky-rocketed because of poor crops.

All over Western and Central Europe the harvest was deficient both in quantity and in quality. The flour produced was much less than millers usually obtained, and overall estimates suggest that the average grain crop was half of normal, and much worse in some areas. The effect on domestic animals was severe: the lack of fodder caused the slaughtering of many animals that could no longer be fed, and both herds and flocks had close to 50% mortality. However, hardly any official recognition of the poor harvests was expressed during the time the catastrophe was developing. Partly this was a deliberate attempt to avoid speculation in wheat, which had been a problem in some of the war-torn years just previously.

Conditions do not seem to have been so bad to the north and east: harvests were normal in East Prussia, in Scandinavia, and in Russia, apparently.

The monsoon came late to India, so that drought was followed by too much rain later in the rice growing season. This blighted the rice crop.

It is probably not a coincidence that there was a plague epidemic that swept the Turkish empire in 1816, and a cholera pandemic began in Bengal in 1816­1817.

The summer of 1816 was hard on Western and Central Europe in particular because the region was trying to recover from the economic dislocation of the 25 years of war that resulted from the French Revolution and the megalomania of Napoleon Bonaparte. The British blockade of Europe had ruined seagoing trade, and industrial production had declined. The British had lost their European markets, too, and the entire region was already in a depressed state economically. Paradoxically, the end of the war brought new problems, as old trading patterns were re-established after a long break. A flood of British-manufactured cotton upset European textile trade, especially in Belgium, Germany and Switzerland. The Europeans had been cut off from sugar made from sugar cane, and a thriving sugar beet industry had grown up: this was devastated as cheaper cane sugar began to flow into Europe once again. Armies of occupation had to be supported by the French, while in all the countries involved in the war, hundreds of thousands of soldiers, and sailors too in Britain, were paid off and remained largely unemployed, and industries geared to a wartime economy were struggling to adjust to peacetime.

In Britain, good harvests in the previous three years had dropped the price of wheat, and the rural districts were depressed. The British Government passed a new Corn Law in 1815 to cut down wheat imports; this was meant to help the agricultural centers, but it was a delicate business because it tended to raise the price of bread in the cities. It also exported hardship to Prussia, which in 1815 had a large wheat surplus it could no longer export to Britain. Paradoxically, the harvest of 1816, better in Prussia than in Western Europe, raised wheat prices and Prussian prosperity while countries to the west were suffering great hardship. Grain prices are a good proxy for cost of living at this time: bread took about half the income of a normal laboring family, even at times of normal prices. It's clear that there was a genuine and severe drop in real income in the years 1816­1817 as grain prices went up. By almost any standard, grain prices in Western and Central Europe were higher in 1816 than any year previously, and were to be for another 50 years. The highs outstripped previous years like 1709 and 1740. Britain, in fact, was able to weather the 1816 crisis better than some countries on the Continent because it has an accumulated surplus of wheat.

Overall, populations that had high levels of employment were able to cope with rising food prices, and suffered dearth rather than famine. Where there was severe unemployment, there was famine. Thus Britain, France, Belgium, Holland, and most of Germany suffered a dearth rather than a famine. The German Rhineland, and a southern belt across Baden and Bavaria, were close to famine. In the Austro-Hungarian Empire hardly any province escaped at least local famine. The problems were accentuated by high levels of inflation, and bread prices were already high in Vienna in 1815 before they rose again to a peak in 1817. 1817 was called the Year of the Beggar in central Europe. The famine was worst in Hungary and Transylvania, where tens of thousands of people starved to death. Austria and the Balkans were not as bad.

There were close-to-famine conditions almost everywhere in Italy. Most of Switzerland suffered dearth, but isolated cantons in the east suffered famine. The worst of all conditions were experienced in Ireland, where wheat, oat, and potato crops had all failed, and general poverty made import impossible.

There was dearth rather than famine in the United States, especially in New England. Cattle died from lack of fodder, and many people suffered great hardships in Vermont.

As deprivation grew, so did discontent. It's well known that hungry people can cause a lot of trouble, whereas genuinely starving people are too weak to do so. Thus the overt, often violent protests over food shortages occurred in regions of dearth rather than regions of famine. Food riots erupted in Britain, France, and Belgium in 1816, and spread to most of western Europe by 1817. They were sometimes violent demonstrations, and sometimes "taxation populaire," that is, mobs forced bakers and millers to sell them cheap bread or flour. In famine areas, society was more in danger of being overwhelmed by beggars than by rioters. Civil unrest was everywhere, however, with theft and arson being the most prevalent crimes: one for subsistence, the other as an outlet for desperation and frustration.

The riots of 1817 were more concentrated and more dangerous than those of 1816, as the scale of the economic disaster became apparent, and bread prices climbed toward their peak. In France, the policy of the monarchy to keep the population of Paris fed at all costs had been noticed, and thousands of beggars had flooded in from outlying districts. By 1817 the population of Paris was 714,000, with 85,000 of those classed as destitute. This was a very dangerous proportion in a society inured to violence for 25 years. which had been kept relatively short of food. The government kept quiet about the multiple reports of violence in and outside Paris, but its policy of pretending that food supplies were close to normal backfired as people realized they had been systematically deceived. Mobs began to use Napoléon's name freely. In Britain there were similar disturbances on a much smaller scale, and the Government felt compelled to suspend the Habeas Corpus Act in February 1817.

Riots on this scale had not been seen since the French Revolution, and governments were well aware of the comparison. In different ways, with different degrees of enthusiasm, and different degrees of success, they attempted to alleviate the situation to try to keep basic foods available to poor people. They generally did this in two ways‹by actions that would keep food prices down, and by public works programs to employ people, keeping purchasing power up. Results varied with the efficiency and will of the various governments. There was a definite contrast between the outcomes in the more highly organized countries (Britain [except for Ireland], France, Germany, Belgium, and the Netherlands) and those in less well administered countries (Italy, Austro-Hungary, Switzerland, and the Balkans).

France abolished tariffs on imported grains in August 1816, and by November was paying premiums to importers. Many of the smaller German states banned the export of cereals, which helped them but made the situation worse in their less fortunate neighbors. In the end, however, famine relief had to come by grain import ‹but where from? There were few areas with grain surpluses‹eastern Europe, the Turkish empire, or North America‹and the only effective way of shipping bulk commodities was by water. But grains could not be shipped from the surplus areas of East Prussia and Russia until the Baltic ports thawed out in the spring of 1817. For example, shipments did not reach Riga down the river until April 19 that year. The crisis began a new grain trade with Russia via the ice-free Black Sea port of Odessa: 600 ships were sent from Mediterranean ports to Odessa in the fall of 1816. The United States exported $23 million worth of cereals and flour to Europe in 1817.

Much of the new imported grain and cereal arrived after the main crisis, however, and meanwhile governments tried to control food prices. The French subsidized bread prices, especially in Paris, where the newly restored Bourbon monarchy was afraid of the power of the mob. In the end the government had to take out massive loans from London and Amsterdam to pay its bills. The government of Belgium and Holland bought Baltic rye, and managed to keep bread prices within reason.

Prussia was different: it was ruled from Berlin, while the western provinces in the Rhineland were worst hit by the bad summer. Promises of rye from the surplus harvested in eastern Prussia were not kept, and the Rhineland had several very hard months early in 1817. It's possible that unseasonable winds kept the ships from moving west as quickly as they were expected to, but the Prussian government censored the subsequent enquiry, and it's most likely that corruption and incompetence at high levels was the real cause of the delay, coupled with a Prussian indifference to the fate of its non-Prussian subsidiary provinces.

The Austrian government was close to bankruptcy, riddled with inertia, and handicapped by the tremendous difficulty of transporting any bulk materials within its large inland empire. The situation was left for each province to work out on its own. In the same way, the British left Ireland to fend for itself.

It was not until the crisis was over, in most areas with the harvest of 1817, that it became obvious how much the civil unrest had been economic rather than political. King Louis XVIII of France, no doubt with great relief, proclaimed an amnesty in August 1817 for those who had been convicted of crimes caused by deprivation: "Our heart has groaned from the severities that justice and the law have ordered against too many persons who, in several parts of the Kingdom, have been involved in criminal disorders through the scarcity and high cost of provisions."

One response to the famine and deprivation was to emigrate away from it. Hundreds of thousands of people did this unsystematically, taking to the roads as beggars and occasional robbers. But many thousands of others left home as deliberate emigrants in 1816 and 1817. Tens of thousands moved down the Rhine from the hard-hit areas upstream into the Netherlands. Switzerland, as a country with limited agricultural production, had long been a country of emigrants: about 40% of the natural increase normally left the country in the 18th century. But as its economy became more urbanized and prosperous late in the century, emigration slowed markedly, and the Swiss exported textiles and watches instead of mercenary soldiers. Most of the Swiss, Alsatian, and South German emigrants heading down the Rhine were bound for the United States. Of course, many of them were penniless, and they added to the urban distress of cities such as Mainz and Amsterdam.

On the other side of the Atlantic, the emigration of people from New England to the MidWest at this time was a major exodus from the region. Maine lost 10,000­15,000 people. You could argue that the opening of the MidWest was a direct result of the eruption.

The crisis of 1816­1817 was to be the last one of its type, although other events followed such as the Irish famine of 1845­1847, and the crisis of 1914­19 that resulted from the bloodbath of the First World War from 1914­1918 coupled with the influenza pandemic of 1918­19.

Death rates rose throughout the regions affected, and the effects were related to the severity of the food crisis. Sometimes they were moderate, as in Britain and France. But the death rates in Italy were higher: in Lombardy in 1817 they were 39% higher than in 1815; in Tuscany 68% higher; 78% in Apulia; 26% in the Tirol; and 53% in Carniola. In Switzerland as a whole they were 56% higher. Despite the naturally high birth rates of the early 19th century, the population of Switzerland, Austria, northern Italy, and Württemberg fell in 1817. These were direct effects of the crisis, because they all recovered to a strong rate of increase by 1819.

It's not clear what the role of disease was: anyone starved of food is more vulnerable to disease, however. There was no widespread disease epidemic in most of Western Europe, but the lower standards of public health elsewhere, coupled with the much greater numbers of itinerant beggars, apparently encouraged disease transmission in Ireland, in Alpine Europe, and in Italy and the Balkans. Typhus is a fever spread by the feces of human lice, and so can be transmitted by dirty clothing and bedding as well as from person to person. An epidemic apparently broke out on both Adriatic coasts in 1816, and ravaged the people on both sides who were already suffering from malnutrition, including the nutritional diseases scurvy and pellagra; then the epidemic spread to the Alps and to all of Italy.

In eastern Switzerland the greatest cause of death seems to have been simply starvation. Typhus followed the food crisis, and reached its peak in Switzerland in the winter of 1817­18.

Typhus struck the Irish in the fall of 1816, and may have caused deaths in the hundreds of thousands, but in Britain it was largely a post-crisis epidemic. It had begun in London in 1816, but reached its peak in 1818. Even so, the death toll was in thousands, rather than hundreds of thousands.

The economic effect lasted longer than the food crisis. The sudden demand for grains from Western and Central Europe led the Russians and Prussians to plant thousands of acres more crops. With the return to normal climate, the demand slackened, and most of these new enterprises went bankrupt. In general, the surplus agricultural production that followed the crisis promoted sharply lower prices and economic stagnation.

The political effects lasted even longer than the economic ones. The upheavals during the subsistence crisis tended to strengthen conservative attitudes, and there was a general movement throughout Western Europe in the early 1820s towards conservative, rather repressive, authoritarian social policies, in Britain, in Germany, and particularly in France. The effects of this were to last thirty years, and when they were finally swept away, it occurred with much more violence and upheaval than one might have expected, in the amazing year of 1848.


Mount Pinatubo lies on the Philippine island of Luzon, about 100 km northwest of Manila. Its 1991 eruption was significant in several ways. It progressed slowly to full activity, so that volcanologists were able to monitor its bahevior and make successful predictions that saved many thousands of lives and billions of dollars by timely evacuation. At the same time, the eruption demonstrated the long-term effects that an eruption can have on a densely populated agricultural region. And climatologists were able to predict, and then confirm, the effect of a major eruption on global weather.

After lying dormant for 635 years, Mt. Pinatubo began preliminary activity on April 2, 1991, with steam explosions. A line of vents 1.5 km long formed high on the mountain, at first emitting mostly steam and ash that stripped vegetation over several square kilometers. Ash fell as far away as 10 km from the vents. The next morning, 5000 people living on the upper slopes within 10 km of the summit were evacuated. A seismometer array was put in place by 10 April, and immediately began to record dozens of small shallow earthquakes a day.

A crash reconnaissance geology program showed that the volcano was surrounded by recent pyroclastic flows and lahar deposits, and therefore had the demonstrated capacity to inflict major damage. The vulnerable area included Clark Air Force Base, leased at the time by the United States from the Philippines. It lay only 25 km east of Mt. Pinatubo, and was built on old pyroclastic flows and lahars from the volcano. The geologists told the American and Philippine authorities of the potential dangers from the volcano, aided by video tapes of the effects of other eruptions.

On June 1 earthquakes began directly under the new vents, and on June 3 a small ash explosion was accompanied by "harmonic tremor" on the seismographs, a characteristic feature of magma moving at very shallow depth. A larger explosion on June 7 sent ash and steam to a height of 7­8 km, and ash eruption on a small scale became continuous on June 9. Philippine authorities now evacuated people within 20 km of the summit (25,000 in all), and the American commander at Clark Air Force base ordered out 14,000 non-essential personnel the next day. All military aircraft and much important equipment was flown out, leaving behind only a few helicopters. It was estimated later that the timely warnings of geologists saved the loss of $4 billion in equipment. An estimated 10,000 local people fled (voluntarily) from the nearby town of Angeles.

On June 12, Pinatubo blew out a very large cloud 19 km high, and small pyroclastic flows swept short distances down the surrounding river valleys. The evacuation zone was extended to 30 km, now affecting 58,000 people.

Explosions became more frequent and more powerful over the next three days. Observation of the volcano became much more difficult as the air filled with dust and ash. By June 15, Clark Air Force base was at times in total darkness in daytime, and the last military personnel left. A climactic explosion took place that afternoon, blasting an eruption cloud 35­40 km high and sustaining it for 11 hours. In this phase, Mt. Pinatubo blew out the equivalent of about 3­5 cu km of dacite magma. Most of it was airborne ash and pumice that blanketed the area to distances of 150 km from the mountain, but pyroclastic flows down the mountain side reached up to 18 km from the base.

Even the geologists were forced out of Clark Air Force Base by this time, to safer positions 40 km to the west. Most of the seismometers on the mountain had been put out of action, but it no longer mattered: the earthquakes under the volcano were now up to magnitude M5 as the summit collapsed to form a caldera 2­3 km across.

The local effects of the eruption were complicated because it coincided with the rainy season in the Philippines. By bad luck Typhoon Yunya passed over the Philippines during the eruption. Many centimeters of ash on the land surface was saturated by heavy rains to form lahars that travelled tens of kilometers outward from the mountain in all directions, and airborne ash rained out in a thick paste that collapsed all the hangars at Clark. More important, the ash collapsed the roofs of dwellings in the region, killing more than 200 people sheltering inside them. River and stream channels near the volcano were filled with tens of meters of debris, and floods ran wild over the new landscape. Meanwhile, the winds of Typhoon Yunya carried ash falls as far away as Singapore and Thailand, 2500 km away.

Ninety thousand people were still in evacuation camps in late July 1991. By November, the number of deaths attributed to the eruption had reached more than 700: 281 in the original eruption, 83 in subsequent mudflows, but 358 from disease in refugee camps. The aftereffects of the eruption continued to pose longterm danger and to represent an economic regional disaster for the region. The economic devastation in the region was very great, accentuated by the fact that the US Air Force decided not to re-establish Clark Air Force Base.

First, the volcano continued low-level activity, with steam and sulfur dioxide emission for more than a year. Much more hazardous, however, were the problems posed by the enormous volumes of ash that had erupted in 1991. It choked valleys for miles around, altering the drainage patterns of rivers, forming new lakes or deep ravines. Every river flood became a mud-flow (lahar). The mud was deposited on the river plains downstream, choking and burying villages and even small towns, and covering roads and fields with mud that became a quagmire after every rain. For miles around the volcano, the only transport that could operate in the rainy season was donkey-cart, boat (between floods), and helicopter. Lahars were often unpredictable, and were killing people in the rainy season of 1992.

The ash essentially put an end to agriculture near the volcano. Although most people moved out, members of the Aeta tribe hung on, eking out a living by hunting, gathering forest products, and digging in the ash for pieces of wood that had been charcoaled in the pyroclastic flows.

The rainy season of 1992, plus the usual typhoons, sometimes led to dangerous explosions as large amounts of rain water percolated through the ash to layers that were still superheated. These explosions, taking place in loose wet sediment, often resulted in severe mudslides that added to the lahars. Civil authorities spent $300 million on emergency check dams in the river drainages around the volcano, but often these dams failed in the floods of July and August 1992. About 70,000 people were still living in evacuation centers and resettlement sites in July 1992.

This situation did not improve markedly in the next few years. Several villages were buried 4­5 m deep in the rainy season of 1992. The ash from Mount Pinatubo took a long time to wash downstream, or to be covered by new vegetation. Each rainy season, and especially each typhoon, continued to generate new, dangerous mudflows all round the volcano. The mudflows of summer 1993 buried several more villages. A levee had been built round one of them, except for one break where a landowner had refused to give permission: the lahar swept through the gap and dumped six feet of mud throughout the village. In late 1995 Typhoon Sybil (Merang) went directly across Mount Pinatubo, causing huge lahars that buried yet more towns and villages.

But Mount Pinatubo also had a global effect. Its eruptive products included large amounts of sulfur as H2S and SO2, and the pumice included anhydrite (calcium sulfate), as at El Chichón. The eruption cloud from Pinatubo was high enough to inject dust and gas directly into the stratosphere: perhaps 19 million tons of SO2 were blown over 20 km high. The H2S and SO2 formed an H2SO4 aerosol in the stratosphere, as had happened in the 1982 eruption of El Chichón, and in three weeks had travelled westward completely round the world to form an aerosol cloud in low latitudes: later it spread poleward.

The eruptive activity continued at lower levels for several months, but periodic explosions continued to eject material 15 km high, and therefore continued to maintain the stratospheric cloud. As late as November, the aerosol cloud was still dense within the tropics, and was beginning to affect higher latitudes, mainly by offering them beautiful sunrises and sunsets. At that time the total volume of pyroclastic deposits erupted had reached perhaps 7­11 cu km, making Pinatubo an eruption rivalled this century only by Cerro Azul (Quizapú) in Central America in 1932, Katmai in Alaska in 1912, and Santa Maria (Guatemala) in 1902.

The stratospheric aerosols from El Chichón had been forecast to cause 0.3° C cooling at the Earth's surface in 1983. This was only just above the background annual variability (about 0.2°C), and was difficult to assess at the time, especially because El Chichón erupted in the same year as one of the most powerful El Niños of the century. Some people suspected that the El Chichón eruption either triggered or amplified the El Niño, but the link could not be established.

The Pinatubo eruption of 1991 produced 3­5 times as much magma, and 2­3 times as much sulfur gases, as El Chichón. Estimates late in 1991 anticipated a global cooling of 0.5° C in 1992­93, large enough to detect above background variation, even against the background of a moderate El Niño. The temperature records to April 1992 confirmed the prediction, and the summer season of 1992 over much of North America was short and cool, breaking records in many localities.The temperature records to April 1992 confirmed the prediction, and many area of the United States had a very cool summer, in some areas the coolest in 80 years. By mid-September, newspapers were reporting that the corn crop was one of the largest ever, estimated at 245 million tonnes, but was not yet harvested: it was so far behind schedule because of the cool summer that there was a serious danger it would be damaged by frost before it could be harvested. In the end, it was not, but it was a close call, and underlines the fact that volcanoes, unlike earthquakes, can have global effects.