When Ephriam Che walked downhill toward Lake Nyos the morning after the disaster, one of the first things he noticed was the color.

The lake was normally blue. Now it was reddish brown.

Then he reached the outlet and noticed something stranger. The waterfall that usually drained the lake was not running. Around him the morning was unnaturally quiet. In an interview years later, Che remembered the absence of the sounds that should have been there — birds, insects, ordinary life.

Farther down the slope he began finding people.

The buildings had not been flattened. There was no burn scar moving through the village, no wall of volcanic ash covering the houses. Families had simply stopped where the night had found them.

Che’s own parents, siblings, aunts and uncles were among the dead.

The official American scientific team that arrived afterward recorded similarly bewildering details. Some survivors awakened six to thirty-six hours after losing consciousness. Oil lamps had gone out even though fuel remained in them. Animals nearby were dead. In one place four of five goats had died; in another house four members of a nine-person family survived while the others did not.

Death followed the land.

Cattle in low places died while animals on higher ground lived.

About 1,700 people and thousands of livestock were killed along the valleys below Lake Nyos on the night of August 21, 1986.

Nothing about the scene looked adequate to explain that number.

The explanation was in the water.

The lake looked normal because the dangerous part was underneath it

Lake Nyos occupies a volcanic crater in northwestern Cameroon. It is small enough to cross without much ceremony but unusually deep, reaching roughly 200 meters.

That depth mattered.

Carbon dioxide of magmatic origin was entering the bottom of the lake and dissolving into the deepest water. Pressure from the water column above kept much of the gas in solution, rather like carbon dioxide remaining dissolved inside a sealed bottle of soda.

The surface could remain comparatively ordinary while the lower lake accumulated an enormous quantity of gas.

When Kling sampled nearshore water before the disaster, he found nothing alarming because the gas-rich water was concentrated much deeper in the lake. The problem was not that he missed an obvious warning at the surface; the relevant information was below the depth he could reach from shore.

Lake Nyos is close enough to the equator that seasonal temperature changes do not reliably overturn and mix the entire water column the way they do in many temperate lakes. Deep water can remain isolated. Carbon dioxide entering at depth therefore had somewhere to stay.

Year after year, the lake could charge itself.

The top still looked like a lake.

The bottom was becoming something else.

There had already been a warning

Two years earlier, another Cameroonian crater lake had killed people.

On August 15, 1984, gas escaped from Lake Monoun, roughly sixty miles south of Nyos. Thirty-seven people died along a low section of road.

The event was baffling enough that volcanologist Haraldur Sigurdsson investigated it. Among the clues was water recovered from depth that was heavily charged with carbon dioxide.

The possibility emerged that a volcanic lake could store gas and then release it catastrophically.

It was an unfamiliar hazard.

Then Nyos demonstrated its scale.

By the evening of August 21, 1986, thunderstorms had passed through the Nyos area. According to the U.S. investigation, conditions had become calm by roughly 9:30 p.m. Most people were doing what people usually do at that hour.

Then came rumbling.

Witnesses close to the lake described perhaps fifteen or twenty seconds of noise. Some went outside. Several later remembered an unusual warm sensation or odors they compared with gunpowder or rotten eggs, although about half of one interviewed group near Nyos reported no unusual smell or taste at all.

One observer on high ground saw a pale cloud rising from the lake and a large wave strike the shore.

People in the valleys generally did not report seeing the cloud that was about to reach them.

A gas heavier than the night air

Carbon dioxide is invisible.

At ordinary atmospheric concentrations it is part of everyday life. Humans exhale it. Plants use it. Industrial systems produce it constantly.

Concentration changes everything.

The Nyos release displaced breathable air across the low ground surrounding the lake. The medical record compiled afterward from hundreds of survivors was consistent with massive exposure to an asphyxiant gas. People lost consciousness. Some remained comatose for hours. More than 800 survivors received medical treatment in the two principal hospitals studied afterward.

Because carbon dioxide is denser than ordinary air, terrain helped determine where it went.

It poured into valleys.

That explains one of the strangest patterns investigators encountered afterward: survival sometimes changed with elevation across relatively short distances. Houses were still there. Crops could remain standing. The geography of death was written less in wreckage than in contours.

The gas followed the land until dilution finally made the air survivable again.

The precise death count varies slightly among later sources. The most defensible wording is about 1,700 people, with some later literature giving 1,746 and other institutional summaries rounding upward or downward. Thousands of cattle and other animals also died.

What actually started it?

This is where popular explanations tend to get more certain than the evidence.

The carbon dioxide came from a magmatic source beneath the region. That part is well supported.

A direct volcanic eruption on the night of August 21 is not.

The U.S. scientific team found no persuasive evidence of a new eruption beneath the lake and no clear seismic trigger. Lake temperatures and bottom sediments did not show the disturbance expected from some of the early volcanic-explosion theories. The 1987 Science synthesis concluded that the bulk of the gas had already been stored in deep lake water before it escaped.

What began the runaway degassing is harder to establish.

Possibilities have included a landslide entering the lake, unusual wind-driven mixing, cooling of surface water, or instability developing internally as the deep water approached saturation. More recent USGS treatment considers a substantial landslide a probable trigger, but the mechanism is still better described as a leading interpretation than as a witnessed fact.

Once gas-rich water began rising, however, the process could feed itself.

Pressure decreases as water approaches the surface. Dissolved carbon dioxide then comes out of solution as bubbles. Those bubbles make the rising mixture more buoyant, carrying still more gas-rich water upward. More pressure is lost. More gas forms.

The lake had not emptied itself

There might have been some comfort in believing the catastrophe had exhausted the danger.

It had not.

Measurements made after 1986 showed that the gas release had mixed the lake substantially but incompletely. Carbon dioxide remained below. More continued entering.

Within years scientists could watch the deep layers recharging. Measurements from 1986 through 2001 indicated continued accumulation, and by 2001 parts of the system were again approaching very high gas saturation.

That produced an unpleasant engineering problem.

The lake needed to be emptied of gas.

Disturbing the lake was also one of the things scientists feared might release the gas.

There were proposals involving tunnels and chemical treatment. None was attractive. Whatever solution was chosen had to remove enormous amounts of dissolved carbon dioxide without destabilizing the water column it was supposed to protect.

The pipe

The idea was to lower a vertical pipe into the gas-rich deep water and bring that water to the surface deliberately.

At first, the water has to be lifted.

Then the physics takes over.

As deep water rises through the pipe, pressure falls. Carbon dioxide begins forming bubbles. The mixture becomes lighter. That accelerates the upward flow, which reduces pressure further and releases still more gas.

Once properly started, the system becomes self-sustaining.

The lake powers its own degassing.

It is one of those pieces of engineering that seems obvious after somebody has survived all the uncertainty required to prove it.

The first experiments were cautious. Researchers tested small pipes during the 1990s. There was genuine concern that deliberate degassing might create enough mixing to trigger the very disaster they were trying to prevent.

The tests worked.

Full-scale degassing at Nyos began in 2001. A pipe reaching roughly 200 meters into the lake produced a fountain tens of meters into the air.

No pump was needed once the flow was established. The carbon dioxide did the work.

The remedy looked strangely like a domesticated version of the catastrophe.

In 1986, gas-rich water had reached lower pressure uncontrollably and initiated a massive release.

Now engineers were allowing gas-rich water to reach lower pressure continuously, through a pipe, in quantities the atmosphere could disperse.

The dangerous process had not been abolished.

It had been given plumbing.

One pipe was not enough

The first permanent installation made Nyos safer, but the initial removal rate was not enough to dispose rapidly of the enormous gas inventory while new carbon dioxide continued entering from below.

Researchers kept monitoring the lake.

Two additional larger pipes were installed in 2011. Long-term studies show that the project substantially reduced the stored carbon dioxide, and a 2020 review of more than three decades of work concluded that the large initial inventory had been brought down far enough that a single appropriately maintained pipe could, in principle, balance the lake’s continuing natural recharge over the long term.

That sentence hides a lot of human persistence.

Scientists and engineers had been returning to these lakes since the 1980s. Funding came unevenly. Equipment had to operate in a remote region. The intervention itself had once been considered potentially dangerous.

And the project was preventative.

Preventative engineering has a peculiar public-relations problem. If it succeeds perfectly, nothing happens.

There is no spectacular rescue photograph. No city is rebuilt. The evidence of success is an ordinary morning.

A fountain where the disaster used to be

There are photographs of the degassing system operating on Lake Nyos.

The visual is almost ridiculous.

A narrow white column rises from the center of an otherwise tranquil crater lake. The USGS describes one fountain as roughly 120 feet high and explains that the flow needs no pump because the carbon dioxide provides the lift.

It resembles a municipal fountain placed in the wrong landscape.

That fountain is doing the exact thing Lake Nyos once did catastrophically, only slowly enough to be safe.

The practical finding was specific: a surface sample could not describe a hazard stored roughly 200 metres down. After 1986, repeated measurements through the full depth of the lake showed carbon dioxide rebuilding over time, and the degassing system was designed around that record.

There is also something unusually satisfying about the final piece of engineering.

The lake could not be ordered to stop receiving carbon dioxide.

So the system was designed to stop allowing it to accumulate.

That distinction is worth keeping.

Final Vestige has encountered the same broad idea in a very different setting in The Cisterns Beneath San Francisco. After catastrophe, the valuable work is often the work that looks unnecessary on every day the catastrophe does not return.

Lake Nyos offers the stranger version.

A lake killed people because it could hold its breath for years.

The answer was to make sure it never had to hold that much again.

---

Sources and further reading

• U.S. Geological Survey: Final Report of the United States Scientific Team
• Science: The 1986 Lake Nyos Gas Disaster in Cameroon
• Medical study of Lake Nyos survivors
• Smithsonian Global Volcanism Program: Oku Volcanic Field / Lake Nyos
• PNAS: Degassing Lakes Nyos and Monoun
• Final Vestige: The Cisterns Beneath San Francisco