Effect Measure

Yesterday was the 12th 22nd anniversary of an environmental catastrophe in Cameroon. On August 21, 1986 Lake Nyos in that West African country belched a huge load of carbon dioxide and suffocated 1700 people as they slept. Like its monoxygenated cousin carbon monoxide, carbon dioxide is “a colorless, odorless gas that can kill you in your sleep,” as the marketing gurus of monoxide detector like to say, but it was true in spades on that fateful day 22 years ago. Carbon dioxide doesn’t kill the same way as carbon monoxide, which binds tightly to the homoglobin in your blood, shutting out the oxygen. But in large enough quantities it is just as effective in killing you. When CO2levels reach 10% you suffocate from lack of oxygen.

Lake Nyos is a volcanic lake, meaning it sits in the crater of a volcano. A brief note in Wired tells what happened:

Magma deep underneath the lake releases carbon dioxide into its depths. Lake Nyos is 690-feet deep, enough for the water pressure to keep the CO2 dissolved in the lake water, rather than letting it bubble up and escape to the surface. And the crater rim towers above the lake, blocking winds which could otherwise stir the surface and create convection currents that would circulate the deep, CO2-saturated water upward to areas of lower pressure. The lack of seasonal variation less than seven degrees north of the equator also contributes to the lake’s placidity.

Volcanic rumbling or other seismic activity could have triggered the sudden release of the gas that deadly night, but there’s no record of any tremors and no evidence that anything shook off the shelves of homes in nearby villages. It’s possible the gas at the lake’s bottom just got so concentrated that even under pressure it came out of solution and formed bubbles. Once the bubbles started rising, a “chimney effect” would have rapidly siphoned huge amounts of gas to the surface.

The gas burst through the surface with a rumble, generating a giant wave that scoured vegetation from the shores. The CO2 cloud was at least 300-feet high, because it suffocated cattle on hillsides that far above lake level. Iron from the deep water oxidized and stained the lake waters with rust.
Then the gas crept down the mountain valleys, invading homes. It extinguished oil lamps and suffocated people in their sleep. Some who were awakened by the loud gas bubble stood up and lived, because their heads were above the invisible gas near the ground. But many who went outside paid with their lives.

Few survived. Those from neighboring villages who discovered the devastation recalled with terror the legends about evil demons living in mountain lakes.(Wired)


Lake Nyos

The Wired link has a slideshow of ten other “Death by Volcano” catastrophes of one kind or another (not mainly CO2 events). But you don’t need a volcano to die from carbon dioxide suffocation. It’s long been known as a workplace hazard, not just because it won’t sustain animal life but because it is heavier than air and can settle in trenches, pits, valleys and other work spaces. The EPA surveyed CO2 workplace fatalities between 1975 and 2000 and found 51 incidents and 72 deaths from accicental discharge of fire extinguishing systems (see here). Beside displacing necessary oxygen, there are probably some toxic effects on the central nervous system, but in mass events like Lake Nyos, suffocation was probably the main mechanism.

Twenty-two Twelve years ago the Lake Nyos catastrophe was front page news. But life goes on and most of us had forgotten about it. CO2 is much in the news again, but not at the parts per hundred levels that killed people in Lake Nyos but in the parts per million levels that contribute to the Greenhouse effect (see our post here, here, here). Lake Nyos is a reminder that even relatively non-toxic gases can kill quickly rather than over a century if the circumstances are just right.

22 years ago yesterday they were just right. Or perhaps more appropriately, horribly wrong.


  1. #1 Frank Mirer
    August 22, 2008

    Permit a toxicologist to make a small point. The mechanism of fatal carbon dioxide exposure described above is incorrect – it’s not from exclusion of oxygen. This is a point of some importance for occupational exposure.

    Carbon dioxide acts as a general anaesthetic. There’s reference to its use in veterinary practice. Stages of anaesthesia progress from euphoria to loss of consciousness to respiratory depression. Respiratory depression causes death at levels where oxygen exclusion would not. At 10% carbon dioxide in air by displacement, oxygen is reduced by 2%, to about 18%, which will support life.

    Gas detectors used for confined space entry don’t detect carbon dioxide.

  2. #2 Rosie Redfield
    August 22, 2008

    2008-1986=22 years ago.

  3. #3 revere
    August 22, 2008

    Rosie: Yikes. Correction made. Thanks. I guess unless the mathematics has Greek letters I’m not so good at it!

  4. #4 Thomas
    August 22, 2008

    Anyone wanting to test the effects of CO2 poisoning has a voluntary test subject in Arthur Scargill:
    “I challenge George Monbiot to test out which is the most dangerous fuel – coal or nuclear power. I am prepared to go into a room full of CO2 for two minutes, if he is prepared to go into a room full of radiation for two minutes.”

    You’d imagine that someone who has worked as a miner should know something about the risks of bad air.

  5. #5 revere
    August 22, 2008

    Frank: Thanks for pointing this out. I did think about this when I wrote the post, which is why I said there was evidence CO2 acts on the CNS but that for Lake Nyos I thought the mechanism was anoxic asphyxia. My reasoning was that the evidence was people were rapidly immobilized and the huge CO2 bolus, being heavier than air, settled on them at high concentrations. Your point about the 10% level is a good one. I looked up the change in atmospheric pressure with altitude that would correspond to a drop in O2 partial pressure of about 10% and it corresponds to a height above sea level of only 3000 feet. That is clearly not sufficient to cause a problem (Denver is a mile high, right?) so one wouldn’t expect a sea level problem at 18% ambient oxygen. You’d probably need somewhere around 20-25% for that.

  6. #6 ECOLOG
    August 22, 2008

    Good article.
    Some adding to the article about of “Volcanic rumbling or other seismic activity could have triggered the sudden release of the gas … but there is no record … no evidence… “.
    There is a record of atmospheric precipitation. It is well-proven
    , that catastrophes 1986 and in 1984 on the lakes of “NYOS” and “MONOUN” were caused by the diminishing of atmospheric precipitation in 1983 in contrast with any other year of the period with 1929 till 1988.

  7. #7 daedalus2u
    August 22, 2008

    If flames were extinguished, that means an O2 level less than ~15%


    That implies a CO2 concentration of more than 40% (0.21/(.79+.21+.4))=0.15 which is acutely fatal with coma in less than 1 minute.

    My guess is that it was essentially pure CO2, with perhaps some H2S (if there was H2S dissolved in the water too but perhaps not because iron in the water is incompatible with H2S being in the water too, H2S might have killed plants too, which CO2 would not). It would displace all the air not mix with it because the CO2 source is essentially a planar source with no velocity gradients.

  8. #8 M. Randolph Kruger
    August 22, 2008


    And be aware there are some 22 sites around the US that are potentially the same. Mostly around Yellowstone, some in Arkansas and Mammoth Lake in CA.