What is epsilon?

I recently posted about the unfortunate failure to examine data properly which contributed to the disastrous loss of life with space shuttle mission 51-L (the Challenger disaster). One of the members of the commission which investigated the disaster was physicist Richard Feynman.

In one of his biographies he tells an interesting story. My memory is imperfect, but it goes something like this. At one point he was meeting with a NASA manager and some engineers. He asked them point-blank, what’s the probability of mission success for a launch? The manager replied something like “certainty.” Feynman protested that no it wasn’t. The probability of success is 1 minus epsilon. What’s epsilon?

He even let them submit their estimates by “secret ballot,” writing them on a slip of paper, but it was immediately obvious whose guess was whose. The engineers gave failure probabilities around 1 out of 100, or even as optimistic as 1 out of 400, but the manager’s estimated failure probability was 1 out of 100,000.

1 out of 100,000! At those odds, you could launch a shuttle every day for 300 years and expect only one mission failure. Feynman’s wasn’t the only bullshit meter flashing red.

He also tells of an interesting discussion with the range safety officer, who had to decide whether or not to include a self-destruct mechanism. About 1 out of 25 solid-rocket launches had failed, but the shuttle was better-than-average rocket technology so he estimated the chance of catastrophic failure at about 1 out of 100, optimistically, which would indeed require the inclusion of a self-destruct mechanism. Management balked at this figure — so the range safety officer altered his estimate to conform with management wishes, then attached a self-destruct mechanism anyway.

We’re seeing the same exact thing happen with the global warming “debate.” Fossil-fuel shills, conservative politicians and pundits, and misguided and ill-informed bloggers are telling us “No problem, don’t do anything. Keep burning fossil fuels, impose no restrictions, no costs. Drill baby drill!”

My bullshit meter is flashing red.

So here’s your chance to answer the question: What’s epsilon?

I moderate this blog with a heavy hand, but for this post I’ll allow anyone to submit an estimate. That does not mean you get to proselytize. Don’t give us your reasons for thinking climate sensitivity estimates are too high, or it’s all the sun, or it’s galactic cosmic rays or sheep albedo, or launch into a diatribe about emails from the climate research unit — attempts to argue why are not the point of this post, and indulging in same will get you sent to the trash-bin. The prohibition against proselytizing goes both ways — don’t give us your reasons for insisting climate sensitivity is high, or it’s not the sun or cosmic rays, or a diatribe about the character assassination of honest climate scientists.

Just give a number. You don’t even have to sign your name (feel free to do so if you wish) — just a number on a virtual “slip of paper.” Have the courage to put your opinion on the record, say outright, what’s your estimate?

What’s the chance that if we continue with business-as-usual, man-made global warming will lead to disastrous climate change? It isn’t zero. It isn’t one. What is epsilon?

I’ll start. 95%.


186 responses to “What is epsilon?

  1. Define “disastrous”.

    My guess as to the probability that man made global warming given BAU will lead to major ecosystem destruction and a major mass extinction of species other than man:

    97% give or take 3%.

    The probability that that will lead to the collapse of our civilization along with world trad, with consequent reduction of the Earth’s population to that which could be sustained approximately 100 years ago:

    60% give or take 20%.

    The probability of species extinction for humans:

    5% give or take 5%.

    [Response: It’s up to each respondent to apply his or her own definition of “disastrous.”]

  2. Ionut Enache


  3. It all hinges on your definition of “disastrous”. If you mean “will cause massive changes in natural ecosystems and major disruptions to subsistence farmers in poor countries” then I think it’s O(99%). If you mean “will cause a collapse of global civilization in the 21st century” then I think it’s more O(10%)

  4. The pedant asks, what qualifies as “disastrous” (and over what time frame)?

    Regardless, I’ll estimate: 85%.

  5. A nit: your 95% chance is, I’m sure, the “chance that if we continue with business-as-usual, man-made global warming will lead to disastrous climate change.”

    Epsilon itself, though, would logically be the complement–5%.

    As to the substance–I’m ruminating!

    [Response: No, I’m definining epsilon as the probability of disastrous climate change, so its complement is the chance we won’t suffer disastrous climate change.]

    • “I’m defining epsilon as the probability of disastrous climate change, so its complement is the chance we won’t suffer disastrous climate change.”

      Roger that–I was working from Feynman’s “1 minus epsilon” when I made that comment.

      Upon rumination, I think I’d define “Disastrous” as applying to any event involving serious loss of life or treasure. In that case, epsilon approaches 1 very closely indeed, since there have already been several such events which are unlikely to have occurred without the observed warming.

      Cumulatively, I’d guess–I emphasize “guess”–that it’s a 95% chance that without AGW they’d have ALL occurred.

      Further, I’m 99.99% certain that warming is going to continue, hence the preconditions for further such disasters are going to increase in the near future.

      Further yet, Dr. David Archer (among others) has pointed out that the long tail of the CO2 curve extends out to as much as 100,000 years, which means this elevated risk will likely (as in “absent significant geoengineering or some as-yet-undreamt-of natural equivalent”) exist for a very, very long time in human terms.

      Put that all together and there’s not much complement left, I’m afraid.

      As to how bad it may get, given all the unknowns, I think the wisest word is Dr. Ladbury’s: “Uncertainty is not your friend.” Or mine. Or anyone’s.

      • Oops–verbal ‘sign error’: “they ALL would have occurred” should read “NONE of them would have occurred.”

        Probably clear from context anyway, but there you are.

  6. Sorry to confuse things right off the bat, but I have two numbers.
    Disastrous for mankind: 35%.
    Disastrous for the environment: 85%

  7. Disastrous?


    Leading to political crises?


    Leading to extinction of species?


    Changing the earth’s face as the anthropocene has seen it so far?


  8. 99% It’s already there.

  9. Daniel J. Andrews

    Depends on length of time too. By 2100, 70%. By 2200, 98% or higher, in other words, virtually certain providing no new mechanisms are uncovered (to steal a caveat from Lord Kelvin).

  10. human extinction: 0.1%
    any mass extinction: 20%

  11. Disastrous to me means climate change alters our fundamental relationship with the climate causing human civilization to have to adapt in ways that changes almost every aspect of society.

    Probability = 90%

  12. I’d say less than a one in a hundred chance of not having disasterous climate change.

  13. Chance that BAU emissions will cause more harm than good: >99%
    Chance that BAU emissions will cause “significant” damage (eg, >1 percent of GDP, extinction of >5% of species, etc): 90%
    Chance that BAU emissions will lead to some kind of major economic collapse, mass starvation, loss of major ecosystem, etc: 15%
    Chance that BAU emissions will lead to species extinction for humans: <1%.

  14. 95%

    I started higher but thought about it and adjusted it to what seemed reasonable. Did Feynman start by giving everyone his own guess? I’d hide that, to avoid biasing the results.

  15. Catastrophic change: negligible
    Moderately bad change: 20%
    No better or worse than now: 60%
    Better than now: 20%
    Hugely better than now: negligible.

    But the funny thing is that I think the chances of dramatically improving our quality of life in the West by doing what it would take to lower carbon emissions very substantively are just about 100%.

  16. Great idea for a poll by the way. Makes one get clear about what he thinks about something very basic.

  17. The wording of the question is a little unclear – are you asking for P or e = 1-P? Both your number and Curtis’s seem to be P.

    Anyway, I think Curtis’s approach is good, and I’ll follow his lead:

    – Species extinction clearly linked to man-made warming; P = 90%
    – Modern civilization collapse clearly caused by man-made warming: P = 15%
    – Human extinction caused directly or indirectly by man-made warming: P = 1%.

    [Response: I’m asking for the probability that business-as-usual will lead to disastrous climate change. I call that epsilon, you can call it P if you like.]

  18. Probability of human extinction: 1 in a million. Sorry, but we are an extremely robust species and fit into too many ecological and climatological niches to be wiped out. We’re like ants in that regard.

    Probability of civilization collapse: sorta like “disastrous,” it depends on how you define “collapse.” The level of civilization in Dark Ages Europe was lower than under Rome, but still much higher than pre-Rome. Let’s define it as a lowering of real per-capita GDP by 20% or more. Epsilon: about 75%.

    • I can’t recall where or when I saw (heard?) it, but humans have been referred to as “king weed”. We are a very weedy and adaptable species and the risk of our own extinction is slight.

      • EVERY species goes extinct; humans will, too.

        Your implied question is “what is the chance of disastrous human-caused climate change being the cause of human extinction?” Factor in possible nuclear winters, ocean acidification, and a host of other factors, and the chances of human extinction in the Anthropecene is five nines.

        Let’s hope we evolve into something else fast.

  19. Considering that Global Warming is already driving many disasters, I’d say it is already an established fact that global warming has led to “disastrous” climate change. So I dispute your assertion that “it isn’t 1.” See for example: [http://www.wunderground.com/blog/JeffMasters/comment.html?entrynum=2008]

    Similar comment relating to Tom’s “major ecosystem destruction and a major mass extinction of species other than man.” Again, these events are already in progress, so it’s hard to argue that epsilon isn’t 1.

    But I think I agree that the probability of humanity’s extinction is rather low. 10% sounds about right to me.

  20. Defining BAU as something similar to IPCC SRES A2, I think 95% chance of disastrous climate change is a good estimate.

  21. 70% regarding to my definition of disastrous.

    90% that we’ll face a temperatur increase by more than 3°C in 2100, but thät’s another question.

  22. This to me is the crux of the debate, and the question that deniers and their fellow travellers are desperately trying to prevent being asked. We know global warming is happening, we know human activity is largely responsible, and we know there will be consequences. The scale of the consequences is far less well constrained.

    [Response: This is dangerously close to proselytizing.]

    My own quantification of epsilon would be that it will approximately linearly increase, from quite close to zero right now, approaching unity after about 100 years if “business as usual” continues that long.

  23. If we continue with business as usual – assuming the standard definition – 99% probability of disastrous.

    Definition of disastrous – say, equivalent to or maybe worse than the period of the black death period (I’ve only a hazy idea about that time). Anyway much worse than WWI and II. Life as we know it is turned topsy turvy – all the normal expectations (war, famine, water shortages, disease, disruptions to lots of things – travel, communication etc) plus some things we haven’t thought much about.

    Probability of dropping below business as usual in the next 30 years – say, 40% to 60%? (Am I being optimistic?).

    • Black death was something like 30-50% mortality globally. WW2 was 2% & WW1 0.6%. (The Spanish Flu of 1918 killed over twice the number that WW1 did suggestiong that world war isn’t half as bad a s dose of flu! Then, some reckon the flu became the killer it was (and only could have) in the human filth of the trenches.)

  24. To perpetuate BAU, the probability that a nation, or a group of wealthy individuals/corporations, acting unilaterally, will attempt to save our lifestyle by killing up to 80% of the earth’s human inhabitants – 29.9%.

  25. The Wonderer

    I recall Feynman’s description of his time on the Challenger commission being in his second autobiography, which I cannot find right now, but I did find this on the web:

  26. Tamino , what do you mean by “BAU” in this context ?
    An example : some argue that the natural depletion of FF could be enough to limit our emissions below a dangerous threshold . But “naturel depletion” can mean “BAU”, in the sense that it is the natural increase of FF extraction costs that could cause the destruction of demand – so it’s “BAU”, but in a depleting world, we have nothing to “do” .

    Independently of whether you personally think it is plausible, or not, would you call this possibility a “BAU scenario” , or not ?

    [Response: I’m afraid you’ll have to decide for yourself what you consider “BAU.” It’s quite understandable that people will disagree.]

      • Kevin, as Tamino says below, the question wasn’t whether you believe or not in this hypothesis- it was just whether you would consider this scenario as “BAU” or not, if it were actually to happen.

        Tamino’s question obviously relies on the total amount of FF that would be “naturally” burnt with BAU – but this quantity will also obviously be finite, so the real point is both
        a) which quantity you assume to be burnt , AND
        b) which climate sensitivity you assume – both issues being of the same importance for the resulting aftermaths.

        Some people do think that coal reserves are not that big. In any case, the question of reserves is NOT addressed by climate science, nor by any “science” I know – so I don’t think it can be considered as scientifically settled; nowhere in IPCC report you will find any substantiation of the validity of scenarios – they’re just “storylines”. So I think everybody is free of having it’s own opinion about the reserves, and mine is as defensible as yours – without contradicting any scientific result.

      • Fair enough, jarch. But my point–cryptic though its phrasing admittedly was–is that depletion per se won’t be enough to prevent more trouble than we already have, if coal use is maintained. And BAU–any BAU that seems plausible off the top of my head, anyway–would, I think, preclude a quick end to coal combustion.

        Just an unsubstantiated opinion, hopefully not too unreasonable.

    • I understand your point, but I would be interested in knowing your answers to more detailed questions that I ask further down in the thread.

  27. In any case, and independently of the answer , my personal guess is
    * it’s extremely likely that our civilization cannot survive the extinction of FF, and more generally the exhaustion of natural resources (say, 99 %), w or w/o climate change.
    * it is extremely unlikely that climate change will constitute the main issue of the coming century ( < 1 %).

  28. 99% that the humans who lived on earth between 1950-2050 will be cursed by future generations for 10s of thousands of years.

  29. 90% – but it really does depend on what you mean by disastorous.

    But I was hoping you were going to be asking what do you think climate sensitivity is? Of which I would have to guess at 3.


  30. Paul from VA

    Different definitions of disasterous (assumed target year is “by 2100”):

    Civilization sucks (defined as each subsequent generation being poorer than the one that came before, vast increases in inequality): 97%
    Civilization collapse (defined as human depopulation of 20% or more from peak): 50%

  31. arch stanton

    I chose to define “disastrous” anthropocentrically: Widespread crop failure/famine, along with costal storm flooding due to sea level rise resulting directly in loss of >1 billion human lives by 2075


  32. I’d use lifespans for a time constraint.
    99.9% disastrous with BAU continuing until my grandchildren’s grandchildren are mature.

  33. >= 95% for disastrous outcome. Consider food safety with 9B population and increased crop failures due to drought/heat.

  34. I depends upon the time frame.
    By 2050 50%.
    I do not believe that business as usual will continue past this date because the evidence will be overwhelming and even Lindsen, Mockton & Co. will have shut up or be ignored.

  35. 80%

    that within the lifetime of my children millions and possibly billions of people will be starving or killed in conflicts fueled by climate change and that my children’s lives will be, at best, signficantly less comfrotable than my own.

  36. 100%
    It’s already happening in some areas, but like a cancer it will spread.

  37. Disaster — 99.5%.

    Even if benefits were somehow remotely comparable to costs (2% chance of that, maybe?), benefits will be concentrated among the rich, while harms will be concentrated among the poorest and most vulnerable. So disaster is highly likely.

    Enough disruption and direct harm to reverse the overall trend to greater wealth and prosperity — 10% — 70% (highly uncertain) — say 40% over the next few centuries.

    Collapse of civilization, barbarism, going a long way backwards in terms of knowledge and human adaptive capacity even as impacts continue for centuries — 5%.

    About the first estimate I (and most economists who have studied the issue) are very confident, and it’s easy to make the case for mitigation based on that alone. The other estimates are highly uncertain, and many analyses ignore them, foolishly in my view.

  38. OK, risk analysis is my day job, so I have to show my work.

    First, define BAU–Laissez-faire wrt increase of energy demand and the sources to supply that demand. No concerted effort to develop and deploy renewable energy sources. No concerted effort to increase third-world development and thereby decrease birth rates.

    Under these conditions, global population crests around 10 billion people sometime before 2050. CO2 continues to increase by 2.5- 3 ppmv/yr. to roughly 500 ppmv over the same period. By 2060-2070, we are looking at a full doubling over preindustrial levels. There is more than enough coal, tar sand, oil shale… to do this.

    Probability of a 20-30% reduction in GDP–90-95%
    Probability of >50% reduction in GDP–~70%
    Probability of catastrophic disruption of global civilization…30%
    Probability of extinction of human species within 3-4 centuries due to climate change~0.

    • “Probability of a 20-30% reduction in GDP–90-95%….”

      Not sure by when you expect this. End of century or 2060-70? or? You mean global GDP presumably, and this would be percentage drop from today?

      • Catastrophes do not occur by clockwork. Rather, the conditions to cause catastrophic failure come into existence, and then the system responds to a shock–perhaps even a minor one–by a cascading sequence of failures. You can see this in the Mortgage crash of 2008 or the Sovereign Debt crash of 2010 or the Collumbia and Challenger disasters,…

        I estimate that the conditions for such a global economic catastrophe will likely be extant by 2050. We may have some respite if the predictions of a lull in solar activity pan out and last a few decades. I would say 2060-2070 is a good bet, but I would be very surprised (were I alive) if it didn’t happen before 2100.

        There are lots of wildcards in such a prediction. China could precede the rest of the world in ecological collapse. This might slow growth in fossil fuels and draw out the scenario. Or Africa could take off economically, increasing the demand for fossil fuels.

        One thing people often neglect in these scenarios is the fact that susceptibility to catastrophic failure is a matter of degree. In a mildly destabilized system, a large (and therefore rare) perturbation would be needed to bring about collapse. It may even be that we can reassert control before the worst of the damage is realized. However, in a system far out over the edge, even a small disruption will cause collapse, and it will likely run its course until all the damage is done. We need to remember that even if we cannot stop climate change, we can certainly make it worse by doing nothing.

  39. Large-scale extinctions (other than human extinction): 99%

    Civilization collapse, defined as returning mostly to pre-industrial population and civilization type, probably due to grain production collapse via precip. extremes/temperature extremes:
    ……By 2050: 10%
    ……By 2075: 30%
    ……By 2100: 60%

    Human extinction: 1%. A few humans are likely to survive pretty much any catastrophe short of a K-T -style impact.

    I’m very much looking forward to your analysis of these stats.

  40. Tony O'Brien

    Depends upon your definition of disastrous. So probability of wars, collapse of our economic system and with it civilization as we know it and massive loss of human life (in the billions).

    By 2050 10%
    By 2100 80%
    By 2150 95%

    It will not be climate change alone, but our responses to the changes caused by climate change. What rises out of the ashes may even be better than what we have now, but the transition will be a horror beyond imagining.

  41. nanny_govt_sucks

    Your keywords are “global warming” and “disastrous” climate change. Since history has shown that warming is beneficial, my estimate is 0.

    If you had said will man-made climate change lead to disastrous climate change, then I would up my estimate since the possibility of global cooling is not excluded.

    • [edit]

      [Response: This post is not about refuting peoples’ claims, it’s just about giving one’s estimate, whatever the reasons or rationale.]

  42. I think that 95% is a good ballpark figure at present — also think that it’s destined to approach 100% asymptotically in the coming decades…

  43. By 2025 or perhaps even sooner, with the arctic ice free in summer and Greenland starting to melt in a runaway fashion, it will be evident to all but the hardcore deniers that we are in deep doo doo due to BAU. But it will be too late to do any useful mitigation by then, so I predict, much like our esteemed BPL does (and apologies in advance if I am putting words into your mouth, BPL), that there will be global catastrophic crop failures by 2035. This will inevitably lead to a worldwide war for resources with climate refugees caught up in the middle of it all.

    Epsilon: 97%

    And my family and friends wonder why I am depressed :-\ On the bright side of things, I expect more people to believe in evolution by then.

    (sorry, new handle just for this post)

  44. 25% (But that’s still enough to be a climate change hawk)

  45. 99.9%.
    Inundating the lowlands of Bangla Desh and China will be a disaster, just for starters. Then there’s the disruption to agriculture in already marginal parts of, for example, Africa. Think about the consequences of Russia’s recent decision to halt wheat exports. Tiny perturbations to the current climate cause millions of deaths, so how could such major changes not be called a disaster? And they are already starting to happen.

    The residual 0.01 represents the possibility of a major asteroid strike before the worst of it.

    Tamino: Why do you think there’s a 5% chance that if we continue with business-as-usual, man-made global warming will not lead to disastrous climate change? Can you suggest how such an outcome might eventuate? I mean, let your imagination run riot. Think outside the square.

  46. 99.9%

  47. Gavin's Pussycat

    It will be disastrous for the victims, with 100% certainty.

    What will happen, and we should be clear about this, is that the climatic disruption will be only the start of the disaster, the trigger; it will hit certain groups of people hardest, people that cannot defend themselves and lack resources to adapt.

    Consider sea-level rise: yes, it is possible, and even somewhat affordable, to protect oneself against even a couple of metres of rise. This is undoubtedly what Manhattan, Greater London etc. will do. But elsewhere, who will pay for it? Seeing how splendidly mankind manages inter-generational solidarity, what hope should we hold for the “inter-elevational” kind?

    So, money won’t be there, and sea defences won’t be up to snuff, and inevitably, Katrina-like disasters will start to unfold. Those that can afford to will move out, those that cannot are stuck.

    My number would be 100%, but not for climate change alone; for BAU amplified by human nature, and worst case, by weapons technology. This cannot be fully understood without systems thinking, and a fair bit of cynicism.

  48. Warmer surface: Epsilon = 100%
    Severely altered ecosystems: Epsilon = 95%
    Human civilization suffering severe crisis (migration and suffering in areas not prone to adaptation): 85%
    Civilization’s General Collapse: 5%

  49. BAU= 100,000,000 premature deaths attributable to direct effects of weather, disease, starvation, migration, flooding or draught- in space of 5 years or less over the next 60 years- mostly in sub-Saharan Africa, South America and sections of Asia.= my definition of catastrophic.


    Benchmark: Common understanding of number of people living with HIV/AIDS globally

  50. Sock Puppet of the Great Satan

    Famines, dislocated human populations: 99.9%
    My risk estimates for BAU:

    Greater than 10% GDP reduction as result of climate change: 90%
    Mass extinction of non-human species (particularly top predators): 85%
    Significant increase in localized conflicts: 75%
    Major regional or worldwide conflict, probably involving nuclear weapons, as direct result of climate change: 35%
    Collapse of civilization to pre-industrial technology levels: 25%
    Human extinction by 2,500 AD: 5%*

  51. Defining “disastrous” as meaning “there will be some events called disasters” (floods; Katrinas, whatever) makes the question pointless. So defining disaster as “significant (> 25%) climate-caused reductions in human population by 2100” I think the chance of disaster is < 25%. It might be substantially less but that feels like a safe-ish upper bound.

    • Pointless because it has already happened?

      • Clearly, Katrina has happened. As have the Pakistan floods, the Thailand floods, etc. And similar or wosre will happen in future. They are “disasters” in the usual sense of the word, but not in the sense we’re looking for. We’re looking for, oh I don’t know, climate-related effects of the scale of the Japan earthquake, happening somewhere in the world about once a month.

      • Thanks for elaborating. Tamino did say it was up to each of us to define ‘disaster’–but if there’s more discussion on that topic needed or wanted, I’d continue it on the Open Thread.

    • The definition of disaster here is important. I put it as somewhat higher, but nowhere near the 90% range that most others seem to place it.

  52. David B. Benson

    The question is not well defined as there appears to be no upper limit on the tme interval. NOr is it easy to separate disasters due to climate from disasters due to overpopulation and overexploitation of resources. But if (some of) the flooding of the past two years can be attributed to global warming, then epsilon = 1 as the disasters have manifestly already occurred.

    • Yes, its very interesting and thought provoking reading the responses, but it might have been better with a defined disaster and a defined time scale. For instance: global population decline of greater than 30% by 2100? Someone refers to an event on the scale of the Black Death. That probably killed 50% of the European population of the time.

  53. 2030 97.5 %

  54. worry, be happy

    I would define the ultimate disastrous climate change for Earth as moving out of the Goldilocks Zone because of BAU.

    My somewhat educated guess for that to happen with BAU is 1/400.
    My random guess (because my educated guess is based on incomplete knowledge) for that same future scenario is 1/4.

    The time period is defined by one of future events:
    a) AGW disaster
    b) some natural disaster (if nothing else, then due to the evolution of our Sun)
    c) some human caused technological disaster or event (for example technological singularity) which prevents an AGW disaster

  55. For “disastrous” I will assume 1930s Soviet and WWII-scale disruption: 98%
    Civilization collapse: 10%
    Extinction events definitions taken from Wikipedia article:
    Widespread (non-human) species extinction, minor: 99.9%
    Widespread (non-human) species extinction, major: 15%
    I think BAU can continue no later than 2050. The “Holy Crap” realization will occur sometime around 2030. Is 20 years too long to bring emissions down drastically? I dunno.

  56. Rick Gregory

    I think we will be shocked by 2 statistics:

    1 The incredibly high percentage of species that are able to either adapt or migrate to accommodate AGW, and
    2 The incredibly high percentage of species that become extinct.

    20% and 80%, respectively, with man just scraping by to fall into the first category would be my optimistic view.

  57. > What’s the chance that if we continue with business-as-usual,
    > man-made global warming will lead to disastrous climate change?
    > It isn’t zero. It isn’t one. What is epsilon?

    .985 +/- 0.05 within three standard generations of the G8 developed nations (being 3 x 30 years)

    • Disastrous – OED definition ‘great or sudden misfortune’ is inadequate for this. I prefer to use an near unequivocal attribute analysis (99% or .99) of the majority (>0.51 or >51%) of weather events that would be classed as ‘extreme’, as of the baseline year of 1987, being attributed to anthropogenic emissions.

  58. This is my estimation. Either we burn to much fossil fuel and the civilization collapse due to the severe reduction of the bioproductivity; either we will exhaust the fossils fuel before destroying the climate but the civilization will collapse in the process. Hence, using the maximum entropy principle the probability should be 50%.

  59. Pete Bondurant

    I’m an electrician so wouldn’t know shit. Tamino is a statistician – so what he said: 95%

  60. I think the chance that millions of people will die and trillions of dollars lost because of climate change under BAU (whether directly due to disasters or indirectly due to famine, wars, dislocation, etc.) is high — perhaps 95%.

    Not asked, but I think the chance of that actually happening because people are stupid enough to continue down that path for another two decades or more is relatively low. If you assess intelligence by reading comments on the Internet then you will get depressed, but if you assess it by talking to actual people in the real world then you’ll realise that most people are completely unaware of the “debate” that’s “raging” online and that the majority actually want their governments to act.

    Having said that, the chance that people will die who didn’t need to die and money will be lost that didn’t need to be lost is pretty much certain, because while we’re not stupid enough to burn all the fossil fuels in the fact of mounting evidence, we weren’t smart enough to start down the optimal path — I believe that point was passed a while ago. The longer we wait, the less optimal the result will be, but at every point in time there will be worse outcomes that can be averted by acting, at least for the foreseeable future.

    I’m ignoring potential tipping points that take the matter out of our hands. If we lose the ability to contain the situation then all bets are off.

  61. I’ll go with 100-(1/infinity)%, since from paleoclimate results we know the sorts of things that will happen. I’d have a different answer if there were some constraint on continued BAU, since realistic ways out of the problem involve some such constraint.

  62. Tony O'Brien

    I see that while I am more pessimistic than others on the climate, the bigger difference is on how we will respond. And in truth I would be much more optimistic if rapid strong action was taken, not just because of the climatic considerations but because it would imply a level of cooperation not currently existing.

  63. Extremely complicated situation — more complicated than any other 100 year period in human history. e = 95% — but warming will be only one of many anomalous (base = 20th c.) developing conditions at the heart of the human and environmental disasters.

    2112 will look very different (but the e for Neil Peart’s prophecy is .0001%).

  64. I’m going to take “disastrous climate change” to mean something analogous to “mission failure”: the very large decline in quality of life due to climate change. To keep it simple, I’ll define this as a reduction in mean life span by more than 10 years from current levels (Russia’s decline in the early 1990s was about 5 years). Also, I’m assuming “business as usual” means something similar to the IPCC A2 scenario.

    First answer, “I don’t know”, but if I found myself compelled to answer by a charismatic Nobel Laureate, I’d say …

    Epsilon for significant regions (representing a total population of 100 million or more) is probably more than 90%. Epsilon for the most developed western economies is probably less than 10%.

    For my part of the world (Canadian Prairies) I’d put epsilon at less than 1%.

  65. It’s not a matter of if, it’s a matter of when.

    BAU with fossil fuel RESOURCES (not just current proven reserves) we can continue on our rate of current consumption of oil and gas O(100) years, coal O(1000) years. More specifically oil O(70) years, gas O(140) years, and coal O(3000) years.

    Data from BGR, IEA, EIA, Shell, ExxonMobil, BP, WEC, USGS, WCA, and EWG.

    It’s immediately obvious, that continued use of fossil fuels, until we run out, leads us to coal as our predominate fossil fuel, in that scenario, summing all fossil fuel types (current production/consumption) and summing all fossil fuel types (total fossil fuel resources), gives O(1600) years.

    I am now working on a rough estimate of total atmospheric GHG concentrations, using the current fraction left in our atmosphere of from one third to one half. But with BAU (continued burning of all reachable fossil fuels at the current ratios (or short term projections out to 2035/2050) to total energy production from all sources. I have not included the release of methane clathrates stored in our continental shelves and other GHG’s stored in permafrost. If one were to include these, then in all likelyhood we would have to seriously consider a PETM-like extinction events (basically 2X global temperature changes at O(10) years to O(100) years.

    My SWAG on total atmospheric GHG concentrations would be 4X < GHG < 8X (pre industrial atmospheric CO2 of say 280 ppmv, yields 1120 ppmv < atmospheric CO2 < 2240 ppmv).

    Under a long term BAU O(1000) years, it's virtually certain to be above a Six Sigma (99.99966%) process.

    Sea Level Rise (SLR) is a virtual given with BAU, O(100) of say one meter, O(1000) of say — 20 meters < SLR < 70 meters.

    So my paper guess would be 99.99966% < p < 100%, with p very much closer to the upper limit of certainty. Call this one 666 Sigma certainty.

    So to use the 1e-5 single flight chance combined with 1e10 hourly flights, suggests 1e5 failures on average, yields an average failure rate of 11.4 years, or O(10) for the return period (meaning that something quite climatic will occur with certainty at decadal timescales under current BAU practices).

    Like I stated at the top, it's not a matter of if, it's a matter of when (assuming current/continued BAU practices).

    We should also consider NASA's vast resources placed behind each and every Shuttle launch (hundreds of millions of dollers) versus the resources placed behind driving your favorite mode of transportation down the road (virtually nil with respect to CO2 tailpipe emissions).

    • Try as I might, I can find no major area of disagreement with this.

      The sole area I might add is the impact of human affairs once things start getting bad. As drought worsens, food resources will become scarce. Coupled with potable water limitations due to flood contamination, I put theater-scale nuclear exchanges in the Mideast and Indian subcontinent at 40%.

      By the time the decadal doublings of ice sheet degradations kick fully in (GIS & WAIS) later this century no one will particularly care (90%).

  66. Michael Hauber

    1% chance of disastrous
    90% chance of regrets

    • Janne Tuukkanen

      1% sounds small, but ‘small’ is relative to the loss, if that’s total.

      Had you a can of 100 pills, where one were deadly poison and 99 placebo. Would you take a random pill for 1000 dollars? For a million? Or, as we are discussing about decades to come, would you hand that can to your son or daughter and say: “Take one dear, the epsilon is almost zero.”

      I take disastrous as something compared to full scale nuclear exchange during the height of the stockpiles, and put epsilon to 15% — it’s close enough to Russian roulette.

      For regrets, 99.9%.

      • Rob Honeycutt

        Nice analogy. I’d suggest it’s a game of Russian roulette with a gun with 20 chambers and 19 of the chambers have bullets.

  67. Just curious for “GDP reduction” estimates : do they apply to the value of the future GDP compared to the current one ? (net loss ?) or do they apply to the reduction of GDP actualized by the assumed economic growth ? methinks that in all “BAU” scenarios, the economic growth is about + 2% /yr, so doubling the GDP in approximately 35 years (100*ln(2)/2 ). Barring any climate effect, the GDP would grow by a factor about 3 in 2050 and 8 in 2100. In any case, it seems unlikely that CO2 production will continue if civilization collapses, so a side question would be, to make things more quantitative : what would be for you the “tipping point” when GDP starts to decrease, and what would be the CO2 concentration, the temperature, and the GDP (compared to the current one) at this point ?

    My own estimates come from the following computation : models based on the known proved reserves give a fossil peak between 2020 and 2030 (rather soon actually), and I don’t expect CO2 having risen more than 20 and 30 ppm , and temperature more than 0.2 to 0.3 °C , at this point. Thus the decrease of civilization is much more likely to happen because of the depletion of FF than because of the CC.

  68. Warren Pattinson

    Tamino you asked what is epsilon. My simple answer is 0.15, meaning the probability of disastrous consequences (parameters undefined but not good for homo sapiens generally) is 1 – e = 0,85 ie. 85%.

  69. 99.5 % that in the next 50 years, disastrous weather related events will occur that result in crop failures, massive human migrations, loss of property, infrastructure and other capital, civil unrest and wars, that would not have happened were it not for the effects of generally warming global climate.

    Why a 0.5% chance that it won’t happen? Some overlooked or underappreciated bit of physics that allows a hiccup to change ocean circulation, causing more heat to be moved into the deep ocean in a way that rebalances the TOA energy budget and temporarily saves our butts. Or something. And I pulled that number out of clear air.

    If you’ve concluded that disastrous events are already occurring due to climate change, evidenced by increased frequencies of these events, and coupled with large populations reliant on everything going just right, then the question seems moot. It’s like asking the crew of STS-51-L, at the moment the wind shear broke the temporary aluminum oxide seal that had plugged the initial burnthrough, and the plume started burning into the hydrogen tank, whether the flight might continue as planned.

  70. Absent any viable of definition of “disaster,” and being reasonably good at putting myself in other people’s shoes, I’ll offer 5% from the perspective of a resident of Fukushima.

    From the perspective of someone who would suffer disastrously if the sea level rose half a meter, I’ll say 99.9%.

    “Disaster” is far too relative a concept for this question to be meaningful.


  71. What is “disastrous climate change”? If you mean an increase in the number of disasters and look in the scientific literature for, for example, projected changes in the number of tropical cyclones, I think many of the commenters here would be surprised.

    Likewise, for “large-scale ecosystem collapse” we don’t need climate change: think of industrial fishing or deforestation in tropical countries. The recent (ongoing) food crisis in the Horn of Africa has as much to do with religious conflict and politics as with drought. In the real world you have population growth, economical and technological development and political systems, all of which determine whether or not a particular event will eventuate into a disaster. One cannot look at only one factor in isolation and say that climate change, and climate change alone, is to blame.

    And in addition to the issue of time horizon mentioned by others, there is also geography. Western countries almost certainly will be able to adapt to BAU climate change this century, but of course this will come at a price. Developing countries might as well, if we share resources and knowledge.

    Objections to the question aside, I would put the number at no more than 20%. Which, from a risk management point of view is still unacceptably high.

  72. Epsilon is One.
    That is if we continue with business as usual then the probability of disastrous climate change is zero.


  73. I’ll take BAU to mean no action until it is obvious to blind freddy that we have a big problem. There will be action then, but I don’t know what it will be.

    With that definition, I think there is about a 50% chance of severe dislocation this century.

  74. I’m going to frame my answer for epsilon with reference to perspectives of time. My interpretation of ‘disaster’ covers both direct climate effects of human emissions, and consequent climatological synergy with other human activities. For ‘business as usual’, I am simply going to assume that humans carry on with their current (ineffective) rate of response until system limits impose themselves on any further human activity.

    By 2050:

    Serious ecosystem(s) damage: 0.95
    Extinction of species: 1.00
    Extinction at level of ‘Great Extinction’ events: 0.01
    Serious and essentially global economic depression: 0.95
    Disintegration of global Western-style societies: 0.20
    Disintegration of global human societies: 0.10
    Extinction of humans: 0.00

    By 2100:

    Serious ecosystem(s) damage: 0.99
    Extinction of species: 1.00
    Extinction at level of ‘Great Extinction’ events: 0.10
    Serious and essentially global economic depression: 0.99
    Disintegration of global-scale Western-style societies: 0.50
    Disintegration of global-scale human societies: 0.25
    Extinction of humans: 0.00

    By 2500:

    Serious ecosystem(s) damage: 1.00
    Extinction of species: 1.00
    Extinction at level of ‘Great Extinction’ events: 0.50
    Serious and essentially global economic depression: 1.00
    Disintegration of global-scale Western-style societies: 1.00
    Disintegration of global-scale human societies: 0.95
    Extinction of humans: 0.05 (estimated before I read SPotGS’s post!)

    By 3000:

    Serious ecosystem(s) damage: 1.00
    Extinction of species: 1.00
    Extinction at level of ‘Great Extinction’ events: 0.95
    Serious and essentially global economic depression: 1.00
    Disintegration of global-scale Western-style societies: 1.00
    Disintegration of global-scale human societies: 0.99
    Extinction of humans: 0.10

    Note: although I am pessimistic about the continuance of global-level societies, I am not discounting that they might continue on a regional level. But to put it into context, my epsilon for disintegration of even regional Western-style societies by 3000 is 0.75. For regional human society of any sort beyond Iron-Age technological level, by 3000, epsilon = 0.50.

    Unlike others I consider that humans are at some risk of extinction, for two reasons:

    1) most humans these days rely on complex economic systems for sustenance, and have lost the cultural knowledge of subsistence (where ‘subsistence’ = completely independent of modern, organised industrial input). Loss of such complex social structure is likely to have spill-over effects on ostensibly more resilient societies.
    2) societies that do retain viable subsistence skills are likely to be vulnerable to the compounding global effects beyond their control, possibly including misguided attempts by future humans to manage global human numbers.

    The only thing likely to alter my estimations of epsilon would be if that last possibility – that groups or governments may attempt to control global human numbers – actually occurs before 2050.

  75. Disastrous climate change 100%
    Reduction of the human race by 50% by 2100 70%
    Species extinction directly attributable to climate change: 99.9%
    Human extinction: 2%
    Civilization collapse to pre-1900 living standards: 60%

  76. As a farmer who knows what is going on in the world food production system I am going for 99.99%

  77. LazyTeenager


    You have not defined what disastrous is.
    Extinction of the human race?
    Societal collapse leading to mass starvation and war and reversion eventually to a middle ages level economy.
    Reversion to a steam train level economy.
    Great Depression version 2 with eventual recovery based on sustainable fuels.
    Collapse of the USA to minor power status with China advancing to conquer the solar system?

    I am going with the steam punk option at 60% and lesser amounts for the other outcomes.

  78. 70% disaster
    95% “what were those idiots thinking???”
    100% “Al Gore’s fault”

  79. Probability of disastrous outcome due to climate change is somewhere around 6% +/- 5% (>25% of organisms destroyed)
    Probability of impacts on population centers: 10% +/- 8%
    Probability of non-disastrous changes in ways of life: 25%

    I am reading the replies here and got confused. For the longest time I thought you were asking for epsilon, where guesses of 95 would mean that there’s a 5% chance of disastrous impacts. Some here even made the mistake of saying an epsilon of 95% and then said the disasters were already occurring.

    • Here’s what Tamino said:

      “I’m defining epsilon as the probability of disastrous climate change, so its complement is the chance we won’t suffer disastrous climate change.”

      What I said (in paraphrase):

      ‘I think there’s a greater than 95% chance that some disasters which have already occurred are down to AGW.’

  80. I estimate the chances of disruption to our civilisation and welfare close to 100%. If not directly harmed by the consequences of AGW, we will probably be confronted with restrictions of our freedoms and rationing of resources.

    Perhaps not quite within the rules for answering the question posed… won’t GDP rise as we have more and more disasters, from inflation due to scarce resources and the need to rebuild/relocate infrastructure? And won’t disasters actually be good for employment as more labour is required for rebuilding?

    Thinking of the oil spill last year… doesn’t it cost more to clean up a barrel of oil than it’s worth? Wasn’t the nett effect an increase in GDP?

  81. From the post: “The probability of success is 1 minus epsilon. What’s epsilon?”

    Therefore epsilon is probability of failure=catastrophe.

  82. “What’s the chance that if we continue with business-as-usual, man-made global warming will lead to disastrous climate change? It isn’t zero. It isn’t one. What is epsilon?”

    disastrous effects:

    20% average increase in intensities and occurrences of hurricanes and typhoons and tornadoes

    20% increase in the occurrence of severe flooding

    20% increase in pollution (e.g. ozone)

    10% increase in desertification

    The probability of all those effects occurring by the end of the 21st century is greater than 90%

    so epsilon = 90%

    Significantly Increased ocean acidification and significantly decreased arctic sea ice by century end are to me virtual certainties under business as usual conditions

  83. Interesting thread… let’s approach it from the perspective of a systems ecologist:
    Serious harm to biodiversity: 100% (already is)
    Loss of genetic and species richeness: 100% (already is)
    Undermines critical ecosystem services: 95% (almost certainly is)
    Negative effects on the material economy: 100% (already is)
    Benefits the material economy: 1% (highly unlikely, balancing costs and benefits)
    Greatly reduces the human ecological footprint: 98% (will do so in the future)
    Ultimately leads to human extinction: 10% (a possibility if critical ecological services are lost completely)

  84. Nominations for the 2012 Bloggies are open: Best Science or Technology Weblog (http://2012.bloggi.es/#technology). Tamino definitely deserves some noms.

  85. Probability of some disaster related to climate? I think >95%. I’m inclined to think the risk of disaster (economic or otherwise) related to sea level rise alone is around that ballpark, and that’s only one of several possible mechanisms by which BAU can seriously mess things up for our descendants.

    Re the talk of significant reductions of human population: I think the risk there is also very high. Not only due to famines & other climate-related disasters, but also due to war. The recent kerfuffle about scientists ‘censoring’ several papers on how to make bird flu far more deadly makes me think that in 20-30 years even fairly basic labs will have the ability to do this. Given this, I think it highly likely that some accident somewhere will result in a human-created super-flu being released into the wild, with tens to hundreds of millions of deaths as a result – the 1918 ‘Spanish Flu’ is estimated to have killed 3-6% of the population of the planet, and the ‘engineered bird flu’ is thought to be much deadlier, with mortality rates well into the double digits. So with a pop of (say) 8 billion, and a highly globalised & mobile world economy, I’d give better than even odds that we’ll see a pandemic with a toll close to or exceeding a billion.

  86. Assumptions:

    The End-Permian mass extinction event was the worst in Earth’s history and was caused by rift volcanism and coincident burning of coal (Earth’s greatest mass extinction caused by coal: study).

    Humans are changing atmospheric and ocean chemistry faster than at any time Earth’s history, releasing sequestered carbon some orders of magnitude faster than the Siberian trap eruption. (Rate of ocean acidification the fastest in 65 million years).

    Earthlike planets are capable of supporting multicellular life for a relatively small fraction of their existence (a href=”http://www.desdemonadespair.net/2011/06/ugo-bardi-man-vs-gaia.html”>Peter Ward).

    Humans have already caused irreversible changes to ocean chemistry that force a transition to anoxic oceans dominated by microbes: (Jeremy Jackson, fast forward to 36:49, “Three scenarios of global behavior and consequences”).


    Probability that humans have already destroyed the biosphere in which the species evolved: 100%.

    Probability that humans can survive the Under a Green Sky scenario in bio-domes for some millions of years, until biodiversity recovers: 0%.

    Probability of extinction of all complex multicellular life by human destruction of the biosphere: 90%.

    Probability of extinction of human species within a few centuries: arbitrarily close to 100%.

  87. Some additional thoughts.

    Suppost this same question were posed at WTFUWT?, what is the probability that their answers would be straight outta Bizzaro World, e. g. as in p (WTFUWT?) = 1 – p (Open Mind)?

    I think is it quite safe to say that p(WTFUWT?) ~ 0.0%.

    (As an aside, it has become quite obvious that denier sites like WTFUWT? and now Dr. Curry’s engage in what I’ll call “sling anything up on the wall and hope that some of it sticks” mentality, devoid of any reason, logic, or physical rigor. That these denier sites far outweigh the climate science literature and real climate science blogs by one to two orders of magnitude (readership and article posts). And that this will only get worse as we approach the IPCC’s AR5. In this real world scenario, I am extremely doubtful that humankind will do anything meaningful in the near-to-intermediate term (meaning before mid-century).)

    On a geological timescale, fossil fuel carbon deposits were stored and created over a time span of approximately 200Myr (coal being the most lengthy part of this timespan, at approximately 170Myr), we are currently using up those carbon stores at a rate of almost a million times (O(1,000,000)) their geologic storage rate.

    I would like to see a whole chapter of the IPCC WG1 AR5 (or perhaps better placed into WG2 or WG3) devoted specifically to this question, the BAU practice carried out long term (O(100) to O(1000)) in terms of risk assessment. Without getting all locked down on economic details and other minutia, simply stated;

    Does Earth’s biosphere benefit, in any way, from continued long term human BAU fossil fuel practices?

    The only possible benefit that I can think of is delaying (or eliminating) the onset of the next glacial period of our current ice age. If the human population is anywhere’s near its current (or expeced 2050) population.

    Finally, being a gambler (and not a very good one at that, considering my losses to date), it is extremely foolish to bet against the house long term. Short of counting cards, and not being caught, the house always has the long term advantage (legally usually set between 1:0.8 to 1:0.98 (house:gambler with 1:1 being an even money chance, or p = 0.5, you need p > 0.5 to win long term (i. e. 1:1+))). You simply can’t collectively beat the house long term (a few people will win big enough to retire, if they wish to do so, at the cost of everyone else (plus operational overhead costs for the house).

    In this case, the house is Mother Earth, currently occupied by approximately 7,000,000,000 (and counting) GHG gamblers. IMHO there is only one possible outcome under BAU practices, humanity loses big time. But then again, estimates of the onset of the next glacial period are somewhere on the order of 10,000 years (O(10000) absent anthropogenic forcings).

    So I’ll admit it, I am a gambling addict, just like humanity is a fossil fuel addict. Once we all face that one basic fact, that we collectively, cannot possibly beat Mother Earth long term, perhaps then, and only then, will we possibly face up to our collective fossil fuel addiction, and collectively go into rehab.

    [Response: Although this isn’t an attempt to argue with someone about his/her expressed opinion, it’s a bit outside the intended purpose of this thread. Therefore I’ve started an open thread. All readers: please use the open thread for discussion (including arguing with people about their reasons) and keep this one just to vote for your estimate of the probability of disastrous climate change under business-as-usual.]

  88. Tamino,

    If possible, please move my last post above to the open thread (with your Response included).

    Thanks for letting it through anyways.

    [Response: Unfortunately I don’t know how (if it’s even possible).]

  89. I assume “disasterous” climate change means extinction of most of mankind and habitat, like if we suddenly got an ice age tomorrow morning (lets pray we can affect the climate to avoid that when it comes eh?)/

    My guess is 1%

  90. Probability that the climate will change to a state that mankind has never experienced in written history by 2150: 97.4%.

  91. I disagree with you that it isn’t 1.
    100% probability that we’ll see disasterous climate change.

  92. Probability of disasterous climate change by 2100 assuming business as usual: 5%

    After that, who knows.

  93. Disaster criteria: No more Santa’s workshop at the north pole.

    Probability: 50% by 2032.

  94. 90% (Disastrous meaning the death of at least 1bn people by 2050).

  95. Why are 0 and 1 excluded? BAU in its current incarnation implies constantly accelerating CO2 emissions. If we’re excluding 1, then I choose 1 – 1/Inf as my chance of catastrophic climate change.

    • To exclude the certainties. Since there are some non-zero chances that something will interfere with catastrophic climate change, such as being bashed to extinction by 10km asteroid next year might be only 1/65e6, 1/Inf is unreasonably small.

      [Response: Please confine critiques of others’ estimates to the new open thread.]

  96. Greg Simpson

    I’d say 10%, but I have pretty high standards for what constitutes a disaster. The drought in Texas is not even close. 10% is scary.

    • The images in the media that came through of the floods in Pakistan in 2011 were terrible, but they were an improvement on 2010. Two consecutive years with agriculture wiped out in a densely populated, nuclear armed country internally besieged by militant religious fanatics.

      If that isn’t a disaster, I really don’t understand what “disaster” means for people. a bit of the 6th avenue el in the top of the head, eh?

  97. I’d say that there’s a nearly 100% chance of a billion excess deaths in the developing world due to climate change by 2100 if we change nothing. I’d say 50-50% that the human population sees so many excess deaths (triggered predominantly by climate-change-linked disasters) that by 2100 that population growth stops.

    It is hazier to me what the impact will be to the industrialized world that will be better able to adapt, but the political strife will be a huge negative effect.

  98. The devil is in the details :) (also deniers might object that any value that is not precisely zero is still ‘looking bad’ for them)

    Anyway – probably something like 20% to have 1e5-1e6 lives catastrophically affected by 2050, but that spikes to 99.99% by 2100, and with more than 50% to have 1e6-1e9 lives catastrophically affected by 2100, and that range itself could go to 90% by 2200 :(

  99. 1% chance of disaster, with disaster defined as climate having a substantial negative effect on agriculture, accounting for:

    – The possibility that climate change could be good for agriculture
    – The robustness of our current agricultural systems

    Meanwhile, the chance of UN-led trade agreements climate agreements leading to agricultural disaster is probably close to 20%, and in fact “green” policies such as corn ethanol subsidies have already caused mass social unrest in many parts of the world.

  100. -Probability of greatly increased frequency of weather disasters leading to widespread suffering: floods droughts, famines 95%
    -Probability of collapse of civilization as we know it (ie, democracy, capitalism, improving quality of life for most people, the trend in reductions in violent deaths) 40%
    -Probability of the end of humans over the next few centuries 0%
    And an answer to a question not asked:

    -Probability of us taking sufficient political action to bend the BAU trajectory in time to avoid the worst disastrous outcomes, about 20% and shrinking with every passing day.

  101. Sock Puppet of the Great Satan


    [Response: I’ve said repeatedly that this thread is not for commenting on peoples’ answers. I have created an open thread for that purpose. Commentary on the opinions expressed here must be placed there.]

  102. Disastrous being:
    o >10% of ecosystems flipping by 2100
    o changes causing > 100M people to migrate by 2100
    o changes resulting in enough famines often enough to alone (not counting resource conflicts) reduce human population by 1% by 2100
    o enumerated increases in resource conflicts resulting in loss of human life
    o measurable reductions in GDP and GNH by at least 1%
    o measurable increases by 5% in disasters over US$1B in damage :

    85-95% based on what we know today.

  103. ‘Disastrous’ in my mind = water shortages for millions and wars for water, agricultural hardship leading to famine in low-subsistence areas and astronomically high food prices in others, the displacement of tens of millions of people over the next two centuries, and the economic gutting of many/most countries trying to adapt to these and other, less obvious or direct consequences.

    Probability –> 85%

  104. I consider that epsilon defined as the chances of disastrous climate change unless GHGs are reduced dramatically as greater than 97% + or – 2% but increasing the more we delay in taking effective action.

    Feynman wrote about his Challenger investigation in a Chapter ‘Mr Feynman Goes to Washington: Investigating the Space Shuttle Challenger Disaster’ in ‘What do You Care What Other People Think?’, and a worthwhile read it is too.

  105. -Probability of droughts, famine and natural disaster in the next 40 years? 100%
    -Probability of those same things even without climate change? 100%
    -Probability that there will be a natural growth and use of technology so that predictions going out to 2200 or, even sillier, 3000 are moot? 100%
    -Probability of climate change caused “disaster” by 2050? 10-15%
    -Probability that some people will still question whether any disaster which does occur is unnatural, even in 2050? 100%

  106. Practically, I think it’s 1, a full 100% certain. But I make a dunk for 1% mercy.
    So 99%

  107. Probability of +100M deaths $1-10^9$


  108. 99% by 2100.

  109. For the business as usual emissions scenario and the expected 3°C to 5°C of global warming, think more in terms of the Great Depression plus the Black Death plus the Mongol Invasion extended over a century or longer.

    Great Depression because a significant part of the world’s economic output will be absorbed dealing with the costs of repairing and mitigating against extreme weather, sea level rise and changing climate conditions instead of being used for higher living standards.

    Black Death because hunger and starvation will result from failing agriculture and tropical diseases will spread into populations that don’t have natural resistance while disease in general will increase among less healthy underfed people.

    Mongol Invasion because the populations of poorer nations that can’t cope with the adverse conditions of climate change at home will migrate (by force if necessary) to cooler and richer lands.

    Probability of all three occurring by 2100 – 85%

  110. John Ramming

    Time frame, 500 years out.
    Specific focus of “disaster”, Sea Level Rise.
    All of Greenland ice gone, 100%,
    West A. Ice gone, 100%
    thus 80′ sea level rise peak.
    chance that all Ice gone, 50%.
    By 2100, humanity working very hard to avoid 240′ SLR, 100%.
    Our imagination not able to see the problems before they hit us, 100%

  111. Susan Anderson

    I’m with the 99% (pun intended). This ponders many human consequences well:
    “Is this the final century of our civilization” aired, IIRC, June 2007
    I’m also baffled that people think 2100 is some kind of endpoint; we are looking at an accelerating continuum.
    Mankind’s survival, I agree, is likely at least for the next few centuries, at the very least in some vestigial locations.

  112. Assuming BAU includes the possibility of making sensible science-based energy policy decisions, I’d say the epsilon is less than 1, by a hair (let us define: hair = 0,5), assuming ‘disastrous’ is defined sufficiently small.(f.e. only the coastal cities and the capability of their surroundings to produce enough food for the dispersing populations is destroyed by 2150.(This assuming Hansens’ scenario of rapid GIS breakdown is mostly correct)

  113. oops, (hair = 0,005)

  114. With BAU considered a static condition Epsilon will be 100%.
    Even that might not necessarily lead to distinction of human existence but existence per se would need to be re-defined.

    In all other cases (E < 100%) you implement some kind of learning curve into BAU which yet needs to be established.

    Looking forward to the combined results from all this.
    But then, the question was a too easy one, wasn't it? :)


  115. Tamino, if we are approaching this as engineers, now it’s time for the next question: “What factor of safety should policy makers apply in decisionmaking regarding climate?”

    Determining epsilon is only one part of the real exercise, and epsilon’s uncertainty is wide. The most important thing is good advice to policy makers, from engineers, on what factor of safety should be regarded as ethical and should be applied. (And of course the reasonable FS may differ among various aspects of the climate problem.)

    [Response: That discussion should be carried out on the open thread.]

  116. 1%. If you have to prove it was from AGW -> 0% (that will be impossible!)

  117. Less than 1%. Climate change is slow. Humans are the most adaptable species on the planet. It’s hard for me to imagine any of the global human welfare metrics (life expectency, GDP per capita, etc.) diminishing due to climate change alone. There will be man made disasters in the future, but I see climate change itself as a relatively low risk item on a global scale. Not that there won’t be lots of individual disasters, but as Stalin would say, those are just statistics.

  118. BAU = CO2 emissions continue at least near today’s until 2080 (thus 700gtC more emitted which is about what has been emitted already).
    1 Billions dead through crop failure – 99%.
    2 Billions dead through collapsed world economy – 92%
    3 Billions dead through war of ‘survival’ – 87%
    4 Extinction of manknid 0%
    Not sure at which point in this progression that civilisation would be adjudged ended – likely it would be at step one.

  119. What’s the chance that if we continue with business-as-usual, man-made global warming will lead to disastrous climate change? It isn’t zero. It isn’t one. What is epsilon?



  120. Al-san,

    The 25% Black Plague death rate was in Europe, not global. Might have hit large regions of Asia as well, but we don’t know.

  121. MightyDrunken

    First I feel compelled to quantify some terms.
    Disastrous is having a significantly bad impact on society. Much greater costs on food, less security, war, wide spread pestilence, etc. Things that humans have survived in the past and will likely survive this time, but would rather not. These pressures will seriously threaten civilisation and maybe topple our current technological age.
    BaU i will take to mean the consumption of most fossil fuels including tar sands and shale deposits over the next 100 years or so.

    Therefore my epsilon is 97%.

  122. Horatio Algeranon

    “Epsilon” is nearly one
    If we burn oil as we’ve done
    And “epsilon” is “oopsolong”
    If we prolong this crazy run.

  123. caveat emptor

    BAU = we keep adding fossil carbon to the atmosphere as fast as we can with very little restriction or regulation
    Disaster = event or series of events clearly caused by climate change that kills at least on the order of 1% of human population (60-80 Million)

    epsilon by 2050 = 25%
    epsilon by 2100 = 75%

    I also think the the probability that we are able AND willing to keep BAU going after 2050 is low.

  124. The BAU scenario is to burn fossil carbon as fast as is economically feasible.

    Another disaster criteria: By 2200, any of (SLR>2m, ocean acidification, further fisheries collapse, a war-inducing drought or flood, stagnation of economic growth, more starvation than 20m/year, significant extinctions or other)

    Epsilon = 1-(.2 * .2 * .8 * .8 * .4 * .1 * .2* .5), or about than 9999 in a 10000

  125. Disastrous for a significant number of humans (and applying my own definition for the latter): 92.5%

  126. Phil Mattheis

    Given the relative late stage of this busy thread, the probability of making an apolitical statement is about 1%. I’ll align with Susan Anderson, at 99% for likelihood of human disaster to follow business as usual (including the chance that we are in early phase of an extended disaster-in-making). There is power in words well chosen and reinforced. Whatever else we call this decade, a (the) predominant meme is the 99%-1% dichotomy.

  127. I’ll define disasterous as >3K deaths from a single event that is generally conceded to be attributed to AGW within the next 30 years. My choice of the number of deaths is related to 9/11 given the magnitude of response to that event.


  128. If by “disastrous” we mean something like “at least 50% more extinctions (per century) than would have occurred with 100% action against AGW”, then I’d say epsilon is 100%. In other words, if we follow BAU (for another 50 years, say) then my best guess at epsilon is 100%.

    As to the probability that we follow BAU for another 50 years or so, well that is a whole new ball game.

  129. Xavier Onnasis

    adopting arch stanton’s definition of disastrous, but extended the time period to encompass the next 100 years, I’d guesstimate epsilon (Tamino’s definition) at 98%.

  130. Although the specific mention of epsilon by Feynman is in the the citation in a post above Feynman describes his findings WRT the Challenger shuttle disaster in another book of his, ‘The Pleasure of Finding Things Out’ in the Chapter, ‘Richard P. Feynman’s Minority Report to the Space Shuttle Challenger Inquiry’..

  131. I think it’s 1-P(ps)-P(mh/gt)-P($$) where:
    P(ps) is the probability that there is a huge political scam perpetrated by all climate scientists who just happen to be left wing crazies to a man/woman….about 1/1000000
    P(mh/gt) is the propability that all climate scientists are suffering from mass hysteria or group think……about 1/10000000
    P($$) is the probability that all mainstream climate scientists are paid such vast sums of money to do their research that they are deliberately lying……about 0

    So if my maths is any good, that makes it about 99.9999989%. I’m under absolutely no illusions, having followed the AGW debate for 25 years I have not seen a single significant finding suggest things aren’t a lot worse than any scientist is willing to predict.

  132. @RayLadbury
    “Probability of extinction of human species within 3-4 centuries due to climate change~0.”

    I think a possible scenario would be catastrophic nuclear war triggered by food/water/habitable land shortages caused by AGW. That could well equate to human extinction in a nuclear winter. Just a thought.

    [Response: Please confine comments on others’ estimates to the open thread]

  133. Disastrous?
    100% – Forced relocation of many 100MM people
    50% – 30% species loss
    100% – Many $trillions of damages.
    100% – > triple # of climate related fatalities/ year
    60% – millions of increased famine deaths
    0% – extinction of human species
    <1% – collapse of civilization. We will carry on.
    75% – collapse of West Antarctic Ice Sheet within 200 years
    100% – disappearance of Arctic summer sea ice. (Best guess, by 2030-35)

  134. Rob Honeycutt

    Can I throw in a spin on the definition of “business” as usual?

    1) FF industry gets is way, undermines all potential competition for other energy sources, and we put maximum carbon into the atmosphere.

    98% chance of full collapse of modern civilization. Human population reduced eventually to 10-20% of our max. All this occurring over the course of the next 200 years.

    2) FF industry has to compete fairly with new energy industry. Level playing field. Let market forces and energy innovation drive the process.

    50/50 chance of the same as above. (Still lousy odds.)

  135. Steve Milesworthy

    (Irony on).

    Disastrous in my definition means a disaster for me. I and most of my nearest and dearest live in England and live on high ground, so 5%.

  136. littlerobbergirl

    Steve- ‘london is burning but i – live by the river’
    you forgot the roving gangs of starving youths/military/both. i live on high ground too, but a damn sight too close to the capital.

    mass extinction (>50% species) 95%
    large arctic methane release and runaway warming of 6 or more degrees 90%
    ocean anoxia 50%
    collapse of western democracy and return of life expectancy to medieval levels 99.99%

    When the limits to growth World3 model is run now with the present as the starting point, it always ends in collapse. I only read this recently, and it scared the shite out of me. run for the hills folks, and teach your daughters how to shoot.

  137. Lewis Cleverdon

    By any humane civilized standard, a body-count due to all impacts of climate destabilization equal to the ~ ten million deaths ascribed to nazi ideology would seem a reasonable measure of the term ‘disastrous’.

    Continuing BAU past the point where time-lagged warming will empower the interactive feedbacks’ outputs to exceed the natural carbon sinks – and thereafter become ‘self fuelling’ – would make a disastrous outcome certain. (At least to the extent that certainty is respectable scientifically).

    Given the decades of continued warming – and the loss of the ‘sulphate parasol – to which we are already committed, I must therefore respectfully disagree with the statement that epsilon is neither nought nor one, and post the figure: 100%.


    Lewis Cleverdon

  138. Under BAU
    Disastrous for Humans Civilisation 95%
    Collapse of Civilisation 85%
    Humans driven back to a pre-medieval society 60%
    Humans driven back to the caves 35%
    Human Extinction 15%
    6th Mass Extinction 75%
    Runaway GH a la Venus 5%

    Under likely levels of CO2 reduction
    Disastrous for Humans Civilisation 75%
    Collapse of Civilisation 50%
    Humans driven back to a pre-medieval society 40%
    Humans driven back to the caves 15%
    Human Extinction 5%
    6th Mass Extinction 60%
    Runaway GH a la Venus 1%

    Under Emergency Response levels of CO2 reduction
    Disastrous for Humans Civilisation 50%
    Collapse of Civilisation 25%
    Humans driven back to a pre-medieval society 15%
    Humans driven back to the caves 5%
    Human Extinction 1%
    6th Mass Extinction 40%
    Runaway GH a la Venus 1%

  139. Man is the animal who adapts our environment to ourselves. This possibility IS zero, and I do not understand why, with virtually no evidence whatsoever, you are all so eager to believe in disaster.

  140. We face what I call the Four Horsement of the New Apocalypse:
    Global Warming
    Peak Oil
    Global Financial Crisis MkII (III etc)
    Chances of advanced Western-style civilisation collapsing due to these in combination by 2100: 99%.
    Chances of Global Warming doing it alone: 95%

  141. > Michael Moon
    > adapts … newtechtire website link
    Well, the link behind his name goes to a real product and it proves to be safe on the highway, he may have a point, though I doubt the assertion that the government will require everyone to buy only that kind of tires.