Scientists hash out the uncertainties of climate sensitivity


In a perfect world, we'd have all the answers and solve the problems of humanity sometime before dinner. Sadly, in real life, dinner is often delayed, or put off altogether in favor of a microwave noodle pot.

On those long, metaphorical evenings, we turn to science as way of narrowing down the number of things we don't know, and helping us push through problems as best we can. Climate science is a great example of this process in action. We know some big, important facts about how Earth's climate works—how adding extra greenhouse gases to the atmosphere causes the planet to get hotter, for instance. Other things are more uncertain, such as exactly how sensitive our climate is to heating by greenhouse gases.

In an upcoming paper in the Journal of Climate, Stephen Schwartz, Ph.D., senior scientist at the Brookhaven National Laboratory, and his colleagues contribute to the ramen-fueled work of inspecting the things we don't know, and offering suggestions for how to get closer to the truth.

Climate sensitivity is an extremely important concept. Climate change tells you what's happening—that the gaseous detritus of modern life is accumulating in the atmosphere and causing the global average temperature to tick upwards. Climate sensitivity, on the other hand, is the information that you need to know in order to make decisions about how best to deal with climate change—how we should alter our behavior, and when.

The data on climate change is pretty unequivocal. Climate sensitivity, however, we're a little more fuzzy on.

"The basic issue is that the things that control climate are complex. We aren't idiots just going around looking at CO2, CO2, CO2," said Gavin Schmidt, Ph.D., a climate modeler at NASA's Goddard Institute for Space Studies in New York, and one of the brains behind, a blog dedicated to explaining the intricacies of climate science.

"There are many different drivers of climate, including ozone and aerosols. We've only just started to relate these things to the policy choices that real policy makers are faced with," he said.

The effects of aerosols are one of the biggest sources of climate confusion. Tiny, submicroscopic particles suspended in the air, aerosols include things like dust, soot from burning furnaces and smog. Like greenhouse gases, they are produced both naturally, and by human activities. They've also been increasing since the Industrial Revolution.

The weird thing: Aerosols can both work to increase the global temperature—black carbon soot, for instance, traps heat the same way a black tar roof does—and also decrease it—other aerosols reflect sunlight away from Earth, cooling the planet's surface.

"There is good reason to think that aerosols are offsetting some of the warming that would otherwise have resulted from increases in greenhouse gases, but the amount of offset isn't well known," Stephen Schwartz said.

Without that information we don't have a clear picture of how sensitive our climate is to increases in greenhouse gases.

"If the aerosol cooling influence is a small fraction of the greenhouse gas influence then the observed warming is obtained with a rather low climate sensitivity. If the aerosol influence is offsetting a large fraction of the greenhouse gas influence, it implies a fairly high climate sensitivity," Schwartz said.

For you and me, that's the difference between having time to make changes that could limit climate change before we experience any major effects, and already being doomed to a seriously climate-altered future.

If the climate sensitivity is low—if the the global average temperature rises 1.5°C for every doubling of CO2 in the atmosphere—then we have about 80 years before we accumulate enough greenhouse gases to commit ourselves to an increase in the global average temperature of 2°C above pre-industrial levels. Many scientists and policy makers have agreed on 2°C as a cutoff point, based on the economic, environmental and societal impacts associated with that level of increase.

On the other hand, if climate sensitivity is high—say a 4.5°C increase for every doubling of CO2—then we already hit the point where the 2°C increase will be inevitable about 35 years ago.

The actual sensitivity isn't known. The IPCC says it's likely to be anywhere between 2° and 4.5°C, with 3°C as the most likely possibility.

Schwartz's paper is focused on increasing the accuracy in the way climate models address the unknown.

All climate models take aerosols into account, he and climate modeler Gavin Schmidt told me. But Schwartz thinks the models are going about it the wrong way. He says that most climate models match very well to observed, historical changes in climate, but that they do so with a wide range of climate sensitivities and aerosol forcing levels. Some of the models may be right for the right reasons, Schwartz said. But others are right for the wrong reasons—balancing out climate sensitivity and the impacts of climate forcings, like aerosols, in a way that doesn't match with what actually happened.

"They certainly can't all have the right sensitivity," he said.

His paper is meant to help narrow down the factors responsible for forcing the global temperature down while greenhouse gases force it up, and he thinks he's done that, pointing to aerosols as the primary perpetrator.

"If the community could constrain the total forcing, up and down, then we could constrain the modelers so they don't have all this latitude to get the right answer for the wrong reason," Schwartz said.

But the paper gives the impression that climate modelers don't realize how important aerosols are, which isn't true, Schmidt said.

"The models do have a range of sensitivities and a range of forcings. The pairs that match the real-world observed climate provide a set of plausible histories for what actually happened, but which pair is closer to reality remains to be seen," he said. "But there is nothing 'too wide' about the assumptions. They cover the ranges of the observational uncertainties."

As modelers around the world prepare for the next update to the IPCC report, due in 2014, they're already working on ways to better account for uncertainties and make the models more accurate, he said.

One method is out-of-sample tests—basically, taking a model that correctly "predicts" observed climate change in the 20th century, and seeing how well it does with other historical time periods. Out-of-sample tests can help weed out models that give right answers for the 20th century, but for wrong reasons.

Ultimately, to really get a grip on how aerosols influence climate, modelers need better observational data, Schmidt said. The best way to track aerosols is by satellite but, so far, no satellite has had the right sort of instruments. That could change late this year when Glory, a new satellite carrying a cutting-edge aerosol sensor, is set for launch.

"It may be that we never get good aerosol data for the 20th Century," Schmidt said. "But we may be better off in the 21st."

Image courtesy Flickr user John LeGear, via CC


  1. Honestly, the biggest problem I have with climate science is the lack of accurate predictions. It seems every time reality doesn’t agree with these models, they end up patching the model, but at the same time they want me to trust that this time the model is right.

    Without accurate predictions (and accurate being a fairly loose term), scientific theories are pretty useless.

    Frankly, it may just be an external perception thing and really the models are only being patched a little bit and the predictions from unpatched models are pretty close, but I haven’t seen a nice graph of global temperatures with a line for each version of each models’ prediction.

    1. That’s my impression too. To be frank, trusting the models doesn’t seem very scientific.

      Two questions I’m quite curious about is how much out-of-sample data is needed to invalidate a model, and how much deviation it takes. Sciences based on observation require some kind of falsifiability criteria.

      Such a criteria needs to account for the nature of the metrics used (world averages). It is obviously much easier to predict a single value over time, than detailed events (carries more information).
      Older sciences such as physics have passed this bar, as they can predict previously unexpected events with great accuracy. I’m not seeing this with climate science.

  2. More of this sort of information is exactly what is needed to back up a more intelligent dialogue. It’s not appropriate to draw a hard line between “climate change activists” and “climate change deniers” when there is actually a hierarchy of positions involved:

    (0) There is no significant warming of the earth.
    (1) The earth has gotten warmer in the last century.
    (2) There is clear evidence that this is due to human action rather than extra-human factors.
    (3) The rate of human-caused change is so great as to result in a crisis unless immediate action is taken.
    (4) The action which is required is the imposition of controls, on a global level, of every human activity producing a significant quantity of greenhouse gases.

    The discussion is not only about positions (0) and (4), you can get on the train but get off at an intermediate stop. The more validated information in particular about stops 3 and 4, the better off we’ll be.

  3. I remember reading a post some time ago about the (good) reasons we should trust the climate scientists even though we may not understand the science behind it ourselves… I mean yes, a majority of scientists agree that anthropogenic climate change is a reality, but this article explained why the average person should believe it despite the feeling (that so many of us seem to get) that they could be making it all up and your average person wouldn’t even be able to tell the difference… anybody else remember that one?

    NB: I understand that there are “obvious” reasons, but sometimes it just needs repeating in a “from first principles” kind of way, and I remember that this article did that…

    1. oohShiny, I don’t know if this is what you meant, but here’s a blog post and a video I already posted in a previous discussion, and which I find useful with regards to why we should ‘believe’ climate scientists (and why we should curtail emissions etc.):

      The blog post:

      The main focus of that blog post is actually something else, but the takeaway points for someone with your question about the believability of climate science are these:

      1.)Climate science is actually a mature science (as opposed to a load of unfounded guesswork) that is unfortunately difficult to explain to a lay audience. The very processes and concepts that make it mature are also what makes it so difficult to explain.

      2.)The very basic fact that CO2 can destabilise the climate system, and under certain circumstances may do so very quickly, is very well understood and pretty much undisputed. It is *not* part of the things that make climate science so difficult to communicate – rather, it’s in every undergraduate textbook.

      3.)In the light of the possibility of total climate destabilisation, the fact that we don’t understand the *exact* threshold conditions for such a destabilisation yet should not encourage us to continue ‘business as usual’. Basically, the less you understand about a really big danger, the more you should be working to *avoid* that danger, rather than blithely hoping for some as-yet-unknown factor that might influence things in your favour. That’s common sense, right?

      The video:

      The reasoning of the vid can be summed up somewhat like this: we are, as non-scientists, basically confronted with a risk assessment. So, what’s the worst that could happen if we choose a certain path of action?

      1.) If we do listen to, and act according to the scientists and activists who say we need to curb our emissions fast, the worst that could happen is probably another, more severe economic crash.

      2.) And what’s the best that could happen if we do listen to them? Well, we could save the climate and possibly end up with a reformed, ‘nicer’ economy in addition.

      3.) So, what’s the worst that could happen if we listen to all the people who think we should *not* act on climate change? A total crash of the ecosystems we depend on for our survival on a much more fundamental level than we depend on the economy.

      4.) Okay, so what’s the best that could happen if we listen to these guys? A continuation of the status quo, pretty much unchanged.

      Now, if you ask me, I think the potential outcomes we end up with if we do believe the scientists and do take action sound better than the ones for choosing the path of disbelief and inaction: the worst case result for taking global action on climate change – economic crash – is bad, but not as apocalyptic as the worst case for not doing so (and remember, the basic fact of the potential of CO2 to destabilise the climate *is* well established, even if certain details remain unknown). And the best case – a greener economy – is nicer than just a continuation of business as usual, too. So, here, too, it seems to me that it would be common sense to listen to the scientists rather than their opponents. If it *should* turn out that all of climate science was wrong, we’d still end up with a more benign result if we do act, than the result we would end up with if it should turn out that the *other* side was wrong after we had heeded *their* advice. Granted, it’s possible that we’d never end up with either of the worst cases – but would it really be sensible to gamble the existence of human civilisation, and of most life on the planet on that possibility? Especially as long as we don’t really understand enough to assess its likelihood?

      And also, I just don’t think it’s particularly likely that a whole branch of science could be so completely wrong about everything they’ve been studying for decades – certainly not about the very basics. And the basics is all you need to ‘believe in’ to be able to do the kind of risk assessment you need to do.

      Whoah… writing all this has taken entirely too big a chunk out of my evening. :-(

      (I feel a bit bad because it seems I’m only ever posting on climate and environmental stuff here, but, basically, I’m quite busy and those are among the very few topics I find important enough to weigh in on at least occasionally.)

      1. It looks like you are trying to apply decision analysis to the problem, which I think is commendable, but you need to consider the probability of the outcomes as well. If we put emission controls in place, it is argued that we are highly probable to suffer economic losses. (Let’s just take this at face value – I’d guess we are just as likely to create economic opportunity, but for this example we’ll ignore this.) The other decision is to not put in emission controls; for the sake of argument let’s say that there is only a small chance that we’ll do catastrophic damage. To figure out which path to take we need to know which is greater:

        significant economic loss times high probability
        catastrophic economic loss times low probability

        If the probability of the catastrophic economic loss is low enough, it doesn’t matter that it is catastrophic. If this is the case, then the deniers are making the correct decision. (according to decision analysis)
        To use decision analysis, we need to understand the probabilities and hopefully this is what climate science can provide. If the probability of catastrophic loss isn’t really that low, we have a case to put in emission controls.

  4. Climate prediction is a largely chaotic process. ‘Chaotic’ in the scientific sense doesn’t mean random and unpredictable though, it means that the system is very very sensitive to the initial conditions.
    If you have a big weather model an you run it, run it again with one of the windspeeds 0.01% out an you’ll get some seriously different weather predicted in a few days time. This is the reason for the limit of the 5-day forecast, and it similarly limits larger scales.
    The issue here is that if we don’t even know the initial conditions – what the climate is sensitive to and by how much – then the accuracy goes down the longer you run your model.

    It’s a knife edge; you need sensitivity to predict small changes, but sensitivity limits how far ahead you can predict.

    This doesn’t quite apply to the largest scale models but it does help explain why historical comparison is important, but can mislead. Your model might fit the last 60 years perfectly but run it a few years into the future and see it’s predictions fail. Accurate measurements are essential to getting these models right so you look at what you don’t know. Aerosols and ozone desperately need better vertical resolution data.

    1. Weather forecasting is very different from climate modelling. Climate is the statistics of weather and looks at large scale spatial and temporal features. Climate models don’t try to do the same things as weather forecast models. Weather forecasting does indeed become extremely difficult to forecast more than about a week ahead because of sensitive dependence on initial conditions. Climate models are different. The climate system doesn’t have the same short term dependencies. Rather, long term processes (greater than annual to decadal to millenial) become crucial. A climate model that can simulate the past 60 years would be doing so from initial conditions using observed forcing. It can only predict the future climate using predicted forcing and could thus only ever be as good as the predicted future forcing. Current climate models are good models of the climate at continental and greater scales. They are not perfect predictors of the climate though. Can you see the distinction?

    2. Weather prediction and climate prediction are two separate issues. Weather prediction is indeed subject to the uncertainties of chaotic systems, because it looks a specific events. Climate change is subject to other uncertainties, but not of chaotic systems because it looks at averages, not specific events.

  5. being attached to someone who works on these models, a couple quick comments:

    “they end up patching the model, but at the same time they want me to trust that this time the model is right.”

    Asinine comment. This is called “refinement” The models are always being refined to reduce the probability error ranges.

    The big issue right now is system feedback. Its insanely complicated and as warming continues how big the feedback signals (permafrost, North atlantic currents. etc.) are going to be is one of the big issues.

    Rate of change is the one to keep an eye on. Anthropogenic factors are affecting the rate of change of warming. As the models stand now, they are working quite well and always being modified as new research becomes available.

    Alosius, spend a bit of time at The GFDL ( is another great source for current information.

    The amount of misinformation out there is staggering. Scientists speak in terms of probabilities and non-science pontificators exploit this to their own means as they are not in anyway accountable.

  6. I don’t like the image on this article. The truth is the Earth or life don’t care about our puny attempts at destroying our environment. I find these anthropocentric displays corny and misleading.

    Reality is Earth, life, evolution couldn’t care less about us humans so we better care for ourselves. Place the image of a human corpse instead.

  7. Personally, I’ve been hoping for a good resource that explains the problem from a scientific standpoint without introducing bias right off the bat. Right now we’ve got ignorant misinformation coming from one side and patronizing noninformation from the other.

  8. i just think the thing that people forget most importantly of all is that prediction is something humans overestimate in importance for long term (it never works) and that a change of the order of under 0.000001% can do a butterfly effect and completely send everything in a completely unexpected direction.

    i personally think that solar variance is the primary driver of climate, and things like aerosols and greenhouse gases modulate the effect, either mitigate or intensify it, for example, in a really low solar minima we would actually want more greenhouse gases, it could even stop an ice age. right now of course the solar variance graph is above what it’s ever been in the last 4000 years. that simple fact alone is why i believe it is the most important input into the system. a lot of climate models also neglect to count the growing use of renewable non-carbon based energy sources into their doom and gloom predictions, at the rate the adoption is going the carbon dioxide levels are going to actually start dropping within 20 years, this will be even faster if ‘carbon farming’ takes off as well as it deserves to. an ironic fact is that carbon makes plants and soil healthier and they produce more yields and require less pest control. and so forth.

    imo the people who are worried so much about CO2 should be the biggest pushers of the view we need to rapidly revegetate the land. it’s not about assuaging guilt, it’s about actually having an effect on things. part of the reason why CO2 has gone up so much in the 20th century is simply the elimination of the normal carbon sinks that used to exist.

  9. It’s not till I talk about this stuff with someone else I realise just how pessimistic I am. Or maybe it’s realistic. It boils down to
    – Climate warming is real
    – It’s probably anthropogenic
    – There’s absolutely nothing that can be done about it. (within the current national and international, political and social systems).

    – Enjoy it while you can
    – Adapt and try to survive the inevitable implications

    Because even though I’ve been waiting for the axe to fall since the early 70s, it’s probably not going to fall on me but on my children and their children. If the warming doesn’t make life difficult, the resource depletion or the pollution will.

  10. @ Julien Couvrer: every means by which one attempts to anticipate the future involves a model. Some are naive and/or intuitive and run on the biological hardware between your ears, some are systematically specified and run on silicon chip computers. We do not have multiple Earths on which to run different climate scenarios in real life, so we’re stuck with models. The question is not can we trust models, but rather: since we have no alternative but to trust models, who has the best ones?

    Among climate scientists, I have no idea who has the best ones. Between climate scientists using well-specified models based on all applicable known physics and…well, anyone else…I think it would be silly not to go with the climate scientists.

  11. @ J. Couvreur again, re: prediction

    Predictions based on climate scientists’ models are not right–“no model is right, some are useful” is a quote I’ve come across many times (don’t remember attribution unfortunately)–but there is every reason to believe that they are better than the alternatives, which are also based on models. This because the alternatives are based on *much less rigorous models*, ranging from (at best) extrapolation from a subset of the data that goes into the climate science models (e.g. correlation of temperatures with sunspot or other solar variance data) to the all too common “reality=the opposite of what my political opponents believe” model.

    As for the prediction of previously unexpected events, you have to understand that climate models predict general climate states, not weather events. Each run of a model is going to have different weather. The average of runs produces an idea about what we can anticipate–but it isn’t exactly a prediction of what will actually happen, because what actually happens is more analogous to a single run of the model. Make sense? So the models can be very useful for examining the factors at play, but when you’re deciding how well they match reality you have to be careful that you are not expecting the model to predict a level of detail that is essentially weather.

  12. @ jwepurchase

    The problem with your comment is that your “for the sake of argument” stipulation (i.e. that catastrophic economic loss is low probability given a business-as-usual emissions scenario) is wrong. Check out the most recent IPCC Working Group 2 report. That’s the one that discusses the likely impacts of impending changes. Just like with the physical models, these predictions are certain to be wrong in the very specific details, but being based on established science they are far better than anything anyone else has to offer. So the best information available is that catastrophic economic damage in the business-as-usual scenario is virtually certain, not “low probability.”

  13. Reasons why I don’t believe the climate propagandists:
    1. Data was massaged – this was admitted by the climate scientists themselves. They did it to produce the charts they were wanting to present to the public.
    2. Data is mostly taken from sites inside or near cities. When a city grows in size and population, the temperature of the city rises. Therefore, the rise in “average global temperature” they are referring to is the average rise of temperature in cities, not the average rise in temperature of the earth. Sensors in the ocean were placed there many years ago and may not be accurate. Such details are never stated by the doomsayers.
    3. Climate scientists tend to leave major variables out of their equation, such as the effect of sun variances have on the earth’s temperature. I have yet to see any study of variance of sun’s energy. There could be enough variation to cause change in earth’s temperature by orders of magnitude greater than man’s ability to change it.
    4. Misrepresentation of data – charts were purposely skewed to indicate that temperature followed CO2 concentration, rather than what the actual data showed – that CO2 concentration followed variance in global temperature.
    There are many other reasons, but to make this short, I have one more:
    5. My stove is a climate change agent. It heats the kitchen and boils water which increases the relative humidity. My central air unit is also a climate change agent. It cools the air inside the house (and sometimes the outside for a moment), and it also reduces the relative humidity. I can also feel wind sometimes when it turns on.

    1. Reasons why tidymas refuses to accept sustainability as a goal:

      1. False. I have seen data in its original state, fresh off the instruments at Mauna Loa and from ice cores.

      2. False. I have examined various data sets but primarily ico core, dendro, and Mauna Loa. All with chain-of-custody known.

      3. Partly false. There are many models, some of which contain more variables than others, but nobody claims that they have accounted for every mouse fart that ever happened.

      4. False. Every data set I have ever seen represented as a chart showed the opposite of what you state.

      And finally:

      Yes, your stove is a climate change agent – unless it uses a biologically derived fuel such as wood, or is powered by a sustainable source such as wind, solar or hydro. Every time you run your stove you kill more of your great-grandchildren, unless you are using carbon-neutral technologies.

      Yes, your air conditioner is a climate change agent, with the same caveats. The outside bits of your AC unit get hot with the energy you’ve removed from the inside bits plus the amount of power added to run the pump that forces gas through the restriction valve. Every time you run your AC you kill more of your great-grandchildren, unless you are using carbon-neutral technologies.

      The greatest economic opportunity of the age is to make carbon neutral technologies available to the masses. If, however, you like being a poor, huddled, downtrodden mass yearning to breathe free, if you like being a subservient tool, if you see your children’s future as being prostitutes and soldiers for the oil barons, you should proudly support the burning of petrochemicals as a sure-fire way to prevent your country from becoming wealthier and healthier. Because that’s what this is really about.

  14. Thanks stanleyk for the thoughtful reply regarding methodology.

    Yes, every prediction involves a model. But in natural sciences such predictions can be shown inaccurate, thus forcing a re-evaluation of the model and theory. A theory and model which is able to provide unexpected and accurate predictions are considered more trustworthy or “better” in some sense.

    From what I read in your response, climate models cannot be meaningfully invalidated, even by observed future climate trends (say 20-30 years from now?).
    Is it fair to say this is a departure from other natural sciences?

    Would you have an example of an other scientific domain where the same methodology was used?
    If this is a new scientific methodology, shouldn’t it be discussed more and evaluated?

    Btw, if a model cannot be invalidated, how can it be corrected or even compared by any objective measure?

    Doesn’t your point on absolute versus relative model correcness (is the model right? vs. which model is best?) exclude the possibility that all models could be wrong, because all climate scientists lack of understanding of one common (yet un-identified) factor?
    An example is light/particle duality before the double slit experiment was discovered. All scientists at the time would have modeled the wrong behavior. But at least, observation would have proven all of their models wrong or flawed.

  15. No, computer climate models absolutely can be invalidated! Over time periods like the ones you mention (20-30 years) and longer, the noise from more-or-less random processes (like weather) and quasi-periodic processes (like el nino/la nina) starts to cancel itself out, and statistically significant climate trends become discernible.

    So, you can look at an average of model runs and compare it to a large enough chunk of observational data with a statistical test. If there’s a big enough difference, you can be fairly confident the model has the physics wrong.

    The issue is that there’s a floor to the uncertainty. That is, even if your model is incredibly accurate, you still wouldn’t expect it to accurately forecast details like, for instance, global temperature anomaly for a particular year. That’s just not what models like this do. But you would certainly be able to test the modeling of the temperature _trend over time_ given enough data to run a statistically meaningful test, and indeed this has already been done for some of the models that were already around a couple of decades ago. They are absolutely mainstream science. They’re just not weather forecast machines.

    It’s also important to remember that the models must contain assumptions about how things like human CO2 emissions will evolve over time. So in the IPCC reports, for instance, you will see model data for different scenarios. When evaluating the models’ performance against observed data, it is important to compare the model data from the emissions scenario that most closely matches what really happened. Seems obvious I know, but people have testified in front of the US congress that some of the older models have been proven wrong, based on comparing observed temperature data to the model output from an emissions scenario that was quite different from what really happened.

    Bottom line: computer climate modeling is traditional natural science.

    I believe that a lot of the people who have a problem with modeling are basically making a mistake analogous to thinking that an accurate coin-toss model should be able to tell you whether your next toss will be heads or tails. A model that gets it right will describe the long-term status of the system well: 50% heads, 50% tails. But A) it’s not like it gets that result by accurately predicting each next toss, and B) after a small number of trials, both observational and model data will vary widely from 50-50; you have to watch longer-term behavior to accurately assess the match.

  16. Personally, I’m going to assume tdidymas is joking and/or trolling, but for the benefit of any readers who took that seriously:
    1. Is false as far as I know, rather than saying “they” did such a thing perhaps t-diddy can tell us who did it, and when? Or are we meant to think that thousands of scientists have all done this? On the other hand people arguing against human-induced climate change have done exactly what t-diddy describes. See Damon and Laut 2004 for examples of people changing data and mislabeling graphs to make it appear that solar variance matches observed temperature change (spoiler alert: it turns out it doesn’t)
    2. Wrong. Subsets of temperature data sets are often examined to watch out for this kind of thing, and sets of rural stations show the same or more warming compared to urban samples. I’m not sure who “the doomsayers” are supposed to be, but the instruments of this science, like any other, are constantly being calibrated, upgraded, or (when no other option is feasible) any systematic error is quantified as well as possible and adjusted for.
    3. Contained three sentences. The first is total bullshit, just take a peek at the IPCC WG1 report. The second may be true, but only if t-diddy hasn’t looked at the scientific literature s/he is claiming to critique. The third is true or false depending on how you read it. Could solar variance of sufficient magnitude cause big changes in earth temperature? OF COURSE. Has the sun been doing anything over the past few decades that corresponds to observed warming? NOPE. (if you think otherwise, you probably have been exposed at some point to the faked data described in the Damon and Laut paper I mentioned above)
    4. I have never heard this one before. All graphs I have seen have always showed that CO2 rises have typically lagged behind temperature rises in paleoclimate data. There are reasons for that, look into it. (another spoiler: none of the reasons is “actually, CO2 doesn’t really function as a greenhouse gas after all!”)
    5. And this was the one that convinced me tdiddy was joking, or…um…something. Otherwise I don’t know wtf #5 is about.

  17. Ok stanleyk, so let me summarize: models can be invalidated but its difficult (exogenous variables) and there hasn’t been enough time passed by yet to gather sufficient data.

    Now, we cannot possibly trust models which require more time to be validated/invalidated, can we?
    So back to my original question: how much out-of-sample data do we need to wait for and what variance in the result would invalidate a model?

  18. Arguing is pointless, ignoring it is foolish. We will eventually need worldwide policies either way.

    Everyone arguing about the model can’t see the forest for the trees.

  19. What’s the term for when a debate is derailed by deliberately focusing on small, technical details that have little overall impact on the topic?

    Well, whatever it is, it should be replaced by “politicised climatology.”

  20. Hmpf – I couldn’t find any video. I think you mean this one…

    How it all ends (YouTube link)

    Give it a watch either way. You’ll love it. It applies a game theory type argument to the question of climate change.

  21. @Anonymous #6 – “The big issue right now is system feedback. Its insanely complicated”

    Exactly right. This is the bait’n’switch the skeptics are pulling here. Arguing that coz we’re unsure what’ll happen next, we must be mistaken when we say serious stuff is happening due to AGW is moronic. It’s like saying that coz I can’t tell you exactly where all the balls will be in 15 seconds’ time, I might well be mistaken in my belief that the cue just smacked into the white ball, and the layout of the balls pool table will look very different imminently.

Comments are closed.