If there are 10 reasons for the widespread mistrust of science, I'd say reasons 1 through 9 are the Mainstream Media and its obsession with "balance" combined with an aversion to providing real information about established facts. In fact, the MSM actively promotes not only mistrust of science expertise, but any sort of expertise in the same way. As with politics, any time a scientist makes a claim about a so-called "controversial" piece of information such as assertions about climate change, evolution, or a new vaccine, the news story has to be capped off with a "Critics say ..." or "Others claim ..." the exact opposite. This reinforces the perception among the average viewer that the experts can't even agree among themselves, so why should I listen to any of them. I would sooner trust Jenny McCarthy or Jim Carrey because they seem like really nice people who are intensely earnest and convincing in their alternative viewpoints (ignoring the fact that these people are actors, and thus very good at delivering an emotional message). The reason for this media practice often seems to be one of these two: (1) always stir up conflict whenever possible because the conventional wisdom is that is what sells air time; and (2) always present opposing views in a futile attempt to counter accusations of "liberal bias," because many scientific discoveries or advances tend to be mistrusted by either religious conservatives (e.g., evolution) or the business community (e.g., climate change), and therefore "liberal."

Its very interesting too see how this discussion has progressed. Up until recently anyone who expressed any skepticism about AGW was accused of being in the pay of the oil companies. Now that there are far too many skeptics that slur has been proven to be a lie. So the next slur is that these people must be anti-science. Actually they seem to be more supporters of the scientific method than the "true believers".Like many I foolishly thought that somewhere between thesis and proof was supposed to be this experimentation stage. Instead all were getting is computer modeling that is being passed off as truth. I recall for the decade before the great recession you had these academic economists getting Nobel Prizes for their computer modeling - which after the fact turned out to be pure BS. But the public is accused of being skeptical and anti science when they apply that observation to the wonderful and sure to be correct modeling of the academics in another field. Please!

On "computer models" that you deride; what researchers start with are smaller submodels that assess physical phenomena. These can be theoretical, empirical, or a mixture of the two. To put it in terms of the scientific method that you claim to be so fond of, their "hypothesis" would be that their models will accurately describe all significant phenomena. Their "prediction" would be taking their model and a set of inputs to predict data. Such data could either be historical or they could take their model to predict future events and compare as new data is obtained. This last step would be the "experiment".

I guess you have a problem with scientists that work computationally. It's a position I can't share, being an engineer that has used computational methods to accurately model aircraft aerodynamics. Somehow worked well even though it was just one of those danged computer models! Trust me, there is nothing inherently wrong with computationally modeling physical phenomena and much of the explosion in technology in the past 40 years is because of it. Now, if you have specific qualms about some of the assumptions of the models in use please expound. But I know you don't.

Well actually I am a trained engineer who has worked in aviation for years. So yes I am well familiar with the importance of computer modeling in both aircraft design and simulator design. I know that the computer is doing what used to be done much more slowly and laboriously by human beings. (Also strangely called computers) What is happening is essentially the manipulation of data so as to be presented in a more useful form. You however,being schooled at some point in the scientific method, should realize more than most the difference between that use of computer modeling and the idea that it is in some way proof of a scientific hypothesis. If you want to discuss lack of scientific knowledge in this country I would cite the common idea that any data processed through a computer is suddenly transformed into "science". I suggest you read my first post about the economists

Well, first, I didn't address economics because it's a soft science. It just isn't the same as subjects built on the hard sciences like physics and chemistry.

Often, these models are not a "manipulation of data to a more useful form". They are applying the laws of nature as we understand them to a complicated system and viewing the results. For example, if I'm using CFD on an aircraft wing all I need to program are the known boundary conditions and hold the system to conservation of mass, momentum, and energy. And I see shocks, expansion waves, vorticity, and other flow phenomena without explicitly programming them in or "manipulating data".

Ok, so back to the climate models; no, computer modeling is not "proof" of a scientific hypothesis in any meaningful way. But proofs in any form are tough to come by in a physical science. Even the hallowed laws of thermodynamics are only laws to the extent that they've been around for a long time and they haven't led us astray yet. But these climate models are testable. So, I ask you again. Do you have any REAL criticism of these models? Because like I've explained, the mere fact that they are computer models does not mean they aren't rigorous applications of our understanding of science.

Liberty:

I have no particular axe to grind in respect of the AGW hypotheses, but have some experience with computer modelling.

There seems to me to be a huge difference between, say, modelling the movement of a mass through a vacuum [using Newtonian mechanics tweaked to account for a bit of Einsteinian stuff...], and modelling - say - the behaviour of the stock market. Doubt you'd disagree with that statement.

The question is surely whether a model is in the first category [the "simple" and repetitive computational application of established science/mathematics to enable development of projections] versus the second category, where, crudely, the variables that define the system are simply unknown - and where the model is part of an attempt to inductively obtain them.

Them we get into the realm of the use of models to run experiments. It can be interesting - for example, I guess we could do experiments that let us explore how stuff would be if there was a different value for c [as in, E=...]. However, using models to run experiments, and then to report the results as empirically valid - we're into fantasy land.

There's nothing wrong with models, when they are models. When they are inventions, that's fine too - but they are no longer models of reality. Which category climate models fall into seems to me to be the essence of the question that needs to be addressed... urgently...

Nick:

I would generally agree with you that using a computer to perform many calculations of well-supported physical theories is different than more complex analysis where the starting values cannot be known *and* outcomes often depend critically on the starting values out to many levels of accuracy. But I don't think that either extreme is very common in computational modeling.

In the example of CFD (computational fluid dynamics) to understand structural response and aerodynamics, the physical laws governing the behavior of fluids (in that case air) are fairly well known (Navier-Stokes). They are also incredibly complicated and approximations must be made to write a computer code. Determining what approximations give the best matches to reality remains an active area of research.

On the opposite end, Global Climate Models (GCMs), bring together many different pieces of physical understanding about the Earth (e.g. atmospheric chemistry, ocean circulation, carbon uptake, etc). As the scientific research has progressed, scientists often had to formulate highly speculative conjectures and then follow them up. But there is the possibility of testing some of the models using ancient data (e.g. the Vostok ice cores and tree rings), and in the last decade the models have been able to roughly do this. It is impossible, and imho unnecessary, to require the models to exactly reproduce the current atmosphere starting with a set of reasonable conditions to "prove" that human beings can drastically alter the global climate. The key is that climate can change radically, and quickly. And that high CO2 levels can contribute to rapid climate change. Studying the impacts of previous ice ages, and changing atmospheric conditions as glaciers receded, has provided evidence that such a correlation exists. And models establish a working hypothesis of the causation. And the physical principles enshrined in those models have been tested in various microcosms.

No, there is no definitive experiment that "proves" beyond any doubt that anthropogenic CC is occurring. The only one we can do is the one all human beings are doing already. But as a scientist who remains skeptical about the details of what the impact of ACC may be, I am also, ultimately, a moral human being. And every piece of evidence in support of ACC, and every reasonable hypothesis based on the current observations, suggests that it may be devastating to the planet's climate. And I believe that we as the species with the power to drastically change our environment do have an obligation to be responsible stewards of our natural resources both for future human beings and for the other creatures we share the planet with. So I believe that caution dictates finding ways to reduce our CO2 output and helping the least privileged amongst us prepare for potentially catastrophic consequences.

Ultimately, it is a moral question. I suppose there is some actuarial calculation to be made balancing the cost of inaction against the potential for disaster...but quite frankly the scientific research has gotten us as far as we need to on this. It's not really up for debate whether climate can change drastically. It's not really up for debate whether human beings can drastically change their environment. And there's very solid evidence that one specific thing humans are doing, increasing the concentration of gases such as CO2 and methane in the atmosphere, can trigger major climatic shifts (there's no real evidence I'm aware of that contradicts this). The extent to which human actions are specifically changing climate right now can only fully be determined after we've let the experiment run for a few thousand years...but that's not the issue. The issue is whether we as a society feel that it's worth sacrificing to avert potential catastrophe...especially when the people who will suffer most are the least privileged and/or have yet to be born. Scientific research has given us enough information...more than we have for most moral dilemmas we face. But it comes down to something that lies outside of the domain of science.

liberty for all deserves a reasonable answer to his question. Yes I do have a real criticism to the models. Its one thing to say they represent the best science we have - its another to ask how good that science is. If it was as good as maintained by its proponents then there would not be raging debates about how the models prediction of the relationship of the rise in CO2 to global temperature has not been met over the last decade. To say (as do the AWG people) that this is just a slight uneveness in the data is at worst a weasel and at best best an admission that there is still a lot of work to make the model match reality more closely. To say something may be the way to bet is not the same as scientific proof.

How has it not been met over the past decade? Look at page 600 of this link: <link>

The black line is observed values, the yellow noise are many simulations that have been run by different computer models and the red line is the average of the computer models. Looks to me like the models do a pretty good job, including in the last 10 years.

Even if the models were perfectly predicting every measurable parameter there'd still be "raging debate" because there are people out there that will deny AGW regardless of the evidence. Existence of "debates" is not proof of faulty modeling.

Well sir I went to your link. The black line representing world temperature is below where it was in the year 2000. Did the worldwide production of CO2 go down in that time? No I believe it went way up. But you'll say 10 years ago the temperature was lower. Quite true, so it does depends where you start measuring from. More to the point, this graph was made up at the end of the period looking backwards. Its sort of what you notice in the hurricane models you see. Every few hours they change. When the hurricane finally comes ashore in Panama City and all the latest models show exactly that, it is good to remember a week ago they were all indicating that it was going to be Beaumont,Tx. The moral seems to me, look at the models, respect them for what they are, but maintain some skepticism. The people who drew up those models know more about the subject than I do but they don't know as much as is being sold to the public.

The black line representing world temperature is below where it was in the year 2000. Did the worldwide production of CO2 go down in that time? No I believe it went way up.

Wow... really? That's your reason for being a skeptic? No one is claiming CO2 is the ONLY variable affecting climate. Suggesting that increasing CO2 levels means that the temperature must increase year on year every year is either scientifically illiterate or highly disingenuous. The models take many of these variables into account, and as I've shown they've done a good job predicting results.

More to the point, this graph was made up at the end of the period looking backwards.

No, what is shown in that chart is an input of information for the start point and the subsequent predicted and observed results. It is not a case of revising the predictions over and over again with new observations. To compare to your hurricane analogy, it would be like having data for a hurricane that already occurred, taking what was known to meteorologists when it was still in the ocean, and taking your knowledge of known physical phenomena predicting it's path; then, getting very good correlation between your model's prediction and what actually happened.

And they HAVE taken their models to predict future events to evaluate them; they've been doing so for two decades with good agreement. Some parameters they predict better than others, but temperature is one thing they've been pretty dead on about.

Slatesurfer:

No substantial argument from me against anything you wrote. I'd add a few small points though:

1.) Note that I am not opposing the AGW hypothesis. However, I note that many people do these days seem to equate asking questions about the validity of its underpinnings with some kind of apostasy.

2.) I recall a lesson from high school math. We were given this sequence: 1, 2, 4, 8... and asked what comes next. Of course, we all said 16. The answer the teacher was looking for was 15 - not that anyone could have reasonably guessed that. The real lesson was that if you have a series of numbers, you can pretty much add anything to the end, and still create a function that models it. The function for 1,2,4,8,15 is not as elegant as the simply doubling function, but that's not the point. This has real implications for computer based modelling where the simulation is itself the experimental system.

3.) I did a good bit of work with neural nets - same principle as above (neural nets are reasonably describable as function approximators). It's possible to train a net with any data set you like, and it is also possible to overtrain - that is, to over-value the lessons of the training data.

4.) You didn't do this, but we guard against exhortations to adopt the "precautionary principle." If we bought into that, we'd all have asteroid shelters in our back yards.

Bottom line: I agree we should buy into the consensus of climate scientists. However, we should continue to be rigorous in demanding careful critique - with empirical techniques whenever possible - that challenge the assumptions. If the science becomes a religion, we're lost.

The efforts of the Copenhagen Consensus economics debate gives us some real insights into the trade offs we are potentially making when we spend vast sums on carbon capture, and we need to be clear we're monitoring the value of the investment.

Nick, you have a reasonable post, but I feel like I should address some of your points;

However, I note that many people do these days seem to equate asking questions about the validity of its underpinnings with some kind of apostasy.

True enough, and I think valid questions need to be asked and addressed. However, some people will just keep asking questions and calling for more study regardless of the science available to use as a stalling tactic to prevent any action regarding climate change. They want to oppose such measures, but want to sound reasonable and can't really find anything wrong in the science being presented. I suspect that businessanalyst in this thread is an apt example of such a person.

The real lesson was that if you have a series of numbers, you can pretty much add anything to the end, and still create a function that models it.... I did a good bit of work with neural nets - same principle as above (neural nets are reasonably describable as function approximators).

My experience with NN is somewhat limited, but my understanding is that it's basically a complicated multi-variable regression tool; basically, it just looks for patterns and tries to reduce error. Like your 2,4,..15 example, yes we can fit a function to any set of data. But that isn't what is going on in the climate models; they aren't simple trend analyses or regressions. Like CFD, they start with physical first principles like conservation of mass, momentum, and energy. As such, when we apply these models and get good correlation with existing data we can have greater confidence in the underling model than if it was merely some sort of curve fit to existing data.

Other than that, I pretty much agree with you. I find it distressing that pretty much the only voices in the public discussion are the ones denying the basic science and the chicken littles saying we need immediate drastic action without analysis; very few people seem interested in a careful risk and cost benefit analysis of various policy options.

Liberty -

Thanks - no discord here. I should have been clearer at the start - I am largely clueless about the fine print of climate models, and the "type" they represent. So, I am not at all looking to be dismissive about them, merely to try to make some small contribution to the characterization of types of models. Also, you're right about NNs (though, doubtless an afficionado would damn us both for over-simplification).

One day, I will retire, and dig into the climate science to a level where I might reasonably know what is real and what might not be. However, by then we'll know... :-)

Are you aware of any forum where smart, calm, knowledgable people have been able to quiz climate change experts on various questions that bother them, and get honest and not-dismissive answers? For example, many people are throwing around stuff like the lack of warming for the last decade, the alleged thickening of the polar ice, etc... (I have no opinions on these things... I cite them as examples).

Nick, Liberty, Business, everyone...

I highly recommend "The Discovery of Global Warming" by Spencer R. Weart. Published in 2003, the title is perhaps a bit anachronistic. But it's a good overview of the progression of the scientific research on anthropogenic climate change. The author is a physicist, who I don't think was directly involved in any of the research he discusses...so I found it to be fairly objective. And perhaps, as a physicist myself, I appreciate his approach to the issue. As someone who spends a lot of time writing computational simulations of data, I'm well aware of both the power and limitation of computer modeling. I'm also generally skeptical about large scale physical simulations b/c I've seen both excellent use and egregious abuse of such things in my own field.

But I think the book does a good job of balancing what parts of the research remain uncertain, and where and how consensus has emerged. It also presents a fair picture of when scientists may have overreached, and when they were too cautious. Overall it's an excellent overview of how the scientific process is carried out throughout a community of researchers, but it also gives a good summary of the state of the research...

Not exactly a forum where you can pick climate scientists brains, but perhaps a starting point about whom to ask.