Scientiific American Says "Chill out over Global Warming"

[velcro]

Seriati wrote:

How exactly does have 4 scientists check another scientists work prior to publishing it advance any goal better than publishing it and having any scientists that want to do so verify or dispute it?

How does having a newspaper editor and fact checker checking articles prior to publishing it any better than publishing it and having any readers that want to do so verify or dispute it?

-It is the reviewers job to check it, even if they are not paid.  Their reputation is on the line, even if they are initially anonymous.  It is not "checking math".  That is a derogatory oversimplification.
-They are intimately familiar with the content and the method, but ideally looking at it from a different angle because of diverse training and background. You seem to assume that all scientists have precisely the same biases and blind spots, and have no incentive to point out errors.
-And most obvious, there needs to be at least some superficial checking before publishing, if the journal wants to have any relevance.

What you are suggesting is, I think, something like this.   The paper is "published", but it is difficult to garner any attention if you are writing it, and difficult to know who to trust if you are reading it.

[velcro]

Seriati wrote

I was also reading several papers that are highly critical of the peer review process itself

Would you mind providing links, if possible?

Greg said

Peer review is imperfect, but it is better than any other approach yet invented to consistently pursue scientific truth at the frontier of human knowledge.

You appear to disagree, and ask for proof of that statement.  While a valid request, it is very difficult to accommodate.

On the other hand, it is trivially simple to disprove.  Just provide a simple example of a system that is better.  Again, this is all in the context of climate change, so please make sure that the better system is applicable to that type of science.

(By the way, "Big Data Analytics" are methods for drawing trends and conclusions from very large sets of data.  It is a tool for visualizing and organizing data, no more, no less.  If anyone were to use such a tool, and were to write a paper based on the results, that paper would be just as prone to error and bias as a paper that simply collected "small" data, graphed it, and drew conclusions. It is orthogonal to peer review.)

[Seriati]

Do you think that's responsive?  Are you asserting that newspaper editing constitutes a verification process for scientific research?   Lol.

There's literally dozens of real verification techniques you could try to critique, why pick a fake one?  I'll get you started, explain how a process that required an experiment be repeated is better or worse than the current peer review process.

How about another, explain how a commercial process that is verified by it's practical and profitable application is less of a scientific advancement than one developed at a university and published through a journal.

Explain how crowd sourcing the verification and testing of theory with the results shared to the group is less of a contribution?  Or less verified?

Peer review came about in a different era of information.  It solved for a problem that was very real - at that time.  Specifically, credibility of verification in a world where a reader didn't have access to the source, the data or the reputation of every paper.  They had to rely on what was in front of them, often with limited access to other materials that may have been written.  Research on a topic took a professional degree in and of itself and was still likely to miss relevant materials.  Frauds were easy to perpetuate cause information was hard to share and move.  Peer review and peer review journals solved for that problem by lending their credibility to the paper, ie endorsing it, so distant readers could know that an expert had confirmed that the study wasn't just shoddy work.  They helped to establish which works were seminal.  Not a guaranty of results but a guaranty of basic craftsmanship.

We are no longer in an age where information is slow or constrained.  We now have access directly to the knowledge, and to the experts an their opinions on the research and we can search with powerful tools to get to it almost immediately.  We are no longer stuck with the stark choice of believing what is presented or rejecting it without knowing anything more, we can now learn as much or as little as we want to make a judgement.  It's not an era where we have to rely on a handful of first lookers to tell us that a paper is real or not, we're able to the limits of our ability, to verify or falsify it ourselves.

The limits today are personal, more than informational.  Not everyone is smart enough or rigorous enough to reach a conclusion on a topic.  Some of us are smarter than the people who wrote the paper some of us are not.  Group bias errors that were previously hidden (because only the group saw behind the curtain) are now open to all and easy to see for those who don't share the bias.  Flaws in methodology become an orthodoxy that has to be defined.  Claims of scientists can become "public facts" and "consensus" far beyond the certainty that the scientists claim.

Greg said

Peer review is imperfect, but it is better than any other approach yet invented to consistently pursue scientific truth at the frontier of human knowledge.

You appear to disagree, and ask for proof of that statement.  While a valid request, it is very difficult to accommodate.

Greg stated it as a fact.  You seem to agree.  Surely, you too did not just accept it as a "fact" without evidence. 

On the other hand, it is trivially simple to disprove.  Just provide a simple example of a system that is better.  Again, this is all in the context of climate change, so please make sure that the better system is applicable to that type of science.

What science would that be?  Is there a climate change experiment that has occurred of which I'm unaware?  Did you invent star travel while I wasn't looking?

(By the way, "Big Data Analytics" are methods for drawing trends and conclusions from very large sets of data.  It is a tool for visualizing and organizing data, no more, no less.  If anyone were to use such a tool, and were to write a paper based on the results, that paper would be just as prone to error and bias as a paper that simply collected "small" data, graphed it, and drew conclusions. It is orthogonal to peer review.)

Lol, if you say so.  Meanwhile the world will go on innovating without your contributions.

[Greg Davidson]

Seriati, it's actually a good question - how to prove the merits of the system of science built on peer review and reproducible results to someone who doesn't want to believe them. Because there are human weaknesses embedded in processes that are executed by humans. Peer review attempts to address this by multiple, independent reviewers. Independence reduces conflicts of interest that may sway an outcome, while the culture of science that Wayward was describing means that one of the main assets of a professional scientist is reputation, and therefore approving research papers that are not reproducible due to methodological failures exacts a relatively higher cost on peer reviewers.

There is some recent research on the  effectiveness of peer review, but it is limited. In this paper there are two posited effects, quality (screening out the methodologically flawed) and impact (predicting which will be the highly cited papers).  For my argument here, the first is the more important effect, because once the flawed papers are screened out I do believe the community (and that all-important reproducibility of results) will do the rest.

But I suspect that this will not be satisfying to someone who believes the following, which is why this is an interesting question to continue to think about:

We are no longer in an age where information is slow or constrained.  We now have access directly to the knowledge, and to the experts an their opinions on the research and we can search with powerful tools to get to it almost immediately.  We are no longer stuck with the stark choice of believing what is presented or rejecting it without knowing anything more, we can now learn as much or as little as we want to make a judgement.  It's not an era where we have to rely on a handful of first lookers to tell us that a paper is real or not, we're able to the limits of our ability, to verify or falsify it ourselves.

This is, in a sense, part of the problem - people who lack the expertise to catch manipulations that can be uncovered by those with professional training (or self-taught but disciplined expertise) still believe in the superiority of their own expertise. Even this might be correctable if they would be capable of listening and understanding when a methodological error underlying their argument is pointed out to them.  But that rarely happens in the current information age. Without the context of a science community, people are free to throw out arguments that are shown to be flawed and yet not take personal accountability for their errors (there are a few other places where the BS is forced to the surface, for example, the Kansas Court where those with methodologically flawed research about voter fraud are being forced to confront their dubious arguments).

Maybe that's why this is so meaningful to me - the fundamentals of good science are the same as the fundamentals of good public policy discussion. And in both cases, those discussions need to be built on a foundation that rewards integrity and accountability. 

 

[Greg Davidson]

Sorry for the short answer, already late for work.

[TheDrake]

One extra thing to add - peer review does a lot of good things besides editing and choosing to publish. It maintains standards in terminology (oh, by the way, this other study already defined that term). It improves clarity in how equations are rendered.

Here's a good honest breakdown of peer review including the following:

Peer-reviewers also have biases. They may be prejudiced against studies that contradict their own research or their preferred beliefs. They may therefore bias the published studies in their favored direction, and may be loath to give a pass to a submission that would directly contradict something they have published. For this reasons editors often allow authors to request or recommend reviewers, or to request that certain people not be asked to be reviewers. Each journal has their own policy. Sometimes an editor will specifically use a reviewer that the authors request not be used, thinking they may be trying to avoid legitimate criticism.

Perhaps the biggest weakness of peer-review, however, is when an entire discipline of peers is lacking in some fundamental way. For example, there are now many journals dedicated to so-called “alternative medicine” (CAM).  The editors of such journals tend to have a pro-CAM bias, and they find reviewers with a pro-CAM bias. So pretty much any pro-CAM article can get published. Some have enough ideological friends at the NLM that they can get approved as peer-reviewed, despite glaring biases in their editorial policy.

But the alternative is what, exactly? There's a reason why I'm probably going to pick a novel that is edited and published rather than self-published. It has gone through a process to tighten it up, clarify, and improve. The same would go for scientific articles. I'm going to pay more attention to something in the New England Journal of Medicine than something on Todd's Body Blog. Somebody has already done some work to ensure I'm not wasting my time.

[Seriati]

Seriati, it's actually a good question - how to prove the merits of the system of science built on peer review and reproducible results to someone who doesn't want to believe them.

That's a strawman.  I didn't ask you to prove the merits of science.  I didn't make any claim of doubt about science, not even about whether the peer review process has merit.

I asked you specifically to verify that your claim that the peer review process is "better than any other approach yet invented to consistently pursue scientific truth" is not axiomatic.  Science will advance with or without a reliance on peer review, it is not inseparable from development.

I even gave you my expectation that you can't do it, that is axiomatic and that you've never seen a verification or falsification of the process.  Effectively, for you - at the time you made the claim - it was literally a tribal claim.  True, without evidence, because it "sounds right" based on your tribal identity. 

Because there are human weaknesses embedded in processes that are executed by humans. Peer review attempts to address this by multiple, independent reviewers. Independence reduces conflicts of interest that may sway an outcome, while the culture of science that Wayward was describing means that one of the main assets of a professional scientist is reputation, and therefore approving research papers that are not reproducible due to methodological failures exacts a relatively higher cost on peer reviewers.

Though I think there's a lot of overstatement there, particularly the idea that there are high costs to a reviewer if an experiment is not reproducible (which honestly is not something that would be able to verify absent it being caused by an identifiable error in process), there's nothing about what you said that I find objectionable.  I agree that is the goal of the system, I dispute that it's always the outcome or result.

This is, in a sense, part of the problem - people who lack the expertise to catch manipulations that can be uncovered by those with professional training (or self-taught but disciplined expertise) still believe in the superiority of their own expertise. Even this might be correctable if they would be capable of listening and understanding when a methodological error underlying their argument is pointed out to them.

Oddly I agree, however, I think what you fail to realize is this in not a one directional error.  Those who "agree" with a study are just as likely in the modern era as those who disagree to have engaged in it.  This is why "debates" on science turn into citation fests where people attach the first google result that has a headline that matches their opinion and then claim they are right.

You can see the effect in the field of climate science (well if you're objective), with the amount of people who dogpile on a topic because of a "consensus" or because "it's settled" without any comprehension that their specific opinion is no more accurate than the opinions on the other side cause they didn't understand the limitations on what the study actually said.  This is why there arguments about those who treat science as a religion, where it's divining incontrovertible truth.  Pretty much anytime "belief" enters into the debate, we've left the realm of science.

But that rarely happens in the current information age. Without the context of a science community, people are free to throw out arguments that are shown to be flawed and yet not take personal accountability for their errors (there are a few other places where the BS is forced to the surface, for example, the Kansas Court where those with methodologically flawed research about voter fraud are being forced to confront their dubious arguments).

And other times they can continue to pretend that their arguments are correct even though they are completely erroneous (like say every global climate change treaty).

Maybe that's why this is so meaningful to me - the fundamentals of good science are the same as the fundamentals of good public policy discussion. And in both cases, those discussions need to be built on a foundation that rewards integrity and accountability.

Integrity and accountability would be great.  But good science and good public policy do NOT have the same fundamental construct.  Science should be about challenging everything to get to the truth, even if it flips everything on its head, even if its wrong and gets disproven next to immediately.

Public policy is about balance and fairness, even if it's not the "optimal" solution as determined by an ivory tower analysis it can be the bets solution for the people involved.  Lots of human choices and free will are not scientifically justifiable, yet perfectly fine.  What's the exact "right" choice of how many trees to have downtown?  Should a community be barred from surrounding an areas with green if the optimal is interspersed?  If one solution is 89% efficient and the other 86% should the latter be illegal?

Honestly, you're deceiving yourself if you believe the left is about "science".  There's plenty of science that touches on uncomfortable issues that the left absolutely chooses to reject as a basis for public policy.  Go back and look at the arguments that occurred around for example transgender military issues, or adoption rights of single parents or gay parents versus bi-gender couples, or look at the reality of how increasing public school budgets rarely increases results (or conversely how the hated statistical testing does).  What's the measure on whether an industrial development that causes pollution is "worth it"?  Is it purely financial, there's nothing scientific about that judgment.

The left is about science to the extent the headline from a study back's its view, same as the right, but since the left is more heavily represented in academia, they tend to have more studies designed to support their views (you don't often see huge headlines about the studies that fail to support a view, instead it gets reconstructed again and again until it says what they want it to say).

[velcro]

Seriati wrote:

Do you think that's responsive?  Are you asserting that newspaper editing constitutes a verification process for scientific research?   Lol.

There's literally dozens of real verification techniques you could try to critique, why pick a fake one?  I'll get you started, explain how a process that required an experiment be repeated is better or worse than the current peer review process.

How about another, explain how a commercial process that is verified by it's practical and profitable application is less of a scientific advancement than one developed at a university and published through a journal.

Explain how crowd sourcing the verification and testing of theory with the results shared to the group is less of a contribution?  Or less verified?

I am not asserting that newspaper editing constitutes a verification process for scientific research.  I was making an analogy to your suggestion that people just publish without any review.  It doesn't work for news, for obvious reasons.  It doesn't work for science, for exactly the same obvious reasons.

I think we all agree that peer review has its benefits and drawbacks.

Peer review plus documentation of repeatability would be better at finding errors, but demonstrating repeatability is time consuming and costly, so less science would get done if this were required.  That is a discussion we can have.

The goal of peer review is to determine whether the results of a scientific study are valid.  Practical and profitable application does not demonstrate any "results" except for a product .  Pet rocks are practical and profitable.  Firestone's vulcanized rubber was practical and profitable, but there were no "results" other than doing this results in that, but we have no idea why.  Edison's method of trying literally thousands of filaments to find the best for a light bulb provided a practical and profitable product, but it did not make any claims about underlying principles, or predict behavior for anything not tested.  In and of itself, it did not advance knowledge, except for a binary decision on a list of materials for a very narrow purpose. 

Applying that standard to actual scientific papers is nonsensical.  Again, for a climate change paper, how do you apply that standard?

Crowd sourcing only makes sense if we have an idea of the skills and reliability of those doing the verification.  At any large scale that is impractical.  At a small scale, that is peer review with repeatability, if possible.

You never provided any links that were requested, by the way.

And this comment:

Lol, if you say so.  Meanwhile the world will go on innovating without your contributions.

was in response to my explanation of big data, and how you misused the term.

Please correct my facts, or acknowledge your error, or just say nothing.   Responding in petulant and demeaning manner just makes you look well, I leave it to each individual to decide.

And as far as innovating without my contributions, how many patents do you hold? 

[Seriati]

I am not asserting that newspaper editing constitutes a verification process for scientific research.  I was making an analogy to your suggestion that people just publish without any review.  It doesn't work for news, for obvious reasons.  It doesn't work for science, for exactly the same obvious reasons.

I find when people cite to "obvious" reasons, its really code for "I believe this to be true but can't explain why," or in other words, this is an axiom and you're a big silly head for not having the same axiom.

Peer review plus documentation of repeatability would be better at finding errors, but demonstrating repeatability is time consuming and costly, so less science would get done if this were required.  That is a discussion we can have.

Repeating an experiment is costly and time consuming, at least in some cases.  In others it could be done more swiftly than completing a peer review. 

However, the real criticism is that less science would get done.  Peer review is literally pulling a scientific research out of their own research to check someone elses.  Even after it's done people with still do the repeatablity tests, usually with a new wrinkle to add some originality.

The goal of peer review is to determine whether the results of a scientific study are valid.

Is it?  You sure about that?

Practical and profitable application does not demonstrate any "results" except for a product .  Pet rocks are practical and profitable.  Firestone's vulcanized rubber was practical and profitable, but there were no "results" other than doing this results in that, but we have no idea why.  Edison's method of trying literally thousands of filaments to find the best for a light bulb provided a practical and profitable product, but it did not make any claims about underlying principles, or predict behavior for anything not tested.  In and of itself, it did not advance knowledge, except for a binary decision on a list of materials for a very narrow purpose.

I think you're out on your own limb with a claim that discovering and developing the vulcanization process, or Edison's experiments did not advance knowledge.  Lol. 

As far as I'm aware the pet rock's only scientific advancements were in the fields of human psychology and marketing, where if you look you can still find case studies and papers on the phenomena. 

Applying that standard to actual scientific papers is nonsensical.  Again, for a climate change paper, how do you apply that standard?

Which standard?

Crowd sourcing only makes sense if we have an idea of the skills and reliability of those doing the verification.  At any large scale that is impractical.  At a small scale, that is peer review with repeatability, if possible.

No, actually the point of crowd sourcing is that it relies on the crowd to correct itself, to point out the flaws in the contrary results.

You never provided any links that were requested, by the way.

And?  Happy to weigh my contributions in the link department against yours.

And this comment:

Lol, if you say so.  Meanwhile the world will go on innovating without your contributions.

was in response to my explanation of big data, and how you misused the term.

Please correct my facts, or acknowledge your error, or just say nothing.   Responding in petulant and demeaning manner just makes you look well, I leave it to each individual to decide.

If you want to provide some facts, I'll be happy to address them.  Maybe you can point out where you think you did so, cause all I say is pretty much unvarnished opinion and demands that I provide facts for you.

And as far as innovating without my contributions, how many patents do you hold? 

Lol, I'll just leave you with a quote on that, "Responding in petulant and demeaning manner just makes you look well, I leave it to each individual to decide."

[Fenring]

Peer review plus documentation of repeatability would be better at finding errors, but demonstrating repeatability is time consuming and costly, so less science would get done if this were required.  That is a discussion we can have.

I think there's some confusion here about what peer review is. It's strictly an assessment of work being professional enough and not making sloppy errors so that it can appear in journals and be understood to be the work of 'serious people'. Nothing about it has anything to do with 'documentation of repeatability'. That isn't a thing. Rather, a group that presents its own findings will usually cite how many times they've repeated their own experiment to make sure their results were consistent and not cherry picked, but real repeatability comes when other teams go ahead and try to reproduce the experiment. This process of having other scientists verify that your work was legit isn't a part of the peer review process, nor are we suggesting it should be. The point is that peer review is about work methods, not about results. Results in hard sciences get checked when others try the same experiment for themselves, which isn't peer review but rather fodder for additional papers on the same subject adding to the general literature.

The goal of peer review is to determine whether the results of a scientific study are valid.

As I mentioned above (and as Seriati hinted) this isn't correct, unless by 'valid' you mean that the way the results seem to have been reached were reasonable. But just to make sure there's no confusion, this doesn't mean the results were accurate in any way. Depending on the type of experiment this can be due to mistakes in first principles, or technical shortcomings of their process, or details in the environment they didn't account for, etc etc. Sometimes this can be caught in review if the methods are blatantly wrong, but a peer reviewer may not be able to catch a methodological mistake in cutting edge field where the team is basically doing new science and are inventing the method itself to do the experiment.

Practical and profitable application does not demonstrate any "results" except for a product.

Tell that to engineers. "It works" is worth a whole lot, not only in terms of commodifying a discovery, but also in realizing what the first principles behind it might be. Invention has always been a significant contributor to scientific discovery in the past, not sure why you'd reduce it to a mere 'product.'

Applying that standard to actual scientific papers is nonsensical.  Again, for a climate change paper, how do you apply that standard?

Seriati's point appears to be that they may be a better way than peer reviewed models to figure out how the Earth's climate works. Actually I'd be interested to hear his take on that, since if peer review isn't ideal for this then I'd like to know what is. I agree with him that the groupthink "do something!" mentality isn't helpful, and likewise with bad climate treaties that are more about populism than science. But in terms of the theory itself I don't have an opinion about the best way to go about studying the problem.

[TheDrake]

No, actually the point of crowd sourcing is that it relies on the crowd to correct itself, to point out the flaws in the contrary results.

Yeah, that's working pretty great to verify the validity of news reports. Crowds reinforce their own stupidity most of the time. Now you could limit this to "anyone with a valid credential" such as an advanced degree in the subject matter. But that doesn't solve your problem of "stifling" innovation.

If you want to test climate change, then the only real test is the predictive value of the models, not peer review. The deniers crowd doesn't seem to have that. They like to point out when the predictions of the people doing real science are off. They like to complain about the measurement techniques. They can't explain away:

predictive climate model correct for 50 year span

And they don't generally advance their own predictions or deploy their own measurement techniques or do anything serious to try to refute mainstream claims. Instead its an endless drone of complaints about how calibration adjustments had been done to old data - tried and true techniques. Then as they've been increasingly boxed into a corner of admitting the global temperature is going up they change tactics and start complaints about whether it is due to human activity. Or if the rate of increase is slower. Or finally that its not such a big deal because the Earth's climate varies anyway. They are pursuing an agenda, in my opinion, not trying to explore and understand the universe.

[Seriati]

Seriati's point appears to be that they may be a better way than peer reviewed models to figure out how the Earth's climate works. Actually I'd be interested to hear his take on that, since if peer review isn't ideal for this then I'd like to know what is.

It depends on what you mean by "climate science."  Do you mean the micro level studies, some of which are experimental but few actually test the climate in any way, which are data points for the macro level studies or do you mean the macro level studies, which are all modeled computer science and not experimental?  I actually think the peer review process in climate science is especially misleading because the macro models are so heavily dependent on judgment.

I agree with him that the groupthink "do something!" mentality isn't helpful, and likewise with bad climate treaties that are more about populism than science. But in terms of the theory itself I don't have an opinion about the best way to go about studying the problem.

I'm not sure that the way we're studying it is wrong, we don't have any real options for true experimentation and we should be erring on the side of caution so long as all our eggs are in one planetary basket.  Where we run into trouble is when we over state the case, and then turn that into policies that don't actually rest on the science because the "science is on our side."  (you can imagine this just like a crusader army, with god on their side, that commits atrocity after atrocity that don't appear in any where in their holy book)

[Fenring]

Ok Seriati, but if you were posed with the problem of being in charge of managing a scientific team, whose mandate was to examine whether humans are damaging the environment significantly in such a way that it will cause global problems, what would your marching orders be? Note that in answering this question the issue isn't strictly to determine what scientific procedures should be implemented (that would be an unfair question in a newish field, and especially since it isn't your field) but rather what would be your timeline for the work? Muddle along for many years and see if new methods develop? Try to do your best to find a short-term result to see if we're in imminent danger? Or put it on the back-burner pending quantum computing and focus on more understandable issues?

In other words, would you focus on the possibility of short-term disaster scenarios, or not? How would you go about determining whether quick work is needed since the very thing to verify is whether there's a short-term problem? It's a bit of a Chicken Little situation to be sure, but lack of having hyper-advanced science knowledge in that area leading to dropping the matter would certainly be a problem if there was, in fact, a short-term problem.

So basically, if you're going to criticize how the science is done on the macro scale situation, how would you do it better?

[DonaldD]

People often get bogged down on climate modelling as if that is the be-all and end-all of evidence of climate change.

But there are many, many metrics by which changes in climate can be measured, physically, including things like glacial retreat, polar sea ice coverage, snow extent, changes to growing seasons, sea level rise, widening of the tropical belt, the frequency of record hot days vs record cold days, and the list goes on.  The focus on models is handy for some because climate models are not well understood, so can be easily attacked, and then can be used as a red herring in order to distract from all the other metrics that are clearly showing warming, especially when taken together.

[Fenring]

People often get bogged down on climate modelling as if that is the be-all and end-all of evidence of climate change.

But there are many, many metrics by which changes in climate can be measured, physically, including things like glacial retreat, polar sea ice coverage, snow extent, changes to growing seasons, sea level rise, widening of the tropical belt, the frequency of record hot days vs record cold days, and the list goes on.  The focus on models is handy for some because climate models are not well understood, so can be easily attacked, and then can be used as a red herring in order to distract from all the other metrics that are clearly showing warming, especially when taken together.

I think where the importance of the models comes into isn't the "what" but the "how". Even if we took it as a given that there is warming, and even that it's due to human contribution, it would take a model to project what sorts of affects various changes in our behavior would have as a result. You can't 'empirical' your way to suggesting that, as an example, lowering greenhouse emissions by 20% would cause exactly X change in the future outlook. So you can look at other data to verify the problem, but not to determine what to do, and whether any particular actions will be effective or not. I don't see how you do that without climate models. My problem with climate science hasn't so much been "do these fools know how to measure temperature" but rather "what do you actually suggest to deal with it, and how do you know it will work?" And the corollary to that is "how much will it cost" and, to that, "will the solution cause more suffering than the problem." 

[DonaldD]

Fenring, are we already at the point where the vast majority of US residents accept that the global climate is warming?  The last I checked, that wasn't the case.

[Seriati]

People often get bogged down on climate modelling as if that is the be-all and end-all of evidence of climate change.

No, the focus is on modelling because its the part that claims predictive power.  The part that claims it can determine causation.

But there are many, many metrics by which changes in climate can be measured, physically, including things like glacial retreat, polar sea ice coverage, snow extent, changes to growing seasons, sea level rise, widening of the tropical belt, the frequency of record hot days vs record cold days, and the list goes on.

Agreed, all good data points to measure what has happened.  All observational.  I do think again people are biased towards what they can easily measure, and tend to believe that the measures are far more accurate than they actually are, but I have little doubt that they are generally taken and reported in good faith.

The focus on models is handy for some because climate models are not well understood, so can be easily attacked, and then can be used as a red herring in order to distract from all the other metrics that are clearly showing warming, especially when taken together.

No, again the reason the models are attacked is because they claim predictive power, and are heavily influenced by how they are constructed.

I agree they are not well understood, that's also why people use them to make claims that exceed their limitations.

[Seriati]

Ok Seriati, but if you were posed with the problem of being in charge of managing a scientific team, whose mandate was to examine whether humans are damaging the environment significantly in such a way that it will cause global problems, what would your marching orders be?

That's quite the mandate.  Am I investigating whether the environment is changing?  Whether its harmful or catastrophic?  Whether its caused by "humans"?  Whether its caused by using steel bladed tools in farming Rutebega?  (Obviously the last is a trifle fascetious, but the possible vectors are next to infinite). 

Note that in answering this question the issue isn't strictly to determine what scientific procedures should be implemented (that would be an unfair question in a newish field, and especially since it isn't your field) but rather what would be your timeline for the work?

Depends on the problem.  If say, you were concerned about the creation of miniature black holes by the Large Hadron Collider and whether they could accidentally tear our planet apart, you'd have had a very specific window (which I think you missed).  If say, you were concerned about GMOs causing uncontrollable mutations and sterility in our children going forward, you'd pretty much have infinity since that ain't happening.  If you were concerned about human consumption and living space eliminating biodiversity on our planet every second wasted is a failure.

Irreversible climate change?  Not all that clear.  Most models are caveat'ed with the all powerful, all other things being equal problem.  ATM I think we know enough to be interested in limiting any further damage and weighing future development against the cost of implementing it.  Of course that means that we ought to do everything in our power to put third world polluting factories out of business.

My rationale self interest says do it by increasing First world production.  You could just as easily favor developing better factories in the third world.  What you can't do, is say sign a Paris Accord massively reducing first world production while agreeing to let the third world increase its production and pollution for 30 more years to make up the production difference... oh wait....

Muddle along for many years and see if new methods develop?

You do realize how many times scientists have effectively had to invent their own tools to even begin to test a theory that they can't directly measure when they start?  Pushing the envelope always requires muddling along for a while.  On the climate, I think you're hinting that the risk of waiting is too high.

That's fine with me, I'm pro environment anyway.  But I'm not going to sign off on "environmental" treaties that hurt the environment just to "do something," nor on policies like the EPA under Obama that weight any environmental harm as too much.  We are going to harm the environment, just by living a modern life.  We have to recognize and acknowledge that fact and actively work to mitigate the harm not just reduce it.  And sometimes we have to make trade offs.

Try to do your best to find a short-term result to see if we're in imminent danger? Or put it on the back-burner pending quantum computing and focus on more understandable issues?

Are you asking what percentage of time to devote to developing science versus pushing the theory envelope?  That's a hugely pressing concern for a single person or a small team, not so clear its a pressing concern when we have hundreds of labs some focused on theory some on practicalities.

In other words, would you focus on the possibility of short-term disaster scenarios, or not?

Maybe, but I think you need a plausible case to verify one way or the other.  I'll ask you a plausible one, how much effort should we be putting into research to try and head off a Yellowstone Caldera event?  Pretty much a super calamity that's guaranteed to happen, should we divert resources from caring for the poor to head it off?  Strip the soldiers of new gear to focus on it?  Aren't we killing the future by not doing so?

I mean the reality is that we could be destroyed by an inbound meteor that's coming in from our polar regions two years from now.  It could be something that's completely avoidable if we devoted 90% of the world's GP to detection and elimination of potential meteor threats.  An extinction level meteor event is something that we have great certainty will happen again.  But even against a very real threat we have no choice but to weight the probability against our current needs.

So basically, if you're going to criticize how the science is done on the macro scale situation, how would you do it better?

I find it kind of depressing that you guys think I'm critical of  science (or at least anymore critical than a critical thinker should properly be), I'm critical of the misunderstanding of science and the arguments by authority that aren't supported by the science.  Computer modelling is a great thing, and very likely the only thing until we discover a way to travel on an interstellar or inter-dimensional basis, if we ever do, but it's still a computer model.

Machine learning has really improved in just the last few years, maybe in once that filters through, and much of the human bias is removed the models will be stronger.

[DonaldD]

Seriati, climate models are used to identify ranges of possible future outcomes based on various input parameters.  They are NOT the basis for determining causation - they do use that causation as one of many bases for their calculations, however.

The causation you are looking for has been demonstrated by several types of empirical evidence, confirming predictions that were made (and fairly accurately) way back in the 19th century.  Primarily, that evidence comes from observed changes in the frequency and energy of downward longwave radiation measured at the Earth's surface, and satellite measurements of upward longwave radiation - and all this predates the first significant computer climate model.

[TheDrake]

Part of where the science takes the brunt of policy is that the policy, for lack of a better word, sucks.

Fossil fuels are better argued against based on respiratory disease than on ice sheet shrinkage. Look at freaking Beijing, those people don't care if it got warmer, they are being asphyxiated.

Stupid advice comes from people who know nothing. Better turn that light off, Charlie, because global warming! When even if every residence in the US went dark for an entire year it wouldn't make a dent in the equations. It's more like a religion than science - the religion of conservation.

The international mechanisms are idiotic photo-ops where everybody promises to "do better" with no actual consequence for not doing better. Make a WTO rule around emissions and countries might pay attention.

Other remedies involve putting taxes on energy producers, but then leave all kinds of credit loopholes where they don't have to reduce anything as long as they trade on a flawed artificial market. Or giving away dollars to help people install inefficient technology that doesn't get us to break-even.

Then there's goofy stuff like the Solyndra loan guarantee. I don't even want to rehash that.

All of which makes me really sympathetic to the people who want to stop all that madness by pretending the Earth isn't getting warmer or that it isn't a problem or that it is unsolvable. I'd love to rally behind a message that says "Global warming is real, it is a threat, and all the proposals suck!"

I'd like to see more talk about sequestration. Let's get Monsanto working on some kind of crazy algae that gnaws through CO2 like cotton candy and sinks to the bottom of the ocean with the carbon. Let's create an international program to advance renewable technology that will enter into the global public domain. Let's scrap environmental rules that delay deployment of wind farms, tidal power, and hydroelectric. Let's do something that isn't half-assed nonsense, and frame it in a way that people aren't giving things up, but gaining them instead. Let's promote telecommuting in some way to reduce the traffic on the roads. Anything more creative and politically viable than "Let's jam up energy producers so everyone can pay more for energy and every good and service that exists, while not really moving the needle on global emissions because nobody else is doing it."

[velcro]

Seriati has not actually answered most of the questions posed to him, simply, deflected, insulted, or ignored.

I politely requested a link to papers he mentioned if possible. His response was that I did not supply links.  Links he never asked for.  This is obnoxious and dishonest.

He told me "the world will go on innovating without your contributions".  I have several patents, demonstrating that I have made contributions, and his insult was contrary to reality.  He has refused to reveal how many he has, which leads me to believe he has none, but refuses to admit it.  So that insult is empty and uniformed. And his lack of ability to admit any fault is plain to see.

He said

when people cite to "obvious" reasons, its really code for "I believe this to be true but can't explain why," or in other words, this is an axiom and you're a big silly head for not having the same axiom.

I said they were obvious reasons because they are obvious.  I did not want to waste time with something so self-evident. If someone besides Seriati says that the reasons why newspapers have editors and fact checkers are not extremely obvious, let me know.  If nobody pipes up, we will all know that Seriati's comment is pure unadulterated rubbish, intended to distract from the weakness of his argument.
 
And finally Seriati wrote

If you want to provide some facts, I'll be happy to address them.  Maybe you can point out where you think you did so, cause all I say is pretty much unvarnished opinion and demands that I provide facts for you.

Funny slip there, I think maybe he meant "all I see".

But as anyone with interest in reality instead of distractions and insults can see, I presented facts, quoted below.

"Big Data Analytics" are methods for drawing trends and conclusions from very large sets of data.  It is a tool for visualizing and organizing data, no more, no less.  If anyone were to use such a tool, and were to write a paper based on the results, that paper would be just as prone to error and bias as a paper that simply collected "small" data, graphed it, and drew conclusions.

This is in response to Seriati saying that Big Data Analytics is an alternative to peer review.  If you know anything about big data, you understand that this is nonsensical.  It is like saying that Excel is an alternative to peer review, because it generates so much knowledge.  Rather than admit that he made an error, Seriati turns to belittling, distortion, and fabricated demands.  I didn't demand facts, I requested that he correct my facts if they were wrong.  I can not think of any honorable, honest reason why an intelligent person would take the words "Please correct my facts" (my actual words) and parrot them back as "demands that I provide facts".

This is argument in bad faith, clear and simple.  Seriati has every right to continue this practice, as he has for years.  I have the right to point it out.

Fenring asked

"So basically, if you're going to criticize how the science is done on the macro scale situation, how would you do it better?"

I asked that. 

Just provide a simple example of a system that is better.  Again, this is all in the context of climate change, so please make sure that the better system is applicable to that type of science.

Here is the response, if you can call it that.

What science would that be?  Is there a climate change experiment that has occurred of which I'm unaware?  Did you invent star travel while I wasn't looking?

I am sure he is intelligent enough to provide an answer, as long as it does not involve admitting any shortcoming.  Or he will respond with sarcasm, distractions, irrelevant questions, and offensive tone, as he did with me.  I hope he does provide a meaningful answer, because I am here to learn new things.  But when I try to learn from Seriati, I just learn how arrogant and obnoxious people can be.

Fenring wrote:

Practical and profitable application does not demonstrate any "results" except for a product. (velcro)

Tell that to engineers. "It works" is worth a whole lot, not only in terms of commodifying a discovery, but also in realizing what the first principles behind it might be.

As an engineer with 30 years experience, I disagree.  "It works" doesn't tell you a damn thing, except it works.  It doesn't tell you how to fix it if it doesn't work.  It doesn't tell you how to make it work better.  It doesn't tell you what conditions it will work under, and what conditions to be avoided.  Now a good engineer will figure these things out, often by trial and error.  And you can get practical and profitable applications at that point. But none of these things, in and of themselves, tell you why it works or how it works.   A really good engineer will try to figure it out, because it can eliminate trial and error, and future risks.  But it is not necessary.  If an engineer takes the step of realizing what first principles are, he is beyond practical and profitable applications, and into science.

Now a scientist, ideally, does not care about anything except how and why it works.  Once you figure that out, all the avenues are open for a wide variety of applications in ways that the engineer, and even the scientist never imagined was possible. 

[Fenring]

As an engineer with 30 years experience, I disagree.  "It works" doesn't tell you a damn thing, except it works.  It doesn't tell you how to fix it if it doesn't work.  It doesn't tell you how to make it work better.  It doesn't tell you what conditions it will work under, and what conditions to be avoided.  Now a good engineer will figure these things out, often by trial and error.  And you can get practical and profitable applications at that point. But none of these things, in and of themselves, tell you why it works or how it works.

We're getting a bit into the weeds on this one, but in practical terms 99.99% of science - physics included - amounts to little more than "it works". The 'why' is a question of going deeper and deeper in what the first principles might be, but to date pretty much all of our theory is (properly) built upon empirical observation and attempts to create theoretical frameworks to explain in more detail how it works, but still not why. A bit of confusion here could be around what "works" means, which doesn't have to be man-made devices. If we're going to consider everything from asteroids to particles as being systems with mechanical properties then what we can come up with right now is "aha! that's the manner in which they work", but still not the 'why' as in the REAL why. Constructing a device that works despite not knowing why it works is huge. Take for instance the EM Drive, which for all intents and purposes is akin to a discovered natural phenomenon. We seem to be able to say it works, but not why. But the realization that it works is huge! It does indeed tell you a 'damn thing', which is that there's something there to solve. Similarly for the guy whose chocolate melted because of microwave radiation. The "it works" (as in, something discernible happened that could be repeated) was the most important thing there, because it points an arrow at us needing to dig deeper there. The simple fact of being able to say a thing works, or it doesn't, speaks volumes about all manner of physical principles. But just to clarify in case we're speaking past each other, I don't mean 'it works' in the sense of putting together a clock and if it's put together foolishly it won't work but if it's done properly it works; that's just an issue of error in method. What I'm talking about is trying to figure out how to make things function that otherwise wouldn't. Some of physics, for instance working with Einstein-Bose condensates, involves engineering projects where devices have to be built to create remarkable conditions, and very often the hack-and-slash method of trying all sorts of things yields a workable answer. One attempt clicks in, and the 'it works!' alerts the team that they've hit something important. Invention as it leads to the backtracking to find the principles behind it, has always been a huge part of science. The fact that physics is becoming so esoteric that this is often not the case any more is often described as being highly problematic and borderline not science at all.

Getting back to the original point, it's non-trivial to create a product or perfect a process that 'works'. Sometimes it just means you did an old process more efficiently, and sometimes it means you made up a new process. Both are important to science and I wouldn't discount either as being highly relevant to the issue of climate science. How, for instance, could we create physical experiments or techniques that 'work' in the climate sciences? Drake gave an example just now of how Monsanto should develop and algae that eats CO2 - well if they did futz around and then find something it might well be a way of learning about how CO2 interactions in the environment work. You never know what deep principle can be gleaned from a commercial venture.

[TheDrake]

As an engineer with 30 years experience, I disagree.  "It works" doesn't tell you a damn thing, except it works.  It doesn't tell you how to fix it if it doesn't work.  It doesn't tell you how to make it work better.  It doesn't tell you what conditions it will work under, and what conditions to be avoided.  Now a good engineer will figure these things out, often by trial and error.  And you can get practical and profitable applications at that point. But none of these things, in and of themselves, tell you why it works or how it works.   A really good engineer will try to figure it out, because it can eliminate trial and error, and future risks.  But it is not necessary.  If an engineer takes the step of realizing what first principles are, he is beyond practical and profitable applications, and into science.

Seconded.

Experimentation can produce lines of inquiry. Let's put thing A into hundreds of growth mediums to see what grows fastest. Now we'll grow thing A in the best one. This is how a lot of drugs are discovered, and why Viagra was eventually sold for its side effects instead of its intended use.

But a more powerful approach is to then take that and understand why thing A grew best under those conditions or why Viagra gives people boners. It is more powerful to be able to predict what a chemical will do by basic principles than to just spray and pray that something works.

In my example, Monsanto would use genetic engineering in a highly sophisticated way based on their knowledge of which genes of which species can be manipulated to increase the conversion rate of CO2. This is not speculative on my part, it has been talked about for many years.

Plants, algae, and other organisms turn CO2 into fuel. Erb and his colleagues reengineered this process, making it about 25 percent more energy efficient and potentially up to two or three times faster.

Erb's new CETCH cycle requires 11 steps to turn airborne CO2 into a chemical called glyoxylate. Each of these 11 steps requires a molecule-transforming enzyme, and each enzyme was carefully selected from the library of 40,000 known enzymes. "Some enzymes are found in the human body, and gut bacteria," says Erb, others are taken "from plants, and microbes that live in the oceans and on the surface of plants."

Not Fake News

Without the basic science and experimentation, they wouldn't have been able to select from that library. Their results will make the next researcher's job even more targeted and refined. The editing and review process will ensure that related studies link via citation to their study, and suggest that the new researchers present their data in a similar fashion making apples-to-apples comparisons of the studies easier. If the new researchers don't live up to the standard of the original, they'll have to keep shopping journals with less prestige and impact until they find a taker. There is a reason why very few engineering companies do any basic research. They wait for academia to take the lead on a method to make memory cells, then they refine the ideas and make them profitable. They can't just wing it and try a bunch of different substrate geometries and hope for the best one to show up and explain their trial and error to stockholders.

[Greg Davidson]

Sorry I am missing much of this discussion - distracted by real life and will continue to be for the rest of the week. First, let me make a more clear concession - there are few rigorous analytic studies addressing peer review itself (such as https://pdfs.semanticscholar.org/dc1e/1acdc1d766a9fb36997d266d8d1b62f86dfb.pdf). The main issue I would be interested in (filtering out really flawed stuff so it doesn't drown the marketplace of ideas in distracting BS) is not the primary focus of the study.  Historically, the system of peer reviewed articles in journals (combined with the efforts within the science community to reproduce results) has been the dominant system worldwide, so the only alternatives are a few like Lysenkoism in the Soviet Union, not a good basis for comparison. My own experience with peer review comes from a decade or so at NASA, where I saw peer review both for basic science grants as well as for larger experimental missions (I was not an astrophysicist myself, but I saw how the process mediated the normal quirks of human behavior). My father edited an economics journal for 40+ years, he was an iconoclast, and so I grew up with a good understanding of the constraints driven by a mainstream conventional wisdom via peer review. But Seriati, I actually thought your question was a good one because I don't have a comprehensive, analytical response that proves peer review + reproducibility is better than a hypothesized alternative.

That's how I agree with one of your points. Let's see if you can do th same:

Do you mean the micro level studies, some of which are experimental but few actually test the climate in any way, which are data points for the macro level studies or do you mean the macro level studies, which are all modeled computer science and not experimental?

What knowledge do you have about macro level studies? Most climate studies are based on a vast number of experimentally collected data points - since you seem to think Big Data is a panacea, how many data points have to go into a study before it counts as Big Data?

Seriati, you also made some hyperbolic claim about how you didn't like climate policies. I am in a rush, but tell me how you feel abut the Montreal Protocol during the Clinton Administration that banned chemicals that depleted the ozone hole. Seemed to have been developed in the same way as all those other policies that you detest.

[TheDrake]

Montreal Protocol was significantly different. It was binding and contained financial penalties, and it called on every nation to reduce CFC by an equal percentage. It didn't let countries "buy" ozone credits from other countries to make their goals. Minimum ozone levels had dropped by 50% over 15 years, a far more obvious and less debatable trend (though the chemical companies didn't think so). Every country is in compliance. Also, reading through, this wasn't during clinton's term as it entered into force in 1989 after agreement in 1987. So that was a Reagan thing, although it later went through revisions.

[Seriati]

Seriati has not actually answered most of the questions posed to him, simply, deflected, insulted, or ignored.

Had to go back and re-read the thread to try and figure out why you seem to think I urinated in your cheerios.  Still don't see it.  I've answered any number of questions, written quite a  bit, and didn't really deflect or ignore anything that I can see.  I was snide in some places, which doesn't cover me in honor, but I certainly didn't start the insult by snideness (have to credit Greg with that with the tribalism claims).  The only question I know of that I didn't answer, or believe was already answered, was this:

I politely requested a link to papers he mentioned if possible. His response was that I did not supply links.  Links he never asked for.  This is obnoxious and dishonest.

I didn't save any links to what I was reading, so to provide them I'd have to try to find them again.  I didn't do this for two reasons:  First, I find the way you respond to me generally to be rude and it doesn't incline me to be particularly cooperative with you, and second, they aren't particularly relevant (they would be if I were trying to prove peer review was invalid, but I've never made that claim, or anything plausibly related to it).

He told me "the world will go on innovating without your contributions".  I have several patents, demonstrating that I have made contributions, and his insult was contrary to reality.  He has refused to reveal how many he has, which leads me to believe he has none, but refuses to admit it.  So that insult is empty and uniformed. And his lack of ability to admit any fault is plain to see.

I didn't refuse, I have zero patents.  Congratulations, you win the internet, do you want to cite to your advanced degrees and tell me how your father can beat up my father (pretty easy of course, my father's pretty old at this point).

Seriously though, congratulations on having patents, but I'm not sure what that has to do with advances related to Big Data analytics.

He said

when people cite to "obvious" reasons, its really code for "I believe this to be true but can't explain why," or in other words, this is an axiom and you're a big silly head for not having the same axiom.

I said they were obvious reasons because they are obvious.  I did not want to waste time with something so self-evident. If someone besides Seriati says that the reasons why newspapers have editors and fact checkers are not extremely obvious, let me know.

Motte and Bailey argument, you said the reasons were obvious for why we needed copy checkers, and that it "doesn't work for science, for exactly the same obvious reasons."  That's quite a different claim.  Pretty much, one could read that to be a statement that the primary function of the peer review process is to correct spelling and fact check on quotes and third party claims, which is literally not the case.  In fact, it's hard to fathom what the "same obvious reasons" could be, when there's no credible claim that the peer review process is to clean up articles and head off defamation claims.

I've noted that when you really get going on trying to "make a point" about how I misconstrue things and dodge, you get highly selective in how you trim your own quotes, often dropping the relevant part to understanding what I was actually responding to.  It makes me wonder if I'm feeding a troll when I respond.

If nobody pipes up, we will all know that Seriati's comment is pure unadulterated rubbish, intended to distract from the weakness of his argument.

Quite the substantive rejoinder.  Were you not the same person that was lecturing me about insults not two quotes above?
 

And finally Seriati wrote

If you want to provide some facts, I'll be happy to address them.  Maybe you can point out where you think you did so, cause all I say is pretty much unvarnished opinion and demands that I provide facts for you.

Funny slip there, I think maybe he meant "all I see".

Hey look, I'm reasonable, I agree, that was a funny slip.

But as anyone with interest in reality instead of distractions and insults can see, I presented facts, quoted below.

"Big Data Analytics" are methods for drawing trends and conclusions from very large sets of data.  It is a tool for visualizing and organizing data, no more, no less.  If anyone were to use such a tool, and were to write a paper based on the results, that paper would be just as prone to error and bias as a paper that simply collected "small" data, graphed it, and drew conclusions.

Well, you presented some facts, not exactly relevant ones, but hey, at least there are a few facts in there.  I think you are misconstruing what error is in this context, but I'm happy to admit that humans are fallible, and that's equally true if they write a paper based on an experiment as if they write it on a derived correllation.  Of course, it's not clear that Big Data papers will be constructed or written in the same way, or that peer review will have a relative context (exactly how will it work, when the reviewer doesn't share the same data pool, analytical tools, or background and the paper isn't about causation?  What exactly are they going to do?).

This is in response to Seriati saying that Big Data Analytics is an alternative to peer review.

I see where you went wrong.  I made that claim in a limited context, one that I was asked specifically to engage in.  To attribute as a substantive argument that I think it would be superior in all cases or somehow could be a complete alternative is your straw man, not my claim.

I was asked for some possibilities against which the peer review process could have been compared for falsification purposes of Greg's claim, I threw a few out of a hat.  Do you think I'm unaware that Big Data analytics serve a different master?  The kind of research Greg's talking about it hypothesis driven, with testing to validate or invalidate, with an ultimate goal of showing causitive relationships.  Big Data is the exact opposite, it's data driven seeking to find correlations without regard to cause, what scientists used to refer to as "spurious" correlations.  It's a completely different model of advancing knowledge, but like any logical puzzle it can't generate new information, it can only reveal information that was already there but was too hard to see (classic research can do both).

If you know anything about big data, you understand that this is nonsensical.  It is like saying that Excel is an alternative to peer review, because it generates so much knowledge.

If you "know anything about big data" you'd understand that comparing it to Excel was beyond absurd.   Lol, I'm guessing your banking on no one else knowing anything about big data.

Rather than admit that he made an error, Seriati turns to belittling, distortion, and fabricated demands.

I'm pretty open to admitting errors.  However, I need to see evidence of the same, rather than just attempts at bullying and belittling, how many patents did you have again?  lol.

I didn't demand facts, I requested that he correct my facts if they were wrong.  I can not think of any honorable, honest reason why an intelligent person would take the words "Please correct my facts" (my actual words) and parrot them back as "demands that I provide facts".

Maybe you should reread the thread.  You chose to jump into an  argument and misconstrue what was going on.  Effectively Greg was making a claim about which camp is more tribally inclined and I was pointing out that he was making a claim that I knew he hadn't verified.  Not because I have a substantive problem with what he was claiming, but as evidence of how he was overlooking his own tribalism.  Even if I'm responding in broken up points, it's still in context of the entire conversation, where you were in fact asserting that I believed in a strawman argument about the overall validity of the peer review process.

This is argument in bad faith, clear and simple.  Seriati has every right to continue this practice, as he has for years.  I have the right to point it out.

Feel free.  I think it would be more convincing to show where I'm wrong, but I agree that's hard work.  Far easier to just to assert it and accuse me of failing to prove otherwise.

Just provide a simple example of a system that is better.  Again, this is all in the context of climate change, so please make sure that the better system is applicable to that type of science.

Here is the response, if you can call it that.

What science would that be?  Is there a climate change experiment that has occurred of which I'm unaware?  Did you invent star travel while I wasn't looking?

I thought that was clear.  In any event, the point is that peer review of climate science, is peer review of a computer climate model.  There's no actual experiment, what there is, is a simulated experiment based on algorithms designed or cribbed by the climate scientist.  If you understand - anything - about how a computer works, you have to understand that they are deterministic.  They can't produce a result that contradicts their inputs.  That means that they are absolutely incapable of returning a result that is inconsistent with the construction of their authors (unless the authors are so incompetent they don't understand how their own data and algorithms will interact).  What we understand of climate forcings mean there is no choice in how they will come out.  What we don't understand about climate is the factor of why they may not be correct.  Effectively, this science is highly prone to systematic bias, because to be qualified you have to have been trained in the bias (and to be clear I don't mean you have to be biased to accept global warming, only that you have to be biased to accept the validity of modelling and the data inputs to include).

What percentage of knowledge on how the climate works that we currently have do you think will ultimately turn out to be correct?  50 years from now, what will have been disproven?  100 years?  2000?  The science is literally the best we have, but so was was leaching at one point.

[Greg Davidson]

the point is that peer review of climate science, is peer review of a computer climate model.  There's no actual experiment, what there is, is a simulated experiment based on algorithms designed or cribbed by the climate scientist.

There's no experimentation in astronomy or evolutionary biology either. Would you accept astrologers or creationists with their non peer reviewed materials as being of similar value to that of actual scientists in those fields?

[Fenring]

the point is that peer review of climate science, is peer review of a computer climate model.  There's no actual experiment, what there is, is a simulated experiment based on algorithms designed or cribbed by the climate scientist.

There's no experimentation in astronomy or evolutionary biology either. Would you accept astrologers or creationists with their non peer reviewed materials as being of similar value to that of actual scientists in those fields?

Well, there is experimentation in astronomy, meaning tests are made using telescopes, EM detectors, and other gadgets in order to examine things that come to the Earth from space. What we can't do is send a probe into the sun yet or into Jupiter, and certain all we have to go on from far away is EM radiation, so that's limiting but still allowing for a remarkable amount of empiricism. And luckily for astronomers there's feedback between that discipline and cosmology, physics, and others, so that the ways in which we come up with hypothesis can come from various angles. For evolutionary biology we can examine evolution in simple organisms but not in species with very long breeding cycles, so that's a hard one to create a data set of in terms of human study. It's not literally impossible, but it will take eons so for now we have to take some of that in terms of being a thought experiment.

A better example would be a field like string theory, where as of yet there's no known experiment that can test individual claims made. And yet even so the claims themselves are merely models as attempts to fit data that we do have, so at its most basic level even these currently untestable fields are inherently based on previous empirical data.

But the problem with climate science and other complex systems analysis (which includes economics) is that there's too much unknown because it's not enough to see how smaller systems work and expand the theory to bigger ones. It's the actual specifics of the real system we have that's the issue, in addition to be able to compute complex systems. In other words, you'd have to completely understand the mechanics of the physical layout of the real system to an outstanding degree of nuance, and then be able to compute within that framework. We're soooo far from that right now, and shy of that I feel like it's not realistic to claim we can model Earth's climate with any kind of predictive power. I have a hard time believing that we'll ever be able to without much more advanced computers, for one thing.

To take a 'science' that's even crazier than this to calculate, imagine Asimov's psychohistory and how insanely advanced our science would have to be to compute that rabbit hole! You'd need to master brain science, consciousness, physics, environmental science, astrophysics, etc etc to be able to pull that one off. But it's really just along a track as compared to climate science, where the more complex the system is, and the more the moving parts are hard to understand (like the brain, for instance), the further away we are from being able to go there. I presume by hypothesis, by the way, that there's such a thing as understanding the brain fully, but anyhow you get the gist of what I'm saying.

[DonaldD]

Those are observations.  They are tests in exactly the same way that observations of changes in long wave radiation observed at the surface and exiting the atmosphere are tests.  And those observations are used in the same way to invalidate or support climate related theories.

[Seriati]

the point is that peer review of climate science, is peer review of a computer climate model.  There's no actual experiment, what there is, is a simulated experiment based on algorithms designed or cribbed by the climate scientist.

There's no experimentation in astronomy or evolutionary biology either. Would you accept astrologers or creationists with their non peer reviewed materials as being of similar value to that of actual scientists in those fields?

First of all no.  Creationists and astrologers are not actually applying the scientific method, and their assertions are not of similar value in that context (on the other hand, to a believer in the power of astrology, what use is an astronomer's paper?  That's a matter of relative context, but I agree for understanding of the universe will live in the scientific world).

Astronomy is a funny case.  I love astronomy, but mostly what I love about it are the stories astronomers tell, there's a lot of writing on the topic that is pretty much speculative hard science fiction.  I think we all realize with Astronomy that we don't have any way to directly prove or test many of the theories, and that something that "doesn't fit with our current understanding" is always cropping up and causing us to rethink what we think we know to our benefit.  Was reading an article on a rapid SuperNova that was causing just that kind of rethink earlier this week.  It's rather that astronomer's make claims about "consensus" or "settled science" and that's perhaps because their discoveries have no political arguments to support or deny (well other than for the flat Earthers I suppose).

Evolutionary biology is another fun one.  It's difficult to test on a macro scale too, though we can run many easy micro tests that support the general theory.  Specifically, we can demonstrate as a local condition that natural selection will weed out gross disadvantages in an animal.  It's a lot harder to demonstrate that a subtle gradual difference will do the same.  I mean think about the moth case study where they showed how they would quickly move between their white and gray conditions depending on environmental factors.  It's stands to reason, but is much harder to prove, that they would move between gray 46 and gray 42 for similar reasons.  The theories necessity requires that any preference be rewarded to the point where eventually the non-preferred genes no longer express (which demonstrates oddly that the white v. dark condition may not be as decisive as it appears given the potential for environmental variation, or that it takes far longer for a previous trait to be selected to extinction in a population).  How long does it take?  Well we have an N=1 situation, which means we have to assume that the length of the history of the Earth is at least sufficient for it to have occurred to the extent of fully forming every existing organism (but that's untestable ATM).

Another thing that has to be true is that the random mutation process has to be capable of generating, over time, mildly preferential traits that will in fact result incrementally in complex structures that are highly adapted to an environment.  We have to assume that this is true ATM, as we haven't data on every organism's mutation history, though with increases in genetic sequencing its certain possible we'll be able to demonstrate that it has a occurred in a population in the future.

I do think it's interesting that the only evolutionary theory that is currently provable - albeit on a limited scale - is a version of intelligent design.  Not the "creationism version" that's reliant on an outside actor, but the simple scientific version that acknowledges that we have through science created mutant organisms by directly manipulating their DNA in ways that nature has not to create organisms that are more (or less) adapted to environments that we have modified.  We've proven that intelligent design can in fact occur (but, it's not possible to prove it did or did not occur in our own past at our current level of technology).

[Seriati]

Seriati, you also made some hyperbolic claim about how you didn't like climate policies. I am in a rush, but tell me how you feel abut the Montreal Protocol during the Clinton Administration that banned chemicals that depleted the ozone hole. Seemed to have been developed in the same way as all those other policies that you detest.

By the by, I'm not ignoring this point, I just don't know a tremendous amount about the specifics of the protocol.  It does look like the original protocol and early amendments were much more science driven than political, but again I have to caveat that by limited knowledge, and thus actually resulted in an environmental good.  I think the future of that may be in question as it looks like the more recent amendments serve a different master requiring harsh cut backs in the first world while allowing China (2nd largest economy in the world) and 3rd world countries to increase production for decades - my prediction, again if the initial look is correct, is that these provisions will be far less successful and actually harm the environment.

[Wayward Son]

But the problem with climate science and other complex systems analysis (which includes economics) is that there's too much unknown because it's not enough to see how smaller systems work and expand the theory to bigger ones. It's the actual specifics of the real system we have that's the issue, in addition to be able to compute complex systems. In other words, you'd have to completely understand the mechanics of the physical layout of the real system to an outstanding degree of nuance, and then be able to compute within that framework. We're soooo far from that right now, and shy of that I feel like it's not realistic to claim we can model Earth's climate with any kind of predictive power. I have a hard time believing that we'll ever be able to without much more advanced computers, for one thing.

Could you elaborate on this idea, because I can't visualize exactly what you are saying.

Certainly climate is a complex system, much like aerodynamics, and the systems are simplified to some extent, like aerodynamics.  But the computer models are getting more precise (as I mentioned in a previous thread, computational resolution have come down from about 500 sq-km back in the first IPPC report (FAR) to 110 sq-km by the fourth IPPC report (AR4)) and are taking into account more systems (mid-70's models were based only on weather forecasting inputs; by the mid-80's, they added influences from clouds, land surfaces, and ice; by the FAR report, the influences of the ocean surfaces was added; and by the AR4 report, volcanic activity, sulphates, ocean depth heat retention, the carbon cycle, aerosol influences, rivers influences, ocean circulation, atmospheric chemistry and interactive vegetation influences were also added.

With all these factors included, why do you think we are "soooo far" from having an adequate model?  What major factors do you think are missing?

And why do you believe the models are wholly inadequate that they do not have "any kind of predictive power?"  When we were designing aircraft in the 1950's, our aerodynamic models were very primitive.  They made huge simplifying assumptions about the interactions of air molecules.  AFAIK, they still do.  But that didn't prevent engineers from designing aircraft heavier than your house from flying without crashing on take-off.  We came pretty close, and had pretty good predictive power.  From the comparisons of the models to historical temperatures and the temperatures we are seeing now, it seems the climate models are coming pretty close, too.  Enough that we can say with a pretty high degree of certainly of what the general outlines of the result of CO2 increase in the atmosphere.

Then there is the most basic model which must be accounted for.  CO2 is a greenhouse gas.  CO2 is increasing in the atmosphere.  Therefore, it is trapping more heat.  Unless there is a mechanism that counteracts this effect, the Earth will warm.  And the Earth is warming.  These facts are not in dispute.  The only thing in dispute is if there is a mechanism that will counteract the heat CO2 is trapping.  And until we come up with a model that shows what this counteraction is--whether it is clouds that reflect more heat than they trap, or the oceans somehow storing the heat for centuries, or whatever--we have to assume that we are warming the planet, and CO2 is responsible.  And, if we don't have an adequate model, we therefore don't know what the effects will be, and so should assume the worst.

There are unknowns, and there are also unknown unknowns, but we can't wait forever to discover everything we don't know before acting.  The Earth is warming.  Something is causing that.  CO2 is by far the most likely suspect.  And we are the ones producing it.  It is in our own self-interest to not continue to increase the concentration of CO2 in the atmosphere, and to find ways to decrease it, lest we change our climate to the point where we find it hard to survive on this planet.  When our best estimates show that what we are doing has no effect, then we can relax and do what we've been doing.  But so far, they have not, and actually show that we are responsible.  Regardless of what might be missing in our models.

[velcro]

Seriati wrote

If you "know anything about big data" you'd understand that comparing it to Excel was beyond absurd.   Lol, I'm guessing your banking on no one else knowing anything about big data.

As described earlier, "Big Data" is a set of tools for manipulating large amounts of data. Those tools can be implemented in Excel quite easily.  The problem is that Excel is not designed to handle that quantity of data.  It is purely a matter of scale.

A snowblower and a snowplow are just different scale, as far as function.  If a snowblower is not a good tool for painting a house, neither is a snowplow.

If Excel is not good for replacing peer review, neither is Big Data.

The comparison is perfectly apt.

[Fenring]

But the problem with climate science and other complex systems analysis (which includes economics) is that there's too much unknown because it's not enough to see how smaller systems work and expand the theory to bigger ones. It's the actual specifics of the real system we have that's the issue, in addition to be able to compute complex systems. In other words, you'd have to completely understand the mechanics of the physical layout of the real system to an outstanding degree of nuance, and then be able to compute within that framework. We're soooo far from that right now, and shy of that I feel like it's not realistic to claim we can model Earth's climate with any kind of predictive power. I have a hard time believing that we'll ever be able to without much more advanced computers, for one thing.

Could you elaborate on this idea, because I can't visualize exactly what you are saying.

I'd be happy to. And I don't think this is an obvious or evident thing to try to define either. By "complex system" I don't merely mean that there are several moving parts. Your example of an airplane also has various moving parts. However in the case of a piece of engineering like that the moving parts have very little fluidity to them (literally gases and liquids), and the fluids that go through it are channeled very carefully in extremely streamlined corridors (without intense variations in contour). But if you want to study the workings of the "machine" called Earth, you need to fully chart the corridors and see how they work down to the finest detail. You can't just know that there's a current going through the Gulf Stream; you need to literally chart every single eddy, current, wave, rock, chasm and peak at the ocean floor level, the heat sources coming from the interior of the Earth, every single measure of pressure in every square millimeter of the planet's oceans and skies...do you begin to see what I mean? Learning Earth's climate function is not simple a question of understanding basic fluid dynamics in a lab (which we don't) so that a simple system, like even a spherical tank, can be understood to satisfaction. No, it's much worse than that. We'd need to be able to take a full grasp of fluid dynamics and then map out the precise layout and contours of the tank called Earth to know its individual contributing factors, and *then* plug the equations (which we don't have now) into THAT system. What I'm talking about is probably hundreds of years away, which is to truly map and chart Earth's climate in real time. This is Star Trek kind of stuff.

What we can do now is simply to look at previous trends, try to draw very, VERY broad generalizations from that, and create models that somewhat follow those past trends, assuming that they'll continue. But the problem is that the past trends cannot be understood in terms of *how*, only in terms of "it happened." It's basically like watching a series of horse races and then modeling the behavior of the various horses on various tracks, but not knowing anything about terrain, horse breeding, jockey health,  etc etc. These models may sort of be able to be made to fit the horse performance in the past, but the predictive power is not going to be any better than what we can do in horse races, which is to set odds. But *this is not science*. It's basically using a casino approach to play the odds in the house's favor so that no matter which horse wins the house will, on average, make winnings. But it *is not* a way to actually determine which horse will win the next race. It doesn't even pretend to claim which horse will win; the spread of odds is merely a buffer in which the house will make more than it loses over time. But in climate science, which works the same way - by seeing what's happened so far without knowing why in precise detail - the whole point of the exercise is in predictive power. Granted, it's still better than nothing to be able to lay down some odds, but actually here's the funny thing: since there are so many more moving parts in play in a planetary ecology than there are in a single horse race, the chance that 'the house goes bust' increases to the point where it would be foolish to take bets on it at all.

I'm not at all arguing, by the way, that we should give up. I think it's good to keep trying as we're doing and see how much better at it we can get. In the case of greenhouse gasses maybe it's possible that it really is as simple as "more CO2 = hotter Earth", and that this is bad for us. But that's a different matter from suggesting exactly how to stop it happening in a way that creates more benefits than harms. What if we create a "CO2" suckers, as Drake supposed we could? Do we know what that would actually do to the entire system? Could it have unintended effects on the global system? It's kind of like what happens when a minor pest is introduced into an ecosystem that is supposedly meant to kill a very harmful other pest. Well these ecological experiments can result in very unexpected side effects, and whether or not they work sometimes it should be clear that there's truly no way to know what will really happen when making a move like this. You do it and then you see. The drawing board will not ever tell you the real result because it's too chaotic, and this is just in the 'tiny' sphere of agricultural science. When talking about making moves on the global scale I can imagine quite a lot in terms of bad things that can happen from immature attempts at geoengineering. Likewise, I can imagine all kinds of harms that could come from adopting CO2 treaties that are well-intentioned but poorly thought out.

I am actually a pretty big environmentalist and hate the idea of polluting, and yet I have been mostly disappointed by the general attitude and approach to the issue of safeguarding the environment by the AGW crowd. Yes, they have met so much opposition that I can see why it would divert a lot of energy just trying to convince the other side. But actually, why bother? The better thing would be to show how to take initial steps towards repairing it that are intelligent and not of a partisan nature. I doubt very many people are "for" pollution and emitting lots of random gasses into the air and waste into the water, so it's probably a bipartisan issue to an extent to try to keep the Earth clean. But it doesn't help that not only has it been a partisan issue anyhow, but that experts are all too keen to trump up and basically fake how good their models and understanding are. The same thing happens in economics analysis, where fancy analyses are made, and then when real life doesn't go how they predicted all of the explanations and the "you don't understand"s come out. Oh, of course you would have been able to predict the final result IF...  

[Seriati]

Fenring, I'd love to say that was a short and excellent summary, but I'm just going to have to go with excellent.   

[TheDeamon]

In other words, would you focus on the possibility of short-term disaster scenarios, or not?

Maybe, but I think you need a plausible case to verify one way or the other.  I'll ask you a plausible one, how much effort should we be putting into research to try and head off a Yellowstone Caldera event?  Pretty much a super calamity that's guaranteed to happen, should we divert resources from caring for the poor to head it off?  Strip the soldiers of new gear to focus on it?  Aren't we killing the future by not doing so?

Yellowstone is a "fun" one in that we potentially have a solution for it already. At least so long as the intervention doesn't  trigger (a likely to be smaller) events on its own.

The "solution" is to make the "hot spot" less hot. Which means deploying geothermal power in that region on a very massive scale.

Of course, colder rocks mean less geothermal activity in Yellowstone,  so the geysers, mudpots, and hot springs would all diminish as well. Flipside is colder rock is more brittle so you could potentially instigate a caldera (collapse) event through doing that.

And the amount of heat involved is immense, it would likely take centuries to (safely) stabilize the hotspot. So if it was due to go off sooner rather than later, we may already be screwed.

[D.W.]

Maybe we can turn it into some sort of death star super weapon and vent the heat faster than that!   Surely we can persuade Trump such a "star wars" weapon is a sound defense expenditure.   

[NobleHunter]

Um... I'd like to point out there's nothing to shoot with a Death Star super weapon. Except us.

[D.W.]

Pffft, the same could be said for modern nuclear arsenals.  Why do you hate progress?  Are you an alien sympathizer? 

[Wayward Son]

But the problem with climate science and other complex systems analysis (which includes economics) is that there's too much unknown because it's not enough to see how smaller systems work and expand the theory to bigger ones. It's the actual specifics of the real system we have that's the issue, in addition to be able to compute complex systems. In other words, you'd have to completely understand the mechanics of the physical layout of the real system to an outstanding degree of nuance, and then be able to compute within that framework. We're soooo far from that right now, and shy of that I feel like it's not realistic to claim we can model Earth's climate with any kind of predictive power. I have a hard time believing that we'll ever be able to without much more advanced computers, for one thing.

Could you elaborate on this idea, because I can't visualize exactly what you are saying.

I'd be happy to. And I don't think this is an obvious or evident thing to try to define either. By "complex system" I don't merely mean that there are several moving parts. Your example of an airplane also has various moving parts. However in the case of a piece of engineering like that the moving parts have very little fluidity to them (literally gases and liquids), and the fluids that go through it are channeled very carefully in extremely streamlined corridors (without intense variations in contour). But if you want to study the workings of the "machine" called Earth, you need to fully chart the corridors and see how they work down to the finest detail. You can't just know that there's a current going through the Gulf Stream; you need to literally chart every single eddy, current, wave, rock, chasm and peak at the ocean floor level, the heat sources coming from the interior of the Earth, every single measure of pressure in every square millimeter of the planet's oceans and skies...do you begin to see what I mean? Learning Earth's climate function is not simple a question of understanding basic fluid dynamics in a lab (which we don't) so that a simple system, like even a spherical tank, can be understood to satisfaction. No, it's much worse than that. We'd need to be able to take a full grasp of fluid dynamics and then map out the precise layout and contours of the tank called Earth to know its individual contributing factors, and *then* plug the equations (which we don't have now) into THAT system. What I'm talking about is probably hundreds of years away, which is to truly map and chart Earth's climate in real time. This is Star Trek kind of stuff.

I don't see how this differs from the fluid dynamics of the atmosphere, Fenring.  Doesn't everything you say, about needing to know every single current, eddy, rock, chasm and peak apply to the air around us?  After all, how heat moves through the air is a huge part of understanding the climate system.

Since we don't have that precise level of understanding, doesn't that imply we have no good formulas or understanding of how airplanes fly?  I mean, how can know how an wing will react in the sky until we know all the eddy currents and downdrafts it would encounter.  How can we be sure a new $100 million jet fighter will take off at all?  Or won't crash when it reaches Mach 5?

Obviously, we do have a pretty good understanding of how wings work in the atmosphere even though we don't have all the details of what the atmosphere is like.  Sure, airplanes still crash when they encounter a strong downdraft while landing , but that doesn't mean we don't have an adequate understanding of the atmosphere to know how airplanes can fly, or how hurricanes develop, or when tornados are likely to occur.

Similarly, we don't need to know "every single measure of pressure in every square millimeter of the planet's oceans and skies" to come up with an adequate model that comes close to what the climate will do.  Because a model that only takes the pressure of every square 500 km of skies may not be that different from one that has every square millimeter.  Sure, it may be quite different, but that's why the results of the models are compared to the historical records.  It there is a major difference, it should show up by the results being way off from what actually occurred.

(Ironically, it wouldn't matter that much even if we knew every square millimeter of pressure in the oceans and skies, because climate is a chaotic system.  Unless you know the initial conditions of a chaotic system, you cannot reproduce the path the system takes.  And there is no way, short of a time machine, to know what the initial conditions of our atmosphere were.)

Many systems, especially chaotic systems like the climate, have what is called "attractor states."  These are relatively stables states of a chaotic system, where the output of the system stays within a certain range.  I believe most of the climate models are showing how our climate reacts within one of these attractor states.  Which means we have a general range of how far temperatures will vary and change over a long period of time.  Which is good enough to make predictions of what average temperatures we can expect a few decades from now if CO2 concentrations double or triple in our atmosphere.  Even if we can't say exactly what the worldwide temperature will be in the year 2068.

Oh, and what equations don't we have right now?  We have the equations for how each cubic millimeter of air reacts with each other, which, when you apply them, shows those cubes combining into currents, downdrafts, etc.  Remember, the formulas in fluid dynamics are approximations of the systems because we can't calculate every cubic millimeter of the fluids.  If we could, we would have better models of the fluid dynamics.  But since we can't, we use these equations which gives us answers which are not exact, but adequate to figure out how a wing will work in the sky.

What we can do now is simply to look at previous trends, try to draw very, VERY broad generalizations from that, and create models that somewhat follow those past trends, assuming that they'll continue. But the problem is that the past trends cannot be understood in terms of *how*, only in terms of "it happened." It's basically like watching a series of horse races and then modeling the behavior of the various horses on various tracks, but not knowing anything about terrain, horse breeding, jockey health,  etc etc. These models may sort of be able to be made to fit the horse performance in the past, but the predictive power is not going to be any better than what we can do in horse races, which is to set odds. But *this is not science*. It's basically using a casino approach to play the odds in the house's favor so that no matter which horse wins the house will, on average, make winnings. But it *is not* a way to actually determine which horse will win the next race. It doesn't even pretend to claim which horse will win; the spread of odds is merely a buffer in which the house will make more than it loses over time. But in climate science, which works the same way - by seeing what's happened so far without knowing why in precise detail - the whole point of the exercise is in predictive power. Granted, it's still better than nothing to be able to lay down some odds, but actually here's the funny thing: since there are so many more moving parts in play in a planetary ecology than there are in a single horse race, the chance that 'the house goes bust' increases to the point where it would be foolish to take bets on it at all.

If the models were only following "past trends," why do they need supercomputers?  Why do they runs the models dozens and dozens of times to see how they differ from one another?  If you only want to match the past trends, why do runs that you already know won't match it?  Just run the one that matches it the best and have done with it.  You'll save a lot of money on computer time that way.

As I said before, they don't compare their results to the past to make it match the past.  For a chaotic system, that would be foolish.  They compare it to the past to make sure they aren't deviating from it too much.  Because if the model is off, it should be off from the past trend, too.

Which means that what the models are doing is modeling the "how" of the system.  They are taking the basic knowledge of the system--the fluid dynamics, the thermal dynamics, the heat from the sun and the reflection of the light, and everything they can think of, create equations for them, and combine them to create a system like our climate system.  Yes, it is rather a hack-and-slash system.  No, it doesn't allow us to reduce all these equations to a nice, simple equation that allows us to see exactly how increasing CO2 by X amount will change the temperature by Y amount.  But it does give us some idea of exactly that.  It does give us a range of how much the system would be changed.  Although we may not be able to wrap our mind around it, the how is in there, in the equations we programmed into the computer.

We just need a supercomputer to give us the answers.

I'm not at all arguing, by the way, that we should give up. I think it's good to keep trying as we're doing and see how much better at it we can get. In the case of greenhouse gasses maybe it's possible that it really is as simple as "more CO2 = hotter Earth", and that this is bad for us. But that's a different matter from suggesting exactly how to stop it happening in a way that creates more benefits than harms. What if we create a "CO2" suckers, as Drake supposed we could? Do we know what that would actually do to the entire system? Could it have unintended effects on the global system? It's kind of like what happens when a minor pest is introduced into an ecosystem that is supposedly meant to kill a very harmful other pest. Well these ecological experiments can result in very unexpected side effects, and whether or not they work sometimes it should be clear that there's truly no way to know what will really happen when making a move like this. You do it and then you see. The drawing board will not ever tell you the real result because it's too chaotic, and this is just in the 'tiny' sphere of agricultural science. When talking about making moves on the global scale I can imagine quite a lot in terms of bad things that can happen from immature attempts at geoengineering. Likewise, I can imagine all kinds of harms that could come from adopting CO2 treaties that are well-intentioned but poorly thought out.

I am actually a pretty big environmentalist and hate the idea of polluting, and yet I have been mostly disappointed by the general attitude and approach to the issue of safeguarding the environment by the AGW crowd. Yes, they have met so much opposition that I can see why it would divert a lot of energy just trying to convince the other side. But actually, why bother? The better thing would be to show how to take initial steps towards repairing it that are intelligent and not of a partisan nature. I doubt very many people are "for" pollution and emitting lots of random gasses into the air and waste into the water, so it's probably a bipartisan issue to an extent to try to keep the Earth clean. But it doesn't help that not only has it been a partisan issue anyhow, but that experts are all too keen to trump up and basically fake how good their models and understanding are. The same thing happens in economics analysis, where fancy analyses are made, and then when real life doesn't go how they predicted all of the explanations and the "you don't understand"s come out. Oh, of course you would have been able to predict the final result IF...

Bad treaties are bad treaties.  Politics is another form of people trying to take advantage of a situation for their benefit.  But we need to keep focused on the essential problem.

CO2, from all we can know, is heating up the Earth.  As the Earth heats, deserts will become larger, glaciers and arctic ice will melt, more floods and wildfires will occur.  Sea levels will rise.  These are not things that are conducive to maintaining a population of over 7 billion people on Earth.  It is in all of our best interest to prevent these things from occurring.

If you don't like a policy, advocate for a better one that will do the job.  If you can't find a better one, or if you can't get enough people to agree to it, support the second best policy.  Or the third.  Or the fifth.  Because we are all sinking in this ship together, and doing something, even if it only helps a bit, is better than doing nothing, or hoping that something will come up in the future that will save us all.  Because doing nothing isn't going to help the situation at all.  And that something that will save you in the future may not show up, or may show up only after you're gone under.

[Greg Davidson]

I think we all realize with Astronomy that we don't have any way to directly prove or test many of the theories, and that something that "doesn't fit with our current understanding" is always cropping up and causing us to rethink what we think we know to our benefit.

Modern astronomy is built on direct tests that substantiate a vast majority of current theory. Of course, the excitement is always on the scientific frontier beyond what is readily testable, but there is an incredible amount of our knowledge of the universe that is substantiated by data that supports theory. As one example, there is an extraordinary amount of insight into the first three minutes after the creation of the universe based on theory substantiated by measurements of cosmic background radiation from the Big Bang

[Fenring]

I think we all realize with Astronomy that we don't have any way to directly prove or test many of the theories, and that something that "doesn't fit with our current understanding" is always cropping up and causing us to rethink what we think we know to our benefit.

Modern astronomy is built on direct tests that substantiate a vast majority of current theory. Of course, the excitement is always on the scientific frontier beyond what is readily testable, but there is an incredible amount of our knowledge of the universe that is substantiated by data that supports theory. As one example, there is an extraordinary amount of insight into the first three minutes after the creation of the universe based on theory substantiated by measurements of cosmic background radiation from the Big Bang

As much as I'm a fan of astrophysics we need to be fair about how much we really "know". What we seem to be good at is developing trees of inference based on some axioms that fit the data, and we can get a good amount of details into those inferential trees. But if the axioms prove to be wrong then it's all bogus and whatever new paradigm takes over scraps the old thing. In physics we can do many tests that give a fair amount of rigor to the basic assumptions, but with astrophysics and cosmology we really are grasping at faint bits of data to make huge conclusions.

Remember that scene in The Princess Bride where Vizzini is working through the logic to whittle the way down to Westley's plan? I more or less think that's what we're into right now in cosmology. Something like - on the one hand, if the expansion began at a single point then it no doubt exploded, in which case the heat density is surely in the cup in front of me; but on the other hand, we can't determine what physics is like at such pressures and so perhaps the cause of the expansion was something else, in which case the heat density is surely in the cup in front of you; but assuming there was no Big Bang, the cosmic microwave radiation is lying to us, to surely the heat density is in the cup in front of me; but what if the same physics that happens inside black holes was happening at the time of the Big Bang, and perhaps the matter emerged from a white hole, in which case the heat density is surely in the cup in from of you. Never go up against a Sicilian when an analogy is on the line! HA! HA!

You get the idea. Any number of tiny blips in the current theory may belie a gaping hole in the current mainstream paradigm, in which case a new theory could blow it right open. One tiny flip in interpretation of data and suddenly a "must have happened" turns into a "couldn't possibly have happened", just like in Vizzini's mental game. Add holographic theory and "information is lost in black holes" turns into "information is possibly not lost in black holes". We know so little about the Big Bang that we don't even know there was one. We don't even know how time would be measured in such a compact space with that kind of energy density; what happens if what we have is conspansion instead of expansion? Does that mean the Big Bang could have occurred infinity ago? How do we measure a changing measuring stick? What if matter can be created and thermodynamics is wrong? What if what we visually see as expansion is a holographic image of something entirely different, that has nothing to do with a universe that blew up out of nothing? All of these, and infinite more possibilities, ought to show that our current cosmological models are basically built on a house of cards that seem sort of reasonable but stand little chance of actually ending up as 'the story.' I'm not being a deconstructionist about this, btw, but I'm leery of claims about how much we know. I don't think we know very much yet! It's impressive compared to 200 years ago, but we've barely begun, I think.

[JoshCrow]

I am actually a pretty big environmentalist and hate the idea of polluting, and yet I have been mostly disappointed by the general attitude and approach to the issue of safeguarding the environment by the AGW crowd. Yes, they have met so much opposition that I can see why it would divert a lot of energy just trying to convince the other side. But actually, why bother? The better thing would be to show how to take initial steps towards repairing it that are intelligent and not of a partisan nature. I doubt very many people are "for" pollution and emitting lots of random gasses into the air and waste into the water, so it's probably a bipartisan issue to an extent to try to keep the Earth clean. But it doesn't help that not only has it been a partisan issue anyhow, but that experts are all too keen to trump up and basically fake how good their models and understanding are. The same thing happens in economics analysis, where fancy analyses are made, and then when real life doesn't go how they predicted all of the explanations and the "you don't understand"s come out. Oh, of course you would have been able to predict the final result IF...

Except in economics, as you well know, the theories have historically been used to make appropriate course corrections (see: Volcker taming inflation) that prompted short term economic pains but addressed long term problems.

Nevertheless, feet-dragging and second-guessing when it comes to the environment "because experts are not perfect" is almost like saying you'd rather have a known terrible thing happen than risk an unknown. It's "the devil we know", in your argument. Except that it's a hell of a devil. If we had another 100 years to study the problem before we took action, I might be there with you. Once you establish that the problem can self-accelerate because of feedback, the urgency increases to the point where dawdling - even with the reason of 'getting it exactly right' - could nonetheless be a catastrophic decision.

I think a better example than economics is nutrition science and 'poor health', which has been blamed on everything you can imagine (fat, salt, sugar, gluten), so we know the science is still rather immature, but there is still a broad agreement that "eating less and getting exercise" has positive health impacts. That fact has never really been shaken despite all the attempts to undermine it. The evidence is just too great. And getting on a treadmill and eating less cake sucks and nobody likes to hear it, but there it is.

[TheDeamon]

I am actually a pretty big environmentalist and hate the idea of polluting, and yet I have been mostly disappointed by the general attitude and approach to the issue of safeguarding the environment by the AGW crowd. Yes, they have met so much opposition that I can see why it would divert a lot of energy just trying to convince the other side. But actually, why bother? The better thing would be to show how to take initial steps towards repairing it that are intelligent and not of a partisan nature. I doubt very many people are "for" pollution and emitting lots of random gasses into the air and waste into the water, so it's probably a bipartisan issue to an extent to try to keep the Earth clean. But it doesn't help that not only has it been a partisan issue anyhow, but that experts are all too keen to trump up and basically fake how good their models and understanding are. The same thing happens in economics analysis, where fancy analyses are made, and then when real life doesn't go how they predicted all of the explanations and the "you don't understand"s come out. Oh, of course you would have been able to predict the final result IF...

Except in economics, as you well know, the theories have historically been used to make appropriate course corrections (see: Volcker taming inflation) that prompted short term economic pains but addressed long term problems.

Nevertheless, feet-dragging and second-guessing when it comes to the environment "because experts are not perfect" is almost like saying you'd rather have a known terrible thing happen than risk an unknown. It's "the devil we know", in your argument. Except that it's a hell of a devil. If we had another 100 years to study the problem before we took action, I might be there with you. Once you establish that the problem can self-accelerate because of feedback, the urgency increases to the point where dawdling - even with the reason of 'getting it exactly right' - could nonetheless be a catastrophic decision.

I think a better example than economics is nutrition science and 'poor health', which has been blamed on everything you can imagine (fat, salt, sugar, gluten), so we know the science is still rather immature, but there is still a broad agreement that "eating less and getting exercise" has positive health impacts. That fact has never really been shaken despite all the attempts to undermine it. The evidence is just too great. And getting on a treadmill and eating less cake sucks and nobody likes to hear it, but there it is.

Except you're ignoring "the other half" of this. Bad treaties are bad treaties. The Kyoto Protocol, all things considered, may very well have resulted in more CO2 emissions and other environmental damage than it mitigated. Ditto for its successor attempts.

Those efforts to reduce "first world CO2" often translated into production moving into the "third world" where things are far less restrictive in practically every way.

Yes encourage greater energy efficiency. Yes, encourage non-carbon emitting power options. Do NOT gut the more carbon efficient economies in favor of growing the more inefficient ones.

DO help those "inefficient" economies in becoming more efficient, you're likely to get more bang for you buck that way as well.

Ideally, we would have created a means for them to either "leapfrog" past the "burn coal" option or employ much cleaner variants(CCS). But that is not what was done.

[JoshCrow]

The Kyoto Protocol, all things considered, may very well have resulted in more CO2 emissions and other environmental damage than it mitigated. Ditto for its successor attempts.

Those efforts to reduce "first world CO2" often translated into production moving into the "third world" where things are far less restrictive in practically every way.

I'm happy to look at evidence for these claims. Ironically, the very uncertainty being used to argue against climate action probably makes it impossible for you or anyone to demonstrate that "Kyoto caused X" or that "efforts to reduce first world CO2" cause production to move to the third world. You're reduced to saying "well, maybe..."

[Fenring]

Except in economics, as you well know, the theories have historically been used to make appropriate course corrections (see: Volcker taming inflation) that prompted short term economic pains but addressed long term problems.

There are some small-term things we definitely know about economics, just like we can do a fairly good job monitoring the PH levels of a fish tank. It's a fluid system but in certain well-defined ways we can make adjustments that work. I wouldn't dispute that. It's the 'entire economy' as a system that I think we don't understand, nor how all the various perturbations add up and create feedback mechanisms with each other. We kind of figure they do, and feel they must, but that's different from actually calculating them.

Re: Volcker, that's a special case, or at least special in the sense of being an inevitable result of a post-wartime boom. The same always happens, which is that post-war inflation and boom cycle goes out of control and the brakes need to put on. Reagan was in a constant battle with the banks and the Fed to stop them from "zooming the economy" as he called it, which he hated them doing. In the 70's and early 80's there was also a serious conflict of theory going in in banking circles, with an insane policy called monetarism somehow gaining traction, and its adherents finally got their way through some black magic and the economy was getting killed. It's true that Volcker laid down the law and championed a different course backed by more traditional Keynesian principles, and that after a few years the recession that Reagan and the Monetarists caused (or even desired) turned back into slow but normal growth again. But it's more complicated than to just say Volcker tamed inflation as actually, if anything, he was against the policies that were initially introduced to totally throttle the economy and that created not just a contraction but a major recession. His goal at that point became to effectively get the bust cycle past him and to go for growth again. However I think I understand what you're trying to argue, which is that it is possible to use definite policy to make definite plans that have effects. However one thing study from that time period should show is the opposite of what I think you want to show: that so-called experts were all over the place in terms of their claims about (a) what could stop the inflation safely, and (b) what could then stop the recession effectively. They didn't know what they were talking about, any of them! And although Volcker was less of a radical compared to the Monetarists, even he was evidently taking gambles and hoping his proposals would work. He himself knew well it was still a shot in the dark. There was no guarantee, and certainly no sense in the market that it was evident to anyone collectively what to do. It would be like if the physics community had no group paradigm and each physicist had his own personal theory that no one else shared. That's a really crappy environment to be claiming that anyone has some special expertise that others should listen to. I mean, look even today at the success rate of 'expert' stock brokers; they barely do better than picking at random! 

Nevertheless, feet-dragging and second-guessing when it comes to the environment "because experts are not perfect" is almost like saying you'd rather have a known terrible thing happen than risk an unknown. It's "the devil we know", in your argument. Except that it's a hell of a devil. If we had another 100 years to study the problem before we took action, I might be there with you. Once you establish that the problem can self-accelerate because of feedback, the urgency increases to the point where dawdling - even with the reason of 'getting it exactly right' - could nonetheless be a catastrophic decision.

No, I agree with you. That's why earlier I asked Seriati how to account for the possibility of a looming crisis. However likewise I don't agree that "but there could be a looming crisis!" is automatically a QED to act as if there is one. Shooting yourself in the foot to prevent being shot doesn't help. And destroying economic health doesn't help climate research since in a recession or depression scenario the budgeting for research would be depressed along with it. Never take away the infrastructure that makes the work possible in the first place.

I think a better example than economics is nutrition science and 'poor health', which has been blamed on everything you can imagine (fat, salt, sugar, gluten), so we know the science is still rather immature, but there is still a broad agreement that "eating less and getting exercise" has positive health impacts. That fact has never really been shaken despite all the attempts to undermine it. The evidence is just too great. And getting on a treadmill and eating less cake sucks and nobody likes to hear it, but there it is.

If you want to use a nutrition analogy we can do that. Remember all the health fads that went on between the 80's and now? Many people, like my grandparents, studiously followed the latest "facts" about what to eat. And you know what? One week the 'experts' were saying "you must stop eating eggs! they will clog your arteries" and then the next week it's "wait! scratch that! they have the good cholesterol!" and then it would flip back to "avoid eggs!" the week after that. And then there's the vitamin E fiasco, and it goes on. Basically they didn't know jack but like all 'experts' pretended to have the keys to the kingdom because that's how they make their living. And then there's actual malfeasance, such as the food pyramid scandal - a true conspiracy to hurt the health of all Americans for money - which apparently the experts accepted as fact because they didn't know enough to be able to say it was false. So while you're totally correct that certain things are gimmes in nutrition, such as "do exercise, have a balanced diet", which is also true in climate science, like "try to pollute less, search for clean energy options." Can't argue with that. However the same danger lies in both where "good ideas" can actually be very bad. In health sometimes people think "I need to turn this around right now! Time to hit the gym!" and then they go full blast doing something new and injure themselves, or else go on some crazy diet which creates a vitamin deficiency, or various other things. You have people who went on Atkins (which took a relatively long time to prove as being dangerous) who cut out all carbs, and only later learned (if they did at all) that you mess up your liver if you force it to synthesize glucose since you're not eating any sugars. "Eat less sugar", how can you go wrong, right? Well you can, actually. A good diet doesn't just mean taking sudden action or implementing the first thing that comes to you. It can be far worse to "do something" if it's the wrong thing than to do nothing and let nature take its course. There are people who eat insanely fatty foods all their life and live to over 100 years old; there's a small town in Italy as I recall with a lot of insanely old people, and they eat in a way that Americans would call unhealthy; but is it? We don't even know! And the same is true with climate science. Some generally good directions include what I mentioned above, but just "doing something" can be far more harmful than letting the market adjust itself. You can already see research developing energy alternatives and some cities and even countries have gone green. And it didn't take a destructive treaty to do this, just the advancing of tech and natural incentive to switch over. And the major antagonistic forces, like Big Oil and Big Coal, will roll over at precisely the right moment and will suddenly be Big Windmill and Big Solar Panel, because you know they aren't going to play dead when a new market emerges. It's true that this does take a bit of a leap of faith, however we're seeing it born out already.

Generally I think the environment movement is a good thing, and the collision between that and big business eventually creates incentive - and a market - for green tech. If either side alone had their way it would indeed be dangerous for all of us.

[Fenring]

And sorry everyone for these walls of text. I try to avoid rambling but sometimes I don't know how to say these things more briefly without risking being unclear. It's a bit of a fault that I'm working on.

[velcro]

In the case of greenhouse gasses maybe it's possible that it really is as simple as "more CO2 = hotter Earth", and that this is bad for us. But that's a different matter from suggesting exactly how to stop it happening in a way that creates more benefits than harms.

The problem is that many people in power are actively preventing us from trying to figure out "exactly how to stop it happening in a way that creates more benefits than harms", because they deny the science.  Or because their donors will make more money in the short term if they do nothing.

[Wayward Son]

Any number of tiny blips in the current theory may belie a gaping hole in the current mainstream paradigm, in which case a new theory could blow it right open. One tiny flip in interpretation of data and suddenly a "must have happened" turns into a "couldn't possibly have happened", just like in Vizzini's mental game. Add holographic theory and "information is lost in black holes" turns into "information is possibly not lost in black holes". We know so little about the Big Bang that we don't even know there was one. We don't even know how time would be measured in such a compact space with that kind of energy density; what happens if what we have is conspansion instead of expansion? Does that mean the Big Bang could have occurred infinity ago? How do we measure a changing measuring stick? What if matter can be created and thermodynamics is wrong? What if what we visually see as expansion is a holographic image of something entirely different, that has nothing to do with a universe that blew up out of nothing? All of these, and infinite more possibilities, ought to show that our current cosmological models are basically built on a house of cards that seem sort of reasonable but stand little chance of actually ending up as 'the story.' I'm not being a deconstructionist about this, btw, but I'm leery of claims about how much we know. I don't think we know very much yet! It's impressive compared to 200 years ago, but we've barely begun, I think.

Fenring, can you name any scientific theory for which this doesn't apply?  Relativity, from which we were able to devise the atomic bomb?  Electromagnetism, from which we created our electrical infrastructure.  Quantum mechanics, from which we created solid-state transistors which enable you to have a computer on your desk?   There is always a possibility that our theories and laws are wrong, that there is something else we're missing, that some new idea will blow the current theory out of the water.

Take gravity.  Newton's laws were the standard for over a century, until Einstein came along and warped space-time.  But you know what happens if you drop out the parts of Einstein's equation that are pretty-much insignificant in our macro world at speeds nowhere near the speed of light?  You get Newton's laws.

Even if a theory is later supplanted by another theory, it doesn't mean the original theory was "wrong."  If it works well, it probably just didn't cover everything, and is still true within the original parameters.  It doesn't necessarily mean it was useless or didn't describe the universe.

Also, don't think so much about how many things a theory or hypothesis can't explain.  Think about how many things it can (and does) explain.  Take the big bang and cosmic background radiation.  List (in your head--too much to do on paper) how many frequencies of electromagnetic radiation exist between 2 and 22 1/cm.  Pretty much infinite, right?  Now list the possible intensities of those frequencies.  Pretty much infinity squared, right?  The curve could look like anything.

Now, look at the chart in the Wikipedia page.  Notice the two curves in the chart, showing the measured spectrum and that predicted by theory.  (Predicted before the spectrum was measured, I might add.)  Notice the error bars.  As the caption reads, "The error bars are too small to be seen even in an enlarged image, and it is impossible to distinguish the observed data from the theoretical curve."

What are the odds that the theory just happened to fit that curve by chance?  What are the odds that there is not some basis of reality that is described by the theory?  What the odds that we found something that is true in this theory?

No, we don't know anything yet.  In 200 years, we're probably going to be laughing at how ignorant we were now.  But that fact doesn't mean we don't know anything.  It doesn't mean Newton's theories weren't good enough to get a good idea of how planets orbit the sun.  It doesn't mean we can't say anything about black holes and other cosmic phenomenon.  We should stay humble, and curious, but not timorous.  Black holes exist.  The universe started with something like a big bang.  And if we are wrong, we are wrong through incredible bad luck, and any reasonable person would believe what we believe based on the data.  And the theories do a pretty damn good job of describing what we see.

IOW, it is science, not guesswork.

[Fenring]

WS,

There's a big difference between math and English-language narratives to explain what's going on. We tend to be pretty good at using maths to calculate observations we've taken and create graphs to chart them. But starting in the 20's the graphs and the common-sense narratives to explain what they mean started walking different paths. We're still collecting data in various ways, but our ability to say what's actually happening has hit a wall and we've had to use long trees of inference to interpret the data. In QM you can take the famous Copenhagen interpretation as an example of that, which has never been verified but whose thought-experiment properties are used to try to figure out what's going on.

What you need to remember is that hard science backed by observation and testing became to retain as a standard in certain fields as of the 20's, and these include cosmology, astrophysics, and some areas in physics. We have no hard tests of string theory; no good tests for the various attempts to explain quantum gravity; no tests for cosmological topology ideas; and so far not even any acceptable tests to even verify there's such a thing as dark matter. There are highly esoteric theoretical fields that mostly try to reverse-engineer the observations we do have into a model that fits them all, but there's no particular reason to expect that any one of these models is actually correct, but we're doing our best as we can. The Big Bang theory is one such among these, which seems like an ok explanation - other than that it's riddled with holes so far and imperfect - but it's one thing to say that we're continuing to explore the theory of the Big Bang, another to positively claim that we know all sorts of things about the Big Bang, as if it's an historically factual event we know happened. We don't.

So we are pretty good at collecting data and using it to good effect, but so far not so good at "what does it all mean." The lack of a unified physics is one of the sticking points here stopping us from being able to say much beyond that certain things work and others don't. Our data from observations will likely not be overturned, but our grand explanations of what they mean almost surely will be. The data can be supplemented with better data; but the actual explanations - yeah, most of them will turn out to be flatly wrong. Not updated, just wrong.