|
Post by sigurdur on Jul 7, 2009 0:36:04 GMT
SoCold: You may be getting your wish. The areospace dept of my local university is now working on a climate model that they say will discredit all AGW models.
The reason the prof that is doing it said he hadn't done it before is that the AGW theory was so silly and modeled so poorly that he never felt it would last any length of time.
The evidence against agw is overwhelming if one thinks about it. There is no hot spot, when one studies levels of co2 in the geological records, they always predict a cold snap. That is one of the stunning things that most people miss. Have you ever seen a graph of co2 and temp combined? The co2 keeps rising towards a peak, and the temp cools, yet co2 keeps increasing. I think it has something to do with that 800 year time lag does it not?
Anyways, that you for that woodfor trees site. I had not realized that the latter 1870's were warmer than our present temps.
|
|
|
Post by socold on Jul 7, 2009 7:19:02 GMT
So he's claiming to know the result before he's done the study? Interesting...
And noone has yet answered why a climate model hasn't already been made to show low climate sensitivity. If it's so easy to do why not put such an argument out there by doing it?
Either it isn't easy to do or noone can be bothered to do it.
|
|
|
Post by poitsplace on Jul 7, 2009 9:00:05 GMT
So he's claiming to know the result before he's done the study? Interesting... And no one has yet answered why a climate model hasn't already been made to show low climate sensitivity. If it's so easy to do why not put such an argument out there by doing it? Either it isn't easy to do or noone can be bothered to do it. They do, but low sensitivity implies low to negative feedbacks. You don't need to do any COMPLEX computer modeling. They just go to something like this... geosci.uchicago.edu/~archer/cgimodels/radiation.html...and find that .6C offsets CO2 levels of 600ppm. Now when you start throwing in wild assumptions of poorly understood feedbacks, it gets kinda complex (obfuscated and needs that big ol' computer.
|
|
|
Post by icefisher on Jul 7, 2009 11:33:11 GMT
It seems you are claiming that models would output low climate sensitivity if only the physics was done correctly. So lets go with this. Can you explain why the physics has not been done correctly? modelling has existed for decades, but not one person has tried to model climate accurately? All of them deliberately fudged it? Given how important skeptics find climate science for the sake of the economy, how come not one of them has bothered to end the whole AGW thing in one move by demonstrating that physics-done-right leads to models showing low climate sensitivity? How come not one of them has bothered to show it is at least a possibility? All it would take was a GCM that showed..a low climate sensitivity. My hunch is that they haven't done it, because it simply cannot be done. Ie there is no such error or bias in the physics which all the skeptics are aware about but those evil AGW scientists put in anyway. It's a myth. You missed the point Socold. Fact is you can't model climate because you cannot model clouds. So why would somebody want to tell you they can? You have fallen into a fallacy. If 12 people approached you with models of the value of the Brooklyn Bridge. . . .would you buy it because 12 people modeled it and nobody did that told you you should not buy it?
|
|
|
Post by socold on Jul 7, 2009 18:02:29 GMT
It's not a matter of whether the GCMs are right or wrong, that's another question for another day.
This question is about whether it is or is not possible to get low climate sensitivities from GCMs.
Consider the ramifications of either:
If it is possible to get a low sensitivity from a GCM then that throws the level of certainty of AGW projections out of the window. Afterall if the GCMs show that it's possible to get low sensitivities then the uncertainty widens to within what the skeptics claim. It would also demonstrate that the GCMs can be "tuned to show anything"
If it isn't possible to get a low sensitivity from a GCM then that throws the whole argument that GCMs can be tuned to show anything out of the window. In fact the greater people insist GCMs can be tuned to show anything the even more remarkable it would be if noone could tune model to show a low climate sensitivity.
On both cases skeptics have a strong motive to demonstrate low climate sensitivity from GCMs.
|
|
|
Post by icefisher on Jul 7, 2009 20:57:24 GMT
It's not a matter of whether the GCMs are right or wrong, that's another question for another day. This question is about whether it is or is not possible to get low climate sensitivities from GCMs. Consider the ramifications of either: If it is possible to get a low sensitivity from a GCM then that throws the level of certainty of AGW projections out of the window. Afterall if the GCMs show that it's possible to get low sensitivities then the uncertainty widens to within what the skeptics claim. It would also demonstrate that the GCMs can be "tuned to show anything" If it isn't possible to get a low sensitivity from a GCM then that throws the whole argument that GCMs can be tuned to show anything out of the window. In fact the greater people insist GCMs can be tuned to show anything the even more remarkable it would be if noone could tune model to show a low climate sensitivity. On both cases skeptics have a strong motive to demonstrate low climate sensitivity from GCMs. Let me see if I can summarize the gist of that argument: - Because skeptics are allegedly motivated to demonstrate low climate sensitivity and because skeptics have not produced a model; therefore, no one can produce a model that shows low climate sensitivity. - Its just another stupid hypothesis from you folks who believe unquestionably in models. Seems to me the primary argument skeptics are making is models are not evidence. Actually producing a model turns the argument into a war of models; in other words a popularity contest. That would lend more credibility to the alarmist argument than it deserves.
|
|
|
Post by poitsplace on Jul 7, 2009 22:06:22 GMT
It's not a matter of whether the GCMs are right or wrong, that's another question for another day. This question is about whether it is or is not possible to get low climate sensitivities from GCMs. Consider the ramifications of either: If it is possible to get a low sensitivity from a GCM then that throws the level of certainty of AGW projections out of the window. Afterall if the GCMs show that it's possible to get low sensitivities then the uncertainty widens to within what the skeptics claim. It would also demonstrate that the GCMs can be "tuned to show anything" If it isn't possible to get a low sensitivity from a GCM then that throws the whole argument that GCMs can be tuned to show anything out of the window. In fact the greater people insist GCMs can be tuned to show anything the even more remarkable it would be if noone could tune model to show a low climate sensitivity. On both cases skeptics have a strong motive to demonstrate low climate sensitivity from GCMs. That's NOT how GCMs work. There is no validated base code. It's not like programs used to crash test virtual cars...which have been created with well understood concepts, simulates something that's been built to known specifications and has been thoroughly tested by crashing THOUSANDS of different cars. Even with that, btw they still end up finding things they didn't expect. No, GCMs essentially say what the programmers want them to say. In fact, they could say for every PPM increase in CO2 the temperature of the earth goes up by 100C. It doesn't have to make any sense for the MODEL. The model its self can show just about any behavior they want as long as they have enough parameters. Now in THIS case what they've done is something stupid. They have (as I have pointed out in the past) assumed that most of the warming of the 80's and 90's was due to CO2 and modeled based on that. They've also attempted to fit the temperatures from the past (to some extent). The problem is that most of the parameters that they've made within the models are technically filled with unknowns. They literally make those values whatever seems to make the model work. Whenever (and this happens in ALL of the IPCC models) they make an assumption of one "forcing" or "feedback" being higher or lower than reality to model the past...the results are skewed in the future. Again, they've been "tuned" to show pretty much a range of warming that the modelers expect. They aren't validation, they're each an individual hypothesis that's been fit as best as possible to reality. Saying that a model of poorly understood processes is accurate is like saying those crazy conspiracy theories are correct. Just because you can figure out a way the scenario could have hypothetically worked (with flawed initial knowledge)...that doesn't mean it happened like that. It's a vicious cycle. They make the model to show their assumptions, toss out any models that don't provide results in the range they wanted, and then claim the models themselves are validation. It's junk science, pure and simple...well, I take that back, it's not even science.
|
|
|
Post by socold on Jul 7, 2009 22:35:49 GMT
Let me see if I can summarize the gist of that argument: - Because skeptics are allegedly motivated to demonstrate low climate sensitivity and because skeptics have not produced a model; therefore, no one can produce a model that shows low climate sensitivity. - Not quite. I reword slightly: Because skeptics are motivated to demonstrate low climate sensitivity and because skeptics have not demonstrated low climate sensitivity with a GCM, therefore this suggests they are unable to. Ie that low climate sensitivities are not possible to achieve using GCMs. Which tells us that a) The models can't be simply tuned to show anything. b) The high sensitivities shown by GCMs are not simply cherrypicked. c) The skeptics are banking on unknowns, not knowns in their attacks on the theory. Ie they don't know why mainstream climate physics is wrong, they are arguing that it will turn out to be wrong somehow. I only focus on skeptics wrt getting low clim sensitivity from models because I imagine if I suggested that decades of scientists haven't managed to produce low climate sensitivity from GCMs the response will be something like "yea that's because they are all biased" or something. No, skeptics are claiming models are wrong for reasons they claim to know. Yet none of them correct for it to demonstrate the difference it will make. A war of models would be science. If the skeptics put up a model that showed low climate sensitivity it would be possible to compare it and narrow down the causes, ie what makes the difference between low and high sensitivity.
|
|
|
Post by socold on Jul 7, 2009 22:52:04 GMT
No, GCMs essentially say what the programmers want them to say. In fact, they could say for every PPM increase in CO2 the temperature of the earth goes up by 100C. It doesn't have to make any sense for the MODEL. The model its self can show just about any behavior they want as long as they have enough parameters. This isn't true though. The model inputs are constrained by results from observation and experiment and the output is validated against the real world too. Also you are doing precisely what I argued you couldn't. You are claiming a GCM can be easily tuned to output anything. Okay, then why hasn't a GCM been tuned to output low climate sensitivity to show that it's at least possible? Wrong way round. The models are built based on understanding of climate and as a result they find high climate sensitivity, which necessarily leads to warming of the 80s and 90s in hindcasts because ghg rose so much during that time period. Then why doesn't someone change those values in a way that causes it to output low climate sensitivity? And if you think there's a conspiracy going on preventing "AGW scientists" doing it, why haven't any skeptics done it? There's been decades to do it and various skeptic scientists are all online writing blog posts. Why haven't any of them fired a torpedo into AGW by demonstrating that models are compatible with low climate sensitivity, not just high? Why are the skeptic scientists leaving the field right open for IPCC reports, etc to point out that all model results lead to high climate sensitivity. This is what is convincing the scientists afterall - that climate as we understand it leads to a result of high climate sensitivity. No matter how you turn that variable or twist that input, it leads to high climate sensitivity. Many people, including me, have built up the impression that low climate sensitivity from models could only be achieved by deliberately putting in false physics or unrealistic parameters in there. Why hasn't someone in the last 30 years proved us wrong? If it could be shown otherwise surely this is the first thing to show in order to convince people, as well as other scientists that climate sensitivity likely isn't high. Why leave this boat floating? None of this matters, even the models that don't reproduce 20th century well have a high sensitivity. Noone has demonstrated a low sensitivity from models.
|
|
|
Post by magellan on Jul 8, 2009 1:40:55 GMT
No, GCMs essentially say what the programmers want them to say. In fact, they could say for every PPM increase in CO2 the temperature of the earth goes up by 100C. It doesn't have to make any sense for the MODEL. The model its self can show just about any behavior they want as long as they have enough parameters. This isn't true though. The model inputs are constrained by results from observation and experiment and the output is validated against the real world too. Also you are doing precisely what I argued you couldn't. You are claiming a GCM can be easily tuned to output anything. Okay, then why hasn't a GCM been tuned to output low climate sensitivity to show that it's at least possible? Wrong way round. The models are built based on understanding of climate and as a result they find high climate sensitivity, which necessarily leads to warming of the 80s and 90s in hindcasts because ghg rose so much during that time period. Then why doesn't someone change those values in a way that causes it to output low climate sensitivity? And if you think there's a conspiracy going on preventing "AGW scientists" doing it, why haven't any skeptics done it? There's been decades to do it and various skeptic scientists are all online writing blog posts. Why haven't any of them fired a torpedo into AGW by demonstrating that models are compatible with low climate sensitivity, not just high? Why are the skeptic scientists leaving the field right open for IPCC reports, etc to point out that all model results lead to high climate sensitivity. This is what is convincing the scientists afterall - that climate as we understand it leads to a result of high climate sensitivity. No matter how you turn that variable or twist that input, it leads to high climate sensitivity. Many people, including me, have built up the impression that low climate sensitivity from models could only be achieved by deliberately putting in false physics or unrealistic parameters in there. Why hasn't someone in the last 30 years proved us wrong? If it could be shown otherwise surely this is the first thing to show in order to convince people, as well as other scientists that climate sensitivity likely isn't high. Why leave this boat floating? None of this matters, even the models that don't reproduce 20th century well have a high sensitivity. Noone has demonstrated a low sensitivity from models. Have you ever read anything contrary to the nearly plagiarized propaganda sources on climate models? A GCM is a hypothesis with hypotheses embedded in them. They are not based on first order physics contrary to what you are implying. Then why doesn't someone change those values in a way that causes it to output low climate sensitivity? Roy Spencer and others have challenged modelers to do just that. Who controls the purse strings?
|
|
|
Post by poitsplace on Jul 8, 2009 4:43:16 GMT
There are numerous reasons you wouldn't...but a better question is why would these idiots create GCMs that show exponential increases in temperature and requiring feedbacks far more powerful than the "forcing" of CO2. That outcome is ENTIRELY IMPOSSIBLE. There is nothing in the climate system to indicate that it's remotely possible. Ice albedo would have been long since exhausted. By the way, some models HAVE been made with low sensitivity climatesci.org/2009/07/07/new-paper-global-and-regional-comparison-of-daily-2m-and-1000-hpa-maximum-and-minimum-temperatures-in-three-global-re-analyses-by-pitman-and-perkins-2009/“…..a climate model that reproduces the observed change in global air temperature over the last 50 years, but fails to quantitatively reproduce the observed changed in ocean heat content, cannot be correct. The PCM [Parallel Climate Model] has a relatively low sensitivity (less anthropogenic impact on climate) and captures both the ocean- and air-temperature changes. It seems likely that models with higher sensitivity, those predicting the most drastic anthropogenic climate changes in the future, may have difficulty satisfying the ocean constraint.”They find high climate sensitivity because that's what they set out to find and programmed into the models. The assumption came BEFORE...hindcasting required tuning. However with such a small sampling of reliable temperature records and the fact that the planet has been warming for the entire time...it is virtually impossible to tell what part (if any) is from CO2. The bottom line though is that inherent to being skeptical of high climate sensitivity...is a skepticism of the models. A perfect example is all of the model runs outputting mid to high temperature projections (mid to high relative to the IPCC runs). The highest models would require exponential increases in CO2 (even assuming high climate sensitivity) but even though we've had exponential output growth...the CO2 has remained linear. At the ever-increasing rate of CO2 uptake even if we did find double the coal reserves we STILL wouldn't manage to do much more than double atmospheric CO2. Guess what...the model has a parameter that's wrong...CO2 uptake. Guess what...so do all the others. At current rates we'd need to find new reserves JUST to double CO2. I won't go into the insane feedbacks in those models because even you admit they're too high. They also assume high feedbacks for the middle runs too. But again we run into problems. Assuming even a linear increase in CO2 and that the 80's and 90's were warmed by those rates of CO2 increase...we come up with a maximum anomaly of about 2C by 2100. But even linear increases require almost exponential feedback to offset the logarithmic nature of CO2. Most of the models are already proven wrong. The ones that are still clinging to being feasible by the skin of their error bars...are likely based on faulty assumptions. Not being completely disproven for a couple years is not validation by any stretch of the imagination. Real (accurate) computer models have to mature over years and with thousands of examples.
|
|
|
Post by socold on Jul 8, 2009 18:37:41 GMT
Then why doesn't someone change those values in a way that causes it to output low climate sensitivity?
Roy Spencer and others have challenged modelers to do just that. Who controls the purse strings? I would be interested in seeing one of these challenges if you have a reference to one at hand (I understand if you don't have links at hand), to see what aspect of the models Roy Spencer is suggesting they change. I am a bit suprised that he would suggest something was done, but not find a way of getting it done, either by himself or someone else. There are plenty of researchers who take public climate models and use them in studies to test this and that. Why can't someone do such a study to demonstrate a low climate sensitivity? We see skeptics spending lots of unpaid time on blog posts, yet none of them can put in the time to demonstrate what would be a strong point?
|
|
|
Post by socold on Jul 8, 2009 19:09:16 GMT
There are numerous reasons you wouldn't...but a better question is why Could you list some of the numerous reasons? In this case relatively low refers to a climate sensitivity of 2.1C in the PCM, which is high by skeptic standards. What if that isn't what they set out to find but is simply what is found when someone does this objectively? On my screen now I have a list of sensitivities from 19 different models, and while there is a wide range, they all fall between 2.1C and 4.4C. None of them below 2C. Is it even likely that these models were all produced fraudulently? The question is why do they all output mid to high climate sensitivities? The importance of this question can be seen in the three possible answers I can think of below. Of course not all answers make as much sense, and I have commented on why I find some less likely than others: answer 1) The models all output mid to high climate sensitivities because this is what current human understanding of how the climate works leads to. Whenever anyone has tried to put together knowledge of how the climate works into a model, high climate sensitivity is what they have found. Noone knows of any fault with this understanding as demonstrated by a lack of contradictory demonstration to date. For me this is the answer that makes the most sense. This doesn't mean the models are right, it just means that however they are wrong noone can pinpoint why. answer 2) Models all output mid to high climate sensitivities because this is what one explaination for how the climate works leads to, but there are other explaination(s). For me this answer is less likely, because it begs the question why haven't the other explaination(s) been tried? The idea that all the models ever implemented just happened to have coincidentally avoided the better explaination(s) seems unfeasible. Furthermore with the validity of climate models being of so much public concern, it seems all the more unlikely that any better explaination(s) wouldn't have been demonstrated by now. answer 3) They all output mid to high climate sensitivities because of a known error. Ie all the models are fraudulent. This one is the least likely answer in my opinion. It just makes no sense for a lot of reasons. It suffers from the problems of answer #2, but also requires belief in a conspiracy across continents and decades.
|
|
|
Post by icefisher on Jul 8, 2009 22:04:34 GMT
A war of models would be science. If the skeptics put up a model that showed low climate sensitivity it would be possible to compare it and narrow down the causes, ie what makes the difference between low and high sensitivity. No it wouldn't work out that way. The people criticizing the modeling approach to climate policy; aren't doing models because they see it as the wrong tool for the job. Most of them believe you need to be able to model the alternatives such as clouds, magnetism, and xray flux. But none of that can be done because of either a lack of understanding of their forcing effects or a lack of understanding of their physical mechanisms. Thus attempting to build an alternative model at this point in time would be somewhat like trying to build a wooden ladder to the moon.
|
|
|
Post by poitsplace on Jul 9, 2009 4:36:16 GMT
(1) If sensitivity is low, modeling is pointless. Natural variation would completely overwhelm the CO2 warming and due to the chaotic nature of the natural systems...it cannot be properly modeled anyway (this is actually the same for high sensitivity models but it's just much worse for low sensitivity ones)
(2) if there aren't powerful feedbacks there's no need to model at all. There aren't substantial secondary affects just straight forward warming. The simple math (which falls short of course) is about all that's needed.
(3) If sensitivity is low...WHO CARES ABOUT CO2? The changes will be trivial, modeling is unnecessary, especially considering answer #1.
The models DO NOT WORK THAT WAY! How do you miss this? If they KNOW how it works they don't need to model the damn thing. The models are based on assumptions. They also don't model low sensitivity for that first reason I mentioned, its useless to bother with it. But the logic is flawed because higher sensitivity is nonsensical. The upper ranges are ENTIRELY IMPOSSIBLE on numerous fronts. (not enough fuel, feedback that is by definition runaway, etc).
The models are just a coded version of the same hypothesis...they prove nothing. Their hypothesis is that CO2 will cause a lot of warming...so thats what they show. The ONLY validation can come from long term prediction. That's why you're NOT supposed to use them for this kind of stuff. It's untested.
We use models in other places because they're immediately verifiable (like crash test simulations). we sometimes use models to test mathematical or statistical concepts...because they simply involve too much math to do...but that math would be verifiable at every single point.
We can TRY to use models on things we don't fully understand...but we can NEVER treat those models as proof.
but again, if global warming doesn't have all sorts of additional feedbacks to obfuscate what SHOULD be a straight forward warming process...then a model simply isn't needed (although in a sense, even the simple math of absorption is a "model")
|
|