6 ms·

I am not a climate change denier by a long shot, but stories like these give a ding to the credibility of climate scientists. > "So we can't throw them out yet

I am not a climate change denier by a long shot, but stories like these give a ding to the credibility of climate scientists. > "So we can't throw them out yet." This might be an elementary view of science, but I think there is a danger here that while everyone is making their models, if anyone is an outlier they go and tweak their model to match the patterns of others: 'Klaus Wyser’s group "switched off" some of the new cloud and aerosol settings in their model, he said, and that sent climate sensitivity back down to previous levels'. That seems to me a questionable reason to "switch off" part of the model - you should create the most accurate simulation possible and trust the output, not tweak the inputs to match literature data.

nerdponx 7y ago
Typically it's not obvious what to switch on and off for "accuracy". If your model produces wildly different results from previous models, either you hit on some really big discovery or you made a mistake. An ethical scientist will try to rule out the latter possibility before accepting the former.
- gmueckl 7y ago
  It goes deeper than that. If you can tweak your new model to correspond to an older one for which you have results available, you must run tests in which you apply these tweaks and verify that you get the same results. If you get different results, one of the models simply isn't doing what you think it's doing and you need to understand that. And in the vast majority of cases, it's likely a mistake in a model, not any kind of new science. One pretty general debugging/analysis strategy for a complex model is to deactivate it piece by piece until you get it to a point where it should reproduce a result that you can verify by other, independent means. E.g. tweak the parameters of the model so that reproduces a result that is known analytically. For instance, whenever it is possible in a model, a good sanity check is to set parameters that must lead to e.g. a perfect conservation of certain quantities.
orev 7y ago
Tweaking the inputs would help one to identify what factor is causing the difference, which allows you to focus on that. It doesn’t mean you permanently leave it out. And it does not make a credibility issue. This is how science works — it is a process that strives to be as accurate as possible, not come up with a static answer and stick to it (that’s what religion does). The ability for science to change and adapt is what makes it so strong. That is the message that needs to always be driven when discussing science; not throwing hands in the air and saying “why do these scientists keep changing their answers? They obviously don’t know what they’re doing”
- justadudeama 7y ago
  These are valid points, and it is important to figure out what changes in the model is causing such a dramatic change and scrutinize it more. > The ability for science to change and adapt is what makes it so strong To me this article is them _resisting_ change, looking for a reason NOT to accept it, because it goes against what they previously said. Maybe this is all to early, and in a year or so we will be seeing 4 or 4.5 degree predictions, but this article describes them trying to change their inputs to match previous answers, not getting new answers.
  - Angostura 7y ago
    Again, that's the way that science tends to operate - a la Kuhn and his Structure of Scientific Revolutions. You have some novel research that seems to contradict well-established consenus. It comes under close scrutiny, because science tends to be conservative. If after giving it a good kicking, its clear that the old consensus is wrong, science takes the wrecking ball to it. But it prefers not to get the wrecking ball out of it can help it.
    - freeflight 7y ago
      > You have some novel research that seems to contradict well-established consenus. It comes under close scrutiny, because science tends to be conservative. Indeed, even people we nowadays consider brilliant minds of their time, encountered plenty of resistance [0] during these very same times because their proposed ideas just seemed too outrageous. [0] https://en.wikipedia.org/wiki/Criticism_of_the_theory_of_relativity#A_Hundred_Authors_Against_Einstein https://en.wikipedia.org/wiki/Criticism_of_the_theory_of_rel...
epistasis 7y ago
I don't know how you could have reached that malicious interpretation of why that part was switched off. To go and impugn all of climate science in the basis of a misinterpretation of motivations would indicate that you have internalized the denialists claims and look to seek to justify them. To point out just how wrong your misinterpretation is, the very next sentence has the "why": > A new research paper co-authored by Zelinka from the Lawrence Livermore National Lab likewise pointed to the role of virtual clouds in determining the results. The key phrase being "role of virtual clouds." The motivation wasn't to slavishly reproduce results from the literature, and to even suspect that it completely outside the bounds of reasoned discourse. It comes across as if you are trying your hardest to find something to nitpick, some small phrase that you can misinterpret in order to raise doubts in those who aren't paying close attention. This is exactly how denialists and propagandists behave.
gbrown 7y ago
I'd propose that has more to do with the headlines than any of the actual science. Model tweaking is exactly how a lot of this works - models are supposed to reflect reality, so part of how we explore them is to find out what models fit past (and predict current/future) reality. This is both intuitive, and based on theoretically sound inference (see Approximate Bayesian Computing for an exploration of the connection between simulation and formal inference).
dls2016 7y ago
This part of climate "science" always makes me put on my Popper hat. The social aspect of being the outlier (grants dry up, career goes away) makes me skeptical of the IPCC reports, just not in the usual direction. I fear reality could be much worse. I can't ding the climate scientists too much, though... I'm not sure there's a better way to study this stuff.
- Angostura 7y ago
  In this case have grants actually dried up or careers gone away? Nope.
  - dls2016 7y ago
    I don't think I implied they did. I'm only saying that humans are social creatures and that scientists are susceptible to peer pressure (even in "hard" sciences with lots of possibility for experiment). I imply no maliciousness, but am trying to voice my fear of the impossibility of knowing whether or not the consensus warming scenario is a "reversion to the social mean", that is, we just have to wait and see if warming is worse that expected. But I'm only a bit surprised this would be such a controversial opinion here.
hyperpallium 7y ago
> Why didn’t they discover that the new number was higher right away? http://calteches.library.caltech.edu/51/2/CargoCult.htm http://calteches.library.caltech.edu/51/2/CargoCult.htm
rriepe 7y ago
This idea, along with the null hypothesis, are simply not honored by this part of the scientific community. Climategate proved this ten years ago. It's a doomsday cult and they'll stick you with the "D" word if you don't narrative-fit. Check out Ned Nikolov's work for a better alternate theory.
cm2187 7y ago
That's the main problem I have with climate science. They don't really have the luxury of proving by experimentation, and I have seen enough perfectly good backtestings in my industry (finance) that go wildely off as soon as they go live, that I have a fundamental mistrust of backtesting as a way to calibrate a predictive model.
- New_Horizon 7y ago
  Especially because they seem to test it by making it predict history, but we don't have anywhere near the same amount of historical data. So what are you testing, really? Just that it accurately finds average temps?
  - bannable 7y ago
    Climate models as far back as the 1970s have predicted the temperature changes in the time from then to now -- those are not predictions that are coming from hindsight, but rather predictions where their effects were observed over the last 50 years. If you are unfamiliar with climate sensitivity, this is a good place to start: https://skepticalscience.com/climate-sensitivity.htm https://skepticalscience.com/climate-sensitivity.htm
- DennisP 7y ago
  From the article: "A recent study compared models as old as 1970 with observations made in the decades since. Some models warmed up too much, and some too little, but 14 of 17 past projections turned out to be consistent with the measured path of global average temperatures." So we actually did an experimental test of the predictions made in 1970, and those predictions turned out to be largely correct. We should probably halt this experiment on ethical grounds.
zyxzevn 7y ago
The climate models are not settled. Sadly the political activists have hijacked the normal scientific process. And this allowed some pseudoscience (with alarming predictions or alarming observations) to mix in too. In some science news I read that the Arctic Ice-sheet is growing back, and that Envisat showed that the oceans were not rising. So very likely the climate is more complex than some alarmed scientists thought.
willyt 7y ago
The reason clouds are starting to be simulated now where they weren't before is probably because we have only recently had the computing power available to do this at the level of detail required. Switching the 'clouds' feature on and off is part of the model verification process. Listen to this podcast[0], it's a in depth technical discussion about weather forecasting models with the actual scientists that create them. They discuss model verification for weather models and they touch on how this also applies to climate modelling. [0]http://omegataupodcast.net/326-weather-forecasting-at-the-ecmwf/ http://omegataupodcast.net/326-weather-forecasting-at-the-ec...

nerdponx 7y ago

Typically it's not obvious what to switch on and off for "accuracy". If your model produces wildly different results from previous models, either you hit on some really big discovery or you made a mistake. An ethical scientist will try to rule out the latter possibility before accepting the former.

gmueckl 7y ago
It goes deeper than that. If you can tweak your new model to correspond to an older one for which you have results available, you must run tests in which you apply these tweaks and verify that you get the same results. If you get different results, one of the models simply isn't doing what you think it's doing and you need to understand that. And in the vast majority of cases, it's likely a mistake in a model, not any kind of new science. One pretty general debugging/analysis strategy for a complex model is to deactivate it piece by piece until you get it to a point where it should reproduce a result that you can verify by other, independent means. E.g. tweak the parameters of the model so that reproduces a result that is known analytically. For instance, whenever it is possible in a model, a good sanity check is to set parameters that must lead to e.g. a perfect conservation of certain quantities.

orev 7y ago

Tweaking the inputs would help one to identify what factor is causing the difference, which allows you to focus on that. It doesn’t mean you permanently leave it out. And it does not make a credibility issue. This is how science works — it is a process that strives to be as accurate as possible, not come up with a static answer and stick to it (that’s what religion does). The ability for science to change and adapt is what makes it so strong. That is the message that needs to always be driven when discussing science; not throwing hands in the air and saying “why do these scientists keep changing their answers? They obviously don’t know what they’re doing”

justadudeama 7y ago
These are valid points, and it is important to figure out what changes in the model is causing such a dramatic change and scrutinize it more. > The ability for science to change and adapt is what makes it so strong To me this article is them _resisting_ change, looking for a reason NOT to accept it, because it goes against what they previously said. Maybe this is all to early, and in a year or so we will be seeing 4 or 4.5 degree predictions, but this article describes them trying to change their inputs to match previous answers, not getting new answers.
- Angostura 7y ago
  Again, that's the way that science tends to operate - a la Kuhn and his Structure of Scientific Revolutions. You have some novel research that seems to contradict well-established consenus. It comes under close scrutiny, because science tends to be conservative. If after giving it a good kicking, its clear that the old consensus is wrong, science takes the wrecking ball to it. But it prefers not to get the wrecking ball out of it can help it.
  - freeflight 7y ago
    > You have some novel research that seems to contradict well-established consenus. It comes under close scrutiny, because science tends to be conservative. Indeed, even people we nowadays consider brilliant minds of their time, encountered plenty of resistance [0] during these very same times because their proposed ideas just seemed too outrageous. [0] https://en.wikipedia.org/wiki/Criticism_of_the_theory_of_relativity#A_Hundred_Authors_Against_Einstein https://en.wikipedia.org/wiki/Criticism_of_the_theory_of_rel...

epistasis 7y ago

I don't know how you could have reached that malicious interpretation of why that part was switched off. To go and impugn all of climate science in the basis of a misinterpretation of motivations would indicate that you have internalized the denialists claims and look to seek to justify them. To point out just how wrong your misinterpretation is, the very next sentence has the "why": > A new research paper co-authored by Zelinka from the Lawrence Livermore National Lab likewise pointed to the role of virtual clouds in determining the results. The key phrase being "role of virtual clouds." The motivation wasn't to slavishly reproduce results from the literature, and to even suspect that it completely outside the bounds of reasoned discourse. It comes across as if you are trying your hardest to find something to nitpick, some small phrase that you can misinterpret in order to raise doubts in those who aren't paying close attention. This is exactly how denialists and propagandists behave.

gbrown 7y ago

I'd propose that has more to do with the headlines than any of the actual science. Model tweaking is exactly how a lot of this works - models are supposed to reflect reality, so part of how we explore them is to find out what models fit past (and predict current/future) reality. This is both intuitive, and based on theoretically sound inference (see Approximate Bayesian Computing for an exploration of the connection between simulation and formal inference).

dls2016 7y ago

This part of climate "science" always makes me put on my Popper hat. The social aspect of being the outlier (grants dry up, career goes away) makes me skeptical of the IPCC reports, just not in the usual direction. I fear reality could be much worse. I can't ding the climate scientists too much, though... I'm not sure there's a better way to study this stuff.

Angostura 7y ago
In this case have grants actually dried up or careers gone away? Nope.
- dls2016 7y ago
  I don't think I implied they did. I'm only saying that humans are social creatures and that scientists are susceptible to peer pressure (even in "hard" sciences with lots of possibility for experiment). I imply no maliciousness, but am trying to voice my fear of the impossibility of knowing whether or not the consensus warming scenario is a "reversion to the social mean", that is, we just have to wait and see if warming is worse that expected. But I'm only a bit surprised this would be such a controversial opinion here.

hyperpallium 7y ago

> Why didn’t they discover that the new number was higher right away? http://calteches.library.caltech.edu/51/2/CargoCult.htm http://calteches.library.caltech.edu/51/2/CargoCult.htm

rriepe 7y ago

This idea, along with the null hypothesis, are simply not honored by this part of the scientific community. Climategate proved this ten years ago. It's a doomsday cult and they'll stick you with the "D" word if you don't narrative-fit. Check out Ned Nikolov's work for a better alternate theory.

cm2187 7y ago

That's the main problem I have with climate science. They don't really have the luxury of proving by experimentation, and I have seen enough perfectly good backtestings in my industry (finance) that go wildely off as soon as they go live, that I have a fundamental mistrust of backtesting as a way to calibrate a predictive model.

New_Horizon 7y ago
Especially because they seem to test it by making it predict history, but we don't have anywhere near the same amount of historical data. So what are you testing, really? Just that it accurately finds average temps?
- bannable 7y ago
  Climate models as far back as the 1970s have predicted the temperature changes in the time from then to now -- those are not predictions that are coming from hindsight, but rather predictions where their effects were observed over the last 50 years. If you are unfamiliar with climate sensitivity, this is a good place to start: https://skepticalscience.com/climate-sensitivity.htm https://skepticalscience.com/climate-sensitivity.htm
DennisP 7y ago
From the article: "A recent study compared models as old as 1970 with observations made in the decades since. Some models warmed up too much, and some too little, but 14 of 17 past projections turned out to be consistent with the measured path of global average temperatures." So we actually did an experimental test of the predictions made in 1970, and those predictions turned out to be largely correct. We should probably halt this experiment on ethical grounds.

zyxzevn 7y ago

The climate models are not settled. Sadly the political activists have hijacked the normal scientific process. And this allowed some pseudoscience (with alarming predictions or alarming observations) to mix in too. In some science news I read that the Arctic Ice-sheet is growing back, and that Envisat showed that the oceans were not rising. So very likely the climate is more complex than some alarmed scientists thought.

willyt 7y ago

The reason clouds are starting to be simulated now where they weren't before is probably because we have only recently had the computing power available to do this at the level of detail required. Switching the 'clouds' feature on and off is part of the model verification process. Listen to this podcast[0], it's a in depth technical discussion about weather forecasting models with the actual scientists that create them. They discuss model verification for weather models and they touch on how this also applies to climate modelling. [0]http://omegataupodcast.net/326-weather-forecasting-at-the-ecmwf/ http://omegataupodcast.net/326-weather-forecasting-at-the-ec...