7 ms·

This is why I'm a bit skeptical about the global warming predictions. AFAIK they're all based on multi-million lines of code models. Not only it's monumentally

by badpun 8y ago

This is why I'm a bit skeptical about the global warming predictions. AFAIK they're all based on multi-million lines of code models. Not only it's monumentally hard to have such large code base without bugs, but also, in case of a software implementing scientific models, finding bugs is extra hard (compared to software which works in a visible way and has millions of users, such as Linux kernel, games, car's embedded software etc.). An effort to have such model verified (let's say to a NASA standard) to be bug-free would probably cost billions of dollars. And that's only bugs, let's not forget that all the models are approximations, plus the numerical methods used all have their quirks and limitations etc. All in all, the problem seems too hard to tackle for humanity right now.

nafey 8y ago
The counterpoint is that we have already seen a steady temperature rise. So even though the specifics of the various simulations might not play out as predicted we can expect temperatures to rise.
- badpun 8y ago
  Knowing that the temperature will rise is not enough to make policy decisions though. You need to predict the increase's magnitude, as well as practical consequences, such as climate change, how much the sea level will rise etc. And for that you need resonable and bug-free models.
  - leeoniya 8y ago
    > how much the sea level will rise etc this will be the least of our problems. we understand very little about nature to have any reasonable prediction model. we don't know the inflection point which will cause massive collapse of major ecosystems we depend on. all we know is that things are changing fast. faster than many non-human organisms are able to adapt.
- femidav 8y ago
  Steady? We have seen 20-years rise after a colder period. Nothing out of order.
magic-chicken 8y ago
How would the global warming predictions all be biaised in the same way ? _All_ the studies are measuring the same tendency : the temperature is rising. The model does not need to be _absolutely_ precise to be right.
- badpun 8y ago
  > _All_ the studies are measuring the same tendency : the temperature is rising. Are you talking about predictions (and not measurements, you don't need a model to measure temperature)? Assuming you are, there's unfortunately a huge problem with modeling (and heavy math and stats-based science in general), in that researchers tend to stop looking for bugs in the model when it returns the results that they expect. In other words, if a bug in the model tells the researcher that Earth's temperature will decrease by 4 C by 2100, he will look over the model until he finds the bug, but if the model tells him that the temperature will increase by 2 C, thus confirming his inner bias, he'll declare it correct and move on to writing a paper based on the "finding". Alternatively, as a thought experiment, imagine if math research were done in the way climate science is done. We would have proofs that are millions of pages long and were never verified by anyone. We would trust in them only because the author says that they are correct. Is this science?
  - magic-chicken 8y ago
    A given prediction can be wrong, an experiment may be biaised, my point is that you choose to ignore that the vast majority of the experiments and measurements point in the same direction. >>Researchers tend to stop looking for bugs in the model when it returns the results that they expect. Again... they are _ALL_ biaised ?
    - badpun 8y ago
      It's not impossible - groupthink (https://en.wikipedia.org/wiki/Groupthink https://en.wikipedia.org/wiki/Groupthink) has happened many times in the past among supposedly most brilliant minds. Also, how many independent, comprehensive models (with codebases) for global climate change are there in the science world?
jplayer01 8y ago
So you'd rather just ignore all scientific results until somebody does a NASA-level code review because there might be bugs? You're basically on par with an anti-vaxxer, and the potential consequences are far worse. Congratulations.
- geokon 8y ago
  He said he's skeptical, not that he thinks it's wrong. I've got to say that this kind of rabid response from climate change proponents kinda prods my inner contrarian. I think at this point the group think is so strong that if anyone came out with evidence against global warming they would commit career suicide by publishing it
  - __s 8y ago
    I agree with your point, but note.. > they would commit career suicide by publishing it Meanwhile some researches create their career by publishing this sort of stuff, even when it isn't well researched
  - jplayer01 8y ago
    > I've got to say that this kind of rabid response from climate change proponents kinda prods my inner contrarian Oh good. So you'd rather see millions of people be displaced. We're fucked as a species when "I'm a contrarian" is a valid reason to disagree.
- badpun 8y ago
  > So you'd rather just ignore all scientific results until somebody does a NASA-level code review because there might be bugs? The bugs, if serious enough, can make it not scientific, in the same way that a paper that has grave errors in it is wrong and thus obviously not scientfic. So, before we do a thorough review of the code, we should treat it with due scepticism. EDIT: this can also be expressed in terms of risk analysis. For typical software, the consequence of a bug is low - most software is commercial, so the impact of the bug will be limited to that company's bottom line (with exceptions of software that can kill, but these people already are serious about bugs). Also, most bugs in most software are either highly evident (the button does not work, you get random segfaults etc.) or have limited impact. On the other hand, the bugs in climate change models, given their "pipeline" nature (wrong result from one module is propagated downstream all the way to the final prediction of expected temperature change), can quite often have severe impact. They can also be not evident at all - they can for example change the final outcome predicition by 1 C. Compound that with the fact these predictions are used to make trillion dollar decisions on global policies, and you can see that the actual damage done by bugs is not unlikely to be in the trillions. And that's why I say it's probably wise to subject the models to extreme scrutiny.
xorcist 8y ago
While reasonable skepticism is healthy, global warming is such a well sudied phenomenon by now that an unreasonable number of independent codebases must have identical bugs in order for it to be false. There's also the fact that we have had a pretty solid grasp about the chemical reactions of greenhouse gases since long before computers, both theoretically and empirically, and we know roughly how much is put out in the atmosphere. Where the models diverge is on far finer points than what is needed to make the basic policy changes that seems to be where we are stuck right now.
- mo3gut 8y ago
  Quite right. In fact, those statisticians in the posted article are surely mistaken. We know that using tricks to hide things is sound science. http://www.realclimate.org/index.php/archives/2009/11/the-cru-hack/ http://www.realclimate.org/index.php/archives/2009/11/the-cr... http://www.realclimate.org/index.php/archives/2009/11/the-cru-hack-context/ http://www.realclimate.org/index.php/archives/2009/11/the-cr... Since it was sound science in 2009, why would it not be now?
- rasteau 8y ago
  > an unreasonable number of independent codebases must have identical bugs Entertaining[0] badpun’s skepticism, it is not necessary they have identical bugs, only that their bugs yield similarly biased results. For example, if a significant number of bugs are identified by their affect on the results, then bugs contributing to “wrong” results might be more likely to be identified and fixed. [0] In the Aristotelian sense.
- aldoushuxley001 8y ago
  Not if the bugs are actually just bad/corrupted data. For example, the main dataset the IPCC is based on appears to contain all sorts of bad data that definitely make me skeptical of the conclusions the IPCC comes to (https://researchonline.jcu.edu.au/52041/ https://researchonline.jcu.edu.au/52041/). If global warming is actually wrong, it's most likely because of bad/corrupt data in the datasets used.

by badpun 8y ago

nafey 8y ago

The counterpoint is that we have already seen a steady temperature rise. So even though the specifics of the various simulations might not play out as predicted we can expect temperatures to rise.

badpun 8y ago
Knowing that the temperature will rise is not enough to make policy decisions though. You need to predict the increase's magnitude, as well as practical consequences, such as climate change, how much the sea level will rise etc. And for that you need resonable and bug-free models.
- leeoniya 8y ago
  > how much the sea level will rise etc this will be the least of our problems. we understand very little about nature to have any reasonable prediction model. we don't know the inflection point which will cause massive collapse of major ecosystems we depend on. all we know is that things are changing fast. faster than many non-human organisms are able to adapt.
femidav 8y ago
Steady? We have seen 20-years rise after a colder period. Nothing out of order.

magic-chicken 8y ago

How would the global warming predictions all be biaised in the same way ? _All_ the studies are measuring the same tendency : the temperature is rising. The model does not need to be _absolutely_ precise to be right.

badpun 8y ago
> _All_ the studies are measuring the same tendency : the temperature is rising. Are you talking about predictions (and not measurements, you don't need a model to measure temperature)? Assuming you are, there's unfortunately a huge problem with modeling (and heavy math and stats-based science in general), in that researchers tend to stop looking for bugs in the model when it returns the results that they expect. In other words, if a bug in the model tells the researcher that Earth's temperature will decrease by 4 C by 2100, he will look over the model until he finds the bug, but if the model tells him that the temperature will increase by 2 C, thus confirming his inner bias, he'll declare it correct and move on to writing a paper based on the "finding". Alternatively, as a thought experiment, imagine if math research were done in the way climate science is done. We would have proofs that are millions of pages long and were never verified by anyone. We would trust in them only because the author says that they are correct. Is this science?
- magic-chicken 8y ago
  A given prediction can be wrong, an experiment may be biaised, my point is that you choose to ignore that the vast majority of the experiments and measurements point in the same direction. >>Researchers tend to stop looking for bugs in the model when it returns the results that they expect. Again... they are _ALL_ biaised ?
  - badpun 8y ago
    It's not impossible - groupthink (https://en.wikipedia.org/wiki/Groupthink https://en.wikipedia.org/wiki/Groupthink) has happened many times in the past among supposedly most brilliant minds. Also, how many independent, comprehensive models (with codebases) for global climate change are there in the science world?

jplayer01 8y ago

So you'd rather just ignore all scientific results until somebody does a NASA-level code review because there might be bugs? You're basically on par with an anti-vaxxer, and the potential consequences are far worse. Congratulations.

geokon 8y ago
He said he's skeptical, not that he thinks it's wrong. I've got to say that this kind of rabid response from climate change proponents kinda prods my inner contrarian. I think at this point the group think is so strong that if anyone came out with evidence against global warming they would commit career suicide by publishing it
- __s 8y ago
  I agree with your point, but note.. > they would commit career suicide by publishing it Meanwhile some researches create their career by publishing this sort of stuff, even when it isn't well researched
- jplayer01 8y ago
  > I've got to say that this kind of rabid response from climate change proponents kinda prods my inner contrarian Oh good. So you'd rather see millions of people be displaced. We're fucked as a species when "I'm a contrarian" is a valid reason to disagree.
badpun 8y ago
> So you'd rather just ignore all scientific results until somebody does a NASA-level code review because there might be bugs? The bugs, if serious enough, can make it not scientific, in the same way that a paper that has grave errors in it is wrong and thus obviously not scientfic. So, before we do a thorough review of the code, we should treat it with due scepticism. EDIT: this can also be expressed in terms of risk analysis. For typical software, the consequence of a bug is low - most software is commercial, so the impact of the bug will be limited to that company's bottom line (with exceptions of software that can kill, but these people already are serious about bugs). Also, most bugs in most software are either highly evident (the button does not work, you get random segfaults etc.) or have limited impact. On the other hand, the bugs in climate change models, given their "pipeline" nature (wrong result from one module is propagated downstream all the way to the final prediction of expected temperature change), can quite often have severe impact. They can also be not evident at all - they can for example change the final outcome predicition by 1 C. Compound that with the fact these predictions are used to make trillion dollar decisions on global policies, and you can see that the actual damage done by bugs is not unlikely to be in the trillions. And that's why I say it's probably wise to subject the models to extreme scrutiny.

xorcist 8y ago

While reasonable skepticism is healthy, global warming is such a well sudied phenomenon by now that an unreasonable number of independent codebases must have identical bugs in order for it to be false. There's also the fact that we have had a pretty solid grasp about the chemical reactions of greenhouse gases since long before computers, both theoretically and empirically, and we know roughly how much is put out in the atmosphere. Where the models diverge is on far finer points than what is needed to make the basic policy changes that seems to be where we are stuck right now.

mo3gut 8y ago
Quite right. In fact, those statisticians in the posted article are surely mistaken. We know that using tricks to hide things is sound science. http://www.realclimate.org/index.php/archives/2009/11/the-cru-hack/ http://www.realclimate.org/index.php/archives/2009/11/the-cr... http://www.realclimate.org/index.php/archives/2009/11/the-cru-hack-context/ http://www.realclimate.org/index.php/archives/2009/11/the-cr... Since it was sound science in 2009, why would it not be now?
rasteau 8y ago
> an unreasonable number of independent codebases must have identical bugs Entertaining[0] badpun’s skepticism, it is not necessary they have identical bugs, only that their bugs yield similarly biased results. For example, if a significant number of bugs are identified by their affect on the results, then bugs contributing to “wrong” results might be more likely to be identified and fixed. [0] In the Aristotelian sense.
aldoushuxley001 8y ago
Not if the bugs are actually just bad/corrupted data. For example, the main dataset the IPCC is based on appears to contain all sorts of bad data that definitely make me skeptical of the conclusions the IPCC comes to (https://researchonline.jcu.edu.au/52041/ https://researchonline.jcu.edu.au/52041/). If global warming is actually wrong, it's most likely because of bad/corrupt data in the datasets used.