Could it be a selection bias?
Climate models are running red hot, and scientists don’t know why
11–20 of 332 posts
Re: Climate models are running red hot, and scientists don’t know why
#12I guess climate scientists don't use version control?
The article isn't very precise, but my understanding was that the run some kind of machine learning based on past data. Adding in the more recent data changed the predictions dramatically. So it's not an issue of versioning (I assume they could reproduce the old results using the old data).
It's not machine learning, it's a physical model. Computers solve a big differential equation that takes in all of the physical things we know about the Earth's climate system and runs them forward in time. The increase in thermal predictions are due to changes in the cloud and aerosol models which were made due to an improved understanding of the physics involved.
Re: Climate models are running red hot, and scientists don’t know why
#13"You are in private mode. Subscribe to continue reading." Ok, that's one more I will not open anymore; seems Bloomberg started using similar "privacy mode" detection as NYT.
Re: Climate models are running red hot, and scientists don’t know why
#14Well, politicians aren't doing anything anyway, whether the models predict catastrophe or CATASTROPHE. Already according to the 3° sensitivity predictions we would need an effort that rivals WW2 to stay below 1.5°. Yet major countries are led by climate denialists.
we would need an effort that rivals WW2 to stay below _2.0_
Re: Climate models are running red hot, and scientists don’t know why
#15"You are in private mode. Subscribe to continue reading." Ok, that's one more I will not open anymore; seems Bloomberg started using similar "privacy mode" detection as NYT.
If you happen to be on Firefox, you can use this extension to get around it:
https://addons.mozilla.org/en-US/firefox/addon/temporary-con...
Re: Climate models are running red hot, and scientists don’t know why
#16...
Re: Climate models are running red hot, and scientists don’t know why
#17Earlier quoted context omitted.
The article isn't very precise, but my understanding was that the run some kind of machine learning based on past data. Adding in the more recent data changed the predictions dramatically. So it's not an issue of versioning (I assume they could reproduce the old results using the old data).
It's not machine learning, it's a physical model. Computers solve a big differential equation that takes in all of the physical things we know about the Earth's climate system and runs them forward in time. The increase in thermal predictions are due to changes in the cloud and aerosol models which were made due to an improved understanding of the physics involved.
The line can be a bit blurry, because they are actually running inverse problems with the PDEs to fit the data. This needs regularization just like statistical ML models. Usually this amounts to solving an iterated least square problem, just like logistic regression. Bayesian methods are becoming more popular, though.
Re: Climate models are running red hot, and scientists don’t know why
#18I guess climate scientists don't use version control?
it isn't that, it is just a badly written article.
Re: Climate models are running red hot, and scientists don’t know why
#19Well, politicians aren't doing anything anyway, whether the models predict catastrophe or CATASTROPHE. Already according to the 3° sensitivity predictions we would need an effort that rivals WW2 to stay below 1.5°. Yet major countries are led by climate denialists.
Can you even imagine how a Total War economy/society based on transitioning to greener energies, cutting emissions and pollutants, and working on improvements to housing and urban planning would end up looking?
Like ignoring the whole climate change thing for a moment (though we absolutely should not do that), just the quality of life improvements to our health and well being this kind of mobilization would bring would be astonishing.
Re: Climate models are running red hot, and scientists don’t know why
#20Could it be a selection bias?
"In all, as many as a fifth of new results published in the last year have come in with anomalously high climate sensitivity." The article suggests that a consensus may well form around the new, higher estimates. So, quite possibly.