Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Even 50-year-old climate models correctly predicted global warming

sciencemag.org

Post reply on HN
Print view View original

171–180 of 242 posts

Re: Even 50-year-old climate models correctly predicted global warming

#171

Post by BLKNSLVR »

mistermann wrote: post #122
BLKNSLVR wrote: post #98

Earlier quoted context omitted.

You missed the umbrella: "It's all part of God's plan, and therefore it is both futile and in aid of the devil to work against it." I think there are some people out there that may want to bring on 'the great flood' so they can actually live through a biblical story, as if they themselves desire to be chosen to build the boat of salvation.

Out of curiosity, I just have to ask for clarification: do you consider this a standard (normal, average) conservative response? If so, upon what is this belief based?

I don't consider it normal or average for a conservative person, but I have come across the "god's plan" part quoted from a politician or two. Sadly. The 'in aid of the devil' I added as my tacit understanding of what "god's plan" means. This was poor judgement on my part, obvious to me in hindsight...

Australia currently has a very conservative government that are in denial of climate change who just happen to be led by a card-carrying Pentecostal. So there's a good chance I'm overly sensitive to such commentary :)

P.S. I think 'conservative', 'liberal', 'progressive', 'left wing' and 'right wing' tend to get in the way of proper discussion of what is actually factual, logical, or at least has basis in valid research.

'Conservatism' isn't helped by it's entanglement with religion and that old quote "A Conservative is a fellow who is standing athwart history yelling 'Stop!'"

This isn't the article quoting a politican, but it gives a clue as to why a politician may say such a thing: https://www.nbcnews.com/news/religion/climate-change-schism-...

Re: Even 50-year-old climate models correctly predicted global warming

#172

Post by jointpdf »

hcknwscommenter wrote: post #126
mistermann wrote: post #103

Earlier quoted context omitted.

> As noted in TFA, they looked at 17 forecasts from 14 models Pardon a little pedantry: >> The researchers compared annual average surface temperatures across the globe to the surface temperatures predicted in 17 forecasts. Those predictions were drawn from 14 separate computer models released between 1970 and 2001. In some cases, the studies and their computer codes were so old that the team had to extract data publ…

TFA article does say. They looked at 17 forecasts drawn from 14 models. You are the one who implicitly believes they are lying about that and only reporting 17 out of some huge unknown number. The reality is that the vast majority (>99%) of climate change research does not produce the type of forecast, or really any forecast at all, that one could do this analysis on. Therefore, the 17 forecast number is actually ver…

Simulation is one way to produce a numerical forecast, but the prediction of any forecast model would never be the result of a single simulation run. Model predictions are always (or should always) be reported as an expected value, or the average over many simulation runs. The expected value of a random variable (e.g. mode output) is probably the most fundamental concept in all of statistics/probability, and many mind-wrenching (read: wrong+incoherent) internet hot takes are rooted in logical fallacies about the science of prediction. The reason that there are more predictions (17) than models (14) is that several of the models (eg Hansen 1981) consider several distinct “scenarios” based on differing assumptions about future levels of GHG emissions.

Secondly, the forecast models evaluated in the study are from the likes of James Hansen, William Nordhaus, and the IPCC (in other words, authoritative sources). So at the very least this adds weight to the hypothesis that the models published by the IPCC are accurate and methodologically sound. Sure, there is some probability that the models and reality corresponded by random chance, but that probability is made ever more diminishingly small as evidence accumulates.

So, I don’t think any of the comments in this thread that accuse the authors of retrospective cherry-picking of climate models are anything but embarrassing examples of pseudo-intellectual contrarianism and/or confusion over basic statistics.

The models considered by the study, their predictions, and the methodology used are all located here and were easily found after 30s of digging: https://github.com/hausfath/OldModels

Re: Even 50-year-old climate models correctly predicted global warming

#173

Post by s1k3b8 »

There were dozens and dozens of "climate models" which predicted everything from global cooling to a fiery wasteland and everything in between. Of course you could go back and find a model that "predicted" correctly because every possibility was predicted.

That's like dozens and dozens of "stock price models" predicting FB stock to go up, down and stay the same. Of course one is guaranteed to be correct.

What is "scientific" about this? It's simply people who want to prove something and then narrowly searching for the data to prove that point and ignoring everything else.

This is so disappointing coming from something calling itself a sciencemag.

Re: Even 50-year-old climate models correctly predicted global warming

#174

Post by jointpdf »

Akababa wrote: post #139
mturmon wrote: post #63

Earlier quoted context omitted.

Your comment contains hypotheticals ("run random models and wait long enough you will always be able to find at least one model that predicted the current observations") that are NOT what the present study did. As noted in TFA, they looked at 17 forecasts from 14 models and found that the temperature predictions of 10 of them were consistent with what has happened in the meantime. And if you adjust for forcing (CO2,…

But we don't know how those 17 were chosen in the first place. There would be survivorship bias if models which match observations get more attention. With that being said, I doubt that survivorship bias is strong enough to explain their combined result of 10/17 models being significant.

Since one of authors posted the paper, we do know:

> “We conducted a literature search to identify papers published prior to the early-1990s that include climate model outputs containing both a time-series of projected future GMST (with a minimum of two points in time) and future forcings (including both a publication date and future projected atmospheric CO2 concentrations, at a minimum). Eleven papers with fourteen distinct projections were identified that fit these criteria.

Re: Even 50-year-old climate models correctly predicted global warming

#176

Post by refurb »

I’m confused as hell to be honest. A few years ago they said there was a global warming pause.[1]. They seemed so sure it was happening that Nature devoted a whole issue to it.

Then they came back and said “actually it never happened”.[2]

Shouldn’t it be pretty straightforward to tell whether there was warming or not?

[1] https://www.nature.com/collections/sthnxgntvp

[2] https://insideclimatenews.org/news/18122018/global-warming-h...

Re: Even 50-year-old climate models correctly predicted global warming

#177

Post by jointpdf »

s1k3b8 wrote: post #173

There were dozens and dozens of "climate models" which predicted everything from global cooling to a fiery wasteland and everything in between. Of course you could go back and find a model that "predicted" correctly because every possibility was predicted. That's like dozens and dozens of "stock price models" predicting FB stock to go up, down and stay the same. Of course one is guaranteed to be correct. What is "sci…

This comment is just completely nonsensical and intellectually dishonest. This is not at all what this paper did, nor what any remotely credible scientific study ever does.

Re: Even 50-year-old climate models correctly predicted global warming

#178

Post by jellicle »

bitlax wrote: post #88
pstuart wrote: post #64

Earlier quoted context omitted.

Lead by example and institute policies that strongly support taking aggressive action. This won't happen while Trump is in office, unfortunately.

Well I'll take that bet. Climate activists have advocated that precisely the opposite should happen: that they should maintain or even increase their growth since they are developing economies.

Climate activists have made the much more subtle point that it is unjust to expect nations which have never industrialized not to industrialize so that rich industrial nations can run their air conditioners with the doors open and their SUVs idling.

Re: Even 50-year-old climate models correctly predicted global warming

#179

Post by mturmon »

whokal wrote: post #144
hcknwscommenter wrote: post #126

Earlier quoted context omitted.

TFA article does say. They looked at 17 forecasts drawn from 14 models. You are the one who implicitly believes they are lying about that and only reporting 17 out of some huge unknown number. The reality is that the vast majority (>99%) of climate change research does not produce the type of forecast, or really any forecast at all, that one could do this analysis on. Therefore, the 17 forecast number is actually ver…

With sincerity, I think you may misunderstand what GP is pointing out. I understood “forecasts” to be a simulation run using a model. In that case, they took 14 models and ran those each at least once to come up with 17 predictions. What GP wants is transparency around how many possible models were there over the last decades from which they chose the 14. This gives the reader a better understanding of whether 14 is…

They did a literature search and chose what was reproducible. This has been covered in TFA and highlighted by a comment nearby: https://news.ycombinator.com/item?id=21709446

Re: Even 50-year-old climate models correctly predicted global warming

#180

Post by nanis »

Some time ago, I got curious about the sources of data in the period where humans were measuring temperature using thermometers. When we talk about predictions from fifty years ago matching today (also I read the linked article and I didn't really see proof that a model that was specified at that time being explicitly benchmarked), it matters how the distribution of sources of surface temperature data varied since then.

I haven't been motivated enough to do another one of these animations, but this video[1] shows locations of all temperature stations in GHCNv3[2] (I see that GHCNv4[3] is out). Notice how where humans measure temperature depends so much on what living conditions humans seek or find acceptable. I find this visualization interesting in terms of the number of modeling questions it poses. If you are interested, it shouldn't take much to replicate something similar with current hardware/software.

As another note, it makes little sense to say model predictions are not statistically significantly different from each other. What is the population from which these models are being drawn? What is the measure of variability among models? The article[4] is behind a paywall, so I can't see what they did.

[1]: https://www.youtube.com/watch?v=h95uvT67bNg

[2]: https://www.ncdc.noaa.gov/data-access/land-based-station-dat...

[3]: https://www.ncdc.noaa.gov/data-access/land-based-station-dat...

[4]: https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2019...

Post reply on HN