Hacker News Reader

Comment by arctangos | Hacker News Reader

The authors don't include some extremely nasty feedback loops the appear to currently be happening. For example, methane release from permafrost melt.

As a result, the simulation appears to be misleadingly optimistic.

- This tool is based on climate interactive's previous one, C-Road.

- in the C-Road simulation when you click on parameters -> assumptions to add parameters about methane release through permafrost melting and human activities. By default, it's set to zero, but even their maximum values don't seem as impactful as I would expect.

- in EN-Road simulation (this one) the methane release parameters are removed! So, no possibility to add any positive feedback loops. It feels like the relation between CO2 and temperature is a simple linear equation, or close to it.

View on HN

While negative feedback loops are big in the media, they are not well understood and difficult to model.

Also, other feedback mechanisms seem to get little attention. Are we only looking to model the feedback mechanisms that reinforce a narrative of cataclysm?

Don't other feedback mechanisms exist that lessen CO2 concentrations, or temperature dissipation?

No. They aren't "only looking at feedback mechanisms that reinforce cataclysm." Modelers include every feedback that they can quantify well enough to include and the ones they don't understand well enough are studied carefully to be able to work out what the uncertainties are.

Scientists 50 years ago predicted that we would be at around 1C (1.8F) of warming by 2020 with a 40% over preindustrial co2 level. That's exactly what we have now. They predicted almost perfectly today's climate decades before it came to pass. That may not seem like a big deal until you understand that global temperatures haven't been this high in 130,000 years.

It isn't being recognized properly yet in the media, but the stunning accuracy of predictions by climate modelers is one of the greatest achievements in the history Earth sciences.

This sounds like you're saying that 50 years ago there was a single climate model, upon which significant consensus among scientists had been reached, that predicted a 1C temperature rise. I've read quite a few articles and discussions on this topic but have never heard this before.

Am I misunderstanding you? If not, do you have a link that discusses this?

Perhaps the Sawyer paper from 1973? https://www.carbonbrief.org/analysis-how-well-have-climate-m...

Surely there was one model that happened to come up with the right prediction, I would assume there were multiple. But my questions is whether there was one model with consensus support (which seems to be the claim), which would be necessary to rule out survivorship bias.

> Surely there was one model that happened to come up with the right prediction

What does that mean? This "model" is someone sitting down and, using what we know of physics, calculating how something will warm up when you add some extra energy (in this case, the earth). You surely don't think that physics works by people randomly guessing answers and then one of them "happened" to be correct?

You don't, do you?!

From the 1975 US National Academy of Sciences/National Research Council Report

...we do not have a good quantitative understanding of our climate machine and what determines its course. Without the fundamental understanding, it does not seem possible to predict climate... [0]

The point is that there were a lot of climate models done in the 70's, and only a few of them were accurate. No one was randomly guessing, but the model you gave was not the consensus.

[0] https://skepticalscience.com/ice-age-predictions-in-1970s-in...

When a thousand hedge funds hire mathematicians and economists to create models of the global economy so that they can predict whether individual stocks will move up or down over different timescales, their task is similar to predicting the weather. Some will beat the market, many won't. Are they guessing?

Maybe that's a bad analogy, but my point is that climate is a chaotic system, and detailed models might be right and might be wrong. It's not that the scientists are guessing, of course the science is methodical - but the performance of the model might not be better than guessing, hence the speculation of survivorship bias. There were lots of climate models over the last century, was there consensus around a paper that is now proven correct? Or were there a hundred models that were off and one that, yes, happened to be correct. I don't know the answer.

I think an individual stock in this analogy would be "Will it rain in Duluth". Climate predictions, to use your economic model, would be more along the lines of "Will stocks continue to rise on the whole over the next 100 years?"

> What does that mean?

It means that if you are dealing with a complex multivariate system where the behavior of individual variables is not perfectly understood, let alone the behavior in an interactive system, and you are building multiple systems to model the environment and output a prediction between 1 and 10, undoubtedly some of the models will be correct (since you are predicting within such a small range), but there is no good way of knowing whether the correct prediction is due to proper modeling, to chance, or to some combination of the two.

> You surely don't think that physics works by people randomly guessing answers...

No, I do not think this.

> ...and then one of them "happened" to be correct?

Yes, in that I am suggesting that this is a possibility.

> You don't, do you?!

Rereading my comment, I see no actual content that suggests I believe this.

I wonder, if you were to reread my comment while keeping in mind the theory of System1/System2 thinking [1], do you still reach the same conclusion?

And for clarity, while this may seem (again, see [1]) that I am antagonizing you, I am actually and sincerely trying to make what I consider to be an incredibly important and overlooked point: the manner in which human beings think, on particular (culture war, identity related) topics, is a core problem that is preventing forward progress on not only climate change, but a wide variety of issues.

I am completely open to the idea that this theory is incorrect, but it seems almost no one who is even willing to acknowledge the possibility that it may be correct - an observable phenomenon which I proclaim further supports the theory.

[1] https://en.wikipedia.org/wiki/Thinking,_Fast_and_Slow#Summar...

Even zero and one dimensional ultra simplified energy balance models come pretty close to approximating how much warming we've had based on the amount of co2 we've released.

Global average temperatures are governed simply by the need for incoming energy to be balanced by outgoing energy.

You're confusing the complexity of weather with the complexity of climate. A lot of the value of increasingly large scale, complex, supercomputer based climate models is the spacial resolution to be able to tell people of particular regions how their local climate might be affected. For instance a lot of work is going into understanding how the monsoon season will be affected.

"Evaluating the performance of past climate model projections"

https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2019...

"We found that climate models – even those published back in the 1970s – did remarkably well, with 14 out of the 17 projections statistically indistinguishable from what actually occurred."

Yes, I read that report and participated in the subsequent discussion [0] where I raised a number of concerns with the study, but wasn't able to find anyone who would seriously and objectively address them.

Frustrated by the unwillingness of anyone to engage in rigorous discussion in that thread, I also contacted the authors to see if they were willing to address some of the concerns. Unfortunately, I did not hear back.

I must confess, this strange combination of group declaration that "the science is in", combined with complete refusal to discuss "the science", leaves me feeling a bit suspicious about the degree to which anyone has actually read the science, or is thinking critically on this matter.

If you could address some of the concerns I raised in [0], I would be delighted.

[0] https://news.ycombinator.com/item?id=21708936

> I raised a number of concerns with the study, but wasn't able to find anyone who would seriously and objectively address them.

This is an inaccurate summary of that thread, in which you raised a simple accusation that the article cherry-picked the models, were swatted down with a citation from the article referenced in the OP, and then retreated into a maze of long, twisty replies that failed to raise any other specific points.

Your comments in this thread have had the same flavor. Consider brevity and specificity.

> This is an inaccurate summary of that thread, in which you raised a simple accusation that the article cherry-picked the models

Ironically, or not, your summary of my summary of that thread is inaccurate.

> "you raised a simple accusation that the article cherry-picked the models" is not what I did. What I did do is ask for an explanation of how we know no cherry picking occurred. Their methodology for article selection was not published, so therefore it cannot be reproduced. I emailed the authors asking for clarification, and they did not reply.

> were swatted down with a citation from the article referenced in the OP

"swatted down", but didn't address my actual question, rather re-reffering to the original ambiguous wording that I was complaining about.

A reproducible methodology could clear up the uncertainty (and please note, my claim is that there is uncertainty not that there is wrongdoing), yet no methodology was offered in the original paper, and the authors did not reply to the request.

In such situations, I adopt a position of "Unknown - more information is required", but obviously others have a much more flexible approach to what they're willing to believe. Although, I wonder if this approach varies depending on the subject - would be interesting to read up on.

> and then retreated into a maze of long, twisty replies that failed to raise any other specific points

"Mazes of long, twisty replies", and failures to "raise any other specific points" are a common consequences when one party in a discussion is unwilling to directly address questions as asked, and the other party's response to that is re-asking the same question.

> Your comments in this thread have had the same flavor.

Indeed they do, as do the replies: unwillingness to directly address questions as asked.

> Consider brevity and specificity.

I will do so.

In return, please consider honesty, epistemic humility, your willingness to acknowledge that ambiguity often exists in written language, and whether a lack of 100% agreement should be interpreted as opposition, as opposed to curiosity and strictness. Sometimes those who appear to be your enemy may actually be some of your best friends.

EDIT: Here's another way to look at it (a better description of my main intent): "conspiracy theorists" and "deniers" are a problem, to some degree, in broad acceptance of climate change messaging, agreed? Might it be a good idea to consider whether there are some flaws in the messaging that could be improved, that would result in reducing the material they have to work with in any influence campaigns?

Your question was answered in the comment underneath it -- the answer is taken from the article in the discussion:

https://news.ycombinator.com/item?id=21709446

Basically: selection criterion was, any published model that had the inputs ("forcings") and outputs (temperature) needed.

Then, incredibly, the paper author chimed in with this offer:

https://news.ycombinator.com/item?id=21709010

And you couldn't respond constructively. If he didn't include a model, point it out.

Instead, you've come to another thread complaining that you had concerns and "wasn't able to find anyone who would seriously and objectively address them".

Pardon the delay in replying, I didn't see this until now.

> Your question was answered in the comment underneath it -- the answer is taken from the article in the discussion:

> https://news.ycombinator.com/item?id=21709446

Well, shit. This is kind one of those timing error, at the time I wrote that the thread was quite young, so I presumed my more substantive complaint (which is immediately under my comment) would have been seen. That'll teach me for being trite, hopefully.

Anyways, here is that comment (and I imagine you won't like my pedantry, but I did note it as such, and I'm not forcing anyone to answer - if they'd like to allow ambiguity/uncertainty to remain, be my guest):

---------------

https://news.ycombinator.com/item?id=21708936

> 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 published in papers, using special software to gauge the exact numbers represented by points on a printed graph.

They compared annual average surface temperatures across the globe to 17 forecasts, from 14 separate models.

But....how many models did they look at, before choosing those particular 14 models?

TFA doesn't say.

Articles written in this style provide rich fodder for conspiracy theorists, particularly because there is no shortage of examples in the past where authoritative, trustworthy organizations have gotten caught in lies. Rare is the climate change article I've read that can't easily have similar holes poked in it.

What's the real truth here? Based on the literal content of this article, no one knows. It is speculation vs speculation.

---------------

> Then, incredibly, the paper author chimed in with this offer:

> https://news.ycombinator.com/item?id=21709010

> And you couldn't respond constructively. If he didn't include a model, point it out.

An obvious lack of thoroughness and reasonableness on my part - it would have been wiser to restate my above question in it's entirety rather than expecting him ti find it. So I'll take the reputation hit on that one.

But then on the other hand, I did email one of the authors directly (after looking through the github repo to see if my questions were answered there), with a much more detailed question, in polite terms, and with mention of the possibility that this sort of ambiguity is open to exploitation by propagandists, but didn't hear back, so far anyways.

UPDATE:

To the author's credit, they did address a question posed to them here: https://news.ycombinator.com/item?id=21709249

Also, his reply to another comment is suggestive that all available models (that fit his criteria) were in fact used: https://news.ycombinator.com/item?id=21708782

I get a fairly trustworthy feel from him after reading those, although I still prefer that these things were written using more defensive wording.

"While negative feedback loops are big in the media"

You mean positive, right? As in self-reinforcing? Negative would mean the opposite, and there are a few of them (greater plant growth etc.) but they seem to be more than cancelled out by the positive (bad) feedback loops.

If they did exist, we would take their effects as baseline assumptions and just ratchet down our opinion about the crisis, if that makes sense. To the extent there are buffers, we just assume their effects, and we'll be unpleasantly surprised when the buffers run out.

Examples: in the 70s we had a great deal of particulate smog, which effectively reduced some of the radiation hitting the Earth and camouflaged the effects of global warming. But instead of saying "hey, global warming must be a big problem since we're noticing X amount of warming EVEN WITH PARTICULATE SMOG", it just let us ignore global warming for longer.

Example: the deep oceans have been an effective heat sink and some amount of the global heat has gone into them. Again, instead of saying "global warming must be a big problem since we're noticing X amount of land warming EVEN THOUGH THERE'S A GIANT HEAT SINK", it just let us ignore global warming for longer.

So, to the extent that there are any buffers, we'll just build them into our baseline of "how bad the problem is", and therefore ALL of our surprises and feedback effects will be seen to be negative.

Reply on news.ycombinator.com