Forums  > Off-Topic  > global warming poll  
     
Page 3 of 5Goto to page: 1, 2, [3], 4, 5 Prev Next
Display using:  

ahd


Total Posts: 30
Joined: May 2017
 
Posted: 2020-01-01 15:05
@strange, yes, a healthy amount of humility is always good. my wife likes this comic https://xkcd.com/793/ for reasons i can't fathom :D (make sure to hover the cursor over the image to read the extra bit). but, as @nt said, not everything requires deep subject expertise to be able to draw useful conclusions. although some things do. the real question is can we distinguish and i think the answer here is yes. your mileage may vary.

re. bathroom witticisms, another favorite: we aim to please. you aim, too, please.

the current poll at the link is 10 no's and 1 maybe. so, probably 10,2,1 if everyone who's expressed an opinion so far were to vote.

happy new year!

Strange


Total Posts: 1649
Joined: Jun 2004
 
Posted: 2020-01-01 16:13
Oh, certainly true that risk management is paramount. The key problem is that risk management in this case, like in most cases, comes at a cost. Here is a situation familiar to you:

Risk Officer: Inflatable Trader, you have a short position in zero-strike floors. It's too big and can take down the bank. Here are my models projections.
Inflatable Trader: You risk people! Your models failed to catch the rogue trader last year and those CDOs in 2008. Now you are trying to use your models to tell me what to do?! Inflation has never fallen below zero in the recent history and CBs will be in front of it!

The incentives are for the trader to keep his upside while the risk officer has relatively low motivation to press the matter. After all, the Inflatable Trader has a chance to get a nice bonus and sail into the sunset, while the risk officer will keep getting his base regardless. You recall the usual outcome of these conversations, right?

The parallels are obvious. All models are imperfect once you start undressing them (I know that for certain, I used to date one). Incentives are always towards getting rich now and punting the outcome to the next guy. Getting people to pay attention is only possible if you get them to panic one way or the other.

'Progress just means bad things happen faster.’

goldorak


Total Posts: 1091
Joined: Nov 2004
 
Posted: 2020-01-01 22:45
@pj. You are asking for a logical link between two not well defined statements.

If you are not living on the edge you are taking up too much space.

chiral3
Founding Member

Total Posts: 5163
Joined: Mar 2004
 
Posted: 2020-01-01 23:07
@strange: didn’t want you to think nobody got the Limelight reference; however, I never did that backroom thing. Funnily I think the Limelight is a shopping mall now. We are getting old.

Nonius is Satoshi Nakamoto. 物の哀れ

pj


Total Posts: 3522
Joined: Jun 2004
 
Posted: 2020-01-02 06:45
> You are asking for a logical link between two not well defined statements

Please, could you expound?

The older I grow, the more I distrust the familiar doctrine that age brings wisdom Henry L. Mencken

goldorak


Total Posts: 1091
Joined: Nov 2004
 
Posted: 2020-01-02 12:55
Dear pj, please define in unequivocal language the following statements:

"climate change := "
"accuracy of climate models :="

This is a necessary prior to any logical analysis of statements as you know it, don't you? Equivocal statements resist any logical analysis as you can see from my previous example.

You were asking the following:

> I am asking the third time, what's the logical connection between the accuracy of climate models and the climate change?

But do not waste your time, it is impossible. I was just trying to explain to you that asking that kind of question as if you held the key that would shut up all the others does not bring you anywhere. I know you can be better than this, come on.

If you are not living on the edge you are taking up too much space.

ebal


Total Posts: 20
Joined: Jul 2009
 
Posted: 2020-01-02 14:23
ahd, as far as I know, any reasonable model predicts the increase of global temperatures by a few degrees at least, which could lead to undesirable consequences of varying degrees so it’s prudent to manage this risk.

As a physicist, I understand that precise forecast is impossible because there are factors we don’t understand, don’t know about , impossible to model because of uncertain societal response etc. However, in the end it’s a simple balance of energy; energy comes in from the sun as sun light and leaves via infrared. Greenhouse gases stop the latter. Extra heat is absorbed mostly by the ocean (as is CO2) but the average temperature and CO2 concentration of the atmosphere also rises that leads to positive feedback; permafrost melting and releasing more greenhouses gases than humans ever did, polar caps melting and reducing light reflection, etc. Sure, there are many more effects, like clouds, but it can go both ways and I don’t think any one of them is even close to offsetting the above.

We’ve had enough observations to know that increasing average temperatures, whatever it means, by a a degree in 200 years but mostly since the seventies, is highly unusual, and rarely happens not on geological scale.

pj


Total Posts: 3522
Joined: Jun 2004
 
Posted: 2020-01-02 14:52
Aha, so the thread title and the poll question are unrelated.

The older I grow, the more I distrust the familiar doctrine that age brings wisdom Henry L. Mencken

goldorak


Total Posts: 1091
Joined: Nov 2004
 
Posted: 2020-01-02 15:10
Have you ever read a book pj? Ever heard of topic / sub-topic?



If you are not living on the edge you are taking up too much space.

chiral3
Founding Member

Total Posts: 5163
Joined: Mar 2004
 
Posted: 2020-01-02 15:49
Ebal is resonating, and kinda my last point before, which I think is *potentially* the Achilles of the argument ahd mentions re temperature causing the increase versus the other way around. The Occam’s razor of this all is that industrialization phucked up the equalibria and, at a minimum, hosed up the characteristic time scales the cycles were orbiting in. I am sure there’s a diff eq with a forcing function that describes this and is non-linear and has chaotic solutions and all the things that all the dynamical systems people want but, at the end of the day, this will never be modeled or figured out: the ticker will punch out another daily data point of temp, min, max, stdev, ... and all the monkeys will scratch their heads because the models don’t work and start punching, fucking and killing each other over what the data point means.

Nonius is Satoshi Nakamoto. 物の哀れ

pj


Total Posts: 3522
Joined: Jun 2004
 
Posted: 2020-01-02 18:52
>Ever heard of topic / sub-topic?
nah, just of bait and switch.
< Added >
The last analogy.
Let's decide whether the put-call parity
holds by establishing whether Black-Scholes model is true in reality.

The older I grow, the more I distrust the familiar doctrine that age brings wisdom Henry L. Mencken

ebal


Total Posts: 20
Joined: Jul 2009
 
Posted: 2020-01-02 21:42
ahd, since you mentioned xkcd, I find the below one pretty illuminating and funny
Average temperature

nikol


Total Posts: 1126
Joined: Jun 2005
 
Posted: 2020-01-02 21:45
What bothers me is that alternative opinion (no Global Warming) is suppressed severely in the public discussion.

Still, there is no Galileo with "And yet it moves", hence, likely, there is no strong evidence for either hypothesis.

rftx713


Total Posts: 125
Joined: May 2016
 
Posted: 2020-01-03 01:14
apologies but it's hard to take you seriously when you say the alternative opinion is suppressed severely. my experience has been quite the opposite. my personal experience of the public discourse has been people falling roughly evenly into 3 camps:

1. denialists: usually supporters of a certain politician who dominates the news cycle; these arguments are usually too full of blatant logical fallacies or rely on research funded by Koch, etc. Debates with this camp usually devolve into identity politics.

2. enlightened centrists / dunning-kruger-ists: usually well educated, working a well paying job, living in a major metro area; these arguments are usually too full of blatant ignorance of their own ignorance and tend to pick (random) overly nuanced views that make them look intelligent without pissing too many people off. if i was a psychologist i might even try to squeeze in an element of rationalizing (as smart people are best at) that the future really will be OK. these debates will devolve into more and more granular intelligence measuring contests (aka a smaller subset of identity politics).

3. believers: usually not supporters of a certain politician who dominates the news cycle; these arguments are usually too full of blatant logical fallacies or rely on Netflix documentaries, etc. These debates also devolve into identity politics.



Anyways, I have no idea what's happening, or not happening, with the climate. I do know that people who have dedicated their life to this are mostly saying we should probably be careful, and I also know from my own education/profession that there are viable options for prudent risk management of this scenario. Not risk elimination - risk management. So I do struggle with those who want to debate it until the end of time. Just like with all of our other existential risks, we should do the best we can to manage it without identity politics hijacking the course and killing whatever effectiveness the risk management would have had. (Nuclear nonproliferation, disease control, upholding the social order, defense against another superpower, the list goes on. Nobody knows what the outcome might be, but nobody disputes that it's smart to dedicate resources proportional to the threat that the experts laid out.)

nikol


Total Posts: 1126
Joined: Jun 2005
 
Posted: 2020-01-03 09:19
In private life, yes, there are those categories. But the public discussion is dominated by the third one driving it to histerya level. This creates severe market distortion similar to the pre-burst stage of the bubble on stock market.

chiral3
Founding Member

Total Posts: 5163
Joined: Mar 2004
 
Posted: 2020-01-03 14:49
This is kinda interesting re scale: it would suggest that from 1999 it took a year's worth of local daily data to correlate with global data (using their methods) but only a day's worth of data from 2012 onward. Said differently, they would interpret this to mean that the larger length and time scales are mixing with the smaller scales, which would suggest that climate = weather. Not sure I can fully appreciate that final leap but the part about the scales is potentially interesting.

Climate change now detectable from any single day of weather at global scale

These top-down (backward) analyses are nice, because they are easy to digest, but lacking, because they are in-sample statistical analysis. On the flip side I am not sure I have the bandwidth to even begin to assimilate the causal analyses. For instance, I understand that wind across the caps plays a big role and the wind has picked up in recent time by 0.5m/s.

Nonius is Satoshi Nakamoto. 物の哀れ

Maggette


Total Posts: 1233
Joined: Jun 2007
 
Posted: 2020-01-03 16:53
Thx chiral3

That's interesting. Thats quite a strong statement in the paper. I will definetly need some time to process it.

Actually even though the scientist say climate is different than weather, I learned that the simulation models are structually very similar.

When I started to fool around with the topic of GCMs and specifically AOGCMs, I did what everbody did: go to wiki (not bad). After that I did a short dive into the blog sphere. That didn't get me anywhere.

One of my Jülich friends recommended a book to me and I really liked it:
Demystifying Climate Models: A Users Guide to Earth System Models (Earth Systems Data and Models Book 2)

I downloaded some data from the IPCC (links in my first post) and again explored it a bit.

And I repeat myself: IMHO it's not that bad. Certainly not as bad as advertised. The prediction global trend out of sample made by modles 20 or even almost 30 years ago is quite good!!! In sample even better... Much better than most macro economics stuff I am aware of.

And of course: the models are worse out of sample. If I use f(x1_t,x2_t,x3_t) to predict y_t ...it helps if my inputs x1_t, x2_t and x3_t are data points, and not predictions themselves. Which at _t-1 they are.


And most important: if it is soooo easy to tweak the simulation models : is there a single model by the "human impact is not important" scientist, that predicts anything right? I am talking in sample here? I am not aware of even one published model that is still out there?!?

I am pretty sure: if you can come up with a good non.human related explaination and have a decent simulation model that can fit past data well and has an ok prediction let's say 5 years out of sample....you might get booed by the established community, but the Heartland Institute and Exxon will pay quite well and push your research.

Ich kam hierher und sah dich und deine Leute lächeln, und sagte mir: Maggette, scheiss auf den small talk, lass lieber deine Fäuste sprechen...

silverside


Total Posts: 1417
Joined: Jun 2004
 
Posted: 2020-01-03 17:10
I haven't got the eloquence to write a well argued post contradicting the above

but at some point I think you have to trust the scientific consensus when it is backed up by more and more events outside historical norms , most obvious of which recently is the wildfires in Australia

otherwise you are aligning yourself with the flat earthers, anti vaxxers and other dangerous extremists.

ahd


Total Posts: 30
Joined: May 2017
 
Posted: 2020-01-04 14:17
What are the relevant timescales for climate? Is 5 years meaningful? Is it like a week of data for a SR = 5 HFT guy who makes so many independent bets in a week that the CLT almost guarantees him a profit if he has any real edge? Or is it like a few years of performance data on a slow trend following system which gets down so few bets that sample size is tiny, noise dominates signal, and sample stats have almost no relation to the population stats which will govern (very) long term performance? I think climate is a long timescale phenomena so climate models have had almost no out-of-sample test. It's almost a tautology, really, when you consider the time separation over the data to which they're fit and the time spanned by the in-sample data. If you fit the parameters of your HFT system to annual bars and then trade it tick-by-tick, how confident of future profitability will you be after making money for the first few minutes?

ahd


Total Posts: 30
Joined: May 2017
 
Posted: 2020-01-04 14:51
also, re. "aligning yourself with the flat earthers, anti vaxxers and other dangerous extremists": i don't think that's fair but i definitely worry that you're right, i.e. i'm very uncomfortable to be contrary to such a widely and strongly held majority opinion. but i think it's wrong just to cave in and say "so many others disagree that they must be right" without understanding where my logic is wrong. i'm a smart guy, i believe i have a good understanding of issues that are definitely relevant, why should i turn my brain off? shouldn't someone be able to say "oh no, this part of your argument is wrong because..." instead of just yelling even more loudly? also, flat earthers ignore observations that directly contradict their data and anti-vaxxers ignore the statistically signifcant results of well-designed experiments as well as ample epidemiological data. no one has pointed out a similar irrational feature of doubting the reliability of climate model forecasts.

chiral3
Founding Member

Total Posts: 5163
Joined: Mar 2004
 
Posted: 2020-01-04 14:55
I don’t know, but I get the impression that, like markets, the characteristic timescales are getting smaller? In markets we have an assumption of underlying process and we can talk about integrating an autocorrelation function to back out time scale or looking at power spectra - if the power spectra goes at ~ f^-n we can say things about the information sharing between long or short distance points in the series depending on n, which is just saying the autocorrelation function falls off fast enough to suggest nothing really long-range is happening. In financial markets we have assumed a model for n=2. I haven’t thought about this in years, but if my memory is correct I understand that, empirically, financial autocorrelation functions have been falling off faster and faster in the last century, which would suggest the logical: that trades used to be slow (tickets) and now they are fast (electronic / HF). Not sure how to think about this in terms of climate, although I am sure someone smarter than me has done some work, and the parallels to physics (fluids, plasmas) are obvious. E.g., how does (local) vortices effect (global) flow?, phase transitions at critical Reynolds numbers, etc.

Anyway, at least as it relates to that paper, I suppose the suggestion is that, up until 2012ish, and this is probably a leap, the autocorrelation function would have suggested very long-range memory, and that has changed, suggesting that weather -> climate.

Nonius is Satoshi Nakamoto. 物の哀れ

Maggette


Total Posts: 1233
Joined: Jun 2007
 
Posted: 2020-01-04 23:43
And I ask again:
we are talking about physical (simulation) models, not statistical/ML black box models. There are overlaps regarding danger of over-fitting, but it's not the same.

The scientific community has models, which in median describe the observed phenomena quite well (in sample) and in parts have an ok 20-30 years out of sample prediction.

I am not aware of any physical model published by "human driven climate change is a hoax" peoples that predicts well in sample. Let alone out of sample. If it's so easy to create simulations that do well in sample and explain away the human impact, where are they?

Computing power is cheap. Should be done by now, no?

Ich kam hierher und sah dich und deine Leute lächeln, und sagte mir: Maggette, scheiss auf den small talk, lass lieber deine Fäuste sprechen...

ahd


Total Posts: 30
Joined: May 2017
 
Posted: 2020-01-05 14:00
@maggette, it's not that there's a different model... it's what phenomena are included. read http://www.sciencebits.com/CO2orSolar and companion articles for a credible, well-reasoned discussion of why ipcc overestimates the strength of the greenhouse effect. in a nutshell, their model ignores the established physics of increasing sunspots --> increasing magnetic field --> decreasing atmospheric cosmic ray bombardment --> decreasing cloud nuclei formation --> less capture of blackbody radiation from surface --> cooler temps. so when they fit their model to the empirically observed (ice core) correlation between [c02] and temp they have to overestimate the strength of the greenhouse factor to account for the ignored solar modulation factor. but this is still, imo, a small issue compared to the in-sample nature of the many fitted parameters and the fact that there are many zeroth order, important feedback mechanisms that are not understood well at all. still no reason to believe the forecasts!

chiral3
Founding Member

Total Posts: 5163
Joined: Mar 2004
 
Posted: 2020-01-05 15:24
Your closer to it than me ahd, but haven’t we gone into a period of time with less sunspot activity? In your link the data presented is older. Hasn’t there been a divergence in that warming has been accelerating with (mean) solar activity decreasing?

Nonius is Satoshi Nakamoto. 物の哀れ

Maggette


Total Posts: 1233
Joined: Jun 2007
 
Posted: 2020-01-05 18:53
That is another model.

@chiral

That's the point: less solar activity implies more cosmic rays (energy form outer space, not the sun) hitting earth. That again some people say does have an effect on cloud formation and would create heating (edited).

I think that has been considered and rejected a while ago.
Look at the propaganda page (but I think that's the official consens)
https://skepticalscience.com/solar-activity-sunspots-global-warming-advanced.htm

Ich kam hierher und sah dich und deine Leute lächeln, und sagte mir: Maggette, scheiss auf den small talk, lass lieber deine Fäuste sprechen...
Previous Thread :: Next Thread 
Page 3 of 5Goto to page: 1, 2, [3], 4, 5 Prev Next