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

Climate models are running red hot, and scientists don’t know why

bloomberg.com

Post reply on HN
Print view View original

261–270 of 332 posts

Re: Climate models are running red hot, and scientists don’t know why

#261

Post by Udik »

KevinEldon wrote: post #240
microcolonel wrote: post #215

Earlier quoted context omitted.

Don't need to bulldoze much of anything to do something though. The inefficiency of buildings is a tiny part of the problem, especially in temperate climates; more mixed use zoning, and the relaxing of single family dwelling zoning, will have a dramatic impact on excess travel. Transport accounts for basically a third of U.S. energy use, though it has been hardly growing over the last few years, which is a good start…

I think we agree. I meant inefficiency in building density not energy use. I was thinking about row houses in San Francisco and brownstones in New York City. Old short and inefficient buildings could be torn down and replaced with tall, dense, efficient structures; but people who own row houses, or rent apartments with high rents, or like the way their neighborhood is now don't seem to want this kind of transformatio…

And over what time period would you finally offset the excess co2 introduced in the environment by tearing down and rebuilding existing buildings? They're already medium-density buildings, with considerable historical and urbanistic value. Seems a crazy idea to me.

Re: Climate models are running red hot, and scientists don’t know why

#262

Post by jknoepfler »

lazyjones wrote: post #250
jknoepfler wrote: post #221

Earlier quoted context omitted.

I'm well aware of the distinction between weather and climate. I was asking if the climate change skeptic would take a bet that the global mean temperature (operationalized in some acceptable way) would decline in 2021 compared to 2020. If they believe that the earth is not, on average, getting hotter, then they should be indifferent to this bet (inter-year variation should be random around a fixed point). I don't th…

> If you wouldn't bet against the climate getting warmer, then I don't think you have any place calling yourself a "climate change skeptic" The climate isn't getting warmer or colder just because next year's temperature does.

sure, but if I believe the climate is getting warmer or colder, I know where to place my bets. feel free to propose a better or more precisely defined bet.

Re: Climate models are running red hot, and scientists don’t know why

#263

Post by SiempreViernes »

samatman wrote: post #129

I'm reminded of Feynman's famous anecdote about the Millikan oil drop experiment: Millikan underestimated the charge of the electron, and over the ensuing years, subsequent measurements slowly but surely converged on the correct value. Why? Because experimenters who got the correct value figured it must be too high, and ran the experiment again until they got a more-erroneous value, one that was nonetheless higher. E…

> I'm afraid that the incentives are to credit models that predict more warming just a bit more than those which predict less; the expected effect would be that the predicted warming effect would get larger over time.

This actually doesn't make any sense! In your valued Milikan example, the problem was subsequent experiments were biased towards lower values, meaning that for climate models the bias would be expected towards under heating models. In other words, the lesson from Milikan is that there is a risk our models aren't heating fast enough.

So by the quoted example we should be worry that our models under predicts, and yet Milikan is stood up to argue the models are too wild in their warming predictions?

Re: Climate models are running red hot, and scientists don’t know why

#264

Post by konamicode »

mondoshawan wrote: post #139
leifdenby wrote: post #47

There is a big observational campaign happening right now that is trying to get to the heart of this called EUREC4A (main website: http://eurec4a.eu and information sheet: http://eurec4a.eu/fileadmin/user_upload/eurec4a/2019/Press/E... ). The key uncertainty between different climate models is whether shallow clouds in the tropics (which in general cool the Earth through radiation) will disappear or not and make our…

Can we get the source for the models?

This is mentioned elsewhere in the comments at least some of the code for climateprediction.net appears to be on GitHub: https://github.com/CPDN-git.

Re: Climate models are running red hot, and scientists don’t know why

#265

Post by cultus »

dls2016 wrote: post #157
cultus wrote: post #17

Earlier quoted context omitted.

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.

> because they are actually running inverse problems with the PDEs to fit the data Could you explain this a bit more? Are you talking about data assimilation?

Yes, that's right. The issue is that systems like these have a lot of free parameters that have to be estimated from the data. This is basically because you have to make assumptions if you don't want to simulate Earth's systems down to the subatomic level. The inferred values of these parameters are usually very important as well for interpretation.

All PDEs are solved numerically by discretizing them into linear or nonlinear systems. If it's convex enough, Tikhonov regularization is often used.

Linear PDEs, where the solution depends only linearly on its derivatives, can be discretized into matrices. This gives a system like A m = d, where A is the discretized PDE, m is the "model" or parameter values, and d is the data. This is usually over or under-determined from the data, so it has to be solved by least squares. Since the resulting system is usually almost singular, regularization is applied, usually by penalizing the L2 norm of the model:

min || A m -d || + r ||m||^2

If it's nonlinear but convex enough, one can just solve it by fixed point iteration, with the above linear equation being solved at every step (Newton's method often doesn't work well). If it's too nonconvex, then MCMC or other Bayesian methods are better, since a single solution is useless if the parameter values could plausibly around widely separated minima.

Re: Climate models are running red hot, and scientists don’t know why

#266

Post by tzs »

lbsnake7 wrote: post #30

I’m a climate change skeptic almost entirely because I don’t trust models. I believe there are too many dynamic things happening in the world for anyone to be able to predict accurately say the next 100 years (solar, earthquakes, volcanoes etc). One way to make me a believer though is an experiment: if I were to give you a 100 year stretch of history and you were able to accurately predict with the models what exactl…

> I’m a climate change skeptic almost entirely because I don’t trust models.

We don't need any climate models to know that Earth is warming. We can tell that just from what we know of: how electromagnetic radiation works, the basic thermodynamics of gases, determining isotope ratios of the carbon making up a sample of CO2, how much solar energy falls on the Earth, the mechanisms by which energy can leave the Earth, the amount of energy leaving Earth as measured by satellites.

That's enough to allow verifying that:

(1) a net heating of the Earth taking place,

(2) this is mostly caused by humans (mostly be emission of greenhouse gases from fossil fuels) (and yes, we can tell the difference because emissions from burning fossil fuels have a different carbon isotope ratio than other sources), and

(3) it is getting more extreme.

All you need climate models for is to try to figure out where the extra heat will end up and what it will do there.

Re: Climate models are running red hot, and scientists don’t know why

#267

Post by jknoepfler »

lazyjones wrote: post #255
jknoepfler wrote: post #249

Earlier quoted context omitted.

On a crude first pass, I will accept any combination of the following datasets, assuming they are updated in 2020 and 2021: HadCRUT4 (from UK Met Office Hadley Centre and the University of East Anglia’s Climatic Research Unit) GISTEMP from the NASA GISS MLOST from the NOAA JMA from the Japan Meteorological Agency I'm open to other datasets. I'm not choosy.

So, you would have lost the bet in 2016 and 2017. Did climate get colder?

no, a noisy measure around a globally increasing trend line went down.

I believe that the noisy measure goes up, on average, because it is (noisily) measuring a warming climate. I'll make that bet every year, and I will come out ahead...

using the trend line, you could set odds at which the betting on a colder 2021 would be the correct bet - but it isn't 50/50, which it would be if the process were random noise.

as a professional poker player I'm not surprised by this argument, but I am saddened.

edit: do you actually not understand the basics of statistical inference? this is like, middle-level high school math.

Re: Climate models are running red hot, and scientists don’t know why

#268

Post by SketchySeaBeast »

giarc wrote: post #137
SketchySeaBeast wrote: post #49

Earlier quoted context omitted.

> If in 40 years, my kids ask me what I did to prevent climate change, I feel like I'll have an answer I'll be proud of. Doesn't it kind of feel like this is the equivalent of a kid in the 70's asking why their parents didn't stop World War 2? We absolutely need to be as vocal as possible, but as individuals, outside of trying to get people with power to affect the change we need, I don't know what we can expect to a…

>You just can't do enough diet modifications, carbon footprint reduction, or recycling to have it matter on a global scale. I used to think like that. I used to think "There's nothing I can do, why do anything." Then I had kids and I realized that I need to be able to tell them I did something. Our world might be a wasteland of burnt forests and flooded cities, but at least I will be able to look my children in the e…

I feel as though it may ring a little hollow.

Re: Climate models are running red hot, and scientists don’t know why

#269

Post by willyt »

cryptica wrote: post #142
Accujack wrote: post #66

Earlier quoted context omitted.

Or that there are multiple inaccuracies in the model which interact, so finding and correcting each one makes progress but still doesn't improve the model overall.

Having built some models to simulate very basic things myself, I've seen first-hand how the slightest deviation of the model from reality can completely invalidate the entire output of the model. For this reason, I'm skeptical of models that try to simulate very complex systems with a lot of variables and feedback loops. Models that have a large number of variables are particularly difficult to get right because diff…

It's not impossible. Listen to this podcast, it's a discussion about weather forecasting models with the scientists that create them. They discuss model verification for weather models and they touch how this also applies to climate modelling. http://omegataupodcast.net/326-weather-forecasting-at-the-ec...

Re: Climate models are running red hot, and scientists don’t know why

#270

Post by adrianN »

nesyt wrote: post #153
rayiner wrote: post #108

Earlier quoted context omitted.

> Total War economy/society > quality of life improvements to our health and well being this kind of mobilization would bring would be astonishing. The impact on quality of life and society would be "astonishing," but they won't be improvements. The Soviets and Chinese tried similar "mobilization[s]" for food production. The idea was to direct society's efforts to producing enough food for everybody. Tens of millions…

Thanks for the straw man. American WW2 mobilization didn't kill ten million people - that's not an intrinsic attribute of a country focusing on something. Think also new deal / public works projects like the Hoover Dam (many died working on the Dam but not because it was a government mobilization project). Just the government spending a bunch of money on environmentally focused projects could happen and could be good…

At least it didn't kill ten million people unintentionally.
Post reply on HN