Simulation Code for Earth's Historical Temperature Record

[optotinker]

TL;DR Summary

Earth Temperature Simulator

I am new to the field of climate change. Are there publicly available simulation code for Earth's temperature history, going back, say 2000 years, with a resolution on the order of years?

Thank you very much.

 

[fresh_42]

Summary:: Earth Temperature Simulator

I am new to the field of climate change. Are there publicly available simulation code for Earth's temperature history, going back, say 2000 years, with a resolution on the order of years?

Thank you very much.

It is extremely difficult to separate serious information from politics when it comes to climate change. I think a good way to search for the former is to look up the keyword Keeling curve. It started in 1958 but I assume that it has been calculated backward. E.g.
https://keelingcurve.ucsd.edu/permissions-and-data-sources/

 

[optotinker]

Thanks.
I saw an old thread somewhat related to my question here:
https://www.physicsforums.com/threads/temperature-in-Earth's-history.631851/
I am looking for something more quantitative, i.e. a simulator that takes into account all factors that influences the temperature. With that, one can play "what if" scenarios.
Is the current consensus on global temperature solely based on various gases in the atmosphere?

 

[anorlunda]

I am looking for something more quantitative, i.e. a simulator that takes into account all factors that influences the temperature. With that, one can play "what if" scenarios.

It's easy to make simulations. But it can be difficult to verify that the model is accurate. With a lot of effort you may be able to show that the model would have predicted the known history, but how could you be confident that "what if" scenarios predicted are trustworthy?

 

[fresh_42]

I am looking for something more quantitative, i.e. a simulator that takes into account all factors that influences the temperature. With that, one can play "what if" scenarios.

What do you have at home, a Cray?

There are plenty of models that fall under your description, and little certainties for such a black and white view on the subject. Climate models require huge computing power and probably dozens of assumptions.

Have you looked on the websites of the major institutes across the globe? I assume that every industrial nation operates such an institute. Here are the homepages of two of them:
https://www.usgs.gov/
https://www.gfz-potsdam.de/en/home/

If you want to seriously examine the subject, then start to look for serious information. The Keeling curve is one such measure and the homepages above two serious institutes. If you find on the internet what you are looking for, then take it as a toy, not a serious simulation.

 

[optotinker]

It's easy to make simulations. But it can be difficult to verify that the model is accurate. With a lot of effort you may be able to show that the model would have predicted the known history, but how could you be confident that "what if" scenarios predicted are trustworthy?

It is indeed a very good question. At this point I am only looking at tools and data. I like simulation because a good simulator should take into account all known effects. It has the ability to synthesize these effects in a way far beyond the ability of a human brain, and its references should contain the key literature for said effects.

And of course there is the issue of trustworthiness. But all scientific work face the same issue at some point.

 

[optotinker]

What do you have at home, a Cray?

There are plenty of models that fall under your description, and little certainties for such a black and white view on the subject. Climate models require huge computing power and probably dozens of assumptions.

Have you looked on the websites of the major institutes across the globe? I assume that every industrial nation operates such an institute. Here are the homepages of two of them:
https://www.usgs.gov/
https://www.gfz-potsdam.de/en/home/

If you want to seriously examine the subject, then start to look for serious information. The Keeling curve is one such measure and the homepages above two serious institutes. If you find on the internet what you are looking for, then take it as a toy, not a serious simulation.

In the old days, we got free code from Universities and used them for serious industrial applications. I am looking for something like this in this field. I would worry about computing capabilities later.

I want to get into the field as a player, not a spectator. The information offered by major institutions to the general public is helpful but it is already someone else's work, not mine.

 

[hutchphd]

I guess a certain hubris is required for every undertaking. All you need to model is the entire planet! But one step at a time.

 

[fresh_42]

It is indeed a very good question. At this point I am only looking at tools and data. I like simulation because a good simulator should take into account all known effects.

The current system of great ocean currents alone is a significant part of such models, and there are hints that e.g. the Gulf Stream is already weakened. If it collapses, then all your data will be worthless. Similar is true for melting of the permafrost regions or a major outbreak of a hotspot. These are extreme examples that may or may not occur.

The more direct components involve dozens of feedback loops, chaos, and last but not least, our own behavior. $CO_2$ emissions are only one relevant parameter. It is the one we possibly could measure best and influence most. We have achieved a reasonable weather forecast of about a week, maybe two. But climate is the collection of all weather systems on earth, and over a period of years, not weeks. So if the weather is already unpredictable for next month, how could we rely on climate predictions?

There are many models, and most of them point in the same direction, which increases confidence. But none of them has been calculated in a browser environment. Java, HTML, and PHP are insufficient.

If you really want sound data, then look at who has to deal with it professionally! Study the balances (over the last decades) of our biggest reinsurance companies! Politicians may lie to us, climate models may predict slightly varying scenarios, but money is unaffected by all those. Maybe not really a scientific method, but at least a reliable one.

 

[optotinker]

The current system of great ocean currents alone is a significant part of such models, and there are hints that e.g. the Gulf Stream is already weakened. If it collapses, then all your data will be worthless. Similar is true for melting of the permafrost regions or a major outbreak of a hotspot. These are extreme examples that may or may not occur.

The more direct components involve dozens of feedback loops, chaos, and last but not least, our own behavior. $CO_2$ emissions are only one relevant parameter. It is the one we possibly could measure best and influence most. We have achieved a reasonable weather forecast of about a week, maybe two. But climate is the collection of all weather systems on earth, and over a period of years, not weeks. So if the weather is already unpredictable for next month, how could we rely on climate predictions?

There are many models, and most of them point in the same direction, which increases confidence. But none of them has been calculated in a browser environment. Java, HTML, and PHP are insufficient.

If you really want sound data, then look at who has to deal with it professionally! Study the balances (over the last decades) of our biggest reinsurance companies! Politicians may lie to us, climate models may predict slightly varying scenarios, but money is unaffected by all those. Maybe not really a scientific method, but at least a reliable one.

Firstly, I definitely don't want to touch on politics or insurance at all.

To the first order, the Earth is just a ball under the Sun. It takes in heat from the sun and radiates heat out into the 4K space. Do we really need to model the weather at each point on Earth in order to get a any idea about global temperature?

 

[fresh_42]

Firstly, I definitely don't want to touch on politics or insurance at all.

To the first order, the Earth is just a ball under the Sun. It takes in heat from the sun and radiates heat out into the 4K space. Do we really need to model the weather at each point on Earth in order to get a any idea about global temperature?

No, probably not. But you need to know the concentration and distribution of several atmospheric gases like e.g. methane and carbon dioxide in order to calculate what is reflected back into space and what is not, plus the amount of energy that is stored in oceans, storms, waves, or simply the atmosphere itself, plus the exchange between all of them, plus the conversion details of the gases from short to long wavelengths of sunlight, plus the albedo effect of human installations and global ice shields.

And this list is only my layman opinion. I'm sure that climate researchers have a longer list. It is not a simple in and out. And we are not only interested in the energy balance, but we also want to know the local consequences.

 

[anorlunda]

I like simulation because a good simulator should take into account all known effects. It has the ability to synthesize these effects in a way far beyond the ability of a human brain, and its references should contain the key literature for said effects.

I don't want to be insulting, but that's just foolish. You should have more faith in the "Garbage in. Garbage out." philosophy.

 

[Vanadium 50]

global temperature

Please define this.

 

[hutchphd]

It is not useful to discuss 60 year old physics unless Please let us not unless you know something new. I have done the simple calculations you should too. Then listen to Carl Sagan who knew the answers 40 years ago.

 

[optotinker]

I don't want to be insulting, but that's just foolish. You should have more faith in the "Garbage in. Garbage out." philosophy.

Well, it is a tool, with its pitfalls and advantages.

 

[optotinker]

Well, it is a tool, with its pitfalls and advantages.

Please define this.

It is what we call a "lumped parameter".

 

[berkeman]

What do you have at home, a Cray?

Exactly. That was my first thought as well.

Do we really need to model the weather at each point on Earth in order to get a any idea about global temperature?

You need to model each Butterfly, right?

 

[optotinker]

It is not useful to discuss 60 year old physics unless Please let us not unless you know something new. I have done the simple calculations you should too. Then listen to Carl Sagan who knew the answers 40 years ago.

I am looking for a simulator, not a Saint :)

 

[hutchphd]

Seriously these calculations are essentially at three levels of complexity. The simple radiative balance equations take a competent physicist a half hour to show the Earth should about around 0C(273K) . The next level, including rudimentary atmospheric effects, takes a day or two to work through and adds a few tens of degrees. The final level takes a lifetime of effort and produces the present best estimates which are proving more accurate than one might wish.
Incidentally Prof Sagan's doctoral thesis explained why Venus is 300C hotter than it should be because of radiant balance alone. He was smart, not Holy.

 

[optotinker]

Seriously these calculations are essentially at three levels of complexity. The simple radiative balance equations take a competent physicist a half hour to show the Earth should about around 0C(273K) . The next level, including rudimentary atmospheric effects, takes a day or two to work through and adds a few tens of degrees. The final level takes a lifetime of effort and produces the present best estimates which are proving more accurate than one might wish.
Incidentally Prof Sagan's doctoral thesis explained why Venus is 300C hotter than it should be because of radiant balance alone. He was smart, not Holy.

Thanks. Good to know.
It would be good to find something between "a day or two" to "lifetime".

 

[hutchphd]

We can all agree on that, including those in the midst of calculations.
Did you know a primary impetus for the first supercomputers was for weather forecasting? The atmospheric models were developed and they could do pretty good predictions but the computers took two days for tomorrow's forecast. Even senators could figure out that supercomputers would pay for themselves. ( The space program has similarly paid for itself many times over simply from weather forecasts)

 

[optotinker]

We can all agree on that, including those in the midst of calculations.
Did you know a primary impetus for the first supercomputers was for weather forecasting? The atmospheric models were developed and they could do pretty good predictions but the computers took two days for tomorrow's forecast. Even senators could figure out that supercomputers would pay for themselves. ( The space program has similarly paid for itself many times over simply from weather forecasts)

Interesting.
I definitely don't want to get into the weather stuff. One will never get out of that rabbit hole.

 

[Kyrre]

>I want to get into the field as a player, not a spectator. The information offered by major institutions to the general public is helpful but it is already someone else's work, not mine.

Sorry, but you are not "a player" in this field as a one-man show. Get a degree and become part of a research group. Or, if you insist on being the lone ranger in your spare time, you have to build on existing literature and narrow your field of research a couple of orders of magnitude. Do a study of dendrochronology, or write a review, don't think you can simulate the whole freaking system with someone else's code and ten thousand dollars of computers alone.

 

[optotinker]

Thank you for your suggestions.

But my understanding is that anyone can be a player in science, any science.

A player, in my definition, is someone who brings new insight into the field by acquiring and analyzing data and playing with models, first hand.

I have worked with many first rate scientists both in academia and in industry. One common feature of them, or rather us, is never to tell anyone "you can't do this". We help and encourage. It is up to him to decide if he can or cannot.

Is this field somewhat different? I am just looking for a tool. I am a bit surprised by the odd responses so far.

 

[berkeman]

But my understanding is that anyone can be a player in science, any science.

Not really. An amateur with little education and little experience and little understanding of the current peer-reviewed scientific literatature has little chance of being a "player" in changing the world. I'm not saying that is you, since I know very litte about your education and professional background. We have Insights articles about many aspects of what you are saying in the quote.

Is this field somewhat different? I am just looking for a tool. I am a bit surprised by the odd responses so far.

Which responses in particular? The response theme that you can't simulate the weather without computing resources comparable to multiple Cray computers?

 

[hutchphd]

Is this field somewhat different? I am just looking for a tool. I am a bit surprised by the odd responses so far.

If you wish to play with this stuff then more power to you. There is no better way to learn. But it is not out of line for folks with presumably broader knowledge to offer friendly advice about the difficulties you may face.
You will not paint like Rembrandt because you find some new brushes and a pantograph.

 

[optotinker]

If you wish to play with this stuff then more power to you. There is no better way to learn. But it is not out of line for folks with presumably broader knowledge to offer friendly advice about the difficulties you may face.
You will not paint like Rembrandt because you find some new brushes and a pantograph.

This field is critical for the survival of mankind, quite unlike stuffs like the string theory. It is only natural, therefore, for anyone to take an interest in this. There are many people like me, I believe.

Almost everyone these days "gets a degree, and work for some institutions". But real discoveries are often made by people with a fresh point of view and an original way of thinking, not by the 9-5 employees with degrees. I would especially like to hear more from the first group here. The employees, on the other hand, already have Nature to publish on.

But of course everyone should be welcomed to express his views, including myself.

 

[Vanadium 50]

But real discoveries are often made by people with a fresh point of view and an original way of thinking, not by the 9-5 employees with degrees.

Really? Name five. In the last fifty years.

 

[fresh_42]

This field is critical for the survival of mankind, quite unlike stuffs like the string theory. It is only natural, therefore, for anyone to take an interest in this. There are many people like me, I believe.

Yes, and that is my greatest fear of all. That someone convinced he is smart invents something to change atmospheric mechanisms and make things worse. There are already horrible suggestions on the table.

We are talking about a system whose description requires hundreds of variables, dozens of differential equations, all with uncertain initial values, many depending on each other, uncounted assumptions, and that takes really huge computing power to solve. And now add a self-proclaimed genius who changes one crucial equation, or worse, adds another crucial equation. How are the odds that it improves such a system?

 

[hutchphd]

This field is critical for the survival of mankind, quite unlike stuffs like the string theory. It is only natural, therefore, for anyone to take an interest in this. There are many people like me, I believe.

The survival of mankind is not going to be affected by one guy in his basement with a computer looking for who knows what. The science is sufficient as it stands to yield a direction.
If you wish to save the planet, the hard work is to motivate the naysayers and the troglodytes. Elect people with the knowledge and courage to act. This is a place where your energies and skills might well yield tangible fruit. Let's get busy.

 

[Astronuc]

To the first order, the Earth is just a ball under the Sun. It takes in heat from the sun and radiates heat out into the 4K space. Do we really need to model the weather at each point on Earth in order to get a any idea about global temperature?

A ball that is rotating (period = 1 day) and revolving around the sun (period = 1 year), with the surface constantly changing. Day (heating) and night (cooling) daily + polar summer and polar winter, clouds, air currents, oceans and their current, . . . .

It is what we call a "lumped parameter".

Insufficient.

I definitely don't want to get into the weather stuff.

So, one wants to ignore the physics?

But my understanding is that anyone can be a player in science, any science.

More like a dabbler. Or a spectator who runs on the field and claims they can score against professionals.

Researchers as accidental witnesses to the flood disaster

14 July is a rainy day. GFZ researcher Michael Dietze is on his way back from a field visit in the southern Eifel with colleagues from Potsdam and the University of Bonn. Heading back north, they quickly realize that what is happening just outside is more than a long heavy rain: The online data view of the Altenahr gauge rises rapidly every 15 minutes, faster than the actual forecast, and faster than the researchers would have liked to, because their own measurements are affected. Flooding of the Ahr is nothing unusual: as part of a research project, Dietze and colleagues had set up several seismic stations on a three-metre-high terrace in the Ahr valley a few weeks earlier – assumed to be at safe distance to hostile flood conditions. They wanted to use them to measure ground motion caused by sediment movement and water turbulence during "regular" floods. Now the level is already one metre above the terrace, the stations are lost.

https://www.gfz-potsdam.de/en/media...ust-find-answers-to-after-the-flood-disaster/

Researchers at the Helmholtz Centre Potsdam - GFZ German Research Centre for Geosciences point at a number of effects that have occurred only rarely in Central Europe so far and have therefore not been taken sufficiently into account.

A lumped parameter approach is insufficient.

I am looking for something more quantitative, i.e. a simulator that takes into account all factors that influences the temperature.

Someone else's work/tool? There is no single 'global temperature'. There are trends (daily, seasonal, annual, decadal?, . . . ), there are variations, and are maxima and minima, there is precipitation (frozen and liquid), there is atmospheric composition and ocean salinity, gas transport, vegetation, . . . . Stored heat affects temperature.

 

[anorlunda]

I like simulation because a good simulator should take into account all known effects. It has the ability to synthesize these effects in a way far beyond the ability of a human brain, and its references should contain the key literature for said effects.

I think that is the source of your misunderstanding. Your description quoted above is a science fiction version of simulation, where an AI can self assemble a model based on all the published data and all the scientific papers from history. We won't have that ability unless and until we reach the Technological Singularity in 2047. But today, we are nowhere close to that.

Where are we today? Today's simulations reflect the skills, the knowledge, and experience of the authors. AI can't do that for us. It takes tremendous effort by skilled people to make a good simulation.

The best simulations are verified by comparing model predictions against real world data. Every time predictions match reality, our confidence goes up a notch. But simulations of the distant future (or alternative histories) can't be verified that way by definition because we can't have the real data.

Simulations that can not be verified should not be trusted to make a basis for important decisions.

 

[optotinker]

A ball that is rotating (period = 1 day) and revolving around the sun (period = 1 year), with the surface constantly changing. Day (heating) and night (cooling) daily + polar summer and polar winter, clouds, air currents, oceans and their current, . . . .

This is the type of discussion I like to see. Thank you very much.

In research, one ignores all physics other than those necessary for solving the problem on hand. For example, a pendulum's period doesn't depend on the electron movement in its various parts. Even air drag can usually be ignored.

But, in a complex system, what is "necessary" and what is not? Here is where simulation comes in. It allows one to play with the parameters and get a quantitative feeling for various factors. Of course everything, including the holy back of the envelope calculation, needs to be verified. Simulation is no exception.

The talk by Carl Sagan above, he mentioned modeling by N-number of research institutes around world. Given the vintage of his talk, I suspect the computing capabilities of "supercomputers" used then could be had for very little money today.

Anyway, I am asking for a tool. Does such a thing exist in the public domain?

 

[Arjan82]

The more direct components involve dozens of feedback loops, chaos, and last but not least, our own behavior. $CO_2$ emissions are only one relevant parameter. It is the one we possibly could measure best and influence most. We have achieved a reasonable weather forecast of about a week, maybe two. But climate is the collection of all weather systems on earth, and over a period of years, not weeks. So if the weather is already unpredictable for next month, how could we rely on climate predictions?

Weather is an initial value problem, climate is a boundary value problem (so essentially a very complex energy balance)

This means that you cannot compare weather predictions with climate predictions. It is a common misunderstanding. Weather is influenced by chaos, climate much less so. This however in no way means that climate predictions are less complex, just differently complex...

 

[Arjan82]

Almost everyone these days "gets a degree, and work for some institutions". But real discoveries are often made by people with a fresh point of view and an original way of thinking, not by the 9-5 employees with degrees. I would especially like to hear more from the first group here. The employees, on the other hand, already have Nature to publish on.

This is a very annoying and persistent myth... Most if not all significant and meaningful discoveries nowadays are made of large groups of scientists working in collaboration, this is especially true for climate science since it involves so many systems that are interacting in complex ways. Each of the systems can already require several PhD studies to understand.

Also, to shave all scientist with the 9-5 mentality brush, like all of 'm are happy with their position and can happily mud around without much progress or effort (which is what you say sounds like to me) is a gross misunderstanding of the field. It is very competitive since there are many people willing to work there, but much less positions and funds...