climate change times are a changin'

[dxmnkd316]

Re: climate change times are a changin'

Going to the all star game next week? Dress warmly - the "polar vortex" is back!!!
http://www.washingtonpost.com/blogs...mmer-return-in-eastern-u-s-next-week/?hpid=z3

Take it to the Minnesotans Arguing About the Weather thread mmmmmmmmmmmmmmkay?

(On a serious note, can we all agree to never say the words polar vortex ever again? I hate these millenialesque buzz words.)

 

[dxmnkd316]

Re: climate change times are a changin'

This.

I'll take 10 over 110 any day.

Agreed. Probably why I think Minnesota > (Insert state here)

 

[FreshFish]

Re: climate change times are a changin'

I must admit, that I've had a change of view since we started our dialog recently. I realized that I am resistant to the prescription of what we "should" do in response to AGW, and that resistance has influenced how I view the data.

That being said, I'd like to suggest a few "moderate" ideas, from least controversial to most.

1. "Climate change" has been occurring for millions of millenia. The geologic record makes this assertion incontrovertible IMHO.
(it is also quite plausible, in time periods shorter than geologic epochs, that variations in solar output related to what we've learned about how the sun's energy output is cyclical, could affect our climate).

2. We have clearcut evidence that human activity affects the "climate" in the record of CFCs and the ozone hole.

3. It is hard to deny that human activity also affects the climate in other ways. There is a very high likelihood that our CO[SUB]2[/SUB] output is affecting the climate.

4. that being said, it is also quite hard to quantify, with any degree of reliability, how much this activity is affecting climate.

5. Either the press is misreporting the studies, or the studies themselves are "slanted," to anyone with a basic familiarity with statistical modeling: given a complex model with multiple inputs that have interactive effects, it is highly implausible that the results will have the precision generally attributed to them (I'll expand on this theme in a subsequent post). ANY model this complex must have a range of likely outcomes. Some of the more "responsible" reporting i've read does indeed provide a range; however, there is far too much precision in what I've read to be plausible.

6. There are still far too many unknowns to have much confidence in any predictions, especially as there is a feedback loop not included in most models: we can indeed change our behavior, and so obviate the projections of the best models by deliberately and consciously altering the input variables (see # 2 above: global cooperation on controls of CFC changed how CFC was affecting the ozone hole).

7. Where we really "part ways" is NOT "whether" human activity affects climate, it is on how we might best respond.

8. The idea of limiting CO[SUB]2[/SUB] output is controversial, impractical, possibly ineffective, and certainly would involve significant hardship and suffering among our most poor and vulnerable. CO[SUB]2[/SUB] output is so deeply embedded and engrained in our daily lives, that the ripple effects of trying to suppress CO[SUB]2[/SUB] output would be economically devastating to those who can least afford it. Food prices would be triply affected: in the planting and growing, the shipping and storing, the preparing....daily travel would become far more expensive, and on and on and on....

9. The idea of limiting CO[SUB]2[/SUB] output is also short-sighted and unimaginative: if the problem is too much CO[SUB]2[/SUB] in the atmosphere, why not just remove it instead?

10. the "technology" of removing CO[SUB]2[/SUB] from the atmosphere is already quite advanced, sophisticated, time-tested, and inexpensive. We can easily do it, and afford it.



If you really, truly believe AGW is a serious problem, and you want to do something immediately effective and proven to work, plant trees.

 

[cF[Authentic]]

OK, I gots ta know. What is cF[Authentic] from? It looks like an OO destructor.

It was a counter strike thing waaaay back in the day. cF was the clan name and Authentic was my nickname.

 

[huskyfan]

Re: climate change times are a changin'

one of my intents in starting this thread was to find out if anyone knew of responsible scientifically creditable sources on global warming. and it perhaps someone with that technology expertise could explain it simply to me. if you name call, deride, make fun of or other such statements, I immediately dismiss your ideas. Likewise I distrust our press, or rather the shabby remains of what was once a great free press, to put out any info without bias. every internet site, blog, etc is so slanted one way or the other. I likewise distrust much of the computer modeling which is often taken as fact.

from my own studies I realize the sun plays a great role in our climate and the geologic history backs this up. I do believe we dump huge amounts of carbon into the air and I hate the thought of just the plain old pollution aspect. but there have been other prehistoric carbon dumps that couldn't be explained by man. more carbon in the atmosphere helps plants and trees grow faster?? does that offset the carbon we produce? growing more food and more responsible food (like not dumping a lot of resources into meat production) could be a great help for the millions of hungry people in the world. and like FF I don't see any quick and easy fixes of our reliance carbon production. a quick jump to electric cars would cause an increase in coal plants to make electricity. the whole political scientific economic humane side of the debate seems to have been missed by most on this thread. thanks for the efforts of several of you who have tried to make this thread a discussion and not a flame throw.

hockeyfights.com looks pretty good right now

time to jump off my high horse and into my eco drive car that automatically shuts of when stopped and instantly restarts when you release the brake pedal, grab the re-useable grocery bags, dress in my organic cotton clothes and go buy some food from Peru or China. and some of that nice French wine. (we are not still boycotting that are we? that's simply uncivilized)

 

[FreshFish]

Re: climate change times are a changin'

[SUP][/SUP]

5. Either the press is misreporting the studies, or the studies themselves are "slanted," to anyone with a basic familiarity with statistical modeling: given a complex model with multiple inputs that have interactive effects, it is highly implausible that the results will have the precision generally attributed to them (I'll expand on this theme in a subsequent post). ANY model this complex must have a range of likely outcomes. Some of the more "responsible" reporting i've read does indeed provide a range; however, there is far too much precision in what I've read to be plausible.

For example, in a simple, one variable "monte carlo" simulation, if your average annual rate of change is 1.5%, and your standard deviation is 3%, after 50 years, the difference between the 25th percentile result and the 75th percentile result is pretty wide. if we start with 1,000 as our index number, after 50 years the 25th percentile result is around 1,800, while the 75th percentile result is around 2,365 or 32% higher. So the middle 50% of simulated outcomes covers a fairly broad range. It's not inconceivable that parts of that range might have relatively benign results while parts of that range might have catastrophic results.

It gets more complicated if you add additional variables, since you generally also include correlation coefficients as well[SUP]1[/SUP].
- with 2 variables there is one correlation coefficient,
- with 3 variables there are three correlation coefficients,
- with 4 variables there are six correlation coefficients,
- with 5 variables there are ten correlation coefficients, etc.

Where do these correlation coefficients come from? typically by regression analyses, but what if the original data set itself is not completely precise? What if there is a change in the underlying drivers, so that an extrapolation from past results is no longer predictive of future results?

Anyway, any decent model will produce a range of outcomes with varying likelihoods assigned to each range. The more variables you introduce, the more ancillary assumptions you have to include.....and many times a slight change in the underlying assumptions might have a significant effect on the longer term outcome.[SUP]2[/SUP]

Long (-winded) story short: any time I read a prediction from a climate scientist that gives a single outcome, I get suspicious right away, because I know that their models do not work that way, they only provide a range of possible outcomes with an associated likelhood of outcomes. Perhaps what is being reported is merely a median result and the essential information is buried in a footnote?





[SUP]1[/SUP] a correlation coefficient describes how much a change in one variable will affect a change in another variable. 100% correlation means the two move together in lock step: e.g., if you increase temperature of a gas, you increase pressure. -100% correlation means the two move inversely: e.g., you decrease the volume of a gas, you increase the pressure. 0% correlation means the two move entirely independently of each other.

[SUP]2[/SUP] in my earlier one variable example, I used an average annual rate of change of 1.5% and a standard deviation of 3.0%. Suppose those were 1.75% and 3.25% instead? then after 50 years, the 25th percentile result is around 2,000 and the 75th percentile result is around 2,720, and the spread has widened to 38% from 32%.

 

[Kepler]

Re: climate change times are a changin'

It was a counter strike thing waaaay back in the day. cF was the clan name and Authentic was my nickname.

That's awesome. Thanks.

 

[5mn_Major]

Re: climate change times are a changin'

I must admit, that I've had a change of view since we started our dialog recently. I realized that I am resistant to the prescription of what we "should" do in response to AGW, and that resistance has influenced how I view the data.

...

10. the "technology" of removing CO[SUB]2[/SUB] from the atmosphere is already quite advanced, sophisticated, time-tested, and inexpensive. We can easily do it, and afford it.



If you really, truly believe AGW is a serious problem, and you want to do something immediately effective and proven to work, plant trees.

Good post. I am usually not an or guy but rather an and guy. Why not take it out of the environment and encourage change to alternatives?

 

[FreshFish]

Re: climate change times are a changin'

Good post. I am usually not an or guy but rather an and guy. Why not take it out of the environment and encourage change to alternatives? [emphasis added]

I'm fine with encouraging changes to alternatives, as long as the alternatives are self-sustaining without subsidies, and particularly as long as the alternatives don't create substantial additional problems in their wake (e.g., people rave about electric cars while ignoring the environmental consequences of battery production). What annoys me is that people want to mandate specific changes without regard to alternatives or consequences.

 

[dxmnkd316]

Re: climate change times are a changin'

more carbon in the atmosphere helps plants and trees grow faster??

It does, but we're talking an entire era, not just a few millenia. Look up the carboniferous era on Wikipedia.

 

[LynahFan]

Re: climate change times are a changin'

It gets more complicated if you add additional variables, since you generally also include correlation coefficients as well.
- with 2 variables there is one correlation coefficient,
- with 3 variables there are three correlation coefficients,
- with 4 variables there are six correlation coefficients,
- with 5 variables there are ten correlation coefficients, etc.

and that's if you only include 2nd order terms...

Long (-winded) story short: any time I read a prediction from a climate scientist that gives a single outcome, I get suspicious right away, because I know that their models do not work that way, they only provide a range of possible outcomes with an associated likelhood of outcomes. Perhaps what is being reported is merely a median result and the essential information is buried in a footnote?

I don't read as much original research as I should, but any time I do, I've been more than satisfied with their statistics. It's the secondary reporting where things seem to break down.

 

[Kepler]

Re: climate change times are a changin'

I don't read as much original research as I should, but any time I do, I've been more than satisfied with their statistics. It's the secondary reporting where things seem to break down.

This. In real peer-reviewed journals (sorry, ARJ), the math holds up, but the popular press is written for people who had trouble with basic algebra, so it's dumbed down so much that the actual content is lost.

 

[WisconsinWildcard]

Re: climate change times are a changin'

Long (-winded) story short: any time I read a prediction from a climate scientist that gives a single outcome, I get suspicious right away, because I know that their models do not work that way, they only provide a range of possible outcomes with an associated likelhood of outcomes. Perhaps what is being reported is merely a median result and the essential information is buried in a footnote?

Good post. And I would say correct. Published papers on this topic (from my experience) give ranges, define assumptions and thoroughly explain the statistics. A paper may be 8-10 pages long but if it is in a top tier journal, the supplemental information can range from 10-60 pages, depending on the discipline. Most people just read the abstract. If you are lucky, they read the original paper. Even then, most people do not take the time to evaluate the claims made in a manuscript in the context of all the information provided (supplemental included).

I used to be part of a stem cell journal club, back when a lot of major advances were being made in induced pluripotent stem cells. A single paper in Science or Nature would range about 7-10 pages but 40+ in supplemental information. I would have to spend 10-20 hours reading and trying to understand all of that just to present the paper to our journal club or my PI (Principle Investigator) would tear me a new one

I think it is important to note that climate models are consistently being updated to fit the data better (its a good thing) and to have more predictive power. You are right to observe that there are many variables that go into climate, many of which we are just beginning to understand their impact. A quick example off the top of my head is that some models predicted greater increases in surface temperatures over the last few decades than what was observed. (One of) The reason for this is "we" underestimated the the ability for the ocean to buffer surface temperatures. It is now a better understood variable and should provide less error as our knowledge base grows. Unfortunately, because the ocean is such a good buffer, they are warming, retaining more CO2 (which you chemist out there should be worried about), are then more acidic and sea life (especially coral and shellfish) are having drastically decreased ranges that they can survive. Among other things of course.

 

[The Sicatoka]

Re: climate change times are a changin'

I think it is important to note that climate models are consistently being updated to fit the data better (its a good thing) and to have more predictive power. You are right to observe that there are many variables that go into climate, many of which we are just beginning to understand their impact.

My concern is with the models, all the models.

To call this system we call Earth a "MIMO" system is quite possibly the understatement of a lifetime. The inputs are vast and interact with each other. The outputs are just as vast and interactive. And any model is just that: a model. It's a representation of a reality.

Creating a model that can reproduce already known responses is difficult. Trying to create a predictive model is akin to wizardry. Until the model is validated (by reality data) it is an educated guess.

And before you rip on me, be aware that I've done modeling and model-predictive control of everything from automotive transmissions to nuclear reactors.

Creating a model is easy; creating a true and accurate model for a massively MIMO system is brutal.

That's why when I see 'actual' not aligning to 'predicted' I start to question the models and everything based on them.

 

[LynahFan]

Re: climate change times are a changin'

My concern is with the models, all the models.

To call this system we call Earth a "MIMO" system is quite possibly the understatement of a lifetime. The inputs are vast and interact with each other. The outputs are just as vast and interactive. And any model is just that: a model. It's a representation of a reality.

Creating a model that can reproduce already known responses is difficult. Trying to create a predictive model is akin to wizardry. Until the model is validated (by reality data) it is an educated guess.

And before you rip on me, be aware that I've done modeling and model-predictive control of everything from automotive transmissions to nuclear reactors.

Creating a model is easy; creating a true and accurate model for a massively MIMO system is brutal.

That's why when I see 'actual' not aligning to 'predicted' I start to question the models and everything based on them.

I look at it the opposite way. Predictive models are easy, since you don't have the right answer to compare yourself against - nobody knows if you're right or not.

Creating a model that fits the data - by which I mean ALL the data - is a whole lot harder. You can spend months tweaking coefficients. And in the end, you have a model that is great at matching the finite set of results that you have, but is useless for predicting results for systems with the slightest bit of difference in conditions - it's overly optimized to just your existing data. As soon as one new data point comes in, well, that's no longer a prediction -it's a new data point that you now have to match and you're right back to square one.

 

[The Sicatoka]

Re: climate change times are a changin'

I look at it the opposite way. Predictive models are easy, since you don't have the right answer to compare yourself against - nobody knows if you're right or not.

Touche.
And that is why weather forecasters almost have the perfect job: they make wild, model based predictions today for tomorrow's weather. We don't know the results until tomorrow. Until then the weather forecasters are always right.

Creating a model that fits the data - by which I mean ALL the data - is a whole lot harder.

Indubitably.

And in the end, you have a model that is great at matching the finite set of results that you have, but is useless for predicting results for systems with the slightest bit of difference in conditions - it's overly optimized to just your existing data. As soon as one new data point comes in, well, that's no longer a prediction -it's a new data point that you now have to match and you're right back to square one.

Here I'd say a model should be designed to replicate historic data (to validate the model) as well as predict or extrapolate expected results. A model is a model of a system through time. If your "model" can only reproduce already known data it's not a model, it's a formula.

 

[WisconsinWildcard]

Re: climate change times are a changin'

My concern is with the models, all the models.

To call this system we call Earth a "MIMO" system is quite possibly the understatement of a lifetime. The inputs are vast and interact with each other. The outputs are just as vast and interactive. And any model is just that: a model. It's a representation of a reality.

Creating a model that can reproduce already known responses is difficult. Trying to create a predictive model is akin to wizardry. Until the model is validated (by reality data) it is an educated guess.

And before you rip on me, be aware that I've done modeling and model-predictive control of everything from automotive transmissions to nuclear reactors.

Creating a model is easy; creating a true and accurate model for a massively MIMO system is brutal.

That's why when I see 'actual' not aligning to 'predicted' I start to question the models and everything based on them.

It is safe to assume you know far more about modeling than I. I would first say (with no intent to rip, your response was sincere and so is mine) that a lot of your criticism takes the form of an argument from incredulity. My understanding is that the messier the system, the messier the model. I would guess we know a fair bit about the physics and chemistry of an automotive transmission or nuclear reactor. I am sure there are still assumptions, error bars, etc but in comparison with a global climate, the models probably come out a bit cleaner.

One of the goals of a good climate model (or most scientific endeavors) is to accurately predict the future. It does not matter what the actual Truth is, the model that survives is the one that is the best predictor. Climate scientists are constantly updating their models so they can do a better job in the future. That is the beauty of the thing, if it is wrong, they adapt. Science does not seek Truth, but instead, tries to find more and more accurate ways to describe the given knowledge base. We will never be "right," just less wrong. An idea lives and dies by its utility, its accuracy to predict and its relationship to current knowledge.

There are many inconsistencies, criticisms and shortcomings of Newtonian physics and General Relativity. However, they are good enough. In many cases, they are the best we have. They have taken us to the moon and help solve some of humans greatest challenges. But again they are not "Right" or "Truth," just less wrong than other ideas we have. Most of the models I have read about have gotten the larger trends correct. Most of the observed data I have seen fits within the original error bars given in the original manuscripts.

The Earth's surface temperature is trending upwards. The seas are rising. CO2 levels are higher than recent times and humans pump a **** load of it into the air. The permafrost is melting and releasing methane (as do cattle). We know (from earth and venus) that CO2 and methane are potent (varying degrees) greenhouse gases. We know that humans have not lived in CO2 levels like this, we did not recently evolve in levels like this. We can tell from the geologic record that large climate swings (be it from volcanoes, asteroid, etc) did not turn out well for the vast majority of things living on this rock. Like I said before, I am not qualified to judge their methods but I can say I have not come across a more predictive model. When I do, I will drop whatever I currently believe and subscribe to that.

Science is based, in part, on qualifiers or qualifications. Not those of the scientists, but that of the statements they make. We do not have perfect medications. Many cause unknown side effects. That worry does not stop me from providing medication when I know it is the best that we have. I have to play to statistics and work with the best we have at the time. I apply that same principle to climate science. I know the models are not perfect, but they are the least wrong we have at this time.

 

[The Sicatoka]

Re: climate change times are a changin'

"Remember that all models are wrong; the practical question is how wrong do they have to be to not be useful." -- George E. P. Box

How wrong do they have to be to not be useful? Did any of them predict the worldwide "flat decade"?

 

[walrus]

What technology allows us to remove carbon dioxide from the atmosphere and what do they do with it once removed?

 

[dxmnkd316]

Re: climate change times are a changin'

Trees.

Burn them.