What Do We Think About Climate Change

In the investment game that is called "curve fitting" and it normally results in a model that is prone to major failure at some point. Absent a good understanding of the fundamentals it becomes dangerous. I don't know how relevant that is here but it seems to me that the models used swing from catastrophe to A OK on a small change in inputs. Soooo...?

 

Hmmmm - not quite the same thing in the case I was talking about in my preamble Zed.

We had a quite good understanding of the fundamental principles at play and models developed that generally were well proven to provide good response to ranges of different, relatively simple test situations, where the expected response could be developed from first principles or by observation.

It was the particular application of these models to specific natural locations where the calibration and validation process became necessary because some parameters in the model had ranges of possible values depending on the nature of the natural system.

Mostly I am referring to hydrological and hydraulic modelling of catchments and rivers behaviour in response to historical or predicted model rainstorm, or time series rainfall events.

 

in my previous I posted a link
Ill post it again so its easy for all to find it
its a in depth answer to your question by climate scientists

http://www.realclimate.org/index.php/archives/2008/11/faq-on-climate-models/langswitch_lang/in#more-527

 

and if you have time check out the history of climate science which Naomi Oreskies presents in her vid the amarican denial of global climate change

best
B

Zed
Ill give you curve fitting is an all round bad way to make more than one or two short term trades but Im not sure it would have much merit in climate modeling once a deviation from the norm is realized
and it has
cause there wouldn't be much to compare
so Im not sure it could be applied

from personal experience
if a graph falls within trend then a further analysis of cyclical variability ( curve fitting ) can be applied to try and predict variations within trends or Fibonacci fan ( which I seldom use )
although I do look at the oscillation in the moving average which is a form of cyclical variability that does work well for me

myself I avoid curve fitting in my trading as it seems to just up and bite you in the *** every other trade
not much profit in that

as for it applying to climate models
Ive never built a climate model so I have no clue
Ild urge everyone to check the link I provided cause its the guys who do build em talking about how they do it and how accurate it ends up being
basicaly first hand stuff and bound to be the best most accurate answer to Mr Muds original question

Jim
I almost feel bad for busting you so hard on that last post of yours but hey
just playin with the pudding
and you did blatantly lie about that graph
anytime you want to have an "honest" discussion Im all for it
just leave the blatant falsehoods out of it and be willing to go over the basics
whats so wrong with that idea eh

Im starting to think you throw bones my way just to see if Ill catch you in your bs
that certainly could have been deliberate it was so bad
anyone knows that the NOAA charts all come with there logo on em
and there was no way I was going to miss that
or that the data was not even from NOAA
as you claimed it to be

 

You might well be quite right there, however it seems to me that we are working with some incomplete and cobbled together data sets then trying to call a knife edge result!

Thing is if the models are that good why the observed variations and the controversy surrounding them. Personally I don't think we are as good we think we are! I was watching Catalyst the other night and they where talking about climate phenomena that all but dictates a large part of our drought cycle... it appears that the "discovery" of this phenomena has been in the last five years! I was amazed that it all seemed so recent!... Anyway, I tend to think that we often over-rate our abilities like the LTCM guys ! LOL.

 

Zed,

My point in trying to understand a bit more about the Climate Change models exactly. Climate is alot more complex than the hydrology and hydraulics I was dealing with and we had all sorts of trouble to get good calibrations and validations to observed data.

How much more difficult therefore must it be for the climate models....

And what does that tell us about their predictions?

Cheers,
Andrew

 

Yes...! OK, I'm with you... we just seem to be surrounded by people who are all so certain about this stuff, that alone worries me! I know we can't model economic factors correctly, plenty think they can but "flawed" is the best description of most models! I'm assuming climate is indeed more complex than economics...!

 

not sure this hole thing will come up but Ill see if it will



nope will have to paste it

prediction looks pretty darn good to me
now thats not to say other predictions are as good as this one
but still
that says a lot for the value of there predictions

oh and Zed
its pretty widely accepted that the market is fear and gread driven
two huge and unpredictable variables

look at just the last few days
GM files for bankruptcy and the market soars

seems to be as someone who is intimately involved in both that the market should be
and does seem to be
far harder to predict

 

Zed,

This from a response to a question in the link from Boston in post 3026 mentions this very issue.

"We can’t do a very good job at estimating the future trajectory of technology for instance, or economic development, and so regardless of how well we understand climate, our ability to predict exactly what will happen is limited. Secondly, we don’t have full information about the current conditions, and so, like for weather forecasts, if there are aspects of climate change that are chaotic, we can’t predict those over the long term. However, it is worth pointing out that the statement does not imply that we can’t know anything about the climate system in the future. We know that if there is a big volcano, the climate will cool - and many aspects of the resulting changes will have been predictable. The same is true for increasing GHGs - the climate will warm. Models can’t tell you exactly what will happen where, but there is a lot they can say."

My concern here is this stuff doesn't give me a great deal of confidence in the models output.

What I would love to see is some graph plots of these model performances over historical records as in

"Then there are the tests of climate changes themselves: how does a model respond to the addition of aerosols in the stratosphere such as was seen in the Mt Pinatubo 'natural experiment'? How does it respond over the whole of the 20th Century, or at the Maunder Minimum, or the mid-Holocene or the Last Glacial Maximum? In each case, there is usually sufficient data available to evaluate how well the model is doing." also from Boston's link

 

well that was deja vu
just posted the graph that fits your exact description

 

Nice graph, which sums it up again. Do understand the consortium however, and how it?s correct. A little bit of this and a little bit of that; geology would help. Everything just transformed, as with the fossil fuels, don't go away. I?m still stumped why some beg to differ.

 

Boston,

Only over 100 years? Is it enough for a climate model?

And the chart is obviously a simplification of much more data and model output coz it looks like plotting points are only about every 10 years.

What happens in between?
And - does it matter?

Cheers,
Andrew

 

Well, Mr. Mudd really has a point here! Most humans, us I think, can not really comprehend more then a generation or so. I have read some 40,000 years ago (or so) the climate did some extrodinary things. Is there a graph? Has Herodotus or Thucydiese recorded the temperature way back then? I wana know!

 

Cpt Mud,

I would like to see models that work over many centuries if not millenia. I think we lack enough data to do that but I don't really know!

Thats kinda the point that I was making first up. I can take a stock and model its price over a short term and come up with a trading model, it will work within reason until conditions change. To have a greater degree of success I can build in the companies fundamentals and factor them into the model. I can then claim I understand fundamentals and the model is good, but in reality the fundamentals and value of every other company in the sector need to be layered over the top of that, then the sectors fundamentals need to be layered over that, then the environment for stocks and the broader economies fundamentals need to be layered over that, then the fundamentals of the people that hold the stock need to be considered (as graphically demonstrated by the hedges funds that threw baby bath water and all, out last year)... anyway... and so it gos. When I have backed up to mars I will have a complete model that will work anytime anywhere because it considers all the options... the only problem is that with each layer the options grow exponentially and very quickly get beyond my (anyones) ability to calculate.... so it is (at least todate!) impossible to achieve. Its easy to appear to have something that works over a short period and even convince yourself and others you have done the fundamental work to support it but... stuff happens, lots and lots of stuff that means attempts of this nature are doomed to fail.

So how can we be ever be confident we have "backed up to mars" and included it all... and if we have should'nt the model work across tens of thousands of years? I just don't see that we could have even come close to doing the work needed... yet.

But then I'm stooooopid , just ask around!

 

Yes, my thoughts... I was busy typing & cooking missed yur post!