itfitzme
VIP Member
- Jan 29, 2012
- 5,186
- 393
You need to do a LONGER time series and REMOVE the periodic components from the TSI..
Just like the yearly and multi-yearly variations are removed from you CO2 graph..
Also -- What did you use to normalize the functions? What was in the denominator?
And why do YOU think that the answer of what is the forcing function needs to correlate exactly with the output temperature? Do you believe that all systems operate like that? Or just extremely simple ones.
TSI has approximated a step function.. Rose for 200 years and then flattened.. Good enough for a step function.. The existence of just ONE INTEGRAL in the transfer function of the behaviour of climate system will turn that step into a nice ramp or hockey stick.. True fact grasshopper...
Want an integral that's IN THE climate black box?? Heat storage in the oceans. That's an integrator function right there.
Create a TOA to Surface imbalance and the temperatures continue to climb. Even with a constant power balance diagram..
At least for awhile..
I don't have to do any of that. The variation that matches, matches and the variation that doesn't doesn't. TSI and CO2 are both known drivers of temp. They are all part of the same system. The sun warms the earth and atmosphere. CO2 absorbs IR. We already know, based on fundamental principles, that the will correlate. The only question is by how much. I can guarantee that when the sun is brighter, there is an immediate response. This is obvious to anyone that has been outside on a summers day and walked from shade to direct sunlight.
It doesn't take a year for the warming to accumulate. Obviously it doesn't, we have summers and winters. And those are yearly CO2, yearly TSI, and yearly temp averages. They don't even need to be yearly. They could be different month numbers for each data set. I am not trying to build a model, not trying to make precise predictions. The absolute values are irrellevant. The only interest is in the variability, the relative changes, indeed, the ratio of those relative changes. More formally, that regression is part of a large tool called "analysis of variance". All we are after here is how thing are varying. Any high frequency components that don't match just don't match. Like I said, Fourier transforms are correlation. Regression is correlation. It is all included.
Any further refinement isn't going to result in an appreciable change. CO2 os coming in at 40x greater as a driver. 2% for Sun, 80% for CO2. 18% is other stuff. There ain't no need to "think that the answer of what is the forcing function needs to correlate exactly with the output temperature". I don't need to think that hitting my thumb with a hammer will hurt in order for the pain and hammer to correlate. I don't need to think that temp is responsive to CO2 and TSI for it to correlate. It either does or does not. Lo and behold, CO2 correlates. Gosh, grasshopper, so does TSI, just not as much.
This is the most fundamental aspects of scientific reasoning. a) if a physical law is true in one location, it is true in all locations b) if something is causal, it is correlated. These two axioms lead to a larger reasoning that if something is causal in one place and coorelated in another, then it is causal in that other place. It is the basis for sanity.
You can build as complicated a model as you like, refine it to monthly observations, time series lag some portion. Maybe volcanic ash will pick up some, other GHGs will pick up some. But that 40:1 is so huge that it is foolishness to believe that TSI is somehow going to miraculously swich with CO2 as being more highly correlated.
The reality is that CO2 isn't not correlated. The null hypothesis is the R^2 for CO2 is much less that R^2 for TSI. The null hypothesis is demonstratably false. Thats the thing. You can go about coming up with some more complex model. It won't change the fact that R^2-CO2 >> R^2-TSI. Some more complicated model that doesn't correlate simply proves the more complex model is wrong.
That is the biggest problem with the denialist argument. The argument being, "Because this complicated model doesn't prove CO2, then the simple model that proves CO2 is false". That is just plain stupid.
Starting from the false premise that mitigation techniques will cost you money, then trying to prove CO2 doesn't cause climate change is faulty reasoning. First off, economics isn't the correct science to apply in atmospheric physics.
But, hey, I found an online regression calculator for you. You are welcome to try. This, though, is the fundamental issue I have with con-deniers... laziness. They are all about watching other do the work while they stand around and complain about the work being done. It is "defining reality by fault finding". They all want to be on charge and tell people what to do instead of actually doing some work themselves.
You ain't presented any "true facts", grasshopper, because facts are specific, countable, measurable observations of the real world.