>But at the same time one can only make predictions with the data that is available. It would stand to reason that some things can be predicted by simple linear regression. There is quite a bit of data that exists - and math can really help out here. This is a decent read on explaining glacier retreat mathematically.
>http://vc.bridgew.edu/cgi/viewcontent.cgi?article=1049&context=honors_proj
>and these is kinda interesting too
>http://www.esciencecentral.org/journals/modeling-of-data-from-climate-science-using-introductory-statistics-2332-2594-1000129.php?aid=47928
> https://ww2.amstat.org/publications/jse/v21n1/witt.pdf
>
>so in part I do think it's fair to say that these computer models are based on an unknown number of unknowns - and frankly it's the best models we have. Perhaps a better way of thinking here would be to decrease our pollution FIRST - then come back in 20 years and see how well we did.

Agreed about pollution.
When I was a kid, driving into Los Angeles usually meant burning eyes from the pollution.
It is easy to breathe there now even with so many more cars.

Linear regression is too simplistic to be useful, given that we are 10,000 years into an inter-glacial period, and those generally last about 10,000 years.
During the Medieval Warm Period, Viking settlers were farming Greenland that is now covered in Ice...

I have not heard anything about how to predict the timing of the next glaciation episode.
If we knew that, we might be able to fight it by increasing the atmospheric CO2.

The sad fact is, the Earth's climate is just not stable, no matter how much we wish it.

A much bigger problem is that we humans are killing the oceans with over-fishing and pollution.
The arguments over the correct level the vital plant nutrient CO2 are distracting us from an obvious environmental disaster.