Climate Policy

[kandrathe]

Sorry, I've started to write a more detailed description a few times, but each time it became clear that I would have to go back and cover more material.

Let me try.

Linear equations are what most people understand and perform from simple math to multivariate statistics and complex calculus. Most people know linear equations as polynomials such as Ax^3 + Bx^2 + Cx + D = 0, or x^3 + y^3 + z^3 = 0. There are very many analytical tools, such as Fourier transforms, which can be applied to linear equations. And, some of them can also be applied piece meal to partial derivatives of non-linear equations.

What is little studied, and only by aspiring math graduate students are the little yellow books (eg. support vector regression analysis) that delve into the wild unknowns of non-linear equations. A nonlinear system is any problem where the variables to be solved for, cannot be written as a linear combination of independent components. They are equations such as the Korteweg–de Vries equation -- , or as you can imagine in climate science the effect of ionizing radiation as it moves through a cloud of water vapor.

There is another problem with climate models. Only a comparatively very few individuals ever learn enough about non-linear equations to be able to understand them, know they are accurate, let alone model them in a computer. Even in todays universities, most undergraduates, and most graduate students will never learn any non-linear equations, let alone the mathematics needed to apply them to real world phenomena. Even then, since the equations are "chaotic" it is hard to insure they fit the behavior you are trying to model. I have much more confidence in actual experimentation, where observed phenomena either fit or do not fit the hypothetical prediction. Models may be useful in suggesting the parameters of experiments, although it is hard to set up a climate for experimentation.