But that seems foolish, and if true, would lead me to question why they are not including sunspot activity. Just from an unbiased point of view, back when I used to do these tests myself, I'd look at all the available variables and test them. I'd drop variables as they proved to not be correlated or significant to the equation through regression analysis. That would lead me to a handful of variables, that when put together, would explain 95% or more of the variation in the lagging indicator that I wanted to impact.
So in the case of temperature, I would of course want to include sunspot activity, CO2 levels, and any other variable for which a common sense correlation could be made. I can't believe that no scientist has thought of this. Surely, there is an analysis of just this thing out there somewhere. If there is, I haven't been able to find it, though. Very strange. |
