Using the Polynomials as Predictors of Global Warming

  The lower degree polynomials make better predictors since the higher degree polynomials tend to diverge from mean value in an interval as one moves father forward in time. In the following example the data from the 40 years prior to 1980 were used predict the subsequent global land anomaly. The black horizontal line is the average for the interval. The blue line is the linear fit and the cyan and brown are the best 2-degree and 3-degree polynomials.

If the change was determined by a stochastic variable the horizontal line would be the best predictor. In this case it seems to be the best "predictor" of prior anomalies. The horizontal line and linear fit seem to be the best long term predictors. The quadratic curve is capable of detecting some curvature in the interval but it can go astray more quickly. Smaller intervals for the fit tend to result in higher curvatures and sometimes produces unlikely results.