The solid black line is the global annual mean and the solid red line is the five - year
lowess smooth, i.e. a non-parametric regression analysis that relies on a k - nearest - neighbor model.
Monthly SSTA (°C) and PDO values were smoothed with a 25 -
month lowess smooth.
The mean log2 ratios for each spot were normalized using
the LOWESS curve - fitting equation on a print - tip specific basis to allow for differences among the four printing pins used during array manufacturing.
The post referred to extrapolates the linear trend, as estimated by
a lowess smooth, to generate «what if» scenarios for the not - too - distant (but not too near) future.
I even detrended both global temperature and AMO nonlinearly (with a «slow»
lowess smooth) and repeated the regression, to see whether AMO might at least account for some of the short - term fluctuations (just as ENSO does).
Here's the Berkeley data (minus the final two «don't belong» data points), together with
a lowess smooth on a 10 - year time scale:
The blue line is
a lowess curve, fitted to the observations (generated using Maple, with default settings, omitting the left-most point).
LOESS or
LOWESS is a locally - weighted fit utilizing subsets of the data to produce a smooth curve.
The moving average is performed with
a lowess filter.
The LOWESS is a modified running - time mean, and its use allows the mean to extend to the beginning and end of the record.
The AMO Index (16) smoothed by locally weighted scatterplot smoothing (
LOWESS)(19) is superimposed (in green).