You will note the noise on the residuals (blue trace) still carries a signal of about half the amplitude of the original
HadCRUT signal, so the residuals are not just noise as had been alleged.
The non-AMO «regressors» (SolVolENSO) show little response within the MLR (which will impact on the implied climate sensitivity) and, unlike F&R (2011), their attribution of these «regressors» accounts for very little of
the HadCRUT signal.
If we detrend
HadCRUT, analogous to removing the DC leaving only the power supply ripple, and subtract this (ENSO, PDO, AMO, SSN, Pinatubo, etc) «hum» from the
signal + noise of UAH temperature measurements, we can also improve our Signal to Noise
signal + noise of UAH temperature measurements, we can also improve our
Signal to Noise
Signal to Noise Ratio.