BLog

ImprintImpressum
PrivacyDatenschutz
DisclaimerHaftung
Downloads 

Noise vs. Pattern in the NOAA Global Temperature Anomaly Curve

I am electrochemist and every now and then it happens that our electrochemical measurement curves are noisy, for example in voltammetry, when very small currents in the µA range and below are measured, or when measurements are made with very high data rates, as in this potential curve with over 1 MS/s. This one of a fast amperometric measurement on a galvanized and passivated steel pin in 5 % saline solution is rather zig-zag, and this was caused by captured noise by the measurement signal.

Noise in a high speed electrochemical potential measurement

The potential curve in the fast amperometric measurement on a galvanized and passivated steel pin in 5 % saline solution is rather zig-zag, and this was caused by captured noise in the measurement signal. With suitable methods one can get the curves much smoother (see Electrochemical methods ...), but that is not the topic here. Anyway, this is what noise looks like in measured time series in terms of irregular fluctuations.

In a BLog post in 2019, I had shown how the time scale of the monthly NOAA time series of global surface temperature anomalies can be decimally converted so that line plots like the following can be readily produced.

Well, this curve is also quite jagged, but just not ziggy. Apart from the obvious long-term temperature increase, I recognize repeating patterns as well as trends extending over months and years, and these are of course anything but irregular fluctuations of the measured values. The interesting question is, what is behind this?

The Peak Inversion

Probably the most interesting pattern is revealed to you when you trace some of the peaks back down and on. Some are symmetrically surrounded by preceding and following peaks upward. It looks like the process that was supposed to lead to a much higher temperature peak was disturbed in the middle, which then led to a temperature collapse = peak inversion. However, the disturbance was not an abort, because it went up again.

Note the mirror symmetries both on the vertical axis and a virtual one on the horizontal axis of the inversion, the latter showing up when the interrupted peak is imagined to continue upward.

The scattering of the global temperature anomalies is certainly not random. The recurring patterns shown above and also the comparison with measured curves where the course is really noisy speak against it.

Copyright © Dr. Rolf Jansen - 2021-08-09 21:13:14

Discussion on Twitter: 1434157391029882888