1/f noise in irradiance records affects our understanding of trends in solar radiative forcing

Abstract:

One of the critical issues in solar irradiance observations is the making of records that are long enough to 1) have sufficient overlap between different missions to determine instrument offsets and 2) to distinguish actual solar trends from instrumental effects. Different approaches have been developed for quantifying the minimal duration of such records or for determing when a perceived trend is real.

Recently, we have found that the noise of total solar irradiance radiometers tends to follow a 1/f scaling, which is also known as flicker noise. 1/f noise stands out by having long-range correlations and properties that are substantially different from the usual (and often implicitly assumed) white noise. Here we show how this 1/f noise deeply impacts our perception of trends in irradiance observations, whose assessment requires much greater care.