Climate change

'Pink noise' and climate change

Print edition :

Spatial distribution of the shortest timescale (in years) at which pink noise behaviour appears. Photo: Physical Review Letters

A new study from Yale University says that pink noise may be the key to separating natural climate variability from climate change that is influenced by human activity.

Pink noise or 1/f noise is a random noise in which every octave (halving or doubling of frequency) contains the same amount of energy as against the more familiar white noise, which has equal intensity across frequencies. Since it is also inversely proportional to the signal frequency, pink noise has more low frequency components, and it is called pink because visible light with such an energy spectrum will have this colour. Electronic currents, earthquakes, star luminosities and brain signals are known to exhibit a type of noise. The Yale researchers John Wettlaufer and Sahil Agarwal and Woosok Moon of Stockholm University found pink noise in the evolution of temperature and other climate variables on decadal time scales both before and after the Industrial Revolution.

The researchers showed that such noise is due to natural processes unrelated to emissions produced by humans. The finding may help researchers understand how this noise can interact with human-generated greenhouse gases to change the earth’s climate. The work has been reported in “Physical Review Letters”.

“A central question in contemporary climate science concerns the relative roles of natural climate variability and anthropogenic forcing—climate change related to human involvement—which interact in a highly non-linear manner on multiple timescales, many of which transcend a typical human lifetime,” observed Wettlaufer. “We find that the observed pink noise behavior is intrinsic to [the] earth’s climate dynamics, which suggests a range of possible implications, perhaps the most important of which are ‘resonances’ in which processes couple and amplify warming,” he added.

The authors analysed two types of datasets: monthly averaged surface temperatures from 1901 to 2012 from the Goddard Institute for Space Studies, and that derived from isotope measurements of ice cores and cave formations for the preindustrial temperature record dating back more than 100,000 years.

Using a method from statistical physics known as multifractal analysis, the team found that the temperature data contained pink noise-like fluctuations. Because pink-noise features were observed in both the pre- and post-industrial datasets, the authors have concluded that such noise must be caused by natural processes, but its origin remains to be deciphered.