Scaling behavior of global mean sea surface temperature anomalies

Scaling behavior of the monthly global mean sea surface temperature (SST) anomalies from Kaplan SST V2 data are studied by multifractal detrended fluctuation analysis (MF-DFA) method. A crossover at time scale of 38 months (≈3.2 years) is identified to separate distinct regimes: small-scale and large-scale, indicating different patterns of scaling behaviors at different timescales. The scaling exponent h(2)=1.42 at the small-scale (e.g., <3.2 years), indicating the time series of SSTA is non-stationary and slightly anti-persistent, which may due to the El Nino/La Niña-Southern Oscillation (ENSO). At the large-scale (e.g., >3.2 years), h(2)=0.89, showing it is stationary and persistent. This property maybe related to the Pacific Decadal Oscillation (PDO). At the same time, the monthly global mean SSTA shows multifractality with the curves of h(q), τ(q) and D(q) depending on the values of q.