The analysis of global surface temperature wavelets from 1884 to 2014

In this paper, we sought to investigate the spatial and temporal variability of the global sea surface temperature in the 1884–2014 period, through wavelet analysis. Results for tropical, subtropical, and extratropical areas indicate important oscillations in the 1–12-month and 1–2-, 2–4-, 4–8-, and 8–12-year spectral ranges, as obtained in other studies which used regional mean SST values instead of global gridpoint data with 2.5o resolution, as used in this study. The intraseasonal (1–12 months) and interannual (1–2 years) oscillations are much more evident in the tropical latitudes. On the other hand, lower frequency oscillations (2–4, 4–8, and 8–12 years) are more dominant away from tropical regions. The faster oscillation scale (1–12-month and 1–2- and 2–4-year spectral ranges) SST variance presents negative trends throughout the entire period, while the values associated with the slower oscillation scales (4–8- and 8–12-year spectral ranges) present increasing trends. The reversal of trends for distinct SST spectral ranges over the whole period suggests energy exchanges between distinct oscillatory phenomena. In addition, the energy increase of slower frequencies suggests the prolonged settlement of low-frequency climatic patterns in the future, providing more inertial climatic patterns. The strongest diminishing values of SST variance associated with the higher frequency oscillations were observed in the Atlantic oceanic basins. The strongest increasing values for the lower oscillations, mainly related to the 2–4- and 4–8-year spectral ranges, were observed in the Pacific basins. In general, the Pacific basins are closely correlated with the global mean (0.55 ≤ r ≤ 0.92), compared with the Atlantic (0.31 ≤ r ≤ 0.75) and Indian (0.12 ≤ r ≤ 0.56) basins.