Time–Frequency Characteristics of Global SST Anomalies in the Past 100 Years: A Metrological Approach

Abstract

To comprehensively explore the characteristics of global SST anomalies, a novel time–frequency combination method, based on the COBE data and NCEP/NCAR reanalysis products in the past 100 years, was developed. From the view of the time domain, the global SST generally showed an upward trend from 1920 to 2019, the upward trend was significant after 1988, and the growth mutation occurred in 1930, according to the Mann–Kendall (MK) mutation test. Moreover, we extracted spatiotemporal modes of SST anomalies’ variability by empirical orthogonal function (EOF) analysis and obtained global spatial EOFs that closely correspond to regionally defined climate modes. Our results demonstrated that El Niño–Southern Oscillation (ENSO) is the typical character for the first mode of SST anomaly EOF, and Atlantic multidecadal oscillation (AMO) for the second. From the view of the frequency domain, our data suggested that there is a multi-period nesting phenomenon in global SST variations, in which the first main cycle with the most obvious oscillation was a 30-year cycle and changed in 20-year cycles, and the second cycle was a 15-year cycle and changed in 10-year cycles through wavelet analysis. As for the perspective of time–frequency characteristics, the dominant period of ENSO in the first mode of EOF is 4 years, obtained through filtering and cross wavelet transform. In addition, SST anomalies will maintain an upward trend for the next 60 months, according to the seasonal autoregressive integrated moving average (SARIMA) model, which has the potential value for predicting ENSO.