Assessment of nonlinear trends and seasonal variations in global sea level using singular spectrum analysis and wavelet multiresolution analysis

Abstract

The main purpose of this paper is to apply the singular spectrum analysis (SSA), based on the phase space, and the wavelet multiresolution analysis (WMA), based on the frequency space, to the weekly time series of global sea level anomaly (GSLA) derived from satellite altimetry data over 1993–2013, in order to assess its nonlinear trends and its seasonal signals. The SSA results show that the GSLA time series is mainly dominated by a nonlinear trend explaining 89.89 % of the total GSLA variability, followed by annual and semi-annual signals with an explained variance of 9.15 and 0.32 %, respectively. For the annual signal, both methods give similar results. Its amplitude is less than 14 mm with an average of about 11 mm, and its minimum and maximum occur in April and October, respectively. The calculation of sea level trend, by both methods, is direct without removing the seasonal signals from the original GSLA time series as the most commonly used in the literature. The global sea level trend obtained from the WMA is about 2.52 ± 0.01 mm/year which is in good agreement with 2.94 ± 0.05 mm/year estimated from the SSA. Furthermore, the SSA method is most suitable for seasonal adjustment, and the WMA method is more useful for providing the different rates of sea level rise. Indeed, the WMA reveals that the global sea level has risen with the rate of 3.43 ± 0.01 mm/year from January/1993 to January/1998, 0.66 ± 0.01 mm/year from February/1998 to May/2000, 5.71 ± 0.03 mm/year from June/2000 to October/2003, and 1.57 ± 0.01 mm/year since January/2004.