Empirical orthogonal function analysis of global TOPEX/POSEIDON altimeter data and implications for detection of global sea level rise

Two years of TOPEX/POSEIDON altimeter data are examined to determine the dominant spatial features and timescales of sea surface height variability in the global oceans and to estimate the rate of global sea level rise. Empirical orthogonal function (EOF) decomposition of 69 cycles of TOPEX altimeter data into the significant modes of variability reveals dominant annual and interannual timescales. The annual modes include the hemispheric-scale changes in steric height due to seasonal heating variations, changes in the strength of the major current systems in the equatorial Pacific, and the reversing monsoonal circulation in the Indian Ocean. The interannual modes capture oscillations in the tropical Pacific characteristic of recent El Ni\~no events. A 2-year history of the change in mean sea level derived from TOPEX altimeter data reveals a rise of 5.2 mm/yr. By analyzing the contribution of each EOF mode to global mean sea level variations, we find that 82\% of the rise in mean sea level is caused by a single interannual mode of variability. Altimeter data spanning only 2 years, however, are insufficient to resolve a complete El Ni\~no-Southern Oscillation (ENSO) cycle which dominates the interannual EOF modes. Thus most of the rise in mean sea level derived from TOPEX altimetry is an artifact of incomplete temporal sampling of interannual variability. When a longer time series of TOPEX altimeter data is obtained and a complete ENSO cycle is observed, a significant reduction in the rate of global mean sea level rise estimated from TOPEX altimetry is expected. Most of the remaining rise in global mean sea level is explained by the annual EOF modes, suggesting a possible connection between sea level rise and changes in the steric component of sea surface height.