154-year record of sea level at San Francisco: extracting the long-term trend, recent changes, and other tidbits

A data adaptive method called ensemble empirical mode decomposition (EEMD) is used to examine the 154-year record of monthly sea level at San Francisco. The mode that is lowest in frequency corresponds to the long-term trend. The next highest mode corresponds to an oscillation with a period of ~100 years and may be related to solar variability. When this mode is combined with the long-term trend, the rate of increase in sea level starts to decrease by ~1980. The next lower mode corresponds to interdecadal time scales and thus includes the Pacific Decadal Oscillation. When combined with the two lower modes, sea level itself starts to decrease by the mid-1990s. These results are consistent with the most recent results from the intergovernmental panel on climate change (IPCC), and may be the first obtained from a tidal record. Prior to conducting EEMD, corrections for glacial isostatic adjustment (GIA) and the inverse barometer (IB) effect were applied. The effect of applying the GIA correction was relatively small, but the IB correction reduced the slope of the long-term trend in sea level by almost 15%. This reduction is due to a long-term increase in the variance of sea level pressure. To determine if the 10-15 year ENSO modulation cycle could be detected from the decomposition we first compared the envelope from the mode associated with ENSO, with the two adjacent modes that were lower in frequency. Spectral analysis revealed no significant maxima in the ENSO mode envelope, but a major peak in the spectrum for the two adjacent modes, with a period of 12.8 years. This is consistent with a local response to El Niño warming for the ENSO mode, but a non-local response for the two adjacent modes. A similar analysis was performed for the Southern Oscillation Index and a spectral maximum was found between 12 and 16 years, consistent with our non-local interpretation of the previous two modes.