Low-frequency variability in coastal sea level from tide gauges and altimetry

Long sea level observations are fundamental for studies of the low-frequency changes in ocean circulation and climate. Tide gauges provide records several decades long, but their locations are geographically sparse and confined to the coastline. To assess the extent to which tide gauge point measurements represent broader-scale conditions in the adjacent ocean, we compare tide gauge records to altimeter data in their vicinity for the period 1993-2008. The analysis is based on detrended annual mean series and focuses on interannual variability. The best agreement between tide gauge and altimeter series is found for mid-ocean islands in the tropical Pacific and Indian oceans and along the coast of western Australia. Elsewhere, along most continental coasts, root-mean-square differences between the two sea level estimates can be substantial when compared to the observed variability or the expected measurement noise, with the worst correspondence found in western South America, Japan, and eastern Australia. Differences in magnitude as well as poor correlations between the tide gauge and altimeter estimates are involved. Although land motions and measurement errors can be a factor, results point to the presence of considerable short spatial structure in coastal sea level at interannual time scales, which can be related to the low-frequency dynamics of coastally trapped circulations, western boundary currents and associated eddy regimes, or response to river runoff. Knowing whether coastal sea level variability becomes more spatially homogeneous at longer time scales must await longer sea level records.