Objective estimates of westward Rossby wave and eddy propagation from sea surface height analyses

Abstract

An objective method to estimate westward propagation of sea surface height (SSH) anomalies is introduced and compared to a subjective technique. Both approaches use time series extracted from global SSH analyses, prepared by optimal interpolation of altimeter data, as described in detail. The cost function used in the objective method calculates the mean along-slope standard deviation, where slope is the reciprocal of propagation speed and defines the angle of linear samples through the SSH time series along each transect. The along-slope standard deviation can be computed as an extension of the Radon transform, as shown. The optimum speed estimate minimizes the cost function, and speeds with mean along-slope standard deviation within 1% of the minimum define the range of uncertainty. The 1% criteria was chosen to give a seemingly reasonable uncertainty range over a variety of cases and does not imply a specific probability that the true minimum is bounded in the uncertainty range. Applied in the central Pacific, objective speed estimates increase from 1 cm s-1 or less at the highest latitudes to above 30 cm s -1.nearer the equator. Subjective speed estimates are similar to, but generally lower than, their objective counterparts, reflecting a bias in subjective interpretation that is likely to vary among different individuals., A more detailed examination of the along-slope minimization over 160°E-95°W focuses on the higher •low-latitude speeds, returning estimates of 48 cm s-1 (45 cm s-1) at 5.5°N (5.5°S) over 1993-2006. The corresponding 42-55 cm s-1 (3865 cm s-1) uncertainty ranges bound most of the predictions from sensitivity test cases with alternate preprocessing filters or different multiyear time windows, although variability is larger south of the equator. The findings are consistent with prior studies and suggest the methods may be useful for other applications over the global ocean.