Investigation of local variations of sea‐surface topography

Abstract

The paper describes an attempt to model the sea‐level response to external effects consisting of atmospheric pressure, air temperature, wind stress, and river discharge. The response is sought in two components — amplitude factor and phase lag, both as functions of frequency (from the low‐frequency domain of 0.1 to 6 cycles per year) — and extrapolated to zero frequency. A zero‐frequency permanent response can then be used to estimate the quasistationary (i.e., long‐term mean) local contributions to sea‐surface topography when the quasistationary departures of the external effects are known. The versatility of three mathematical techniques (cross‐spectral analysis, Munk‐Cartwright's weighting function approach, and a response method based on least‐squares spectral analysis) is investigated. In this context, both the input and output series must be considered noisy, with signal‐to‐noise ratio as low as 0.5. The last technique is used to evaluate the permanent response of sea level at selected eastern Canadian ports. From these responses the relative local variations of partial sea‐surface topography are evaluated. Comparison of our results with overland leveling between different pairs of ports shows a consistently good agreement. This agreement indicates that the proposed technique of partial sea‐surface topography evaluation represents a viable approach to this long‐standing geodetic problem.