Statistical characterization of Geosat altimetry noise: Dependence on environmental parameters

[1] Sea surface height (SSH) measurements from the Geosat satellite altimetry mission are analyzed to statistically characterize their noise process, which is defined here as the small-scale (<50 km), timevarying component of the altimetry data. This analysis is one component of a larger effort to extract meaningful information about seafloor roughness from the small-scale fabric of the altimetric gravity field. Both small-scale seafloor topography (e.g., abyssal hills) and SSH noise significantly influence the measured gravity fabric. Hence, a global, quantitative understanding of SSH noise is required to infer what portion of the gravity fabric is associated with the true geoid. It is also anticipated that the SSH noise analysis will prove useful in designing improvements to altimetry processing algorithms. The SSH data are first processed by differencing nearest-neighbor tracks, to minimize the static component of variability, and then by high-pass filtering, to remove variability associated with mesoscale oceanic circulation. Autocovariance analysis reveals that SSH noise can be decomposed into uncorrelated (white noise) and correlated, semiperiodic components. Three parameters are measured which characterize the variance of both components and the correlation length of the correlated component. The global distribution of uncorrelated noise variance is closely correlated to significant wave heights, directly linking the uncertainty of the altimeter return time to the roughness of the sea surface. The variance and correlation length of the correlated component of noise can be associated with three separate environmental parameters: sea ice, precipitation rates, and upper atmospheric variability associated with the subtropical jet stream. As widely differing as these environmental processes are, it is hypothesized that features like sea ice, rain storms, and atmospheric turbulence can cause sudden perturbations in the return time of the altimetry signal, creating decaying oscillations associated with the onboard tracking algorithms.