Theodore Durland (Oregon State University, USA); Steven Jayne (Woods Hole Oceanographic Institution, United States)

Salient results from our 2017-2020 OSTST grant, entitled "Beyond mesoscale eddies: Ocean dynamical signals and mapped SSH":
Our present knowledge of ocean dynamics has benefitted tremendously from 2+ decades of satellite altimetry and the Archivage, Validation et Interpretation des donnes des Satellites Ocanographiques (AVISO) data sets that have merged the irregularly spaced measurements of multiple altimeters into gridded products that are used by virtually the entire international oceanographic research community. Knowledge of the dominant oceanic dynamics has advanced to the point where we now can consider refinements to the covariance functions and optimal interpolation algorithms that might reduce mapping errors and admit additional dynamical features that are adequately sampled by satellite altimeters, but poorly represented in the current AVISO products. Three distinct dynamical processes motivate and will serve as case studies to guide our work on this.

In our current OSTST project we have been reconciling observations from satellite altimetry with numerical process model output, and some significant challenges have arisen that motivate the present proposal. Strong evidence for 30-35 day barotropic Rossby waves radiating northward from Tropical Instability Waves (TIWs) was previously published by the PI. Our more detailed recent analysis of the gridded AVISO product shows that SSH variability coherent with the TIWs can be found throughout most of the North Pacific; as far poleward as the Gulf of Alaska and the Aleutian Islands. The implication that this wide-spread coherence is due to the previously identified barotropic Rossby waves is supported by the coherence patterns, with one puzzling exception. The observed SSH variance in this period band decreases by a factor of about 100 in the 20-40° latitude band, and it increases again after the waves propagate through the region and reach higher latitudes. We followed several conjectures for explaining this behavior, but we were unable to reproduce this localized minimum with process models or come up with a plausible way of rationalizing its existence. We only recently discovered that the minimum is produced by the spatial dependence of the time scales incorporated in AVISO’s objective mapping algorithm. An independent mapping of the along-track data with a spatially uniform mapping timescale produces much better agreement with model predictions and physical reasoning.

The above example highlights a particular limitation resulting from the assumptions that go into the AVISO gridded altimetry product. Another potential weakness concerns assumptions about SSH propagation speeds. At each location, the AVISO mapping scheme assumes a propagation speed based on theory and observations. While this may represent the dominant propagation characteristics at most latitudes, there are important regions of the world oceans (the Southern Ocean and the Kuroshio Extension, for example) where simultaneous eastward and westward propagation can be both predicted and observed, but is not handled consistently in the a priori autocovariance function used for the AVISO mapping.

We have discussed these issues with the personnel at CLS responsible for the AVISO/DUACS processing, and have developed a collaborative plan to work toward improving the AVISO mapping algorithm. Our part of the effort will entail dynamical modeling and independent analysis of available along-track altimetry data to quantitatively describe and understand ocean dynamics presently not represented in the AVISO mapping autocovariance functions and subsequent work to consider different ways of handling such signals in an objective mapping to produce an improved SSH product.