Depth dependence of westward-propagating North Atlantic features diagnosed from altimetry and a numerical 1/6 degree model

Lecointre, A.; Penduff, T.; Cipollini, P.; Tailleux, R.; Barnier, B.. 2007 Depth dependence of westward-propagating North Atlantic features diagnosed from altimetry and a numerical 1/6 degree model. Ocean Science Discussions, 4 (6). 817-853.

Full text not available from this repository.

Official URL: http://www.ocean-sci-discuss.net/4/817/2007/osd-4-...

Abstract/Summary

A 1/6 numerical simulation is used to investigate the vertical structure of westward propagation between 1993 and 2000 in the North Atlantic ocean. The realism of the simulated westward propagating signals, interpreted principally as the signature of first-mode baroclinic Rossby waves (RW), is first assessed by comparing the simulated amplitude and zonal phase speeds of Sea Level Anomalies (SLA) against TOPEX/Poseidon-ERS satellite altimeter data. Then, the (unobserved) subsurface signature of RW phase speeds is investigated from model outputs by means of the Radon Transform which was specifically adapted to focus on first-mode baroclinic RW. The analysis is performed on observed and simulated SLA and along 9 simulated isopycnal displacements spanning the 0–3250m depth range. Simulated RW phase speeds agree well with their observed counterparts at the surface, although with a slight slow bias. Below the surface, the simulated phase speeds exhibit a systematic deceleration with increasing depth, by a factor that appears to vary geographically. Thus, while the reduction factor is about 15–18% on average at 3250m over the region considered, it appears to be much weaker (about 5–8%) in the eddy-active Azores Current, where westward propagating structures might be more coherent in the vertical. These results suggest that the often-made normal-mode assumption of many WKB-based extended theories that the phase speed is independent of depth might need to be revisited. They also suggest that the vertical structure of westward propagating signals could significantly depend on their degree of nonlinearity, with the degree of vertical coherence possibly increasing with the degree of nonlinearity.