Posted by Olaf Boebel on May 09, 2001 at 08:01:20 (UTC):
In Reply to: Dynamics of the ARC meanders posted by Ricardo P. Matano on May 08, 2001 at 23:52:16 (UTC):
Dear Ricardo,
Thank you for the valuable comments. Let me comment briefly:
>The Hovmoller diagram of the DYNAMICS section is quite interesting
>although I couldn't follow the rationale behind the statement that these
>propagating patterns are associated with eddy shedding instead of wave
>propagation.
This can be seen in the animation at the end of the dynamics section, but a sample image
is given in the still image just below the Hofmoller diagram too. Please note the two
cyclonic eddies, one containing a float. The quasi-stationary meanders grew and shed
cyclonic eddies to the north, i.e. into the area between the AC proper and the ARC. Each
cyclonic streak you see in the Hovmoller diagram was traced to such an event in the
animation. In some events floats confirm the closed cyclonic circulation in such an eddy.
Please note that the latitude of the Hovmoller diagram (37.5) is at the northern edge of
ARC meanders. The Matano et al., GRL 1998 paper figure 4 shows a similar plot at 40S.
Based on our new results, that latitude seems to far south to capture the ARC.
>Do the authors mean that actual eddies propagate westward? It seems that the westward
>speed of a self-propagating eddy is no match for the strong mean flow derived from the
>LADCP.
Yes, the eddies do indeed propagate westward. The trick is that they are positioned to the
north of the ARC, that is outside of the strong eastward flow.
>Couldn't it be that the interaction of the ARC with the ridge near 40 E generates
>westward propagating waves that -sometimes- trigger eddies?
The eddies we observe here are clearly formed at the respective meander (see animation),
in a way they jump "back", i.e. west, from meander trough to meander trough.
MODAS SSH shows a dramatic increase in eddy activity when the ARC interacts with
the Southwest Indian Ridge, but the EKE appears to spread downstream, in a
southeastward direction, rather than upsteam. So far I haven't looked at this interaction in
detail though.
>The westward propagation of cyclones in the Hovmoller diagram seem to
>progress well beyond the Agulhas Plateau (is it correct?). I surmise
>that the apparent discontinuity over the Agulhas Plateau (~ 25 E) is
>caused by a deflection to the north of the mean flow and do not
>necessarily mean that the westward propagation stops there. If so: is
>there any evidence that the westward propagation of perturbations in the
>ARC trigger Agulhas eddies?
This is very difficult to answer, since the Retroflection region itself is very variable, and
it is hard to picture a mean relative to which one would discuss the phase propagation.
Studying the animation, I have the impression that the westward propagation of cyclones stops at the Agulhas Plateau, probably caused by the strong topographic control it exerts. While cyclonic eddies are shed from T1, but they appear to propagate northwestward towards Port Elizabeth, rather than to the west. In fact, the movie shows these cyclonic signals to cross the Agulhas and form large meanders cyclones on the INSHORE side of the current, generating large Agulhas Meanders, similar to Natal Pulses, only that they are not formed near Durban but locally. Nevertheless, those inshore cyclones appear to propagate downstream as well and trigger ring shedding, by cutting off the Retroflection loop.
I should emphasize that this is a topic of discussion between the co-authors of this paper, and that this description is not endorsed by all of us. In fact the MODAS version used herein does not capture those parts of the ARC over areas shallower than 1000m, so a large part of the AC signal might be just off limits, meaning that cyclones could show up because the large AC signal has been masked.
Noting that much of this discussion is based on a detailed study of the animation, I will add an ftp link to enable the reader to download of a zipped version of the movie. Using QuickTime will enable you to scroll back and forth and stop to look at details.
Cheers, Olaf