Posted by John Wilkin on December 06, 2001 at 11:20:48:
In Reply to: Re: Invited Commentary posted by Mauricio Mata on December 04, 2001 at 04:57:35:
When the EAC separates near or north of 32S, cyclonic circulation
immediately to the south is evident in results M Bowen, WJ Emery and I
have (**refs below) obtained by applying the maximum cross correlation
(MCC) method to AVHRR imagery in this region. We have processed 8
years of data (over 20,000 images) to get apparent surface
displacements, and hence velocity, from image pairs. The method can
reveal flows close to the coast where altimetry is not available (MCC
doesn't suffer from the geoid reference issue). We certainly see
northward flow at the coast when separation is well to the north -
this flow meets and joins the separated EAC forming a cyclonic
cell. At these times there is often an anticyclonic feature still
further south (a detached eddy). However, the cyclonic circulation
does not survive averaging over the 8 years, i.e. the time mean flow
is still weakly anticyclonic everywhere. Interestingly, the time mean
of the MCC data tends to show separating flow in two locations: just
south of 32S and just north of 36S.
I intepret this as the time average of two typical circulation modes:
Events with strong cyclonic circulation south of the northern
separating EAC are short-lived, and are really just a transition
between two situations. The cyclonic pattern is swept away relatively
rapdily by a southward advancing EAC that migrates the separation
point southward. A strong anticyclonic circulation forms with
separation further to the south, and with significant northward flow
(retroflection) well offshore. This pattern subsequently pinches off
and sheds an anticyclone that migrates southwestward (and can linger
for an extended period). Once the eddy sheds, the EAC separation is
abruptly moved to the north and the cyclone forms. The cycle can then
recur.
This sequence can be seen in Plate 2 of Walker and Wilkin (1998)
[J Geophys Res, 103, 12869-12883].
The same result as the MCC data for the time average of this cycling
(two zones of offshore flow, at 32S, and 36S) is evident, somewhat, in
the CARS climatology of mean dynamic height 0/2000db (Slide 10). The
patterns are a bit more clear in the temperature at 250 m in the CARS
paper (Ridgway, Dunn and Wilkin, submitted to J. Atmos. Oceanic
Tech.).
[** There are two papers in the works on the MCC results. One is to
appear in 2002 in vol 53(1) of "Ocean Dynamics" (a Springer journal);
the other is in review at JAOT (an AMS journal)]