Posted by Deirdre Byrne on December 06, 2001 at 15:32:02:
In Reply to: ASTTEX posted by Johann Lutjeharms on November 28, 2001 at 22:54:37:
Dear Johann and Olaf,
Sheesh, give a girl a break! And now more seriously:
1. Upcoming experiments in general:
I do agree that the coming few years will give us an unprecedented
chance to look closely at the time-varying nature of the Agulhas
system, in particular its interocean exchanges. For my part, I
regard this in no small part as a community effort -- we will all
gain if these experiments are coordinated and implemented well, and
we will all lose if they are not. Speaking for the ASTTEX PIs, we
are quite pleased with this ongoing debate and feel it can only
strengthen the resulting experiments.
2. Olaf asks (these are short-answer question, so I reply to them
first!):
>Do SAMW rings really have such high salinity? (Fig 3b.)
SAMW rings are characterized by very low stratification -- this means
they have a salinity minimum in their uppermost layers. As one progresses
down through the water column, however, surrounding salinities drop rapidly,
while the SAMW stad remains constant. As a result, right near the bottom
of the SAMW layer, they reach a relative salinity *maximum* -- and this
is the layer I've shown in Fig 3b.
> All the mapped area that is not covered by circles (i.e. data) is
> essentially an extrapolation of your data, isn't it?
Correct.
> It might be good to have a separate statistics for T(z)/S(Z) GEM-ETTA
> estimates with eta and tau falling into the range set by the climatology, and
> those lying outside.
Yes, good idea. We will obviously be much less certain of a thermohaline
profile if the SSH and tau values fall outside of anything ever observed
in situ.
>Figure 7: How do you explain that some IES do show a high (expected) correlation
>(e.g. 59) with Topex and others don't (e.g. 62)?
Three reasons: One, the correlation is high when there are a lot of strong,
coherent mesoscale features in the timeseries. SN 59 was positioned so
that its record contained a lot of such features. SN 62, deployed at
only 980 m on the continental slope, was not. Two, the correlation at SN 62
improves when I use the in situ tide record instead of a global model --
the model does not perform well on the slope. I didn't have time to do that
for this poster. Three, there appear to be other processes (many topographic
waves?) acting in this location which are not present at the deeper mooring
sites. These results are the reason why I hope to keep the ASTTEX moorings
at 2000 m or greater.
> Why is the high correlation sometimes not co-located with the PIE (i.e. #56)?
I would say partly for the same reasons given above. SN 56 was deployed
at ~1900 m, again on the continental shelf. My guess is that the majority
of strong and coherent mesoscale variability observed at this location was
propagating along the cont. slope from the SE to the NW -- so that the
larger signals in the record of SN 56 were even more clearly expressed
in the slightly deeper waters just SW of it. In addition, there is the
same possibility of a relatively more poorly performing tidal model the
shallower one goes, which could help explain the better correlation in
the deeper water to the SW and the poorer correlation to the NE.
> Figure 16: Why is the barotropic low-frequency signal so strong at 58 and not
at 61 and 62, which brace the location of 58? Is this the main path for the
barotropically intense eddies?
First, SN 61 was deployed *on* the Walvis Ridge. I do not expect any significant
barotropic signal to survive this ridge. SN 62 was, as I mention above, at only
980 m depth, well inshore of the eddy corridor. So yes, SN 58 is in the right place
for barotropically intense eddies, but there is a very wide span between SN 61
and SN 62 so this statement is not particularly exact. There was another mooring
deployed at 30S, midway between SN 58 and SN 62, but the IES sensor flooded upon
deployment. I would have expected that record to show a strong barotropic signal
from the passage of eddies as well.
> Figure 21: Isn't it rather a decrease in intermediate depth salinity at day 325
> than an increase? Dark purple seems less salty than the
>blue hue on the plot. If so, what about your explanation in the following paragraph?
I am afraid you've read the scale backwards -- increasing salinity is UP on the
color bar (lighter colors tending toward red), so the intermediate depth salinity is
INCREASING at the point mentioned in the text, from ~34.3 psu to ~34.4 psu (very
roughly). I'm sorry I didn't flip the orientation of the color bar -- it's
very confusing when we are all used to thinking of more dense water on the bottom
and less dense water at the top.
3. Re: ASTTEX mooring positions:
Johann has said:
> I suggested [the moorings] could go one or two [Topex/Jason] lines
> further south (or closer to the source). Will de Ruijter seemed to
> agree with me. Perhaps other colleagues might disagree with this advice
> and have good reasons for doing so. I would be very interested to hear
> what others think about the optimal location for the line of ASTTEX moorings.
We have considered seriously all of the comments I have received about
the advisability of repositioning the ASTTEX mooring line. Our objective
s to measure as best we can the fluxes of Indian Ocean central and
intermediate waters and their associated anomalous heat and salt transports
into the South Atlantic. Our intended analysis method relies on there being
some distinguishing characteristics to the thermohaline profiles of water
from the Agulhas -- in differential T/S characteristics or in the
stratification (T(z) or S(z)). Thus I think the ASTTEX moorings should be
located as close as possible downstream of the Agulhas Retroflection without
ever being within it. This should minimize the mixing that has occured,
which will give us the best chances of being able to distinguish Indian
Ocean water masses from Atlantic ones. This proximity will also help insure
that we capture the widest percentage of the swath of water coming from the
Retroflection.
However, we are not only interested in the presence of thermohaline
anomalies, we're interested in their fluxes through the region.
At the other end of the scale, Olaf thinks we should push the mooring
location further *west* than was our original plan. His worry is that
we will not be able to reliably measure these fluxes because the
time-varying component of the transport will be so much larger than
the mean. My response to this is based in part on the information we
gained from the BEST experiment, the data from which data we have been
re-analyzing in this poster. The annual average baroclinic transport
relative to 1400 dbar between 37.4S, 12.3E and 34.2S, 17.4E was 5 Sv,
and at 30 S was about 14-15 Sv. Note that the ASTTEX array will fall
somewhere between the two BEST mooring lines.
It may be that below the thermocline, the transport is weak enough that
our direct flux estimates are not above the noise level, but I feel
quite sure that for central waters (roughly, 6-16C), we'll be able to
make reliable direct estimates. And central waters form the most
important component of the exchange. I say direct estimates because I
think we can fall back on some additional assumptions, if necessary.
For example, given that the ASTTEX moorings are at no time within the
Retroflection itself, we can safely assume (in my opinion) that any
Indian Ocean waters we detect at the array are participating in an
interocean exchange and will not be heading back into the Indian Ocean
any time soon. Thus should our direct transport estimates turn out to
be noisy or unreliable, we could estimate a "bulk" or overall flow rate
through the Cape Basin at different density horizons and fall back on
simple temperature and salinity flux calculations based on the observed
profiles and these bulk rates. Obviously this would be less than ideal,
and Olaf's concerns are important ones. In summary, though I believe
that an understanding of the time-varying signals of the water masses
downstream of the Retroflection, and their relationships to the passage
of different features and to the gyre-scale variability in the South
Atlantic Ocean (and in the Indian Ocean -- I do hope Will will be measuring
and monitoring conditions there) will be the most important contribution
ASTTEX makes, and this requires proximity to the Retroflection -- i.e.,
we'll be deploying in the Cape Cauldron somewhere.
You'll note the mooring positions in the poster abstract are those from
the original proposal, with the majority deployed along Topex/Jason
groundtrack 235. At this point, I am fairly strongly committed to moving
the ASTTEX moorings at least one track further east, to groundtrack 057.
The positions we're considering at present are shown in Fig. 1 of the
poster (in the section labelled "Background and Scientific Goals").
However at this point, I'm waiting hear the opinion of our altimetry
expert, PI Donna Witter. She's going to run a simulation or two and
take a closer look at the MODAS output. We need to be able to reliably
distinguish whether the features we observe in the array have detached
from the Retroflection, and to be able to track them out of the Cape Basin
(assuming they actually leave it). When we look at track 057, we'll also
consider groundtracks 133 and 209, the next two further east of 057.
I'll update the project website when we come to a new consensus, and
I'll be sure to email an update to anyone who has weighed in here.
4. Cyclones in the Cape Basin:
Johann asked:
> We now know that there also are cyclones that come from the continental
> shelf and that seem to move at right angles to the path of the Agulhas
> rings (see e.g. Richardson and Garzoli, DSR special KAPEX issue, submitted).
Johann, you ask a very good question here. As you know, at the time
the experiment was proposed, we did not have any knowledge about these
coherent features. I have watched the MODAS/KAPEX float movie over
and over to try and get a better "handle" on these features. I am
not at this time not certain of their significance in the interocean
transport equation; I'm not even aware of any information about the
origin of the water masses in these features -- are you? Their path
in the Cape Basin does not suggest, at least to me, a direct origin
in the Retroflection and to me it even seems possible that they might
have something to do with the Benguela upwelling system rather than the
Agulhas, although perhaps they're too far offshore for that. In any
case, the ASTTEX array was designed to capture 100% of features of the
typical size of Agulhas rings, and 60% of weak features (~80 km diameter
and ~10 cm anomalous sea surface expression), a point you mention briefly
in your comments. If the cyclones are at least this energetic, and
if they contain some kind of water mass anomaly, we can hope to resolve
them with the array. That said, I could say more definitely if I had some
scale estimates for them and some in situ T/S profiles from inside one
or more of these cyclones -- do you know of any such?
5. Meyer experiment:
Johann, a question for you; given that Lisa Beal is not instrumenting
the main part of the Agulhas Current but only the Undercurrent, what is
your reasoning for encouraging Alan Meyers to measure the Agulhas Return
instead of augmenting the line of Undercurrent Experiment Instruments,
or even an independent line somewhere downstream of that but upstream
of the Retroflection? Both Alan and Lisa mentioned this development to
me, but did not elucidate.
Thank you both for your insightful comments! I look forward to our
continuing discussion.