Look for signature of Weddell-Scotia confluence waters

[DaniJonesOcean]

Mike mentioned that the western part of the transition class looks like the Weddell-Scotia confluence waters. There, the signal off the shelf is injected at mid-depth (a few hundred metres). There's a freshening, associated with CDW.

Mike is going to send some papers, to give us a sense of the specific signal that we could look for.

[DaniJonesOcean]

We could do an east-west split to look for this signature.

[DaniJonesOcean]

Just followed up with Mike, since I hadn't heard from him.

[DaniJonesOcean]

From Mike:

The Weddell-Scotia Confluence is fascinating… it is a region of reduced mid-layer stratification (compared with the ACC to the north or the Weddell Gyre to the south) that sits close to the topography of the South Scotia Ridge. For a long while it was thought that this reduction in stratification is due to vertical mixing caused by the topography (e.g. https://journals.ametsoc.org/view/journals/phoc/10/10/1520-0485_1980_010_1584_twsc_2_0_co_2.xml), but then Tom Whitworth published a key paper that showed it is due to shelf water spilling off the tip of the Peninsula and then sinking and flowing eastward:

https://doi.org/10.1016/0967-0637(94)90046-9

In the attached, Figure 2 shows nicely the reduced stratification above the SSR, which is evident especially at the theta-maximum (around 500m depth) or S-maximum (a bit deeper). I think it’s accepted now that the WSC doesn’t extent as far north as the dashed line in Figure 1 shows, but they only had the data they had back then.

I don’t know how sensitive your automated classification methods are, but if they could place boundaries between the ACC waters and Weddell Sea waters (“clouds” in Figure 4) from the WSC waters (e.g. stations 51, 126 etc in that theta-S plot) then you might be able to automatically extract the spatial extent of the WSC…? From a quick look at the plot you sent, it might be that your technique is already doing this, but perhaps the eastern extent of the WSC is not well known so I’m not sure whether it’s picking up other things also beyond e.g. 20W….

[DaniJonesOcean]

This will perhaps be harder to do in the K=3 model, but maybe it will show up on the western edge of the transition class?

[DaniJonesOcean]

Incorporated ok