Choice of vertical coordinate / depth range of importance

[adele-morrison]

When I presented this at CFP group meeting last week, we had some useful critical feedback on the choice of vertical coordinate. i.e. We're showing correlations in depth space on the 1000m isobath. The problems with this are:

1. The time mean heat transport in the depth range that we are saying is important is actually northward in both the dense and surface regimes (left panel here):

The scatterplots show that there are some temporal reversals of the heat transport in all regimes, when the CSHT switches direction to southward. But I'm not sure we can argue that those depths are what's important if the mean heat transport in those layers is actually northward.

2. For areas where the shelf break is relatively shallow, e.g. as shown below, I struggle to understand what cross-slope heat transport at 800-1000m depth actually means, because the isopycnals there intersect the slope just to the south of the 1000m isobath.

I'm not exactly sure what the solution to this issue is, but I'm sure a reviewer will also raise it, so I think we need to think about it. Possibly we could change to a density coordinate instead? Then look at the correlations in the CDW density range? Or would it be more meaningful to look at the correlations near the depth of the shelf break (but this varies along the isobath)?

[willaguiar]

I was talking to Adele, and we come up with one test we could do in addition to the suggestions of the above post:

The vertical profile of mean CSHT might be misleading, as the vertical resolution at the top is much higher than at the bottom ( i.e., in 900m we have 1 cell with a very + CSHT, but between 200m and 500m we have several tiny cells with negative CSHT). This asymmetry would give the impression that the CSHT is northward on the plot, when the sum is actually negative/southward. We can try to redistribute the z layers homogeneously to see if the plot changes, or make the plot in W/m(dzt)

[willaguiar]

Last week we talked about binning the daily CSHT and ASC speed into sigma layers. CSHT in sigma0 is now finished at: /g/data/v45/wf4500/ASC_project_files/Cross_slope_heat_transport/OM2_IAF/daily_rho/* The ASC speed without the 3deg binning is at: /g/data/x77/wf4500/ASC_project/ASC_speed/daily_rho/OM2_IAF_XYgrid/ The ASC speed with the 3deg binning is at: /g/data/nm03/wf4500/ASC_project/model_data/access-om2/binned_ASC_speed/daily_rho/

---

Number of bins Below is a plot of the CSHT with different number of sigma_bins. The increase in number of bins with resolution only happens at layers denser than 1027 kg/m3, where most of the heat transport is… Based on these plots I chose to use 108 bins (109 bin edges).

Target_bin_edges = np.append(np.append(1020,np.append(np.arange(1023,1027,.15),\
                                 np.linspace(1027,rho0_contour.pot_rho_0.max().values+0.2,80))),1030)

---

Quick validation Below is a plot of the CSHT(sigma).sum(‘sigma’).cumsum(‘contour’).mean(‘time’)and CSHT(z).sum(‘z’).cumsum(‘contour’).mean(‘time’), to check if they match:

Below is a plot of the long-term mean CSHT in sigma and z layers (3deg binning and no-3deg-binning). It seems that the problem we had before with the CSHT at deepest parts of the 1km isobath being represented with only a few contour points (e.g., in -150) is gone, since the maximum density is mostly the same along the contour. We do have some variability in the lighter densities along the contour, but the 3deg-binning seem to better capture it….

Because I chose the density bins to not be equally spaced in the layers (coarse bins for sigma lighter than 27 kg/m3), we end up having that sharp separation in 27 kg/m3.

[willaguiar]

Another thing to keep in mind is that the minimum density where we have heat transports vary a lot seasonally (circumpolar sum HT below). I think that would create a big seasonality in the CSHT at lighter layers, and could increase seasonal correlations in those layers. We can explain that issue on the MS.

We also have the option to analyze the surface as one thick density bin ( e.g., one bin for all sigma0< $target\ sigma$. $target\ sigma$ could be for example 27.5 or 27.7), in the hope to smooth all the seasonality due to surface density changes. caveat is that maximum depth of these density layers vary a lot along the contour, so we have to be careful which value we chose, to not get the CDW (below)

Fig below is maximum seasonal depth of sigma0 targets along the contour ( grey = 27.7, black = 27.5).

[adele-morrison]

Does 27.5 always exist everywhere? If so, that sounds like a good density cutoff for the lightest layer.

We may need to coarsen the denser bins also if there are currently density bins that don't exist for all time in a longitude bin. I guess it's ok having density bins that don't ever exist in a longitude bin though, because they will just have no correlation. e.g. For DSW densities in West Antarctica.

[willaguiar]

Make sense. on the 27.5, it does exist everywhere, but not all the time

Perhaps it might be worth to calculate the correlations too before deciding on the sigma bin coarsening ( I mean, we might not even have a good correlation in these light density layers with strong seasonality)

[willaguiar]

A compromise could be to analyze the results in 3 big sigma bins: Surface waters, CDW and DSW. Those bins can be defined based on the mass transport along the contour (Last panel of figure 1 here).