Defining sigma levels according to water mass properties

[willaguiar]

In the last meeting we redefine the CSHT($\sigma$) binned into $\sigma$ levels that reflects the water mass properties along the 1km isobath. CDW/DSW cutoff should be based in the maximum cumulative northward heat transport, while AASW/CDW should be based in a combination of maximum temperature + age from a TS diagram , taking into account a density range to classify the CDW.

This issue is for us to discuss these definitions. tasks are:

• Analyze mass transports integrated per regime
• Analyze TS diagrams averaged per regime
• Repeat the above analysis in large lon_bins along contour (i.e., $20^o$ lon bins)

[willaguiar]

• Analyze mass transports integrated per regime
• Analyze TS diagrams averaged per regime

Here are some TS diagrams for the data along the isobath, using monthly data for the year 2000 ( to get summer and winter). If we go this route, I think a $\sigma_0$ of 27.6 for the Surface and Reverse regime, and a $\sigma_0$ of 27.7 for the deep regime would work to separate the surface waters from the CDW

For the CDW/DSW sigma limit, the mass and sigma transports peak at $\sigma_0$= 27.7 for the surface regime, $\sigma_0$= 27.8 for the deep regime, and the Reverse regime doesn’t seem to have northward transport of DSW so we consider everything below 27.6 as CDW.

Let me know your thoughts on this option!

[willaguiar]

Based on the discussion, I am defining the following edges for the sigma levels.
$\sigma_0$ = [20, 27.6, 27.7,27.75, 27.8, 27.84, 27.9]
here is how they look like plotted over the cumsum of Heat transport and mass transport.

outputs for the CSHT in these levels are here:
/g/data/v45/wf4500/ASC_project_files/Cross_slope_heat_transport/OM2_IAF/WMbins_daily_rho/*
outputs for the ASC speed on these levels are here:
/g/data/v45/wf4500/ASC_project_files/ASC_speed/daily_rho/WMbins_daily_rho/*
(updated on the read-me already)

[adele-morrison]

Here's what the mean cross-slope heat transport (plus zonal convergence) looks like in the new density bins.

Seems like it would make sense to focus on correlations in the 27.8-27.84 density layer, where all 3 regimes have time mean southward heat transport. Maybe for the Dense regime, we could consider the layers above this also.

Interesting comparing with the T/S diagrams higher up in this thread. The southward heat transport only occurs on the very very densest side of the CDW peak.

[adele-morrison]

Here is the average speed in the different density bins in each regime:

I'm thinking this speed barplot and the above CSHT barplot could go in the new Figure 1.

Also, here is the time average density on the 1000m isobath showing where these different density ranges sit. Density ranges can of course exist for shorter time periods but not show up in the average (e.g. the dense bin especially).

[adele-morrison]

I think we may run into problems calculating correlations in these density bins, because the 2 densest bins (CDW 27-8-27.84 and DSW 27.84-29) don't exist at all times:

The CDW layer may be ok once we bin into larger longitude bins, but I think the dense bin will be problematic. I'm not exactly sure what to do about this.

[ongqingyee]

Here are some TS diagrams for the data along the isobath, using monthly data for the year 2000 ( to get summer and winter). If we go this route, I think a σ0 of 27.6 for the Surface and Reverse regime, and a σ0 of 27.7 for the deep regime would work to separate the surface waters from the CDW

I was going to plot similar T_S plots for specifc regions around Antarctica and overlay density in the colors instead of Age. Is the script to do the binning into 3deg bins Binning_ASC_speed.py , or is there a different one that you used for the TS and Age @willaguiar ?

[willaguiar]

Is the script to do the binning into 3deg bins Binning_ASC_speed.py , or is there a different one that you used for the TS and Age @willaguiar ?

Yes! (for the ASC speed).
the ASC speed and CSHT should be binned into 3deg bins already in the netcdf.... if the goal is to bin TS, you can check the code I used to do that fig here, cells 7-9

[ongqingyee]

Is the script to do the binning into 3deg bins Binning_ASC_speed.py , or is there a different one that you used for the TS and Age @willaguiar ?

Yes! (for the ASC speed). the ASC speed and CSHT should be binned into 3deg bins already in the netcdf.... if the goal is to bin TS, you can check the code I used to do that fig here, cells 7-9

I wanted to bin rho so that I could overlay density on the TS diagrams you already plotted. Will use the Extracting_TS notebook then. Thank you!

[adele-morrison]

Here are the annual correlations in the CDW layer with 10deg longitude bins. You can see even with annual averages there are longitude bins missing where there are gaps in the annual time series where this density range doesn't exist for some years.

The different colours are the different regimes. I'm wondering if perhaps it's not that helpful to average over different regimes? All regimes seem to have low and high correlations in different places. Perhaps an alternative could be to focus just on the correlations in this one density layer of interest and show spatial variations with longitude rather than regimes?

[willaguiar]

I do like the idea of not focusing specifically in the regime averages, as I think It will help point to regions of interest where there is a higher correlation + southward heat transport.

Do you think it would be worth to make a line plot like that one for the mean CSHT /ASC speed too? Just so we can see if those high correlations ( e.g., around -140) align with southward transport locally?

[willaguiar]

Update on water masses definition

Last hackathon @taimoorsohail was able to produce a data with maximum and minimum densities with non-zero CSHT as f(contour, time). I updated his code to separate the AASW/CDW/DSW with the definitions we discussed before. Here are a few plots on how that looks like.....

PS: We talked a lot about how we would define these waters last time, so let me know in case that is not the definition we agreed upon. Definitions are based on the 3deg binned daily output.

Code for the images and definitions

masks for Water mass definition are on /g/data/v45/wf4500/ASC_project_files/{CDW/AASW/DSW}_mask.nc

[adele-morrison]

So that means that we won't be able to do correlations in the DSW layer, because there are very few longitudes where the DSW layer exists at all times. This isn't a problem if we only want to focus on the CDW layer. But is problematic if we wanted to show other layers.

[ongqingyee]

So that means that we won't be able to do correlations in the DSW layer, because there are very few longitudes where the DSW layer exists at all times. This isn't a problem if we only want to focus on the CDW layer. But is problematic if we wanted to show other layers.

But DSW is not always present around Antarctica since its formation is localised. Is that sufficient justification as to why we can't do a circumpolar correlation in the DSW layer but we can in the other layers?

[willaguiar]

But DSW is not always present around Antarctica since its formation is localised. Is that sufficient justification as to why we can't do a circumpolar correlation in the DSW layer but we can in the other layers?

I think so... I also think we are mostly interested in the Southward CDW heat transport, so just doing circumpolar average correlations for the DSW layer (There should be no NaNs circumpolarly) should be enough ( and then we can justify any correlation with the result of Morrison et al 2020)

[taimoorsohail]

So that means that we won't be able to do correlations in the DSW layer, because there are very few longitudes where the DSW layer exists at all times. This isn't a problem if we only want to focus on the CDW layer. But is problematic if we wanted to show other layers.

I think if we stick to fixed density surfaces, this is a natural outcome of that approach. I agree, I don't see an issue with this - it will be easy to justify that DSW does not exist everywhere, and we could stick to circumpolar averages for DSW.

[adele-morrison]

Yep, agreed, I don't think it's an issue.

[willaguiar]

Here is some early analysis, in bar plots like Adele did, but with varying density layers....

It seems that the ASC speed pretty much agrees with the fixed previous density definition for CDW. But the heat transport changes a lot in the CDW layer, becoming Northwards in all regimes, except the reverse one.

The cumulative sum of the time-mean CSHT+ZC show that when shifting from the previous fixed density definition to the now-variable CDW density definition, the heat transport changes a lot, primarily Around Weddell sea, and the other places where we see the trough in the density curtain. The variable density definition in these regions probably ends up selecting different waters at different time (sometimes AASW sometimes DSW)

That means that we have to think a little more on how we define CDW for the correlations. Should we stick with the fixed density thresholds and figure out what to do with the vanishing density layers? Or else?

[fabiobdias]

Quick summary from today's meeting:

1. We talked about time-mean x time-varying, spatially-varying x spatially-constant density thresholds and also on Wilton's post above - I think we agreed that removing the Weddell from the analyses as it doesn't have CDW present would be justifiable;
2. Adele's proposed to look at the CSHT binned into 10deg bins and choose density thresholds for each bin based on the CSHT profile;
3. Wilton also suggest to do again the correlations on the annual mean first (with the new density thresholds defined from (2)).

I have go ahead and binned the daily CSHT in density coordinates (from /g/data/v45/wf4500/ASC_project_files/Cross_slope_heat_transport/OM2_IAF/daily_rho/) into the 10deg longitudinal bins. The binned data can be found here:
/g/data/hh5/tmp/access-om/fbd581/ASC_project/Lon_binning/