[H2] DSW formation in 3 simulations: 01deg_ryf9091 x 01deg_ryf9091_DSW x 01deg_ryf9091_5mdz

[willaguiar]

This issue compares the DSW formation in 3 experiments

1 - 01deg_ryf9091 : The control experiment for the easterlies project.

2 - 01deg_ryf9091_DSW : Our control experiments, the only difference from 1 is that min depth for the bathymetry is 21m instead of 10m

3 - 01deg_ryf9091_5mdz: New experiment with reduced vertical resolution ( 60 levels, minimum dz = 5m).

Read H2 for more information

[adele-morrison]

Minor suggestion: Let's change the experiment name 01deg_jra55v13_ryf9091_40m to match whatever we choose for the upper cell thickness. e.g. if we use 10m, let's call it 01deg_jra55v13_ryf9091_10m.

[willaguiar]

We have by now ran 8 years of 01deg_ryf9091_DSW and 5 years of 01deg_ryf9091_5mdz. These are the changes we found until now. Here are some results (Registering what I've shown in the meeting of April 28th)...

DSW formation along the shelf for classes for 32.6 kg m^-3 and 32.5 kg m^-3 : It seems that coarsening the upper cells to 5m dz decreases DSW formation by ~45% in both density classes(01deg_ryf9091_DSW vs 01deg_ryf9091_5mdz). Changing the minimum depth for the bathymetry does nothing tho (01deg_ryf9091_DSW vs 01deg_ryf9091).

Maximum density of transformation on the shelf (Average of 3 maximum points), is also smaller by 0.1kg m^-3 to 0.05 kg m^-3:

Regionally, the DSW formation seem to shut off in the Weddell Sea mostly in 01deg_ryf9091_5mdz:

Finally for comparison, here is the distribution of WMT along the shelf for both experiments.

Next on: Check how bottom age changes between the experiments to see differences in DSW export regionally, remake SWMT for separate basins (Weddell, Ross, Adelie, Prydz bay), check the overturning (e.g., ~65^o)S, and mass transport across the 1000m isobath.

[PaulSpence]

These are really nice results. Shutdown Weddell DSW by making the upper layer 5m. What happens in the Wedell? Sea ice?

[willaguiar]

Posting here a few updates on our old and new DSW experiments.

Calculation changes: Mass transports across the 1000m isobath now are calculated using sigma0, and daily transport fields.

Additional experiments:

• Additional experiment (5mtop) in which the only change from control (aka 21mbath) , is that the topmost z-cell has 5m thickness , but the remaining cells are the same. This reduces the st_ocean from 75 to 72 levels.
• Additional experiment (_Strongrestoring), where salt restore increased (max_delta_salinity_restore = 5, before it was0.5)

Part 1 - SWMT on the shelf

---

top to bottom: control (aka 21mbath),5mdz, 5mtop, CM2atm, noSaltRestoring, StrongRestoring

I marked some key values of sigma0. For reference, on the control case, waters below 4000m in the southern ocean, that are commonly associated with AABW, have sigma0>27.84 in circumpolar zonal means. We can see that the decrease in the transformation towards denser waters is similar between 5mtop and 5mdz. Transformation towards densest water classes in CM2atm actually increases, and neither turning off the Salt restore, or increasing it makes much difference.

[willaguiar]

Part 2 - Volume transport across 1000m isobath

---

top to bottom: control (aka 21mbath),5mdz, 5mtop, CM2atm, noSaltRestoring, StrongRestoring left: all isopycnal bins, right: zoomed into the peaking values

The similarity in SWMT is maintained when analyzing the transport across the 1km isobath for waters with sigma0>27.84. 5mdz, and 5mtop have decreased volume transport, while CM2atm has an increase of ˜2 Sv. In Wilma's poster we could see that CM2atm also has younger ages in Weddell slope when compared to the control case ( So this increase in export likely happens in Wedell Sea)

What to do next:

• Check the surface of the Weddell Sea in CM2atm. Do we have changes in sea ice cover/production in the winter?