Generate SE-NEMO configurations using local-ME s-cordinates

[oceandie]

1. An SE-NEMO configuration employing localised Multi-Envelope (ME) s-coordinates is generated.
2. The Antarctic shelf is treated using a particular setting of the ME, i.e. we use one setting of the ME for the Antarctic shelf and another setting for the other shelves (hereinafter they are named global shelves).
3. This model will use ME s-coord near the coast (up to isobath 2600m in Antarctica and 1650m in the rest of the Globe) while it will use the standard eORCA025 z-partial steps grid everywhere else.

[oceandie]

All the developments and code can be found in this repo .

[oceandie]

CONFIGURATION OF ENVELOPES FOR GLOBAL SHELVES

An SE-NEMO configuration employing localised Multi-Envelope (ME) s-coordinates is developed. This configuration uses 4 envelopes to discretise the domain excluding the shelves of Antarctica, which will be discretised using a specific setting of ME described later.

The configuration, named ME-300, uses the following setting for the envelopes:

ENVELOPE	MIN DEP	MAX DEP	RMAX	NUM LEV
1	8m	300m	0.2	37
2	108m	1650m	0.2	13
3	608m	2000m	0.03	7
4	5750m	5750m	n.a.	18

After some preliminary sensitivity test, this seems to be a quite good setting (especially the max depth of the upper envelope) able to describe quite faithfully the real bottom topography and not intruducing too many step-like structures.

The localisation works with the same approach employed for improving Nordic-OVF in orca025. The following is the mask showing where the model will use ME s-coordinate (red), standard orca025 z-partial step coordinate (blue) and a linear combination of the two to transition from one type to another (green):

As we can see, ME s-coord are used in regions where the topography is shallower than 1650m .

[oceandie]

The following are selected cross sections showing how the localised ME s-coord looks like.

ME-300

1) Gulf stream region

2) Gulf of Mexico

3) Labrador Sea

4) Norwegian sea

5) European North West Shelf

6) North Sea

[oceandie]

CONFIGURATION OF ENVELOPES FOR ANTARCTIC SHELF

The SE-NEMO configuration employing localised Multi-Envelope (ME) s-coordinates uses 4 envelopes to discretise the Antarctic shelf. Envelopes and model levels for Antarctica are set as follows (ME-800 setting):

ENVELOPE	MIN DEP	MAX DEP	RMAX	NUM LEV
1	8m	800m	0.2	42
2	358m	2600m	0.2	13
3	858m	4500m	0.03	7
4	5750m	5750m	n.a.	13

This setting is chosen to have model levels that nicely follow the Antarctic shelf genlty sloping up to 800 m.

The following is the mask showing where the model will use ME s-coordinate with the ANTARCTIC setting (red), standard orca025 z-partial step coordinate or local ME s-coord with ME-300 setting (blue) and a linear combination of the two to transition from one type to another (green):

As we can see, ME s-coord are used in regions where the topography is shallower than 2600m . Also, the transition zone (green) is slightly wider than the transition zone used in the ME-300 setting for the global shelves.

[oceandie]

The following are selected cross sections showing how the localised ME s-coord looks like.

ME-800

1) Antartctic peninsula

2) Weddel Sea

3) Ross Sea

4) Bellingshausen Sea

[oceandie]

HORIZONTAL PRESSURE GRADIENT ERRORS

To get a sense of the HPGE associated with our local-ME s-coord grid, a one month run where the model is initialised with a stable stratification only function of depth is conducted.

For this run, we use the same initial profile used for HPGE test in the Europoean north-west shelf (see EGU-2021 presentation ). This profile is considered to be representative of winter off-shelf stratification. Of course, it cannot be considered representative of locations other than the North Atlantic. However, it is characterised by a quite strong gradient in corrispondence of the seasonal thermocline, which makes it a good initial test.

1) MAX (in the vertical and time) HPGE for GLOBAL SHELVES:

The blue line identifies the 300m isobath, i.e. the maximum depth of the upper envelope in the ME-300 setting.

2) MAX (in the vertical and time) HPGE for ANTARCTIC SHELF:

The blue line identifies the 800m isobath, i.e. the maximum depth of the upper envelope in the Antarctic shelf.

[oceandie]

OPTIMISING THE ENVELOPES FOR HPGE

Previous HPGE test showed that spurious numerical currents are generated off the 300m and 800m (in Antarctica) isobaths: this means that HPGE are driven mainly by the rmax of the second envelope.

Therefore, we decide

1) to apply a global rmax of 0.15 and 0.1 to the upper and the second envelope, respectively. In the case of the upper envelope, the slightly smaller rmax is just a precaution.

2) to optimise the second envelope for HPGE with the following itearative smoothing process:

After 3 iterations, we have (max HPGE = 0.011 m/s) :

The following are selected cross-sections to show how they look like after optimisation of the second envelope:

ME-300

1) Gulf stream region

2) Gulf of Mexico

3) Labrador Sea

4) Norwegian sea

5) European North West Shelf

6) North Sea

ME-800

1) Antartctic peninsula

2) Weddel Sea

3) Ross Sea

4) Bellingshausen Sea

[oceandie]

Just for refernce, the following are the HPGE developed by two James' hybrid s-z configurations:

1) hsz-39_ztaper (max HPGE = 0.036 m/s):

2) hsz-51_ztaper (max HPGE = 1.42 m/s):

[jdha]

I just need to double check which SZT-51 you have: p3 or p15

[oceandie]

I just need to double check which SZT-51 you have: p3 or p15

Not sure - I picked up the one in /gws/nopw/j04/jmmp_collab/se-orca025/se-nemo-domain_cfg, but there is not p3 or p15.

[jdha]

I think this task is complete - although we drifted into the SZT at the end which has since been updated (probably should open a separate issue for this)