Adding warm layer and cool skin models to NEMO

[jwhile]

Adding the Takaya (2010) warm layer model and Artale (2002) cool skin model to NEMO.

The code is written so that it can be ran independently or alongside standard NEMO.

This code should have no impact on the results from standard NEMO, it only adds new output fields.

[jwhile]

A ticket wiki has been opened at https://forge.ipsl.jussieu.fr/nemo/wiki/ticket/1586

The tests I plan to conduct are:

1. The SETTE tests to ensure that I have not broken the standard NEMO functionality.
2. A test of the new NEMO 3.6+ code and a run of my Vn3.4 code that I have been using in my own work. This test will ensure that the two sets of code give the same results.

[jwhile]

the branch containing my changes is at

branches/2015/dev_r5656_Met_Office_3_diurnalSST/NEMOGCM

This branch contains changes to the code and also to the shared namelist_ref (see the wiki page).
Additionally the field_def.xml routine has also been updated to hold the new cool skin and warm layer variables.

[jwhile]

Cool skin from ORCA025 test

[jwhile]

warm layer from ORCA025 test

[jwhile]

Cool skin ORCA2

[jwhile]

Surface temperature ORCA2

[jwhile]

As shown in the wiki the code has been passed through the SETTE tests and passed. The code has also been compared to the output from rev 5656 of the trunk and produces identical results.

Additionally the following tests have been conducted.

1. The new code was run for a day and compared to results from the same day from the Met Offices Vn3.4 diurnal system. Both of these runs were on the ORCA025 grid in diurnal only mode (i.e ln_diurnal_only=.TRUE.) The same fluxes were used in both experiments. The results from both runs were identical, with the warm layer and cool skin results shown in the following images (units are degees C)

Cool skin

Warm layer

1. To demonstrate that the diurnal code could be ran with the rest of NEMO switched on, I reran the ORCA2_LIM_OBS SETTE test, but with ln_diurnal=.TRUE. This ran and did produced warm layer and cool skin skin values along with the NEMO outputs. Unfortunately this run only had daily averaged forcing so the cool skin and warm layers are not realistic, but do exist. Images of the cool skin and top layer model temperature for the end of the run are shown below (units are degrees C):

Cool skin

Top layer temperature

[jwhile]

Andrea,
Could you please look at the above and fill in the reviews wiki https://forge.ipsl.jussieu.fr/nemo/wiki/ticket/1586/Review. Please let me know if you think there are any issues.

[jwhile]

As per the review, a new chapter has been added to the NEMO documentation. This chapter (the DIU chapter) describes the cool skin and warm layer models used.

In the review I was also asked to consider merging the namdiu namelist parameters into other pre-existing namelists. However, as DIU is a separate module from most of NEMO I feel it is important that it has its own namelist; however short. Having a separate namelist will also allow it to be extended more easily; additional options could include switching on/off the cool skin/warm layer models individualy and providing more user control over the parameters.