20 nov. 2015: discussion on 2015 workplan proposal within the NEMO community

If you’re interested in the perspectives of NEMO development, you are welcome to join the videoconference :
Thursday 20 november at 2:30 Paris time.
The NEMO System team members will be connected to discuss and answer your questions on the 2015 workplan proposal in order to prepare the meeting of NEMO Developer's Committee taking place 4-5 december.
This workplan proposal is available here: https://forge.ipsl.jussieu.fr/nemo/wiki/2015WP
The videoconference will stay open from 2:30 to 4:30.
If you intend to join, please send a mail to System Team for connection information.

Participants:

Name	Institution	Role
Olivier Marti	LSCE, France	NEMO user and/or developer
Kim Serradell Maronda	Earth Science dept, BSC, Spain	NEMO user and/or developer
Miguel	Earth Science dept, BSC, Spain	NEMO user and/or developer
Oriol Tinto	Earth Science dept, BSC, Spain	NEMO user and/or developer
Koenigk Torben	Rossby Centre/ SMHI, Sweden EC-Earth-consortium	NEMO user and/or developer
Uwe Fladrich	SMHI/Rossby Centre, Sweden EC-Earth-consortium	NEMO user and/or developer
Jason Holt	NOC, UK	Developer's Committee member: shelf modeling expert
Luisa D’Amore	UNINA, Italy
Helen Hewitt	Met-Office	Developer's Committee member: consortium expert
Julien Le Sommer	CNRS	Developer's Committee member: consortium expert
Joel Hirschi	NOC	Developer's Committee member: consortium expert
Silvia Mocavero	CMCC	NEMO System Team
Tomas Lovato	CMCC	NEMO System Team
Simona Flavoni	CNRS	NEMO System Team
Sébastien Masson	CNRS	NEMO System Team
Martin Vancoppenolle	CNRS	NEMO System Team
Clément Rousset	CNRS	NEMO System Team
Christian Ethé	CNRS	NEMO System Team
Nicolas Martin	CNRS	NEMO System Team
Claire Lévy	CNRS	NEMO System Team
Emanuela Clementi	INGV	NEMO System Team
Clément Bricaud	Mercator	NEMO System Team
Julien Paul	Mercator	NEMO System Team
Tim Graham	Met-Office	NEMO System Team
Andrew Coward	NOC	NEMO System Team
Gurvan Madec	NOC & CNRS	NEMO System Team
George Nurser	NOC	NEMO System Team
Hedong Liu	NOC	NEMO System Team

Introduction (Claire Lévy):

Objective of the meeting: discuss 2015 workplan.

NEMO System Team develops the NEMO reference through a yearly workplan, i.e. a list of actions, with PIs..., and working groups.
Actions listed in workplan = code developed, whereas Working Groups discuss to make scientific and/or technical choices and set up appropriate development actions.
There is only 1 NEMO reference, so that there must be some coherency in all this.

At this point, discussions took place within NEMO System Team to elaborate the proposal.
Genral context:

• we now have the Development Strategy” document, finalised last june, identifying for mid-term what we agreeded on, and what remains in discussion
• 2015 will be a “special year”, because we needed a STABLE release in May for CMIP6

Objective of this meeting: get the inputs and comments from the NEMO community on the 2015 workplan proposal. For now, it is on the wiki, in 2 parts:

• the so called “Shared Actions”, core of the proposal, as the 2015 contribution to the evolution of NEMO defined in the Development Strategy, to be achieve by the NEMO System Team (experts), i.e. the “minimum 1 man-year” per institution of the Consortium
• the “per institution” pages: additional work, with additional manpower, to achieve some developments of specific interest for an institution of the consortium

As first overview, the Shared actions pages lists 7 paragraphs:

• AGRIF
• HPC
• Configuration Manager
• System Simplification
• Wave coupling
• NEMO robustness
• Publications in GMD

Open questions are:

• Do you find this list appropriate regarding both general strategy and the needs of your projects?
• Do you find this list realistic (i.e. feasible with the known manpower)?

Discussion

• What about the developments, updates... in the coupling interface, needed for CMIP6? They are not listed in 2015 workplan[Uwe Fladrich, EC-EARTH]

Those are expected to be finalised now (i.e. for the Merge Party in december 2014), or at the latest very beginning of 2015. Planning is to have the nemo_v3.6_STABLE release ready and fully validated for CMIP6 mid-2015.

• Could we have some details on HPC actions in 2015? What about HPC-2 action (MERCATOR - M. Chekki)? Are there some plans to look in details at relative performance of routines in NEMO?[Kim Serradell Maronda, BSC]

1. Chekki will be PI of an action on HPC still to be defined.
   BSC is more then welcome to contribute (to be discussed between BSC and S. Masson, leader of HPC working group)

• What about OpenMP? [J. Holt]

Concerning OpenMP, a first development branch exists, implementing OpenMP in the GYRE reference configuration. Work is planned to generalise it for NEMO (HPC-1 - CMCC-S.Mocavero).

• How and when is the validation for CMIP6 going to take place? [Uwe Fladrich, EC-EARTH]

There are 2 target configurations (with a fewdecinations for each): ORCA1 and ORCA025. A first set of validation will take place during the first semester of 2015.
For ORCA1, 2 focal points/actions are identified:
- ORCA1_LIM3_PISCES: Action "ROBUST-2" - NEMO ready for ESMs-CMIP6: ORCA1_LIM3_PISCES (C. Lévy); and some work on ORCA1_CICE at NOC (A. Coward).
- ORCA025: Actions "ROBUST-3" Met-Office and "ROBUST-4"at Mercator.
Comment by J. Le Sommer: there are 2 levels of testing: a "functionnal" level, and some work need to asses solutions. For the second one, a dedicated session will take place dunring the next DRAKKAR meeting, and an ocean medel development group meeting (CLIVAR) will take place afterwards also in Grenoble (in order to define a common forcing strategy).
Comment by Uwe Fladrich: EC_EARTH will continue to assess results and send feedback through tickets

• What about the list of reference configuration? Could there be a coupled configuration (i.e. ESM-like) in the list? [Uwe Fladrich, EC-EARTH]

The list of reference configurations is relatively small today and should remain so: the role of reference configuration are: to help new users setting up a first use of NEMO (all inputs files are available, results of functionnal tests are available), and allow through the SETTE tool, to assess and validate a new development/release of NEMO. Thus, reference configurations are an important amount of work for System Team. This work should not be increased too much: main responsibility of System team is to develop the platform.
Aside from the reference confiurations, there is the alternative to set up a "Shared configuration": shared by a community, but no support from System Team.
It is a very relevant idea to have tools to assess coupling interface in NEMO. S. Masson gives some information on the ongoing work: in 3.6_STABLE, coupling interface will we compiled, even id not used (using fake OASIS routines), and a kind of "fake_ESM" is envisaged to be able to set up preliminary tests at run-time.
Comment from U. Fladrich: EC_EARTH would be interested to take responsibility of an ESM shared configuration with NEMO (to be discussed further between EC-EARTH and System Team)

• What about the diagnostics when using variable volume layers? [O. Marti]

Indded, all diagnostics should be carefully reviewed when activating vvl since layers thickness now change every time step. This is going to be critical for users: concerns not only on-line diagnostics and outputs (done in the reference by System team), but also all the off-line diagnostics build by users. Need to send information, and maybe training for users.

• We (LSCE) are going to develop the accounting of water isotopes in NEMO. Could this task be added in the workplan? [O. Marti]

Interesting to share this information within the community, and see if other groups are interested. If additional tasks where to be envisaged in workplan, need to come with additional manpower in System Team. To be further elaborated with System Team.

• I'm very interested in action "ROBUST-5 CMCC": OBC with BDY for biogeochemistry, when will it be ready? [J. Holt]

20. Lovato, PI of this action, says will be ready next march. J. Holt says it's perfect.