Bug-fix for concatenation along lead-time of multi-member experiments + Update of NEMOVAR hard-coded period

Take advantage of autosubmit ability to rerun 4 times a sim job that would have failed for platform reasons : not deleting rcf which is already deleted; remove the restart directory that is going to be replaced

On our machines, in /cfu/software/ two versions of CDF tools are installed.
Oriol said that CFU has been using 2.1. 
After I added an export statement that redefines the PATH environmental variable,
it works fine now:

export PATH="$PATH:/cfu/software/cdftools2.1"

autosubmit_*.conf variables explanation added

on ithaca with UNKNOWN job status

compute the ocean heat content over various
basins/regions, layers, only in the mixed layer
depth, excluding the mixed layer depth ...
More information in the README

No commit message

No commit message

- "setupexp.sh" has been added which is useful for preparing binaries at expid folder at HPC (essential to run this once after running "create_exp.py")
- in the templates of ifs-only, nemo, ecearth and ecearth3, the following options are newly introduced:
  * MODEL (newly introduced into expdef)
  * %HPCUSER% other than $USER
  * BINPATH (at HPC binaries for the experiment will be called from this PATH)

No commit message

that contains the bash functions allowing a homogeneous
on-line and off-line ocean post-processing

- while the convention of picking IC from disk has kept similar as ecearth v2 templates... 

No commit message

No commit message

No commit message

- a time stamp intorduced while bundling err & out log files

Being able to submit to PDC, taking into account there is no scheduler in Ellen and then serial jobs have to be done wihtin a new HPCQueue: PSQueue, which check remote shell jobs to a ps wrapper.

- introduced chunk no. into the name of zipped restart file
- format for logfiles has been changed from tar to tar.gz